Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Have you lost track of developments in generative linguistics, fi nding yourself unsure about
the distinctive features of Minimalism? Would you like to know more about recent advances in
the genetics of language, or about right hemisphere linguistic operation? Has your interest in
narrative drawn you to question the relation between stories and grammars? The Cambridge
Encyclopedia of the Language Sciences addresses these issues, along with hundreds of others.
It includes basic entries for those unfamiliar with a given topic and more specific entries for
those seeking more specialized knowledge. It incorporates both well-established fi ndings and
cutting-edge research as well as classical approaches and new theoretical innovations. The volume is aimed at readers who have an interest in some aspect of language science but wish to
learn more about the broad range of ideas, fi ndings, practices, and prospects that constitute
this rapidly expanding field, a field arguably at the center of current research on the human
mind and human society.
Patrick Colm Hogan is a professor in the Department of English and the Program in Cognitive
Science at the University of Connecticut. He is the author of ten books, including Cognitive
Science, Literature, and the Arts: A Guide for Humanists and The Mind and Its Stories: Narrative
Universals and Human Emotion (Cambridge University Press, 2003).
For both range and depth of exposition and commentary on the diverse disciplinary angles
that exist on the nature of language, there is no single volume to match this fi ne work of
reference.
Akeel Bilgrami, Columbia University
The Cambridge Encyclopedia of the Language Sciences is a very welcome addition to the field
of language sciences. Its comprehensiveness is praiseworthy, as is the quality of its entries and
discussions.
Seymour Chatman, University of California, Berkeley
Th is ambitious and comprehensive work, and the very high quality of the editors and contributors, ensure that it will be a valuable contribution to the understanding of language and
its uses, for both professionals and a more general audience.
Noam Chomsky, Massachusetts Institute of Technology
GENERAL EDITOR
Patrick Colm Hogan
University of Connecticut, Storrs
Barbara Lust
Cornell University
William Croft
University of New Mexico
Lee Osterhout
University of Washington
Lyle Jenkins
Biolinguistics Institute
James Pustejovsky
Brandeis University
Howard Lasnik
University of Maryland
Deborah Cameron
University of Oxford
Loraine Obler
City University of New York
Nigel Fabb
University of Strathclyde
William OGrady
University of Hawaii
Susan Pintzuk
University of York
Ronald Geluykens
University of Oldenburg
Eleanor Rosch
University of California, Berkeley
Margaret Harris
Oxford Brookes University
Jay Rueckl
University of Connecticut
Zoltn Kvecses
Etvs Lornd University
Mark Turner
Case Western Reserve University
CONTENTS
List of Entries
A Note on Cross-References and the Alphabetization of the Entries
Preface: On the Very Idea of Language Sciences
Acknowledgments
page xi
xv
xvii
xxiii
William Croft
12
23
James Pustejovsky
35
Florian Coulmas
46
Lyle Jenkins
Acquisition of Language
56
Barbara Lust
65
ENTRIES
77
List of Contributors
941
Index
953
ix
ENTRIES
A
Abduction
Absolute and Statistical Universals
Accessibility Hierarchy
Acoustic Phonetics
Adaptation
Ad Hoc Categories
Adjacency Pair
Age Groups
Aging and Language
Agreement
Agreement Maximization
Alliteration
Ambiguity
Amygdala
Analogy
Analogy: Synchronic and Diachronic
Analyticity
Anaphora
Animacy
Animal Communication and Human
Language
Aphasia
Areal Distinctness and Literature
Art, Languages of
Articulatory Phonetics
Artificial Languages
Aspect
Auditory Processing
Autism and Language
Autonomy of Syntax
B
Babbling
Basal Ganglia
Basic Level Concepts
Bilingual Education
Bilingualism, Neurobiology of
Bilingualism and Multilingualism
Binding
Biolinguistics
Birdsong and Human Language
Blended Space
Blindness and Language
Bounding
Brain and Language
Brocas Area
C
Cartesian Linguistics
Case
Categorial Grammar
Categorization
Causative Constructions
C-Command
Cerebellum
Charity, Principle of
Childrens Grammatical Errors
Chirographic Culture
Clitics and Cliticization
Codeswitching
Cognitive Architecture
Cognitive Grammar
Cognitive Linguistics and Language
Learning
Cognitive Linguistics, Language Science,
and Metatheory
Cognitive Poetics
Coherence, Discourse
Coherence, Logical
Colonialism and Language
Color Classification
Communication
Communication, Prelinguistic
Communicative Action
Communicative Intention
Comparative Method
Competence
Competence and Performance, Literary
Compositionality
Computational Linguistics
Concepts
Conceptual Blending
Conceptual Development and Change
Conceptual Metaphor
Conduit Metaphor
Connectionism and Grammar
Connectionism, Language Science, and
Meaning
Connectionist Models, Language
Structure, and Representation
Consciousness and Language
Consistency, Truth, and Paradox
Constituent Structure
Constraints in Language Acquisition
Construction Grammars
Contact, Language
Context and Co-Text
Control Structures
Conversational Implicature
Conversational Repair
Conversation Analysis
Cooperative Principle
Core and Periphery
Corpus Callosum
Corpus Linguistics
Creativity in Language Use
Creoles
Critical Discourse Analysis
Critical Periods
Culture and Language
Cycle, The
D
Deconstruction
Defi nite Descriptions
Deixis
Descriptive, Observational, and
Explanatory Adequacy
Dhvani and Rasa
xi
List of Entries
Dialect
Dialogism and Heteroglossia
Diff usion
Digital Media
Diglossia
Discourse Analysis (Foucaultian)
Discourse Analysis (Linguistic)
Discrete Infi nity
Disorders of Reading and Writing
Division of Linguistic Labor
Dyslexia
E
Ellipsis
Embodiment
Emergentism
Emergent Structure
Emotion and Language
Emotion, Speech, and Writing
Emotion Words
Emplotment
Encoding
nonc/Statement (Foucault)
Essentialism and Meaning
Ethics and Language
Ethnolinguistic Identity
Event Structure and Grammar
Evidentiality
Evolutionary Psychology
Exemplar
Exemplar Theory
Extinction of Languages
F
Family Resemblance
Feature Analysis
Felicity Conditions
Field (Bourdieu)
Film and Language
Filters
Focus
Foregrounding
Forensic Linguistics
Formal Semantics
Forms of Life
Frame Semantics
Framing Effects
Frontal Lobe
Functional Linguistics
G
Games and Language
Gender and Language
Gender Marking
Generative Grammar
Generative Poetics
Generative Semantics
Generic- and Specific-Level Metaphors
xii
H
Habitus, Linguistic
Head-Driven Phrase Structure Grammar
Hippocampus
Historical Linguistics
Historical Reconstruction
Holophrastic Stage, The
Homologies and Transformation Sets
I
Icon, Index, and Symbol
Ideal Speech Situation
Identity, Language and
Ideology and Language
Idioms
Idle Talk and Authenticity
Ijtihd (Interpretive Effort)
I-Language and E-Language
Illocutionary Force and Sentence Types
Image Schema
Implicational Universals
Indeterminacy of Translation
Indexicals
Inequality, Linguistic and
Communicative
Infantile Responses to Language
Information Structure in Discourse
Information Theory
Innateness and Innatism
Integrational Linguistics
Intension and Extension
Intentionality
Internal Reconstruction
Interpretation and Explanation
Interpretive Community
Intertextuality
Intonation
Irony
L
Language, Natural and Symbolic
Language Acquisition Device (LAD)
Language Change, Universals of
Language Families
Language-Game
Language-Learning Environment
Language of Thought
Language Policy
Laws of Language
Learnability
M
Mapping
Markedness
Market, Linguistic
Marxism and Language
Meaning and Belief
Meaning and Stipulation
Meaning Externalism and Internalism
Media of Communication
Memes and Language
Memory and Language
Mental Models and Language
Mental Space
Merge
Metalanguage
Metaphor
Metaphor, Acquisition of
Metaphor, Information Transfer in
Metaphor, Neural Substrates of
Metaphor, Universals of
Meter
Methodological Solipsism
Methodology
Metonymy
Minimalism
Mirror Systems, Imitation, and
Language
Modality
Modern World-System, Language and
the
Modularity
Montague Grammar
Mood
Morpheme
Morphological Change
Morphological Typology
Morphology
Morphology, Acquisition of
Morphology, Evolution and
List of Entries
Morphology, Neurobiology of
Morphology, Universals of
Motif
Movement
Music, Language and
N
Narrative, Grammar and
Narrative, Neurobiology of
Narrative, Scientific Approaches to
Narrative Universals
Narratives of Personal Experience
Narratology
Nationalism and Language
Natural Kind Terms
Necessary and Sufficient Conditions
Negation and Negative Polarity
Network Theory
Neurochemistry and Language
Neuroimaging
Number
O
Occipital Lobe
Optimality Theory
Oral Composition
Oral Culture
Ordinary Language Philosophy
Origins of Language
Overregularizations
P
Parable
Paralanguage
Parameters
Parietal Lobe
Parsing, Human
Parsing, Machine
Passing Theories
Performance
Performative and Constative
Perisylvian Cortex
Perlocution
Person
Philology and Hermeneutics
Phoneme
Phonetics
Phonetics and Phonology, Neurobiology
of
Phonological Awareness
Phonology
Phonology, Acquisition of
Phonology, Evolution of
Phonology, Universals of
Phrase Structure
Pidgins
Pitch
Poetic Form, Universals of
Q
Qualia Roles
Quantification
Quantitative Linguistics
R
Radical Interpretation
Reading
Realization Structure
Rectification of Names (Zheng Ming)
Recursion, Iteration, and
Metarepresentation
Reference and Extension
Reference Tracking
Register
Regularization
Relevance Theory
Religion and Language
Representations
Rhetoric and Persuasion
Rhyme and Assonance
Rhythm
Right Hemisphere Language Processing
Role and Reference Grammar
Rule-Following
S
Schema
Scripts
Second Language Acquisition
Self-Concept
Self-Organizing Systems
Semantic Change
Semantic Fields
Semantic Memory
Semantic Primitives (Primes)
Semantics
Semantics, Acquisition of
Semantics, Evolution and
Semantics, Neurobiology of
Semantics, Universals of
Semantics-Phonology Interface
Semantics-Pragmatics Interaction
Semiotics
Sense and Reference
Sentence
Sentence Meaning
Sexuality and Language
Signed Languages, Neurobiology of
Sign Language, Acquisition of
Sign Languages
Sinewave Synthesis
Situation Semantics
Socially Distributed Cognition
Sociolinguistics
Source and Target
Specific Language Impairment
Speech-Acts
Speech Anatomy, Evolution of
Speech-Language Pathology
Speech Perception
Speech Perception in Infants
Speech Production
Spelling
Spreading Activation
Standardization
Standard Theory and Extended Standard
Theory
Stereotypes
Story and Discourse
Story Grammar
Story Schemas, Scripts, and Prototypes
Stress
Structuralism
Stylistics
Stylometrics
Subjacency Principle
Suggestion Structure
Syllable
xiii
List of Entries
Synchrony and Diachrony
Syntactic Change
Syntax
Syntax, Acquisition of
Syntax, Evolution of
Syntax, Neurobiology of
Syntax, Universals of
Syntax-Phonology Interface
T
Teaching Language
Teaching Reading
Teaching Writing
Temporal Lobe
Tense
Text
Text Linguistics
Thalamus
Thematic Roles
Theory of Mind and Language Acquisition
Tone
Topicalization
Topic and Comment
Traces
xiv
Transformational Grammar
Translation
Truth
Truth Conditional Semantics
Two-Word Stage
Typology
Verbal Reasoning
Verbal Reasoning, Development of
Verifiability Criterion
Verse Line
Voice
Voice Interaction Design
Wernickes Area
Word Classes (Parts of Speech)
Word Meaning
Word Order
Word Recognition, Auditory
Word Recognition, Visual
Words
Working Memory and Language
Processing
Writing, Origin and History of
Writing and Reading, Acquisition of
Writing and Reading, Neurobiology of
Writing Systems
V
Vagueness
Verbal Art, Evolution and
Verbal Art, Neuropsychology of
Verbal Display
Verbal Humor
Verbal Humor, Development of
Verbal Humor, Neurobiology of
X
X-Bar Theory
Cross-references are signaled by small capitals (boldface when implicit). They are designed
to indicate the general relevance of the cross-referenced entry and do not necessarily imply
that the entries support one another. Note that the phrasing of the cross-references does not
always match the entry headings precisely. In order to minimize the disruption of reading,
entries often use shortened forms of the entry headings for cross-references. For example, this
process involves parietal structures points to the entry Parietal Lobe. In some cases, a
cross-reference may refer to a set of entries. For example, architectures of this sort are found
in connectionism alerts the reader to the presence of entries on connectionism generally,
rather than to a single entry. Finally, a cross-reference may present a heading in a different
word order. For example, the target entry for here we see another universal of phonology would be listed as Phonology, Universals of.
In general, entries with multiword headings are alphabetized under their main language
term. Thus, the entry for Universals of Phonology is listed as Phonology, Universals of. The
main exceptions to this involve the words language and linguistic or linguistics, where another
term in the heading seemed more informative or distinctive in the context of language sciences
(e.g., Linguistic Market is listed as Market, Linguistic).
xv
WHAT IS SCIENCE?
Philosophers of science have often been concerned to defi ne a
demarcation criterion, separating science from nonscience. I
have not found any single criterion, or any combination of criteria, compelling in the sense that I have not found any argument that, to my satisfaction, successfully provides necessary
and sufficient conditions for what constitutes a science.
In many ways, ones acceptance of a demarcation criterion is
guided by what one already considers to be a science. More
exactly, ones formulation of a demarcation criterion tends to
be a function of what one takes to be a paradigmatic science or,
in some cases, an exemplary case of scientific practice.
The advocates of strict demarcation criteria meet their mirror opposites in writers who assert the social construction of
science, writers who maintain that the difference between
science and nonscience is simply the difference between distinct positions within institutions, distinct relations to power.
Suppose we say that one discipline or theory is a science and
another is not. Th is is just to say that the former is treated as
science, while the latter is not. The former is given authority in
academic departments, in relevant institutions (e.g., banks, in
the case of economics), and so on.
Again, this is not the place for a treatise on the philosophy
of science. Here it is enough to note that I believe both sides
are partially correct and partially incorrect. First, as already
noted, I do not believe that there is a strict, defi nitive demarcation criterion for science. However, I do believe that there is
xvii
Preface
In putting together a volume on science, then, I have tried
to incorporate the insights of both the more positive views
of science and the more social constructionist views. Put in
a way that may seem paradoxical, I have tried to include all
approaches that fit the loose criteria for science just mentioned.
I believe that these loose criteria apply not only to paradigmatic
sciences themselves but also to many social critiques of science that stress social construction. I have therefore included a
wide range of what are sometimes called the human sciences.
Indeed, the volume could be understood as encompassing the
language-relevant part of the human sciences which leads to
our second question.
Group Dynamics
Society
Mind
Individual Interactions
Mental Representations
Intentions
WHAT IS LANGUAGE?
Like science, ones defi nition of language depends to a great
extent on just what the word calls to mind. Ones view of language is likely to vary if one has in mind syntax or semantics, hearers or speakers, dialogues or diaries, brain damage
or propaganda, storytelling or acoustic phonetics. A fi rst
impulse may be to see one view of language as correct and the
others as false. And, of course, some views are false. However,
I believe that our understanding of language can and, indeed,
should sustain a great deal of pluralism.
In many ways, my own paradigm for human sciences is cognitive science. Cognitive science brings together work from a
remarkable array of disciplines literally, from Anthropology
to Zoology. Moreover, it sustains a range of cognitive architectures, as well as a range of theories within those architectures. Thus, it is almost by its very nature pluralistic. Of course,
some writers wish to constrain this pluralism, insisting that one
architecture is right and the others are wrong. Certainly, one
can argue that particular architectures are wrong. However,
perhaps the most noteworthy aspect of cognitive science is
that it sustains a series of types of cognitive architecture. In
Cognitive Science, Literature, and the Arts (2003), I argued that
these types capture patterns at different levels of analysis. Thus,
all are scientifically valuable.
More exactly, we may distinguish three levels of cognitive
analysis: bodies, minds, and groups or societies. These levels stand in a hierarchical relation such that bodies are more
explanatorily basic than minds, and minds are more explanatorily basic than groups or societies. Lower levels provide
causally necessary principles for higher levels. Minds do not
operate without brains. People without minds do not interact
in groups. In other words, lower levels explain higher levels.
However, higher-level patterns provide interpretive principles
for understanding lower levels (see interpretation and
explanation). We explain the (mental) feeling of fear by reference to the (bodily) operation of the amygdala. But, at the
same time, we understand amygdala activity as fear because
we interpret that activity in terms of the mental level.
In the analysis of cognition, the most basic, bodily cognitive architecture is provided by neurobiology. However, due
to the intricate particularity of neurobiology, we often draw
on more abstract associative models at this level. These models serve to make the isolation and explanation of patterns
less computationally complex and individually variable. The
xviii
Body
Associative Networks
Brains
Preface
the impairment of (mental) syntactic capacities by reference
to (bodily) lesions in that area. For our purposes, the crucial
part of this analysis is its implication that language includes
all three levels and that the sciences of language should
therefore encompass brains, associative networks, intentions,
mental representations, individual interactions, and group
dynamics. Th is takes us to our third question.
xix
Preface
this part very narrowly as including only phenomena that are
necessarily bound up with oral speech, sign languages,
or writing. More theoretically, one might defi ne this part as
comprising neurobiological, mental, or social phenomena that
occur only in connection with distinctive properties of speech,
signing, or writing.
Certainly, an encyclopedia treating language will focus
on phenomena that are inseparable from speech, sign languages, and/or writing and on such distinctive aspects of natural language as syntax. However, here, too, I believe it would
be a mistake to confi ne language sciences within a narrowly
defi ned domain. Therefore, I have adopted a looser criterion.
The volume centrally addresses distinctive properties of natural language. However, it takes up a wider range of phenomena
that are closely connected with the architectural or, even more
importantly, the functional features of speech, sign languages,
and writing.
There are several cognitive operations for which speech,
signing, and writing appear to have direct functional consequences. One is referential the specification, compilation,
and interrelation of intentional objects (see the entries on
reference). Here I have in mind phenomena ranging from
the division of the color spectrum to the elaboration of causal
relations. A second area is mnemonic the facilitation and partial organization of memory (see, for example, encoding). A
third is inferential the derivation of logical implications. A
fourth is imaginative the expansion and partial structuring
of simulation. One could think of the fi rst and second functions as bearing on direct, experiential knowledge of present
or past objects and events. The third and fourth functions bear,
rather, on indirect knowledge of actual or possible objects and
events. Two other functions are connected with action rather
than with knowledge. The fi rst is motivational the extension
or elaboration of the possibilities for emotional response (see
emotion and language). A fi nal area is interpersonal the
communication of referential intents, memories, inferences,
simulations, and motivations.
In determining what should be included in the volume,
I have taken these functions into account, along with architectural considerations. Thus I see issues of interpretation
and emplotment (one of the key ways in which we organize
causal relations) as no less important than phonology or syntactic structure. Of course, we have more fi rmly established
and systematic knowledge in some areas than in others. Thus
some entries will necessarily be more tentative, and make reference to a broader variety of opinion or a more limited research
base. But that is not a reason to leave such entries aside. Again,
the purpose of an encyclopedia of language science is not to
present a compilation of well-established particular facts, but
rather to present our current state of knowledge with respect to
explanation and understanding.
In keeping with this, when generating the entries (e.g.,
Phonology, Syntax, Neurobiology of Phonology,
Neurobiology of Syntax, Acquisition of Phonology, and so
on), I have tried to be as systematic as possible. Thus the volume includes some topics that have been under-researched
and under-theorized. For example, if neurobiology does in fact
provide a causal substrate for higher levels, then there should
xx
Preface
particularly language science. I have tried to follow this philosophy throughout the volume. Specifically, I have sought to
present a range of theoretical ideas (as well as more theoryindependent topics), placing them together in such a way as
to encourage a mutual sharpening of ideas and insights. To
borrow M. M. Bakhtins terms (1981), I have not set out to provide a monological source of authoritative discourse. Rather, I
have sought to present a heteroglot volume with which readers may interact dialogically (see dialogism and heteroglossia) hopefully, to produce more intellectually adequate
theories later. Toward this end, I have encouraged authors to
be open about their own judgments and attitudes. There is a
common view that a piece of writing is biased if the speaker
frankly advocates one point of view. But, in fact, the opposite is
the case. A piece of writing is biased if a speaker acts as though
he or she is simply reporting undisputed facts, when in fact he
or she is articulating a partisan argument. Being open, dialogical, and multivocal does not mean being bland. Indeed, insight
is more likely to be produced through the tension among ideas
and hypotheses that are clearly delineated in their differences.
Th is is no less true in the language sciences than elsewhere.
Indeed, that is one reason why this volume treats language sciences, not the science of language.
Patrick Colm Hogan
xxi
ACKNOWLEDGMENTS
First of all, I must thank Phil Laughlin, who (inspired by the exemplary MIT Encyclopedia of
the Cognitive Sciences) suggested the project initially and invited me to make a more formal
proposal. Without Phils initial idea and subsequent encouragement, this volume would not
now exist. Eric Schwartz took over from Phil, then Simina Calin took over from Eric; both were
supportive and helpful, as were the several editorial assistants, most recently April Potenciano,
Christie Polchowski, and Jeanie Lee. Regina Paleski and Mark Fox ably shepherded this complex project through the production process; Phyllis Berk worked with devotion on copy editing the manuscript; and Robert Swanson took on the tough job of indexing.
The members of the editorial board kindly provided comments on the list of entries and
suggested possible authors. They also served as second readers for most of the entries. I am
indebted to them all. It is difficult and unrewarding work, but extremely valuable. Some entries
were evaluated by specialists not on the editorial board. I am deeply grateful to the following
scholars who agreed to read and comment on entries: J. Abutalebi, E. Ahlsn, A. Aikhenvald,
S. Anderson, A. Atkin, S. Barker, J. Beall, D. Beaver, H. Bejoint, H. Ben-Yami, A. Berger,
D. Bickerton, A. Bilgrami, S. Blackmore, J. Blommaert, C. Bowern, E. Brenowitz, J. Bybee,
J. Carroll, T. Deacon, M. DeGraff, J.-L. Dessalles, A. Edgar, C. Elgin, R. Ferrer i Cancho, J. Field,
H. Filip, D. Finkelstein, J. Forrester, R. Gibbs, R. Gibson, R. Giora, R. Gleave, K. Gluer-Pagin,
M. Goral, M. Hashimoto, J. Heath, D. Herman, R. Hilpinen, J. Hintikka, K. Hoff man, K. Holyoak,
P. Hutchinson, J. Hyun, P. Indefrey, M. Israel, K. Johnson, M. Johnson, J. Kane, P. Kay, A. Kibort,
S. Kiran, C. Kitzinger, W. Labov, B. Lafford, C. Larkosh, A. Libert, P. Livingston, K. Ludwig,
M. Lynch, J. Magnuson, G. Marcus, R. May, J. McGilvray, A. Mehler, S. Mills, D. Moyal-Sharrock,
K. Oatley, B. OConnor, L. Pandit, B. Partee, J. Pennebaker, P. Portner, C. Potts, J. Robinson,
S. Rosen, S. Ross, J. Saul, R. Schleifer, M. Shibatani, R. Skousen, S. Small, W. Snyder, M. Solms,
F. Staal, P. Stockwell, L. Talmy, H. Truckenbrodt, J. P. Van Bendeghem, W. van Peer, S. Wheeler,
and L. Zhang. Thanks to M. Cutter for help with the illustrations.
For some time, a graduate assistant, Karen Renner, took care of many secretarial duties.
Th is work was supported by the English Department at the University of Connecticut, with
some added funding from the University of Connecticut Research Foundation. The support of
the English Department was due to the kindness and commitment of our department head,
Bob Tilton without his help, this project would not have been possible.
Work for the entry on Ad Hoc Categories was supported by National Science Foundation
Grant BCS-0212134 and DARPA Contract FA865005-C-7256 to Lawrence W. Barsalou.
The entry on Dyslexia was prepared with support from a British Academy Research
Readership to Margaret J. Snowling.
Preparation of the manuscript for Speech Production was supported by NIDCD A-93 and
NIDCD grant DC-03782, both to Haskins Laboratories.
Research for Paisley Livingstons entries benefited from fi nancial support from the Research
and Postgraduate Studies Committee of Lingnan University, Hong Kong.
xxiii
1
LANGUAGE STRUCTURE IN ITS HUMAN
CONTEXT: NEW DIRECTIONS FOR THE LANGUAGE
SCIENCES IN THE TWENTY-FIRST CENTURY
William Croft
butterfly) or a grammatical construction (such as the ModifierHead construction for English noun phrases) emerges as a convention when it becomes a regularly used means for solving the
recurrent coordination problem of referring to a specific experience that is to be communicated.
Linguistic convention actually operates at two levels: the
grammatical level of words and constructions, at which the
speakers intentions are formulated; and the phonological level
of the articulation and perception of the sounds that make up
the grammatical units. Th is is the phenomenon described as
duality of patterning in language (Hockett 1960). One could
imagine in principle that linguistic convention possessed only
one level: perceivable sounds (or gestures or written images,
depending on the medium), corresponding to part (i) in the
defi nition of convention, that directly conveyed the speakers
intention (the recurrent coordination problem) as a whole, corresponding to part (v) in the defi nition of convention. These
exist in interjections with specific functions such as Hello and
Thanks. However, most linguistic expressions are complex,
consisting of discrete, meaningful units. Complex linguistic
expressions evolved for two reasons: First, the number of different speaker intentions to be communicated grew to be indefinitely large; and second, a speakers intended message came to
be broken down into recurrent conceptual parts that could be
recombined to produce the indefi nite variety of messages.
Again, one could imagine that each conventional linguistic unit consisted of a unique sound (gesture, image). But languages have distinct meaningful units that are made up of
different sequences of the same sounds: bat, sat, Sam, same,
tame, time, etc. Th is system has evolved for the same two reasons: the increasing number of meaningful units (even the
recurring ones) necessary to convey the indefi nitely large number of speaker intentions, and an ability to break down a sound
signal (or gesture, or image) into parts that can be recombined
as a sequence of sounds (or gestures or images). Thus, the duality of patterning characteristic of human language has evolved
to accommodate the huge number of speaker intentions that
people want to convey, and to exploit the facts that intentions
can be broken down into recombinable conceptual units and
that the medium of expression can be broken down into recombinable units as well.
Language is therefore a joint action that operates simultaneously at four levels (Clark 1996). The higher-numbered levels
are dependent on the lower-numbered levels; the individual
actions of the interlocutors are given in italics:
(4) proposing and taking up a joint project (joint action);
(3) signaling and recognizing the communicative intention;
(2) formulating and identifying the proposition;
(1) producing and attending to the utterance.
content words, such as nouns and verbs. In this way, the speaker
has transformed the unique whole of the original experience
into parts that can be expressed by language.
Th is is not the end of the verbalization process. Content
words denote only general categories of parts of the experience to be verbalized. In order to communicate the original
experience, the speaker must tie down the categories to the
unique instances of objects, events, and so forth in the experience, and the speaker must assemble the parts into a structure representing the original whole that the speaker intends
to verbalize. That is to say, corresponding to the categorizing
step in verbalizing the parts of the experience, there is a particularizing step that indicates the unique parts; and corresponding to the steps of propositionalizing and subchunking
are integrative steps of structuring and cohering, respectively
(Croft 2007). These latter three steps give rise to grammar in the
sense of grammatical constructions, inflections, and particles,
and the semantic commonalities among grammatical categories across languages.
The particularizing step takes a category (a type) and selects
an instance (token) or set of tokens, and also identifies it by situating it in space and time. For object concepts, selecting can
be accomplished via the inflectional category of number, and
via the grammatical categories of number and quantification
(three books, an ounce of gold ). For action concepts, selecting
is done via grammatical aspect, which helps to individuate
events in time (ate vs. was eating), and via agreement with subject and/or object, since events are also individuated by the
participants in them (I read the paper and She read the magazine describe different reading events). Objects and events can
be situated in space via deictic expressions and other sorts of
locative expressions (this book, the book on the table). Events
and some types of objects can be situated in time via tense and
temporal expressions (I ate two hours ago; ex-mayor). Events
and objects can also be situated relative to the mental states
of the interlocutors: The article in the book indicates that the
particular object is known to both speaker and hearer, and the
modal auxiliary in She should come indicates that the event
exists not in the real world but in the attitude of obligation in
the mind of the speaker.
The structuring step takes participants and the predicated
event in a clause and puts them together, reassembling the
predicate and the argument(s) into the subchunk from which
they were derived by propositionalizing. Grammatically this is
a complex area. It includes the expression of grammatical relations in what is called the argument structure of a predicate, so
that She put the clock on the mantle indicates which referent is
the agent (the subject), which the thing moved (the object), and
which the destination of the motion (the prepositional phrase).
But it also includes alternative formulations of the same event,
such as The clock was put on the mantle (the passive voice construction) and It was the mantle where she put the clock (a cleft
construction). The alternative constructions function to present the information in the proposition in different ways to the
hearer, depending on the way the discourse is unfolding; they
are referred to as information structure or discourse function.
Finally, the cohering step takes the clauses (subchunks) and
reassembles them into a whole that evokes the original whole
linguistic comparison of the meanings expressed by grammatical categories and constructions. The typological approach to
grammar has constructed conceptual spaces for a number of
semantic domains using techniques such as the semantic map
model (see Haspelmath 2003 for a survey of recent studies) and
multidimensional scaling (Croft and Poole 2008).
To sum up, the usage-based/exemplar model can be applied
to both phonological patterns in words and grammatical structures in constructions. A speakers knowledge of language
is the result of the interplay between two learning processes.
One learning process is the tallying of statistical regularities of
tokens of words and constructions with a particular phonetic
realization, performing a particular communicative act in a
specific social interaction. The other is the organization of these
tokens into categories and the formation of generalizations that
allow the reuse or replication of these grammatical structures
to solve future coordination problems in communication.
conventions of the speech community. Each speakers systematic knowledge of his or her language is different, because of
differences in the range of language use to which each speaker
is exposed.
SUMMARY
The scientific study of language in its pragmatic, cognitive,
and social context beginning in the latter half of the twentieth century is converging on a new perspective on language
in the twenty-fi rst century. Linguistic conventions coordinate
communication, which in turn coordinates joint actions. The
fragility of social interaction by individuals leads to creativity,
variation, and dynamism in the verbalization and vocalization
of language. Individual linguistic knowledge (the linguistic
system) reflects the remembered history of language use and
mediates processes of language change. The continually changing structure of society, defi ned by common ground emerging
from shared practices (joint actions), guides the evolution of
linguistic conventions throughout its history. Human history
in turn has spawned tremendous linguistic diversity, which
reflects the diversity of human social and cognitive capacity.
But the unchecked operation of contemporary social forces is
leading to the destruction of speech communities and the mass
extinction of human languages today.
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Austin, J. L. 1962. How to Do Things with Words. Cambridge: Harvard
University Press.
Bolinger, Dwight. 1980. Language, the Loaded Weapon. London:
Longmans.
Bratman, Michael. 1992. Shared cooperative activity. Philosophical
Review 101: 32741.
Bybee, Joan L. 2001. Phonology and Language Use. Cambridge
Cambridge University Press.
. 2003. Mechanisms of change in grammaticalization: The role
of frequency. In Handbook of Historical Linguistics, ed. Brian Joseph
and Richard Janda, 60223. Oxford: Blackwell.
Bybee, Joan L ., and Sandra A. Thompson. 1997. Th ree frequency effects
in syntax. In Proceedings of the 23rd Annual Meeting of the Berkeley
Linguistics Society, ed. Matthew L. Juge and Jeri O. Moxley, 37888.
Berkeley : Berkeley Linguistics Society.
Chafe, Wallace. 1977. The recall and verbalization of past experience. In Current Issues in Linguistic Theory, ed. Peter Cole, 21546.
Bloomington: Indiana University Press.
Chambers, J. K., and Peter Trudgill. 1998. Dialectology. 2d ed.
Cambridge: Cambridge University Press.
10
11
with a speakers (or hearers) semantic and contextual knowledge. Here, we review some of what we have learned about the
psychology of linguistic form, as it pertains to sounds, words,
and sentences.
SOUNDS
12
13
14
Using the lexical entry cat as an example, imagine a connectionist system in which all the semantic features associated
with cat, such as [whiskers], [domestic pet], and so on (which
are also shared with all other conceptual lexical entities bearing those features, such as <lion>, <dog>, etc.), are directly
associated with the phonological units that comprise its word
form /k/, /ae/, /t/ (which are likewise shared with all other
word forms containing these phonemes) by means of individual association links that directly tie individual semantic features with individual phonological units (Rueckl et al. 1997).
One important consequence of this hypothetical arrangement
is that individual word forms do not exist as free-standing representations. Instead, the entire lexical entry is represented as
a vector of weighted links connecting individual phonemes to
individual lexical semantic and syntactic features. It logically
follows from this model, then, that if all or most of the semantic
features of the word cat, for example, were destroyed or otherwise made unavailable to the processor, then the set of phonological forms /k/ /ae/ /t/, having nothing to link to, would
have no means for mental representation, and would therefore
not be available to the language processor. We will present
experimental evidence against this model that, instead, favors
models in which a full phonological word (e.g., /kaet/) is represented in a localist fashion and is accessible to the language
processor, even when access to its semantic features is partially
or entirely disrupted.
Several of the most prominent theories of morphology and
lexical structure within formal linguistics make explicit claims
about modularity of meaning and form. Ray Jackendoff (1997),
for example, presents a theory that has a tripartite structure, in
which words have separate identities at three levels of representation form, syntax, and meaning and that these three levels
are sufficient to encode the full array of linguistic information
encoded by each word. His model provides further details in
which it is proposed that our ability to store, retrieve, and use
words correctly, as well as our ability to correctly compose
morphemes into complex words, derives from a memorized
inventory of mapping functions that picks out the unique representations or feature sets for a word at each level and associates these elements with one another in a given linguistic
structure.
While most psycholinguistic models of language processing have not typically addressed the mapping operations
assumed by Jackendoff, they do overlap significantly in terms
of addressing the psychological reality of his hypothetical tripartite structure in the mental lexicon. Although most experimental treatments of the multilevel nature of the lexicon have
been developed within models of language production, as will
be seen, there is an equally compelling body of evidence for
multilevel processing from studies of language comprehension
as well.
The most influential lexical processing models over the
last two decades make a distinction between at least two levels: the lemma level, where meaning and syntax are stored,
and the lexeme level, where phonological and orthographic
descriptions are represented. These terms and the functions
associated with them were introduced in the context of a computational production model by Gerard Kempen and Pieter
15
16
SENTENCES
On the surface, a sentence is a linear sequence of words. But in
order to extract the intended meaning, the listener or reader
must combine the words in just the right way. That much is
obvious. What is not obvious is how we do that in real time,
as we read or listen to a sentence. Of particular relevance to
this essay are the following questions: Is there a representational level of syntactic form that is distinct from the meaning
of a sentence? And if so, exactly how do we extract the implicit
structure in a spoken or written sentence as we process it? One
can ask similar questions about the process of sentence production: When planning a sentence, is there a planning stage
that encodes a specifically syntactic form? And if so, how do
these representations relate to the sound and meaning of the
intended utterance?
For purely practical reasons, there is far more research on
extracting the syntactic form during sentence comprehension (a process known as parsing ; see parsing , human) than
on planning the syntactic form of to-be-spoken sentences.
Nonetheless, research in both areas has led to substantive
advances in our understanding of the psychology of sentence
form.
SYNTAX AND SEMANTICS. A fundamental claim of a generative grammar is that syntax and semantics are clearly distinct. A fundamental claim of a cognitive grammar is that
syntax and semantics are so entwined that they cannot be easily
separated. Th is debate among linguists is mirrored by a similar
debate among researchers who study language processing. A
standard assumption underlying much psycholinguistic work
is that a relatively direct mapping exists between the levels of
knowledge posited within generative linguistic theories and
the cognitive and neural processes underlying comprehension
(Bock and Kroch 1989). Distinct language-specific processes
are thought to interpret a sentence at each level of analysis,
and distinct representations are thought to result from these
computations. But other theorists, most notably those working
in the connectionist framework, deny that this mapping exists
(Elman et al. 1996). Instead, the meaning of the sentence is
17
18
CONCLUSIONS
Collectively, the evidence reviewed in this essay indicates that
psychologically relevant representations of linguistic form
19
20
Byrd, Dani, and Elliot Saltzman. 2003. The elastic phrase: Modeling
the dynamics of boundary-adjacent lengthening. Journal of
Phonetics 31: 149 80.
Caplan, David. 1995. Issues arising in contemporary studies of disorders of syntactic processing in sentence comprehension in agrammatic patients. Brain and Language 50: 325 38.
Caramazza, Alfonzo, and Edgar Zurif. 1976. Dissociations of algorithmic and heuristic processes in language comprehension: Evidence
from aphasia. Brain and Language 3: 57282.
Chistovich, Ludmilla A 1960. Classification of rapidly repeated speech
sounds. Akustichneskii Zhurnal 6: 39298.
Chomsky, Noam. 1957. Syntactic Structures. The Hague: Mouton.
. 1965. Aspects of the Theory of Syntax . Cambridge, MA : MIT
Press.
Chomsky, Noam, and Morris Halle. 1968. The Sound Pattern of English.
New York : Harper and Row.
Delattre, Pierre, and Donald Freeman. 1968. A dialect study of
American Rs by x-ray motion picture. Linguistic s 44: 29 68.
Delattre, Pierre C., Avin M. Liberman, and Franklin S. Cooper. 1955.
Acoustic loci and transitional cues for consonants. Journal of the
Acoustical Society of America 27: 769 73.
Dell, Gary S. 1995. Speaking and misspeaking. In An Invitation to
Cognitive Science: Language. Cambridge, MA : MIT Press.
Draper, M., P. Ladefoged, and D. Whitteridge. 1959. Respiratory muscles in speech . Journal of Speech and Hearing Research 2: 16 27.
Dudley, Homer. 1940. The carrier nature of speech. Bell System
Technical Journal 14: 495 515.
Elman, Jeff rey L. 1990. Representation and structure in connectionist models. In Cognitive Models of Speech Processing, ed. G. T. M.
Altmann, 22760. Cambridge, MA : MIT Press.
Elman, Jeff rey L., Elizabeth A. Bates, A. Karmiloff-Smith, D. Parisi, and
K. Plunkett. 1996. Rethinking innateness. Cambridge, MA: MIT
Press.
Emmory, Karen. 2002. Language, Cognition, and the Brain: Insights
from Sign Language Research. Mahwah, NJ: Lawrence Erlbaum
Associates.
Ferreira, Fernanda. 2003. The misinterpretation of noncanonical sentences. Cognitive Psychology 47: 164 203.
Ferreira, Fernanda, and Charles Clifton, Jr. 1986. The independence of
syntactic processing. Journal of Memory and Language 25: 348 68.
Fillmore, Charles. 1968. The case for case. In Universals of Linguistic
Theory, ed. E. Bach, 180. New York : Holt, Rinehart, & Winston.
Fowler, Carol A. 1986. An event approach to the study of speech perception from a direct-realist perspective. Journal of Phonetics
14: 328.
. 1995. Speech production. In Speech, Language, and Communication, ed. J. L. Miller and P. D. Eimas, 2961. New York : Academic
Press.
. 1996. Listeners do hear sounds not tongues. Journal of the
Acoustical Society of America 99: 1730 41.
Fowler, Carol, Susan Napps, and Laurie Feldman. 1985. Relations
among regular and irregular morphologically related words in the
lexicon as revealed by repetition priming. Memory and Cognition
13: 24155.
Franklin, S., J. Lambon Ralph, J. Morris, and P. Bailey. 1996. A distinctive case of word meaning deafness? Cognitive Neuropsychology
13: 1139 62.
Frazier, L. 1987. Sentence processing: A tutorial review. In Attention
and Performance XII: The Psychology of Reading, ed. M. Coltheart,
330. Hillsdale, NJ: Lawrence Erlbaum Associates.
. 1995. Representation in psycholinguistics. In Speech,
Language, and Communication, ed. J. L. Miller and P. D. Eimas, 127.
New York : Academic Press.
21
22
(2)
3
THE STRUCTURE OF MEANING
(3)
James Pustejovsky
AB
A
1 INTRODUCTION
Semantics is the systematic study of meaning in language. As a
discipline, it is directed toward the determination of how humans
reason with language, and more specifically, discovering the patterns of inference we employ through linguistic expressions.The
study of semantics has diverse traditions, and the current literature is quite heterogeneous and divided on approaches to some
of the basic issues facing the field (cf. semantics). While most
things in the world have meaning to us, they do not carry meaning in the same way as linguistic expressions do. For example,
they do not have the properties of being true or false, or ambiguous or contradictory. (See Davis and Gillon [2004] for discussion
and development of this argument.) For this and other reasons,
this overview essay addresses the question of how linguistic
expressions carry meaning and what they denote in the world.
Where syntax determines the constituent structure of a sentence along with the assignment of grammatical and thematic
relations, it is the role of semantics to compute the deeper
meaning of the resulting expression. For example, the two sentences in (1) differ in their syntactic structures (through their
voice), but they mean essentially the same thing; that is, their
propositional content is identical.
(1)
23
1.2 Compositionality
Because semantics focuses on how linguistic expressions come
to have meaning, one of the most crucial concepts in the field
24
2 LEXICAL MEANING
Semantic interpretation requires access to knowledge about
words. The lexicon of a grammar must provide a systematic and
efficient way of encoding the information associated with words
in a language. lexical semantics is the study of what words
mean and how these meanings are structured. The lexicon is
not merely a collection of words with their semantic forms, but
rather a set of structured objects that participate in larger operations and compositions, both enabling syntactic environments
and acting as signatures to semantic entailments and implicatures in the context of larger discourse.
There are four basic questions in modeling the semantic
content and structure of the lexicon: 1) What semantic information goes into a lexical entry? 2) How do lexical entries
relate semantically to one another? 3) How is this information exploited compositionally by the grammar? 4) How is this
information available to semantic interpretation generally?
The lexicon and lexical semantics have traditionally been
viewed as the most passive modules of language, acting in the service of the more dynamic components of the grammar. This view
has its origins in the generative tradition (Chomsky [1955] 1975)
and has been an integral part of the notion of the lexicon ever
since. While the Aspects-model of selectional features (Chomsky
1965) restricted the relation of selection to that between lexical
items, work by McCawley (1968) and Jackendoff (1972) showed
that selectional restrictions must be available to computations
at the level of derived semantic representation rather than at
deep structure. Subsequent work by Joan Bresnan (1982), Gerald
Gazdar et al. (1985), and C. Pollard and I. Sag (1994) extend the
range of phenomena that can be handled by the projection and
exploitation of lexically derived information in the grammar.
Natural Entity
Physical
Abstract
Individuated
inanimate
animate
rock
human
Mental
Experiential
Figure 1.
Recently, with the convergence of several areas in linguistics (lexical semantics, computational lexicons, type theories),
several models for the determination of selection have emerged
that put even more compostional power in the lexicon, making
explicit reference to the paradigmatic systems that allow for
grammatical constructions to be partially determined by selection. Examples of this approach are generative lexicon theory
(Pustejovsky 1995; Bouillon and Busa 2001), and construction grammar (Goldberg, 1995; Jackendoff 1997, 2002). These
developments have helped to characterize the approaches to
lexical design in terms of a hierarchy of semantic expressiveness. There are at least three such classes of lexical description,
defi ned as follows: sense enumerative lexicons, where lexical
items have a single type and meaning, and ambiguity is treated
by multiple listings of words; polymorphic lexicons, where lexical items are active objects, contributing to the determination of
meaning in context, under well-defi ned constraints; and unrestricted sense lexicons, where the meanings of lexical items are
determined mostly by context and conventional use. It seems
clear that the most promising direction seems to be a careful
and formal elucidation of the polymorphic lexicons, and this will
form the basis of our subsequent discussion.
Lexical items can be systematically grouped according to
their syntactic and semantic behavior in the language. For this
reason, there have been two major traditions of word clustering, corresponding to this distinction. Broadly speaking, for
those concerned mainly with grammatical behavior, the most
salient aspect of a lexical item is its argument structure; for those
focusing on a words entailment properties, the most important
aspect is its semantic class. In this section, we examine these
two approaches and see how their concerns can be integrated
into a common lexical representation.
Once the basic semantic types for the lexical items in the
language have been specified, their subcategorization and
selectional information must be encoded in some form. The
argument structure for a word can be seen as the simplest specification of its semantics, indicating the number and type of
parameters associated with the lexical item as a predicate. For
example, the verb die can be represented as a predicate taking
one argument, and kill as taking two arguments, while the verb
give takes three arguments.
(5)
a. put< AGENT,THEME,LOCATION>
b. borrow<RECIPIENT,THEME,SOURCE>
a. die(x)
b. kill(x,y)
c. give(x,y,z)
25
a. x :physical[fall(x)]
b. x :animate[die(x)]
In the sentences in (8), it is clear how rocks cant die and men
can, but it is still not obvious how this judgment is computed,
given what we would assume are the types associated with
the nouns rock and man, respectively. What accomplishes this
computation is a rule of subtyping, , that allows the type associated with the noun man (i.e., human) to also be accepted as
the type animate, which is what the predicate die requires of its
argument as stated in (9b) (cf. Gunter 1992; Carpenter 1992):
(10) [human animate]: human animate
2.3 Decomposition
The second approach to the aforementioned lexical specification is to defi ne constraints internally to the predicate itself.
Traditionally, this has been known as lexical decomposition.
Since the 1960s, lexical semanticists have attempted to formally model the semantic relations between such lexical items
as the adjective dead and the verbs die and kill (cf. Lakoff [1965]
1970; McCawley 1968) in the sentences that follow.
(11) a. John killed Bill.
b. Bill died.
c. Bill is dead.
Assuming that the underlying form for a verb like kill directly
encodes the stative predicate in (11c) and the relation of causation,
generative semanticists posited representations such as (12).
26
The process consists of the building activity itself, while the State
represents the result of there being the object built. Grimshaw
(1990) adopts this theory in her work on argument structure,
where complex events such as break are given a similar representation. In such structures, the process consists of what x does to
cause the breaking, and the state is the resultant state of the broken item. The process corresponds to the outer causing event as
discussed earlier, and the state corresponds in part to the inner
change of state event. Both Pustejovsky and Grimshaw differ
from earlier authors in assuming a specific level of representation for event structure, distinct from the representation of other
lexical properties. Furthermore, they follow J. Higginbotham
(1989) in adopting an explicit reference to the event place in the
verbal semantics. Recently, Levin and Rappaport (2001, 2005)
have adopted a large component of the event structure model for
their analysis of verb meaning composition.
Although chair and rock are both physical objects, they differ
in their mode of coming into being (i.e., agentive): Chairs are
man-made; rocks develop in nature. Similarly, a concept such
as food or cookie has a physical manifestation or denotation,
but also a functional grounding pertaining to the relation of
eating. These apparently contradictory aspects of a category
are orthogonally represented by the qualia structure for that
concept, which provides a coherent structuring for different
dimensions of meaning.
27
In German, both word orders are ambiguous, since information about the grammatical case and gender of the two NPs
is neutralized.
Although there is often a correlation between the grammatical relation associated with a phrase and the meaning
assigned to it, this is not always a reliable association. Subjects
are not always doers and objects are not always undergoers
in a sentence. For example, notice how in both (22a) and (22b),
the NP the watch is playing the same role; that is, it is undergoing a change, even though it is the subject in one sentence and
the object in the other.
(22) a. The boy broke the watch.
b. The watch broke.
28
Figure 2.
language of types, we can express the rule of APPLY as a property associated with predicates (or functions), and application
as a relationship between expressions of specific types in the
language.
(24) Function Application:
If is of type a, and is of type a b, then () is of type b.
On this view, verbs such as believe introduce an intensional context for the propositional argument, instead of an
extensional one. In such a context, substitution under identity is
not permitted without possibly affecting the truth value (truth
conditional semantics). Th is is an important contribution
to the theory of meaning, in that a property of opacity is associated with specific types within a compositional framework.
One potential challenge to a theory of function application
is the problem of ambiguity in language. Syntactic ambiguities
arise because of the ways in which phrases are bracketed in a
sentence, while lexical ambiguity arises when a word has multiple interpretations in a given context. For example, in the following sentence, the verb treat can mean one of two things:
(26) The doctor treated the patient well.
Either 1) the patient is undergoing medical care, or 2) the doctor was kind to the patient. More often than not, however, the
context of a sentence will eliminate such ambiguities, as shown
in (27).
(27) a. The doctor treated the patient with antibiotics. (Sense 1)
b. The doctor treated the patient with care. (Sense 2)
In this case, the interpretation is constructed from the appropriate meaning of the verb and how it combines with its
arguments.
Now consider the sentences in (29), where there is a combination of a some-NP and an every-NP.
(29) a. Every student saw a movie.
b. Every cookie was eaten by a student.
Unfortunately, however, most instances of adjectival modification do not work so straightforwardly, as illustrated in (35).
Adjectives such as good, dangerous, and fast modify polysemously in the following sentences.
(35) a. John is a a good teacher.
b. A good meal is what we need now.
c. Mary took a good umbrella with her into the rain.
That is, the road is dangerous in (36a) when one drives on it,
and the knife is dangerous in (36b) when one cuts with it.
Finally, the adjective fast in the following sentences acts as
though it is an adverb, modifying an activity implicit in the
noun, that is, programming in (37a) and driving in (37b).
(37) a. Mary is the fastest programmer we have on staff.
b. The turnpike is a faster road than Main Street.
29
Similarly, as presented in previous sections, the classification of verbs appears to reflect their underlying relational
structure in fairly obvious ways.
(39) a. Mary arrived.
b. John greeted Mary.
c. Mary gave a book to John.
30
John swept.
John swept the floor.
John swept the dirt.
John swept the dirt off the sidewalk.
John swept the floor clean.
John swept the dirt into a pile.
The semantics of such a verb should determine what its arguments are, and how the different possible syntactic realizations
relate to each other semantically. These cases pose an interest-
3.5 Presupposition
In computing the meaning of a sentence, we have focused
on that semantic content that is asserted by the proposition.
Th is is in contrast to what is presupposed. A presupposition
is that propositional meaning that must be true for the sentence containing it to have a proper semantic value (Stalnaker
1970; Karttunen 1974; Potts 2005). (Stalnaker makes the distinction between what a speaker says and what a speaker
presupposes.)
Such knowledge can be associated with a word, a grammatical feature, or a syntactic construction (so-called presupposition triggers). For example, in (45) and (46), the complement
proposition to each verb is assumed to be true, regardless of the
polarity assigned to the matrix predicate.
(45) a. Mary realized that she was lost.
b. Mary didnt realize that she was lost.
(46) a. John knows that Mary is sick.
b. John doesnt know that Mary is sick.
Conversational presuppositions, on the other hand, are implicated propositions by virtue of a context and discourse situation.
The response in (50b) conversationally implicates that I am not
hungry (Recanati 2002); conversational implicature).
(50) a. Are you hungry?
b. Ive had a very large breakfast.
3.6 Noncompositionality
While semantic theory seems to conform to the principles of
compositionality in most cases, there are many constructions
that do not fit into the conventional function application paradigm. A phrase is noncompositional if its meaning cannot
4 DISCOURSE STRUCTURE
Thus far we have been concentrating on the meaning of single
sentences. But no sentence is really ever uttered outside of a
context. Language is used as a means of communication and is
as much a way of acting as a means of representing (Austin 1975;
Searle 1969). In this section, we briefly survey the major areas
of research in discourse semantics. We begin by examining the semantic models that have emerged to account for
dynamic phenomena in discourse, such as intersentential
31
Notice that the anaphoric link between the quantifier and the
pronoun in (57a) is acceptable, while such a binding is not possible within a larger discourse setting, as in (58) and (59).
(58) a. Every actress came in.
b. *She said she was happy.
(59) a. Every actress came in.
b. *She said hello.
So, in a larger unit of semantic analysis, a bound variable interpretation of the pronoun does not seem permitted.
Now notice that indefi nites do in fact allow binding across
the level of the sentence.
(60) a. An actress came in.
b. She said hello.
32
is added to the listeners interpretation state so that the listener can use the quantifier to help understand future utterances. In this way, the meaning of a sentence is interpreted
dynamically.
The dynamics of discourse, of course, involve more than the
binding of anaphors to antecedents across adjacent sentences.
Every utterance is made in the context of a common ground
of shared knowledge (presuppositions), with a communicative intent, and in a particular time and place (cf. discourse
analysis, communicative intention). Just as sentences
have internal structure, with both syntactic and semantic
dependencies, discourse can also be viewed as a sequence of
structured segments, with named dependencies between them.
For example, the sentences in (62) form a discourse structured
by a relation of narration, implying temporal sequence (Dowty,
1986).
(62) a. John entered the room.
b. He sat down.
In (63), on the other hand, the two sentences are related by the
dependency of explanation, where (63b) temporally precedes
and explains (63a).
(63) a. Max fell.
b. John pushed him.
Theories of discourse relations, such as rhetorical structure theory (Mann and Thompson 1986), segmented discourse representation theory (SDRT) (Asher and Lascarides 3),
and that of Hobbs (1985) attempt to model the rhetorical functions of the utterances in the discourse (hence, they are more
expressive of discourse structure and speaker intent than discourse representation theory [DRT], which does not model
such parameters). For the simple discourses above, SDRT, for
example extends the approach from dynamic semantics with
rhetorical relations and their semantic values, while providing a more complex process of discourse updates. Rhetorical
relations, as used in SDRT, carry specific types of illocutionary force (cf. Austin 1975; Searle 1969, 9), namely, explanation,
elaboration, giving backgrounds, and describing results.
5 CONCLUSION
In this essay, I have attempted to outline the basic components
for a theory of linguistic meaning. Many areas of semantics were
not touched on in this overview, such as issues relating to the
philosophy of language and mind and the psychological consequences of various semantic positions. Many of the accompanying entries herein, however, address these issues directly.
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Apresjan, J. D. 1973. Synonymy and synonyms. In Trends in Soviet
Theoretical Linguistics, ed. F. Kiefer, 17399. Dordrecht, the
Netherlands: Reidel.
Asher, Nicholas, and Alex Lascarides. 2003. Logics of Conversation.
Cambridge: Cambridge University Press.
Austin, J. L. 1975. How to Do Things with Words. Cambridge: Harvard
University Press.
33
34
diglossia, culture and language, digital media, literacy). This essay is focused on the issue of the media of com-
4
SOCIAL PRACTICES OF SPEECH AND WRITING
Florian Coulmas
INTRODUCTION
Language is constitutive for human society. As a social fact
it cannot be thought of in the abstract, for the medium of
communication is what allows it to serve social functions. The
nature of the social relationship that exists by virtue of language
partially depends on the externalization of language, that is, on
how it is transmitted from one actor to another as speech, writing, sign, or Braille. The anatomy of speech organs (cf. Liberman
and Blumstein 1988) provides the biological foundation of
human society in the most general sense, which is why oral
speech is considered fundamental for socialization both in the
phylogenetic and ontogenetic sense. But unless we study human
society like that of other primates from the point of view of physical anthropology, other forms of language externalization must
also be taken into account as communication potential from
the beginning. There are two reasons for this. One is that the
invention of writing (sign language, Braille) cannot be undone.
The other, which follows therefrom, is that writing has brought
about basic changes in the nature of human communication. It
brought in its wake a literate mindset that cannot be reversed.
Research about language in literate societies is carried out by
researchers who, growing up, were socialized into a literate
world organized by and large on the basis of literate principles.
It is not fortuitous, therefore, that social practices of speech and
writing are dealt with here under one heading.
The scientific enterprise in general, linguistics in particular, is
a social practice involving speech and writing. Even the investigation of unwritten languages happens against the background
of literate society and by means of the tools developed for what
Goody (1977, 151) felicitously called the technology of the intellect. For the language sciences, it is important to keep in mind
that it is not just the technicians who use a tool to do what they
need to do and want to do, but that the tool restricts what can
be done. This holds true for the hardware, that is, the writing
implements, as well as for the software, the code or the writing
systems.
The social aspects of speech and writing encompass a wide
range of topics many of which are dealt with in other entries of
this encyclopedia (sociolinguistics, discourse analysis,
35
36
digitalized contents would require some 750 gigabyte (GB) storage space, which easily fits on an external hard disk the size of
a small book. As compared to print, digital information storage
thus reduces the necessary physical space by a factor of 50,000.
Again, this is a coarse measure only. There is considerable
variation in the bytes per book page, both in actual fact and
in model calculations, but the correlation between information amount and storage space of print media and digital media
transpires from it.
In sum, as clay was followed by paper and paper by digital storage media, information density per square centimeter
increased exponentially, while the weight and size of written
records decreased. It became, accordingly, ever easier to store
and to transmit written text with many consequences for reading and writing behavior, for reproducing text, and for the role
texts play in everyday life. What the history of writing shows is
that new technologies do not always replace old ones. Rather,
the new supplements the old and often transforms its use. For
instance, parchment was marginalized by paper but for centuries never completely driven out; print never replaced handwriting and has not become obsolete by texts typed on a cell phone
keypad. Advances in writing technology have greatly expanded
the repertoire of tools that humanity has acquired for handling
information in the form of written language. The fact that old
technologies continue to be used side by side with new ones
testifies not just to inertia and path dependency but also to the
different properties of the materials used. For centuries after
the introduction of paper, it was considered too feeble a material to hold contracts and other important documents, which
were preferably executed on parchment. Similarly, although it
is technically possible to keep birth registers as electronic files
only, birth certificates on paper continue to be issued. One of the
reasons is that digital information storage is subject to physical
decay and technical obsolence not less but much more than predecessor media.
Writing has made record keeping for posterity and accumulation of knowledge possible. However, with the introduction of
every new writing material, the storage problem that it seemed to
solve became more acute. An archive of records on paper takes
up much less space than one for clay tablets, but it is beset by
dangers that pose no threat to baked clay: dust, humidity, insects,
fire, and water. Theft, too, is a greater threat to libraries than to
clay tablet archives and a greater threat to computers than to
libraries. Keeping books in a usable physical state requires more
work than keeping clay tablets. The same kind of relationship
holds between libraries and digital data archives. Much more
can be stored, but preservation for future use becomes ever more
difficult as time intervals of technical innovation shrink. Only a
few specialists are able to handle data stored with software that
ceased to be produced 20 years ago, whereas a book would last
for centuries. The problem of preserving and organizing the everswelling flood of information remains unsolved, and at the same
time many traditional libraries and collections of documents fall
into decay. Technology has hugely expanded human memory,
but it has not yet eliminated the risk that many parts of the heritage committed to writing will disappear forever.
To guard against collective memory loss, the United Nations
Educational, Scientific and Social Orgnization (UNESCO) has
37
Government
Writing was used early on to extend the reach of authority and to
mediate the relationship between ruler and ruled. Monuments
that embody power, such as the Rosetta Stone inscribed with a
decree to commemorate the reception of rulership by Pharaoh
Ptolemy V on March 27, 197 b.c.e. (Parkinson 1999, 29), as well
as stelae appealing with regulations to the literate public, were
erected throughout the Ancient Near East. Their inscriptions
were drafted by scribes who created the first bureaucratic states.
The Egyptian vizier was responsible for the collection of taxes,
the maintenance of archives, and the appointment of officials.
The skills of the scribe afforded privilege in Egyptian society. As
one of them put it, the scribe is released from manual tasks; it
is he who commands (Goody and Watt 1968, 37). A thousand
years later, Mencius (372289 b.c.e.) made the same point in
China: Some labour with their hearts and minds; some labour
with their strength. Those who labour with their hearts and
minds govern others. Those who labour with their strength are
governed by others (Book of Mencius, quoted from Lloyd and
Sivin 2002, 16). These observations bring to light the connection
between literacy and social hierarchy that persists to this day.
Wherever societies obtained the art of writing, the literati
were close to the powers that be, but the Chinese institutionalized literacy like no other culture. In Confuciuss day, literacy was
already the preeminent mark of the gentleman, and as of the second century b.c.e., the study of the Confucian classics gradually
became institutionalized as the key to attaining public office and
influence. The civil service examination system was employed
with few adjustments from the Han Dynasty (206 b.c.e.220 c.e.)
until the closing stages of the Qing Dynasty at the beginning of
the twentiethth century. It was based entirely on the study of
texts. To prepare for the highest degree, students spent up to 20
38
Cult
In nonliterate societies religious, order and social order are
merged without clear division. Writing tends to introduce a
fracture between spheres, although the differentiation of the
sacred and the profane may take a long time to complete. The
Ten Commandments, written by God and given to Moses, were
a code of conduct regulating the relations between God and the
people, as well as the people among themselves. The spheres of
spiritual and worldly power were only beginning to be separated
in antiquity, a process to which writing and the ensuing interpretation of texts contributed a great deal. Of Moses legendary
Ten Commandments stone tablets no archaeological evidence
remains, but the Aoka edicts are a tangible example of the
closeness of cult and social order. Engraved in rocks and stone
pillars that have been discovered in dozens of sites throughout
northern India, Nepal, Pakistan, and Afghanistan, these edicts
served to inform the subjects of King Aokas (304232 b.c.e.)
reforms and make them lead a moral life according to the truth
of Buddhism. Moral precepts hewn in stone on public display
once again raise the question of the degree of literacy. It must
have been high enough to disseminate the message throughout the vast realm. In the rich spiritual world of the Indian subcontinent, much emphasis was always given to the importance
of oral tradition. Competing with other religious movements,
Buddhism established itself by rigidly regulating monastic life
and assembling a canon of scriptures. At the time of Aoka,
Buddhism was an institutionalized religion that spread from
northern India to other parts of the subcontinent, first and subsequently to South and Southeast Asia, as well as Central Asia
and China.
The history of Buddhism and its emergence as a major world
religion is a history of translation. Translation means writing, turning one text into another and all it involves: exegesis,
doctrine, scholasticism, and schism. As Ivan Illich and Barry
Sanders (1988: 52) have argued, there is no translation in orality, but only the collaborative endeavor to understand ones
partner in discourse. Writing eternalizes spiritual enlightenment (e.g., the word of God), which must be preserved with
the greatest care and does not allow alteration at will. Of all
religions, Buddhism has produced by far the largest corpus
of sacred texts. Significantly, different canons resulted from
translation, the Pali canon, the Tibetan canon, and the Chinese
canon. With translation came schism, and with that the delimitation of religious sects and districts. In the evolution of other
book religions, translation had similar consequences. These
other religions share with Buddhism the vital importance they
attach to scriptures. The holy book is what distinguishes world
religions from other cults. Their legitimacy is derived from the
revelation embodied in the scriptures, which are considered the
true source of enlightenment. The major world religions vary in
the role they assign sacred texts, in how they make use of them
for purposes of propaganda and the regulation of life, but the
reverence accorded to writing is a feature they share, as is the
question of translation.
Translation is a problem for two reasons that can be labeled
authenticity and correspondence. First, the idea of authenticity of the word of God does not agree with its transposition into
another form. Some book religions are very strict in this regard.
According to believers, God himself chose the Arabic language
for his final testament, the Quran, and created every one of the
letters of the Arabic alphabet. Consequently, the only true version of the holy book of Islam is the Quran in Classical Arabic.
39
40
languages or varieties of the same stock. In multilingual environments like medieval Europe, where Latin was the only written language, or in present-day India, cultivated languages are
often reserved for writing and formal communication, while vernacular varieties are used in informal settings. A similar division
is found between linguistic varieties of the same stock, where
one is defined by reference to a corpus of classical texts, such as
Classical Arabic, whereas the others fluctuate without artificial
restriction.
Writing introduces an element of art and artificiality into the
history of language. Every language is the collective product of
its speakers, but a written language is more clearly an artefact
than a vernacular, and the script that it uses more clearly yet.
Historically, the diffusion of scripts coincided in large measure
with that of religions, a connection that is still visible today.
Chinese characters arrived in Japan together with Buddhism.
The spread of Roman traces the expansion of both the Roman
Empire and the Catholic Church, while Orthodox Christianity
uses Cyrillic. Armenian Christians have their own alphabet
designed in the fifth century by St. Mesrob. Estrangela is the script
of the Syrian Church. The Hebrew square script is the script of
Judaism, the Arabic alphabet that of Islam. Many other examples
could be added; clerks were churchmen (Coulmas 1996, 435 f.).
The historical interconnectedness of writing and religion is one
of the reasons that scripts tend to function as symbols of identity, but ethnic, national, and political identity are also readily
expressed by means of a distinct script or even slightly different
orthographic conventions. As Peter Unseth (2005) has pointed
out, there are clear sociolinguistic parallels between choosing scripts and languages. Because of the visibility and the artificial nature of writing, however, the choice of scripts is generally a
more deliberate departure from tradition in that it involves conscious planning.
Schooling
Language is a natural faculty, writing an artifact. That is the reason why children acquire language, but not writing, without
guidance. The difficult art of writing requires skills that must be
taught, memorized, and laboriously practiced. The place to do it
is school. For writing to be useful to the community, conventions
have to be established, individual variation curtailed, norms set.
Collective instruction following a curriculum is a more efficient
way to achieve this than is private tutoring. Already in antiquity,
school became, and still is, the institution that most explicitly
exercises authority over the written language by controlling its
transmission from one generation to the next.
With schooling came the regimentation and the decontextualization of language. Because in writing the continuous flow
of speech has to be broken down into discrete units, analytic
reflection about the units of language was fostered. As writing,
language became an object of investigation and normalization.
Both grammar and lexicon are products of writing. This is not to
deny the grammaticality of oral speech or that oral people have
a mental lexicon. It just means that the notions of grammar and
lexicon as we know them are entirely dependent upon writing.
At school, units of writing had to be practiced with the stylus, the
brush, or the pen, mechanically through repetition without any
41
Economic Organization
The critical importance of writing for economic processes
in early civilizations is best documented for Mesopotamia.
It is widely agreed now that in Sumer, number concepts and
numerical representation stood at the beginning of writing
that evolved into cuneiform (Schmandt-Besserat 1992). The
overwhelming majority of archaeological finds from Ur, Uruk,
Nineveh, and Babylon are records of economic transactions
kept in clay tablet archives by accountants of the palace administration (Nissen, Damerow, and Englund 1990). The Sumerians
and their Akkadian successors were the first to develop bookkeeping into a sophisticated technique of balancing income and
expenditure. Hammurabi, King of Babylon (17281686 b.c.e.),
created the first commercial code, 282 articles written on a large
stone monument, which was erected in a public place for all to
observe. Every business transaction was to be in writing and
signed, usually with a seal, by the contracting parties. At the
time, the public-sector economy was far too highly developed
and too complex to function without writing. Tribute quota lists
had to be kept, rations for laborers involved in public works calculated, inventories recorded. Deeds were issued in duplicate
and receipts stored for future reference. Large-scale trading,
often involving credit and futures, had to be regulated and
overseen by the bureaucracy, consisting of a huge scribal class
charged with creating and servicing these documents.
Ancient Mesopotamia is the paradigm case of the close interconnectedness of economy and the institutionalization of writing, but if economic behavior is understood in the wide sense of
human adaptations to the needs and aspirations of society at a
given moment in history, this interconnectedness can be identified in every literate society. Complex socioeconomic systems of
managed production, distribution and trade, taxation, and credit
did not evolve in cultures that had no writing. Yet the nature of
42
pertinent than ever (Levine 1994) and can explain processes such
as the accelerated spread of English. The globalization of markets and the information technology revolution offer ever more
people the opportunity to enhance their human capital and, at
the same time, compel them to do so. However, the commodification of written language (Heller 2003) and the new forms and
uses it takes on in the new media have consequences, some of
which become apparent only as technology spreads.
Conceptualizing written language as transaction cost brings
market forces into view. Reducing transaction costs is considered a quasi-natural precondition of economic growth, which
partly explains the spread of trade languages. Once in place and
controlled by relevant agents, their use is less costly than translation. In todays world, this principle seems strongly to favor the
further spread of English. However, the effects of technological
innovation are, as always, hard to foresee and even to assess
when it unfolds in front of our eyes. When the Internet ceased
to be a communication domain reserved to the U.S. military,
partly due to the available software at the time, it seemed to be
poised to become an English-only medium. But it turned out
that as the technology caught on, the share of English in cyberspace communication rapidly declined. The new information
technology made it much easier for speech communities big and
small around the world to communicate in writing, a possibility
eagerly exploited wherever the hardware became available. For
some communities this meant using their language in writing for
the first time. In others it led to the suspension of norms, blurring, in many ways yet to be explored, the traditional distinctions
between spoken and written language.
David Crystal (2001) suggests that Netspeak or computermediated language is a new medium, neither spoken language
nor written language. The implications of Internet communication for literacy in the electronic age are only beginning to be
explored (Richards 2000). New multilayered literacies are evolving, responding as they do to the complementary and sometimes
conflicting demands of economic rationality, social reproduction through education, ideology, and the technical properties of
the medium. These developments open up a huge new area of
research into how, since the invention of writing, the range of linguistic communication options has been constantly expanding.
CONCLUSION
Writing is a technology that interacts with social practices in
complex ways, exercising a profound influence on the way we
think, communicate, and conceptualize language. Since it is an
artifact, it can be adjusted deliberately to the evolving needs
of society, but it also follows its own inherent rules that derive
from the properties of the medium. This essay has analyzed the
tension between the properties of the medium and the designs
of its users from two points of view, technological advance and
institutionalization. Harnessed by institutions, the technology
of writing is made serviceable to the intellectual advance of
society and modified in the process, sometimes gradually and
sometimes in revolutionary steps. Three consequences of writing remain constant: segmentation, linearization, and accumulation. The linear and discrete-segmental structure that all
writing systems both derive from and superimpose on language
43
44
45
5
EXPLAINING LANGUAGE: NEUROSCIENCE,
GENETICS, AND EVOLUTION
Lyle Jenkins
46
Explaining Language
based on the idea that language evolved from gestures, an idea
that is represented by a number of approaches to the evolution
of language. After that, we present an example of the comparative approach to evolution in biolinguistics and discuss language
and the neurosciences, focusing on leftright asymmetries of the
language areas to illustrate this research. Finally, we discuss the
genetics of language, using the studies of the FOXP2 gene system
to show how studies of language phenotype, neurobiology, and
molecular biology, as well as comparative and evolutionary studies with other animal systems, are being carried out.
Work on the principles of efficient computation governing the
application of the operation of Merge seem to suggest an asymmetry, namely, that the computational principles are optimized
for the semantic interface, not the sensorimotor interface. This is
because conditions of computational efficiency and ease of communication conflict, as is familiar from the theory of language
parsing. This in turn has led Chomsky to suggest that externalization of language, and hence communication, was a secondary
adaptation of language. This would mean that language arose
primarily as an internal language of thought. Supporting
this idea is the existence of sign languages, which develop in
a different modality but in other respects are very similar to spoken language (Kegl 2004; Pettito 2005).
These design features of language have led Chomsky to propose the following scenario for the evolution of language. Several
decades ago Chomsky suggested, on the basis of results from nonhuman primate studies, that the higher apes might well have a
conceptual system with a system of object reference and notions
such as agent, goal, instrument, and so on. What they lack, however, is the central design feature of human language, namely,
the capacity to deal with discrete infinities through recursive
rules (Chomsky 2004, 47). He proposed that when you link that
capacity to the conceptual system of the other primates, you get
human language, which provides the capacity for thought, planning, evaluation and so on, over an unbounded range, and then
you have a totally new organism (Chomsky 2004, 48).
Thus, let us assume that some individual in the lineage to modern humans underwent a genetic change such that some neural
circuit(s) were reorganized to support the capacity for recursion.
This in turn provided the capacity for thought and so on over an
unbounded range. This in itself provided that individual and any
offspring with a selective advantage that then spread through the
group. Thus, the earliest stage of language would have been just
that: a language of thought, used internally (Chomsky 2005a, 6).
At some later stage, there was an advantage to externalization, so
that the capacity would be linked as a secondary process to the
sensorimotor system for externalization and interaction, including communication (7).
The evolutionary scenario just outlined is derived from design
principles suggested from work on human languages. Many other
evolutionary scenarios have been proposed that assume that
communication or other social factors played a more primary
role. These include accounts involving manual and facial gestures (Corballis 2002), protolanguage (Bickerton 1996), grooming (Dunbar 1998), and action and motor control (Rizzolatti
and Arbib 1998). However, the two kinds of accounts are not
incompatible and may represent different aspects or stages in
the evolution of language. We cannot review all of these research
47
48
Explaining Language
postulated that this region held the key to what he called the sensible Sprachcentrum (sensible/cognizant language center).
Claudio Cantalupo and William D. Hopkins (2001) report
finding an anatomical asymmetry in Brocas area in three great
ape species. They obtained magnetic resonance images (MRI)
(neuroimaging) from 20 chimpanzees (P. troglodytes), 5
bonobos (P. paniscus), and 2 gorillas (G. gorilla). In humans,
Brodmanns area 44 corresponds to part of Brocas area within
the inferior frontal gyrus (IFG). This area is larger in the left
hemisphere than in the right. Furthermore, it is known that
Brocas area is vital for speech production (with the qualifications discussed earlier).
Although the great apes were known to have a homolog of
area 44 on the basis of cytoarchitectonic and electrical stimulation studies, no leftright anatomical asymmetry had been
shown. Cantalupo and Hopkins found a pattern of morphological asymmetry similar to that found in the homologous area in
humans. This would place the origin of the asymmetry for the
anatomical substrate for speech production to at least five million years ago.
Since the great apes exhibit only primitive vocalizations, these
authors speculate that this area might have subserved a gestural
system (see earlier discussion). They note the presence in monkeys of mirror neurons in area 44 that subserve the imitation of
hand grasping and manipulation (Rizzolatti and Arbib 1998).
They also observe that captive great apes have a greater righthand bias when gesturing is accompanied by vocalization. Hence
in the great apes, the asymmetry may have subserved the production of gestures accompanied by vocalizations, whereas, for
humans, this ability was selected for the development of speech
systems, accompanied by the expansion of Brodmanns area 45
(which, along with Brodmanns area 44, makes up Brocas area)
(Cantalupo and Hopkins 2001, 505).
However, additional studies of Brodmanns area 44 in
African great apes (P. troglodytes and G. gorilla) call into question whether the techniques used in Cantalupo and Hopkins
study were sufficient to demonstrate the leftright asymmetry.
Chet C. Sherwood and colleagues (2003) found considerable
variation in the distribution of the inferior frontal sulci among
great ape brains. They also constructed cytoarchitectural maps of
Brodmanns area 44, examining myeloarchitecture and immunohistochemical staining patterns. When they studied the IFG
of great ape brains, they found a poor correspondence between
the borders observed in the cytoarchitectural maps and the borders in the surface anatomy (e.g., sulcal landmarks). There were
similar findings for human brains in an earlier study (Amunts et
al. 1999). Sherwood and colleagues conclude that in the study
by Cantalupo and Hopkins, it is unlikely that the sulci used to
define the pars opercularis coincided with the borders of cytoarchitectural area 44 (2003, 284). In general then, macrostructure is a poor predictor of microstructure.
Sherwood and colleagues also point out that even if humanlike asymmetries of the inferior frontal gyrus and of the planum
temporale are confirmed, these gross asymmetries will not suffice to explain the unique neural wiring that supports human
language (284). To that end, comparative studies of microstructure in humans and great apes are needed. For example, a computerized imaging program was used to examine minicolumns
49
50
Explaining Language
the zebra finch, the Scharff laboratory is using RNA to downregulate FoxP2 in Area A, a striatal region important for song learning
(Scharff and Nottebohm 1991). It is known that young male zebra
finches express more FoxP2 bilaterally in Area X when learning
to sing (Haesler et al. 2004). They will then be able to determine
whether the bird is still able to sing normal song as well as copy
the song of an adult male tutor.
Weiguo Shu and colleagues (2005) found that disrupting the
Foxp2 gene in mice resulted in impairing their ultrasonic vocalization. In addition to their more familiar sonic vocalizations,
mice also make ultrasonic sounds, for example, when they are
separated from their mothers.
In order to study the effect of disrupting the Foxp2 gene on
vocalization, these researchers constructed two versions of
knockout mice. One version had two copies of the defective
Foxp2 gene (the homozygous mice) and the other version had
one copy of the defective Foxp2 gene, as well as one gene that
functioned normally (the heterozygous mice). The homozygous
mice (double knockout) suffered severe motor impairment,
premature death, and an absence of ultrasonic vocalizations
that are normally produced when they are separated from their
mother. On the other hand, the heterozygous mice, with a single working copy of the gene, exhibited modest developmental
delay and produced fewer ultrasonic vocalizations than normal.
In addition, it was found that the Purkinje cells in the cerebellum, responsible for fine motor control, were abnormal. It is concluded that the findings support a role for Foxp2 in cerebellar
development and in a developmental process that subsumes
social communication functions (Shu et al. 2001, 9643).
Timothy E. Holy and Zhongsheng Guo (2005) studied the
ultrasonic vocalizations that male mice emit, when they encounter female mice or their pheromones. They discovered that the
vocalizations, which have frequencies ranging from 30100 kHz,
have some of the characteristics of song, for example, birdsong.
In particular, they were able to classify different syllable types
and found a temporal sequencing structure in the vocalizations.
In addition, individual males, though genetically identical, produced songs with characteristic syllabic and temporal structure.
These traits reliably distinguish them from other males. Holy
notes that these discoveries increase the attractiveness of mice
as model systems for study of vocalizations (White et al. 2006,
10378).
We have focused on the FOXP2 gene here because there is no
other gene affecting speech and language about which so much
information is available that bears on questions of neurology and
evolution of language. However, the search is underway for other
additional genes.
For example, genetics researchers have also discovered a
DNA duplication in a nine-year-old boy with expressive language
delay (Patient 1) (Fisher 2005; Somerville et al. 2005). Although
his comprehension of language was at the level of a seven-yearold child, his expression of language was comparable to that of
only a two-and-a-half-year-old.
The region of DNA duplication in Patient 1 was found to be
on chromosome 7, and interestingly, was found to be identical to
the region that is deleted in Williams-Beuren syndrome (WBS).
Patients with WBS have relatively good expressive language but
are impaired in the area of spatial construction. Lucy Osborne,
51
52
Explaining Language
levels in adult canaries varied with the season and correlated
with changes in birdsong. These facts suggest regulatory control
in development and in the adult brain.
(1980, 216). Not only has that proven to be the case, but with the
explosion of knowledge in many areas, including (comparative)
linguistics, comparative neuroanatomy, evolutionary development, comparative genomics, to take just a few examples,
biolinguistics promises to be a fascinating field for decades to
come.
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Amunts, K., A. Schleicher, et al. 1999. Brocas region revisited: Cytoarchitecture and intersubject variability. Journal of
Comparative Neurology 412: 31941.
Bickerton, Derek. 1996. Language and Human Behavior. Seattle: University
of Washington Press.
Boeckx, Cedric. 2006. Linguistic Minimalism. Oxford: Oxford University
Press.
Bookheimer, Susan. 2002. Functional MRI of language: New approaches
to understanding the cortical organization of semantic processing.
Annual Review of Neuroscience 25 (March): 15188.
Braun, Allen. 2003. New findings on cortical anatomy and implications
for investigating the evolution of language. The Anatomical Record
Part A: Discoveries in Molecular, Cellular, and Evolutionary Biology
271A (March): 27375.
Buxhoeveden, D. P., A. E. Switala, et al. 2001. Lateralization of minicolumns in human planum temporale is absent in nonhuman primate
cortex. Brain, Behavior and Evolution 57 (June): 34958.
Cangelosi, Angelo, and Domenico Parisi, eds. 2001. Simulating the
Evolution of Languge. New York: Springer.
Cantalupo, Claudio, and William D. Hopkins. 2001. Asymmetric Brocas
area in great apes. Nature 414 (November): 505.
Carroll, Sean B. 2005. Evolution at two levels: On genes and form. PLoS
Biology 3 (July): 115966. Available online at: www.plosbiology.org.
Chomsky, Noam. 1959. A review of B. F. Skinners Verbal Behavior.
Language 35.1: 2658.
. 1980. On the biological basis of language capacities. In Rules
and Representations, 185216. New York: Columbia University Press.
. 1995. The Minimalist Program. Cambridge, MA: MIT Press.
. 2004. The Generative Enterprise Revisited: Discussions with Riny
Huybregts, Henk Van Riemsdijk, Naoki Fukui and Mihoko Zushi.
Berlin: Mouton de Gruyter.
. 2005a. Some simple evo-devo theses: How true might they be for
language? Paper presented at the Morris Symposium, SUNY at Stony
Brook.
. 2005b. Three factors in language design. Linguistic Inquiry
36.1: 122.
. 2006. Language and Mind. Cambridge: Cambridge University
Press.
Chomsky, N., and H. Lasnik. 1995. The theory of principles and parameters. In The Minimalist Program, N. Chomsky, 13127. Cambridge,
MA: MIT Press.
Christiansen, Morten H., and Simon Kirby, eds. 2003. Language Evolution.
New York: Oxford University Press.
Corballis, Michael C. 2002. From Hand to Mouth: The Origins of Language.
Princeton, NJ: Princeton University Press.
Cunningham, Daniel J. 1892. Contribution to the Surface Anatomy of the
Cerebral Hemispheres [Cunningham Memoirs]. Dublin: Royal Irish
Academy of Science.
Deacon, T. W. 1989. The neural circuitry underlying primate calls and
human language. Human Evolution 4 (October): 367401.
Dunbar, Robin. 1998. Grooming, Gossip and the Evolution of Language.
Cambridge: Harvard University Press.
Evans, Patrick D., Nitzan Mekel-Bobrov, et al. 2006. Evidence that the
adaptive allele of the brain size gene microcephalin introgressed
53
54
Kohler, Evelyne, Christian Keysers, et al. 2002. Hearing sounds, understanding actions: Action representation in mirror neurons. Science
297: 8468.
Lai, Cecilia S. L., Simon E. Fisher, et al. 2001. A forkhead-domain gene
is mutated in a severe speech and language disorder. Nature 413
(October): 51923.
Liberman, Alvin M., and Ignatius G. Mattingly. 1985. The motor theory
of speech perception revised. Cognition 21: 136.
Liberman, Mark. 2004. Hi lo hi lo, Its off to formal language theory we
go. Language Log (January 17). Available online at: http://itre.cis.
upenn.edu/~myl/languagelog/archives/000355.html.
Lieberman, Philip. 2006. Toward an Evolutionary Biology of Language.
Cambridge: Belknap Press of Harvard University Press.
Lu, Min Min, Shanru Li, et al. 2002. Foxp4: A novel member of the Foxp
subfamily of winged-helix genes co-expressed with Foxp1 and Foxp2
in pulmonary and gut tissues. Mechanisms of Development 119,
Supplement 1 (December): S197S202.
Lyon, Caroline, Chrystopher L. Nehaniv, et al., eds. 2006. Emergence of
Communication and Language. New York: Springer.
Maynard Smith, John, and Ers Szathmry. 1998. The Major Transitions
in Evolution. New York: Oxford University Press.
McMaster, Geoff. 2005. Researchers discover cause of speech defect.
University of Alberta ExpressNews (October 25). Available online
at: http://www.expressnews.ualberta.ca.
Nishitani, N., and R. Hari. 2000. Temporal dynamics of cortical representation for action. Proceedings of the National Academy of Sciences
USA 97: 91318.
Niyogi, Partha. 2004. Phase transitions in language evolution. In
Variation and Universals in Biolinguistics, ed. L. Jenkins, 5774.
Amsterdam: Elsevier.
. 2006. The Computational Nature of Language Learning and
Evolution. Cambridge, MA: MIT Press.
Noonan, James P., Graham Coop, et al. 2006. Sequencing and analysis of
Neanderthal genomic DNA. Science 314 (November): 111318.
Nowak, Martin A. 2006. Evolutionary Dynamics: Exploring the Equations
of Life. Cambridge: Harvard University Press.
Nowak, Martin A., and Natalia L. Komarova. 2001. Towards an evolutionary theory of language. Trends in Cognitive Sciences 5 (July):
28895.
Pennisi, Elizabeth. 2009. Neanderthal Genomics: Tales of a Prehistoric
Human Genome. Science 323. 5916: 86671.
Pettito, Laura-Ann. 2005. How the brain begets language. In
Cambridge Companion to Chomsky, ed. J. McGilvray, 84101.
Cambridge: Cambridge University Press.
Petrides, M., and D. N. Pandya. 1994. Comparative architectonic analysis of the human and the macaque frontal cortex. In Handbook of
Neuropsychology, ed. F. Boller and J. Grafman, 1758. Amsterdam:
Elsevier.
Poeppel, David, and Gregory Hickok. 2004. Towards a new functional
anatomy of language. Cognition 92 (May/June): 112.
Rizzolatti, Giacomo, and Michael A. Arbib. 1998. Language within our
grasp. Trends in Neurosciences 21 (May): 18894.
Roll, Patrice, Gabrielle Rudolf, et al. 2006. Srpx2 mutations in disorders of language cortex and cognition. Human Molecular Genetics 15
(April): 11951207.
Scharff, Constance, and Sebastian Haesler. 2005. An evolutionary
perspective on FoxP2: Strictly for the birds? Current Opinion in
Neurobiology 15 (December): 694703.
Scharff, C., and F. Nottebohm. 1991. A comparative study of the behavioral deficits following lesions of various parts of the zebra finch song
system: Implications for vocal learning. Journal of Neuroscience 11
(September): 2896913.
Explaining Language
Shapiro, K. A., L. R. Moo, et al. 2006. Cortical signatures of noun and verb
production. Proceedings of the National Academy of Sciences USA 103
(January): 164449.
Sherwood, Chet C., Douglas C. Broadfield, et al. 2003. Variability
of Brocas area homologue in African great apes: Implications for
language evolution. The Anatomical Record Part A: Discoveries
in Molecular, Cellular, and Evolutionary Biology 271A (March):
27685.
Shu, Weiguo, Julie Y. Cho, et al. 2005. Altered ultrasonic vocalization in
mice with a disruption in the Foxp2 gene. Proceedings of the National
Academy of Sciences USA 102 (July 5): 96438.
Shu, W., H. Yang, et al. 2001. Characterization of a new subfamily of
winged-helix/forkhead (fox) genes that are expressed in the lung and
act as transcriptional repressors. Journal of Biological Chemistry
276.29: 2748897.
55
6
ACQUISITION OF LANGUAGE
Barbara Lust
INTRODUCTION
How does the child, within the first few years of life, come to
understand and to produce Hop on Pop or Cat in the Hat
(Dr. Seuss)?
(1) We all play ball
Up on a wall
Fall off the Wall (Dr. Seuss [1963] 199l, 1013)
(2) We looked!
And we saw him!
The Cat in the Hat!
And he said to us (Dr. Seuss [1957] 1985, 6)
How does the child come to know the precise sound variations that distinguish hop and pop or wall and ball, the
multiple meanings of words like play or in, the reference
of pronouns like we or him, the meaning of quantifiers (see
quantification) like all, the functional categories determining
noun phrases like the, the idioms like play ball, the systematic order and structure necessary for even simple sentences,
the infinite recursion of sentences allowed by the coordinate
connective and, the infinite possibilities for propositions based
on the manipulation of even a few sound units and word units,
and the infinite possibilities for meaning and truth, which
Dr. Seuss continually demonstrated?
Even more stunning, how does the child both attain the infinite
set of possibilities and, at the same time, know the infinite set of
what is impossible, given this infinitely creative combinatorial system? Just as there is no limit to what is possible, there is no limit to
what is impossible in language, for example, (3), or an infinite set
of other impossible word and structure combinations, for example,
(4). This provides an essential paradox for language learners: How
do they both acquire the infinite possibilities and, at the same time,
the constraints on the infinite set of impossible expressions?
(3) *Ylay
*lfla
(4) *Play all we ball
*in cat the hat
56
Acquisition of Language
Shes going to leave without understanding the particular context in which this sentence is used, and there is an infinite set of
these contexts possible. The challenge of developing an infinitely
productive but constrained language system embedded in the
infinite possibilities for pragmatic determination of interpretation
remains a central challenge distinguishing machine language and
the language of the human species. All areas of cognitive science
must be invoked to study this area. Finally, especially with regard to
language acquisition in the child, developmental psychology must
be consulted critically. The child is a biologically and cognitively
changing organism. Indeed, an understanding of commonalities
and/or differences between language acquisition in child and adult
requires expertise in the field of developmental psychology.
In addition, the fact that the human species, either child or
adult, is capable not only of single language acquisition but also
of multilanguage acquisition, either simultaneously or sequentially, exponentially increases the complexity of the area (see
bilingualism and multilingualism). The fact that these
languages may vary in their parametric values (see parameters) and be either oral or visual further complicates the area.
THEORETICAL FOUNDATIONS
Not surprisingly, then, the field of research involving the study of
language acquisition is characterized by all the complexity and
variation in theoretical positions that characterize the field of linguistics and the language sciences in general. The area of language
acquisition research reflects varying approaches to the study of
grammar, that is, variations regarding viewpoints on what constitutes the end state of language knowledge that must be acquired,
for example, generative or functionalist approaches. At
the same time, it is characterized by variation in approach to the
study of language acquisition, in particular, ranging from various forms of logical to empirical analyses, and including disputes
regarding methodological foundations and varying attempts
at explanatory theories (ranging from rationalist to empiricist
types). (Lust 2006, Chapter 4, provides a review.)
The field is led by a strong theory of what is necessary logically for a strong explanatory theory of language acquisition in
the form of Noam Chomskys early proposal for a language
acquisition device (LAD). This model not only involved
explication of what needed to be explained but also spawned
decades of research either for or against various proposals bearing on components of this model. Yet, today, no comprehensive
theory of language acquisition exists, that is, no theory that would
fully account for all logically necessary aspects of a LAD.
At the same time, decades of research have now produced an
explosion of new discoveries regarding the nature of language
acquisition. These empirical discoveries, combined with theoretical advances in linguistic theory and with the development
of the interdisciplinary field of cognitive science, bring us today
to the verge of such a comprehensive theory, one with firm scientific foundations.
varying positions in the field on these focal issues, then exemplify a range of research results, displaying crucial new discoveries in the field. Ill conclude by formulating leading questions for
the future. For those interested in pursuing these topics further, I
close with selected references for future inquiry in the field.
Focal Tensions
Classically, approaches to the study of language acquisition have
been categorized as nativist or empiricist (see innateness and
innatism). These approaches, which correspond generally to
rationalist or empiricist approaches to epistemology (see Lust
2006 for a review), have been typically associated with claims
that essential faculties responsible for language acquisition in
the human species involve either innate capabilities for language
or not. However, now the debates have become more restricted,
allowing more precise scientific inquiry. No current proposals, to
our knowledge, suggest that nothing at all is innate. (See Elman
et al, 1996 for example). No one proposes that every aspect of
language knowledge is innate. For example, in the now-classic
Piaget-Chomsky Debate (see Piatelli-Palmarini 1980), Chomsky
did not deny that experience was necessary for a comprehensive
theory of language acquisition. On the other hand, Jean Piaget
proposed an essentially biological model of cognitive development and an antiempiricist theory, coherent with Chomskys
position in this way.
Rather, the current issue that is most central to the field of
language acquisition now is whether what is innate regarding
language acquisition involves components specific to linguistic
knowledge, for example, specific linguistic principles and
parameters, or whether more general cognitive knowledge or
learning principles themselves potentially innate can account
for this knowledge. This issue corresponds to the question of
whether organization of the mental representation of language
knowledge is modular or not. Proponents of the view that specifically linguistic factors are critical to an explanation of human
language acquisition generally work within a linguistic theory of a
modular language faculty, that is, a theory of universal grammar (UG), which is proposed to characterize the initial state
of the human organism, that is, the state of the human organism before experience. This current descendant of Chomskys
LAD provides specific hypotheses (working within a generative grammar framework) regarding the identity of linguistic
principles and parameters that may be innately, or biologically,
determined in the human species (e.g., Anderson and Lightfoot
2002). Proponents of the alternative view generally assume a
functionalist theory and a model of cultural learning, that is,
a model where culture in some ways provides the structure of
language, potentially without specific linguistic printiples, with
only general learning principles (e.g., Van Valin 1991; Tomasello
2003; Tomasello, Kruger, and Ratner 1993).
57
58
Acquisition of Language
also be empirically verified, that is, be a veridical account of how
the child actually does use surrounding data. Those approaching the logical problem of language acquisition are formally
diagnosing the properties of syntactic and semantic knowledge
that must be acquired, and assessing varied formal learnability
theories that may possibly account for the childs mapping from
actual input data to this knowledge. (e.g., Lightfoot 1989). (See
also projection principle.)
Discoveries
Fortunately, not only have the last decades seen continuous
theoretical developments in areas of linguistics regarding the
nature of language knowledge, and continuous sharpening
of the debate on the nature of language development, but the
field of language acquisition has also produced a wealth of new
empirical discoveries, all of which promise to inform the crucial
questions and debates in the field, and to eventuate in a more
comprehensive theory of language acquisition than has yet been
available. Ill exemplify only some of these here, all of which are
foci of intense current research. (See Lust 2006 for a more comprehensive review.)
THE FIRST 12 MONTHS. Informed by developed methodologies
for investigating infant perception of language, we now know
that neonates show early sensitivities to language variations
and categorize these variations (e.g., Ramus et al. 2000; Mehler
et al. 1996, 1988). For example, given varying sound stimuli from
Japanese and varying sound stimuli from English speech, newborn infants distinguish variation from Japanese to English significantly more than variation within either English or Japanese,
including speaker variation. In this sense, the infant is seen to
be categorizing variation across languages. (See infantile
responses to language.) More specifically, we also know
that formal aspects of language (phonetic, phonological, and
syntactic) begin to develop at birth, even before language comprehension, and provide the foundations for the appearance of
overt language production and comprehension in the child at
about 12 months of age. Now we also know something of how
that development proceeds even during the first 12 months of
life.
In the area of acquisition of phonology, we have discovered
the very fine, precise, and extensive phonetic sensitivities of the
newborn, sensitivities that appear to be exactly the right ones
to underlie all potential cross-linguistic phonological contrasts,
for example, contrasts in voicing or place and manner features
that characterize speech sounds. Once again, these sensitivities
are categorical (as in categorical perception; see speech perception in infants). We know that by about 6 months, these
sensitivities reveal appropriate cross-linguistic selection and
modulation, and by 12 months, this process is nicely attuned to
the childs first productive words of their specific language (e.g.,
Werker 1994; see Jusczyk 1997 and de Boysson-Bardies 1999 for
reviews). For example, while the 12-month-old infant acquiring Hindi will have maintained sensitivity to Hindi contrasts in
aspiration, the infant acquiring English will show diminished
response to such distinctions, which are not linguistically contrastive in English. Although labials appear in first words across
Given the age of the children and the nature of the words
tested, it is clear that children are engaged both in word segmentation and in long-term storage of these formal elements, even
without their semantic content. Moreover, like the acquisition of
phonology, childrens sensitivities to the language-specific structure of words begins to show refinement after the sixth month.
American 9-month-olds, though not 6-month-olds, listened
longer to English words than to Dutch words, while Dutch
9-month-old infants showed the reverse preference (Jusczyk
et al. 1993).
Simultaneously, and in parallel, in the area of syntax, infants
have also been found to be laying the formal foundations for
language knowledge even within the first 12 months. They are
carving out the formal elements that will form the basis for the
syntax of the language they are acquiring. Precise sensitivities
to linear order, as well as to constituent structure, have
been discovered in these first few months of life. For example,
infants as young as 4 months of age have been found to distinguish natural well-formed clause structure, like (6a) from
non-natural ones, like (6b), in stories read to them (where /
represents clause breaks through pauses, and the stimuli are
matched in semantic and lexical content) (Hirsh-Pasek et al.
1987).
(6) a. Cinderella lived in a great big house/ but it was sort of dark/
because she had this mean, mean, mean stepmother
b. in a great big house, but it was/ sort of dark because she
had/ this mean
59
In all these cases, both phonological and syntactic development does not reduce either to simple loss of initial sensitivities
or to simple accrual or addition of new ones, but a gradual integration of a specific language grammatical system.
More recently, research has begun to reveal even more precisely how these early sensitivities are related to later language
development, thus foreshadowing a truly comprehensive theory
of language acquisition (Newman et al. 2006; Kuhl et al. 2005).
BEYOND FIRST WORDS: LINGUISTIC STRUCTURE. Continuous with
advances in our understanding of early infant development, we
are now also seeing a potential revolution in our understanding of the early stages of overt language acquisition, that is,
those periods within the first three years of life, where the child
is beginning to overtly manifest language knowledge in terms
of language production and comprehension. Child language in
these early periods has traditionally been referred to as holophrastic or, a little later, telegraphic (see two-word stage)
in nature. Even in these early periods, numerous studies are now
revealing childrens very early sensitivity to functional categories
(FCs) in language, that is, to grammatical elements that function
formally to a large degree, often with little or no semantic content. These functional categories play a critical role in defining
constituent structure in language, and in defining the locus of
syntactic operations. Thus, the evidence that infants and toddlers are accessing these FCs in their early language knowledge
begins to provide crucial data on the foundations for linguistic
systems in the child. (Such FC are reflected in different ways
across languages. In English they are reflected in determiners
such as the, auxiliary verbs like do, complementizers like
that, or inflection of verbs, for example.)
Early research had revealed that young children perceive and
consult functional categories, such as determiners, even when
they are not overtly producing them regularly in their natural
speech. For example, L. Gerken, B. Landau, and R. Remez (1990)
showed that 2-year-olds recognize the distinction between
grammatical and ungrammatical function words, contrasting
these, for example (8a) and (8b), in their elicited imitation of
these sentences. Gerken and Bonnie McIntosh (1993) showed
that 2-year-olds used this knowledge in a picture identification
task, discriminating between (9a) and (9b), where grammatical
forms facilitated semantic reference.
(8) a. Pete pushes the dog
b. Pete pusho na dog
(9) a. Find the bird for me
b. Find was bird for me
More recently, a wide range of experimental researchers working with expanded infant testing methods have replicated these
results, and also revealed similar functional category distinctions
even in younger children. For example Yarden Kedar, Marianella
Casasola, and Barbara Lust (2006) showed that infants as young
as 18 months also distinguish sentences like (9a) and (9b) in a
preferential looking task, and again, object reference is facilitated by the grammatical form. Precursors to these functional
60
Acquisition of Language
(in the does too clause, where the second clause does not
state what Bert did), but also competence for construction of the
multiple interpretations allowed in this ambiguous structure (as
in 11ad), and constraint against the ungrammatical possibilities (as in 11 ei). (See Foley et al. 2003 for an example.) In other
words, they evidence early control of and constraints on empty
category interpretation and on other forms of ellipsis. Studies
of Chinese acquisition show similar results (Guo et al. 1996).
Without structure dependence and grammatical computation
over abstract structure underlying such sentences (e.g., reconstructing the verb phrase [VP] in the second clause), children
would not be expected to evidence this competence (see Foley
et al. 2003). All interpretations are pragmatically possible.
(11) Oscar bites his banana and Bert does too.
Possible Interpretations:
a. O bites Os banana and B bites Bs banana.
b. O bites Os banana and B bites Os banana.
c. O bites Bs banana and B bites Bs banana.
d. O bites Es banana and B bites Es banana.
Impossible Interpretations
*e. O bites Os banana and B bites Es banana.
*f. O bites Bs banana and B bites Os banana.
*g. O bites Bs banana and B bites Es banana.
*h. O bites Es banana and B bites Os banana.
*i. O bites Es banana and B bites Bs banana.
As is the case in this study of VP ellipsis acquisition, empirical research results in the Principles and Parameters framework
mutually inform theoretical development; they contribute to and
help to resolve theoretical debates on the representation of this
area of linguistic knowledge.
Evidence for early parameter setting with regard to linear
order in natural language is now provided by a wide range of
cross-linguistic studies showing very early language-specific
sensitivity to the clausal head direction and directionality of
RECURSION in the language being acquired (e.g., Lust, in preparation; Lust and Chien 1984). Very early differentiation of the
pro drop (i.e., argument omission wherein subjects may not be
overtly expressed) possibilities of a language have been attested
across languages (e.g., Italian and English, Spanish and English)
(Valian 1991; Austin et al. 1997, after Hyams 1986). In fact, children have been found to critically consult the complementizer
phrase, and the subordinate structure domain in order to make
this differentiation (Nuez del Prado, Foley, and Lust 1993).
FROM DATA TO GRAMMAR. Our understanding of how the infant,
from birth, consults the surrounding speech stream is now also
expanding quickly. Research has revealed very fine sensitivities
to particular aspects of the speech stream (e.g., Saffran, Aslin,
and Newport 1996), and research has begun to isolate the precise role of particular cues in this process, for example, STRESS,
phonotactic constraints, and statistical distributions (e.g, Johnson
and Jusczyk 2001). Various forms of bootstrapping may be available to the child (phonological or prosodic or syntactic bootstrapping, for example). Here, bootstrapping generally refers to that
process by which the child, in the initial state, might initiate new
61
62
Acquisition of Language
Cohen Sherman, Janet, and Barbara Lust. 1993. Children are in control.
Cognition 46: 151.
Crain, Stephen. 1991. Language acquisition in the absence of experience. Behavioral and Brain Sciences. 14.4: 597650.
Crain, Stephen., Takuya Gloro, and Rosalind Thornton. 2006. Language
acquisition is language change. Journal of Psycholinguistic Research
35: 3149.
Crain, Stephen, and Diane Lillo-Martin. 1999. An Introduction to Linguistic
Theory and Language Acquisition. Malden, MA: Basil Blackwell.
Crain, Stephen, and Cecile McKee. 1985. Acquisition of structural restrictions on anaphors. In Proceedings of the Sixteenth Annual Meeting
of the North Eastern Linguistics Society, ed. S. Berman, J. Chloe, and
J. McDonough, 94110. Montreal: McGill University.
Crain, Stephen, and Mineharu Nakayama. 1987. Structure dependence
in grammar formation. Language 63: 52243.
Crain, Stephen, and Paul Pietroski. 2002. Why language acquisition is a
snap. Linguistic Review 19: 16384.
Crain, Stephen, and Kenneth Wexler. 1999. Methodology in the study
of language acquisition: A modular approach. In Handbook of First
Language Acquisition, ed. William C. Ritchie and Tej K. Bhatia, 387
426. San Diego, CA: Academic Press.
deGraff, Michele. 1999. Creolization, language change and language
acquisition: A prolegomenon. In Language Creation and Language
Change, ed. M. deGraff, 146. Cambidge, MA: MIT Press.
Demuth, Katherine. 1994. On the underspecification of functional categories in early grammars. In Syntactic Theory and First Language
Acquisition: Cross-Linguistic Perspectives. Vol. 1. Ed. B. Lust, M. Suner,
and J. Whitman, 11934. Hillsdale, NJ: Lawrence Erlbaum.
deVilliers, Jill, Tom Roeper, and Ann Vainikka. 1990. The acquisition of long-distance rules. In Language Processing and Language
Acquisition, ed. Lynn Frazier and Jill deVilliers, 25798. Dordrecht, the
Netherlands: Kluwer.
Edelman, Shimon. 2008. Computing the Mind: How the Mind Really
Works. Oxford: Oxford University Press.
Eisele, Julie, and Barbara Lust. 1996. Knowledge about pronouns: A
developmental study using a truth-value judgment task. Child
Development 67: 3086100.
Elman, Jeffrey L., Elizabeth A. Bates, Mark H. Johnson, Annette
Karmiloff-Smith, Domenico Parisi, and Kim Plunkett. 1966. Rethinking
Innateness. Cambridge, MA: MIT Press.
Flynn, Suzanne, and Gita Martohardjono. 1994. Mapping from the initial state to the final stage: The separation of universal principles and
language specific principles. In Syntactic Theory and First Language
Acquisition: Cross Linguistic Perspectives, ed. Barbara Lust, Magui
Suner, and John Whitman, 31936. Hillsdale, NJ: Lawrence Erlbaum.
Foley, Claire, Zelmira Nuez del Prado, Isabella Barbier, and Barbara
Lust. 2003. Knowledge of variable binding in VP-ellipsis: Language
acquisition research and theory converge. Syntax 6.1: 5283.
Friederici, A. D. 2000. The developmental cognitive neuroscience of language: A new research domain. Brain and Language 71: 658.
. 2002. Towards a neural basis of auditory sentence processing.
Trends in Cognitive Science 6.2: 7884.
Gair, James, Barbara Lust, Lewala Sumangala, and Milan Rodrigo. 1998.
Acquisition of null subjects and control in some Sinhala adverbial
clauses. In Studies in South Asian Linguistics: Sinhala and Other South
Asian Languages, 27185. Oxford: Oxford University Press.
Gentner, Dedre, and Susan Goldin-Meadow, eds. 2003. Language in
Mind. Cambridge, MA: MIT Press.
Gerken, L., B. Landau, and R. Remez. 1990. Function morphemes in
young childrens speech perception and production. Developmental
Psychology 26: 20416.
Gerken, Louann, and Bonnie McIntosh. 1993. Interplay of function morphemes and prosody in early language. Developmental Psychology
29.3: 44857.
63
64
7
ELABORATING SPEECH AND WRITING:
VERBAL ART
Patrick Colm Hogan
65
66
67
68
69
70
71
SHAKESPEAREAN INDIRECTION
As the preceding discussions have been rather abstract, it is
valuable to end with a more developed example. Consider
Shakespeares Sonnet 97:
How like a winter hath my absence been
From thee, the pleasure of the fleeting year!
What freezings have I felt, what dark days seen!
What old Decembers bareness every where!
And yet this time removed was summers time,
The teeming autumn big with rich increase,
Bearing the wanton burthen of the prime,
Like widowed wombs after their lords decease:
Yet this abundant issue seemd to me
But hope of orphans, and unfathered fruit,
For summer and his pleasures wait on thee,
And thou away, the very birds are mute;
Or if they sing, tis with so dull a cheer
That leaves look pale, dreading the winters near.
72
73
74
Ong, Walter J., S.J. The jinnee in the well wrought urn. In The
Barbarian Within and Other Fugitive Essays and Studies, 1525. New
York: Macmillan.
Pratt, Mary Louise. 1977. Toward a Speech Act Theory of Literary Discourse.
Bloomington: Indiana University Press.
Shakespeare, William. 2006. The Sonnets, ed. G. Blakemore Evans.
Cambridge: Cambridge University Press.
Stockwell, Peter. 2002. Cognitive Poetics: An Introduction.
London: Routledge.
Tsur, Reuven. 2006. Kubla Khan: Poetic Structure, Hypnotic Quality and
Cognitive Style A Study in Mental, Vocal and Critical Performance.
Amsterdam: John Benjamins.
Walter, Henrik. 2001. Neurophilosophy of Free Will: From Libertarian
Illusions to a Concept of Natural Autonomy. Trans. Cynthia Klohr.
Cambridge, MA: MIT Press.
Abduction
A
ABDUCTION
Albert Atkin
Abduction is a form of reasoning first explicated by the nineteenth-century philosopher C. S. Peirce. The central concept
he wishes to introduce is that of generating new hypotheses to
explain observed phenomena partly by guesswork or speculation. In his early work, Peirce tried to explain abductive reasoning, as distinct from deductive and inductive reasoning, by
reference to syllogistic form. For instance, the following schema
is an example of deductive reasoning:
All the beans in the bag are white
These beans came from this bag
Therefore, these beans are white
Language universals are statements that are true of all languages; for example, all languages have stop consonants. But
beneath this simple definition lurks deep ambiguity, and this
triggers misunderstanding in both interdisciplinary discourse
and within linguistics itself. A core dimension of the ambiguity
is captured by the opposition absolute versus statistical universal, although the literature uses these terms in varied ways.
Many textbooks draw the boundary between absolute and statistical according to whether a sample of languages contains exceptions to a universal. But the notion of an exception-free sample
is not very revealing, even if the sample contained all known
languages: There is always a chance that an as yet undescribed
language, or an unknown language from the past or future, will
provide an exception.
It is impossible, in principle, to survey all languages of our species. If we nevertheless want to make claims about all languages,
only two routes are open: a priori deduction of necessarily true
statements or statistical extrapolation from empirical samples to
the entire set. Absolute universals can then be defined as those
that are necessarily true, statistical universals as those that are
extrapolated from samples.
Absolute Universals
For statements to be necessarily true, they must follow from a
priori assumptions. The assumptions that linguists make are
diverse and heavily debated. An example is the assumption that
words consist of morphemes, that is, minimal form-meaning pairs. If one accepts this, then it is necessarily true that all
languages have morphemes, and there cannot be exceptions.
Why? Suppose someone claims to have discovered a language
without morphemes. One can of course simply analyze the language without mentioning morphemes, but obviously that cannot challenge the universal just because one can always defend
it by reanalyzing the language with morphemes. The only true
challenge would be to show that analyzing some data in terms
of morphemes leads to structures that are in conflict with other
assumptions, for example, that form-meaning pairs combine
exclusively by linear concatenation. The conflict can be illustrated
by languages with morphologies like the English plural geese,
where the meanings plural and goose do not correspond to linear
77
On any of these options, the universal remains exceptionless: On solution (i), no language has morphemes; on solutions (ii) and (iii), all languages have morphemes. As a result,
absolute universals can never be falsified by individual data.
Their validity can only be evaluated by exploring whether they
are consistent with other absolute universals that are claimed
simultaneously.
Absolute universals can also be thought of as those aspects
of ones descriptive metalanguage often called a theoretical framework that are necessarily referred to in the analysis
of every language, that is, that constitute the descriptive a priori.
Depending on ones a priori, this includes, apart from the morpheme, such notions as distinctive feature, constituent (see constituent structure), argument, predicate (see predicate
and argument), reference, agent, speaker, and so on. In some
metalanguages, the a priori also includes more specific assumptions, for example, that constituents can only be described by
uniform branching (all to the left, or all to the right), or only by
binary branching, and so on.
The status of absolute universals is controversial. For many
linguists, especially in typology and historical linguistics, absolute universals are simply the descriptive a priori,
with no additional claim on biological or psychological reality. The choice between equally consistent universals/metalanguages for example, among options (i), (ii), and (iii) in
the previous example is guided by their success in describing
structures and in defining variables that capture distributional
patterns, an evaluation procedure comparable to the way in
which technical instruments for analyzing objects are evaluated
in the natural sciences. In the morphology problem, typologists
would most likely chose option (ii) because it allows for defining
a variable of stem-internal versus affixal plural realization that
has an interesting distribution (suggesting, for example, that
within-stem realization is favored by a few families in Africa and
the Near East).
In generative grammar, by contrast, absolute universals
are thought of not only as descriptively a priori but also as biologically given in what is called universal grammar: they are
claimed to be innate (see innateness and innatism) and to
be identical to the generalizations that a child makes when learning language. Thus, if the morpheme is accepted as a universal,
that is, a priori term of our metalanguage, it will also be claimed to
be part of what makes languages learnable (see learnability)
and to be part of our genetic endowment. An immediate consequence of such an approach is that something can be claimed as
universal even if it is not in fact necessary in the analysis of every
language. For example, even if some language (e.g., the Rotokas
language of Bougainville) lacks evidence for nasal sounds, one
could still include a distinctive feature [ nasal] in Universal
Grammar. Rotokas speakers are then said to have the feature as
part of their genetic endowment even if they dont use it.
78
Statistical Universals
Accessibility Hierarchy
In response to this problem, typologists seek to ensure representativity of a sample not by random selection within families but
by exhaustive sampling of known families, stratified by area. In
order to then control for unequal family sizes, one usually admits
only as many data points per family as there are different values
on the variables of interest (Dryer 1989; Bickel 2008).
Samples that are not based on random sampling do not
support parametric inference by statistical tests. An alternative
to this is randomization methods (Janssen, Bickel, and Ziga
2006): The null hypothesis in these methods is that an observed
preference can be predicted from the totals of the sample (e.g.,
that an observed 90% postnominal relatives in VO [verb-object]
languages could be predicted if 90% of the entire sample had
postnominal relatives) not that the sample stems from a population without the observed preference. Extrapolation to the total
population (the entire of set of human languages) can then only
be based on plausibility arguments: If a preference significantly
deviates from what is expected from the totals of the observed
sample, it is likely that the preference holds in all languages. A
key issue in such argumentation is whether the tested variables
are sufficiently unstable over time so that a present sample can
be assumed to not reflect accidental population skewings from
early times in prehistory (Maslova 2000). In response to this,
typologists now also seek to test universals by sampling language
changes instead of language states a move that is sometimes
called the dynamization of typology (Greenberg 1995; Croft
2003).
While the number of proposed statistical universals is
impressive the Universals Archive at Konstanz has collected
more than 2,000 (Plank and Filimonova 2000) very few of them
have been rigorously tested for independence of area, family,
and time.
Balthasar Bickel
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bickel, B. 2008. A refined sampling procedure for genealogical control.
Sprachtypologie und Universalienforschung 61: 2233.
Croft, W. 2003. Typology and universals. 2d ed. Cambridge: Cambridge
University Press.
Cysouw, M. 2005. Quantitative methods in typology. In Quantitative
Linguistics: An International Handbook, ed. G. Altmann, R. Khler, and
R. Piotrowski, 55478. Berlin: Mouton de Gruyter.
Cysouw, M, ed. 2008. Special issue on analyzing The World Atlas of
Language Structures. Sprachtypologie und Universalienforschung 61.
Dryer, M. S. 1989. Large linguistic areas and language sampling. Studies
in Language 13: 25792.
Greenberg, J. H. 1995. The diachronic typological approach to language. In Approaches to Language Typology, ed. M. Shibatani and T.
Bynon, 14366. Oxford: Clarendon.
Haspelmath, M., M. S. Dryer, D. Gil, and B. Comrie, eds. 2005. The World
Atlas of Language Structures. Oxford: Oxford University Press.
Hauser, M. D., N. Chomsky, and W. T. Fitch. 2002. The faculty of language: What it is, who has it, and how did it evolve? Science 298: 1569
79. This paper and the response by S. Pinker and R. Jackendoff 2005
launched an ongoing debate on the nature and extent of absolute universals in generative grammar.
Hawkins, J. A. 2004. Efficiency and Complexity in Grammars.
Oxford: Oxford University Press.
Janssen, D., B. Bickel, and F. Ziga. 2006. Randomization tests in language typology. Linguistic Typology 10: 41940.
Maslova, E.. 2000. A dynamic approach to the verification of distributional universals. Linguistic Typology 4: 30733.
. 2008. Meta-typological distributions. Sprachtypologie und
Universalienforschung 61: 199207.
Newmeyer, F. J. 2005. Possible and Probable Languages: A Generative
Perspective on Linguistic Typology. New York: Oxford University Press.
Nichols, J. 1992. Language Diversity in Space and Time. Chicago: The
University of Chicago Press.
Pinker, S., and R. Jackendoff. 2005. The faculty of language: Whats special about it? Cognition 95: 20136.
Plank, F., and E. Filimonova. 2000. The Universals Archive: A
brief introduction to prospective users. Sprachtypologie und
Universalienforschung 53: 10923.
Plank, F., ed. 2007. Linguistic Typology 11.1 (Special issue treating the
state of typology.)
ACCESSIBILITY HIERARCHY
Edward L. Keenan and Bernard Comrie (1972, 1977) introduce
the accessibility hierarchy (AH) as a basis for several crosslinguistic generalizations regarding the formation of relative
clauses (RCs).
AH: SUBJ > DO > IO > OBL > GEN > OCOMP
The terms of the AH are main clause subject, direct object, indirect object, object of pre- or postposition, genitive (possessor), and
object of comparison. Keenan and Comrie cross-classified RCs
along two parameters: 1) the head noun precedes or follows the
restrictive clause (RestCl), and 2) the case of the position relativized, NPrel, is pronominally marked or not. In (1), from German,
the RestCl, underlined, follows the head in (1a) and precedes it
in (1b). In (2a,b), from Hebrew and Russian, NPrel is pronominally marked but not in English.
(1) a. der Mann, der in seinem Bro arbeitet
the man, who in his
study is+working
the man who is working in his study
b. der in seinem Bro arbeitende Mann
the in his
study working man
the man who is working in his study
(2) a. ha-isha
she Dan natan la
et ha-sefer
the-woman that Dan gave to+her acc the-book
the woman that Dan gave the book to
b. devuka, kotoruju Petr ljubit
girl,
who(acc) Peter loves
the girl who Peter loves
79
Accessibility Hierarchy
(4) a. Manolotra (m+aN+tolotra) vary ho anny vahiny aminny
lovia vaovao ny tovovavy
offers (pres+act+offer) rice forthe guests onthe dishes new the
young+woman
The young woman offers rice to the guests on the new dishes
b. ny tovovavy (izay) manolotra vary ho anny vahiny aminny
lovia vaovao
the woman (that) offers rice tothe guests onthe dishes new
the young woman who offers rice to the guests on the new dishes
c. *ny vary (izay) manolotra ho anny vahiny aminny lovia
vaovao ny tovovavy
the rice (that) offers tothe guests onthe dishes new the
young+woman
the rice that the young woman offers to the guests on the new
dishes
The first four words in (4c) claim that the rice is doing the
offering a nonsense. Malagasy does not, however, have an
expressivity gap here since it has a rich voice system allowing any major NP in a clause as subject. The form of offer that
takes theme subjects is atolotra, recipient subjects tolorana,
and oblique subjects anolorana. (5a,b) illustrate Theme and
Instrument RCs.
(5) a. ny vary (izay) atolo-dRasoa ho anny vahiny aminny lovia
vaovao
the rice (that) offered-by+Rasoa forthe guests onthe new
dishes
the rice that the young+woman offers to the guests on the
new dishes
b. ny lovia vaovao (izay) anoloran-dRasoa vary ho anny
vahiny
the dishes new (that) offered-by+Rasoa rice for the guests
the new dishes on which the young woman offered rice to the
guests
80
Acoustic Phonetics
causativizing a transitive verb, its agent argument may surface as
an IO (Jai fait manger les pinards aux enfants I made-eat the
spinach to the children). Lastly, S. Hawkins and Keenan (1987)
show psycholinguistically that recall of RCs formed on high positions on the AH was better than recall of ones formed on low
positions.
One interesting modification to the hierarchy generalizations concerns syntactic ergativity. Keenan and Comrie noted
that Dyirbal (Dixon 1972) relativizes absolutives intransitive
subjects and transitive objects, but not transitive subjects. A
verbal affix (antipassive) derives intransitive verbs from transitive ones with the agent as subject, hence relativizable. Mayan
languages such as Jacaltec (Craig 1977, 196) are similar. This is
an elegant solution to the requirement that agents be relativizable, analogous to Bantu applicatives or Austronesian voices
affixes.
(7) a. x s watxe naj hun-ti
asp 3abs 3erg make cl:man one-this
He made this
b. naj x watxe n hun-ti
cl:man asp 3abs make ap one-this
the man (who) made this
Edward L. Keenan
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Comrie, Bernard. 1976. The syntax of causative constructions: Crosslanguage similarities and divergences. In The Grammar of Causative
Constructions, Syntax and Semantics 6, ed. Masayoshi Shibatani, 261
312. Amsterdam: Academic Press.
Craig, Colette. 1977. Jacaltec. Austin: University of Texas Press.
Dixon, Robert M. W. 1972. The Dyirbal Language of North Queensland.
Oxford: Cambridge University Press.
Hawkins, Sarah, and Edward L. Keenan. 1987. The psychological validity
of the accessibility hierarchy. In Universal Grammar: 15 Essays, ed. E.
L. Keenan, 6089. London: Croom Helm.
Keenan, Edward L. 1975. Variation in Universal Grammar. In Analyzing
Variation in English, ed. R. Fasold and R. Shuy, 13648. Washington,
DC: Georgetown University Press.
Keenan, Edward L., and Bernard Comrie. 1972. Noun phrase accessibility and Universal Grammar. Paper presented at the Annual Meetings
of the Linguistic Society of America, Atlanta.
. 1977. Noun phrase accessibility and Universal Grammar.
Linguistic Inquiry 8.1: 6399.
ACOUSTIC PHONETICS
Like the rest of linguistics, acoustic phonetics combines description with theory. Descriptions are images of acoustic properties
and quantitative measures taken from these images, and theory
accounts for the way in which a sounds articulation determines
its acoustics.
Description
The three most commonly used images of speech are the waveform, spectrum, and spectrogram (Figure 1). The waveform displays differences in sound pressure level (in pascals) over time
(Figure 1a, d), the spectrum differences in sound pressure level
Acoustic Phonetics
(in deciBels) over frequency (Figure 1b, e), and the spectrogram
differences in frequency over time (Figure 1c, f,); darkness indicates the sound pressure level at particular frequencies and
moments in the spectrogram.
The images in Figures 1ac differ from those in Figure 1df
in every conceivable respect: Sound pressure level varies more
or less regularly and repetitively, every 0.0073 second, in the
vowel [i] (as in heed), while in the fricative [s] (as in see), it
instead varies nearly randomly. The vowel is thus nearly periodic, while the fricative is decidedly aperiodic. This difference
gives the vowel a clear pitch, while making the fricative instead
sound noisy.
A single cycles duration in a periodic sound is its period (T);
the distance it travels in space is its wavelength (). As measures
of a single cycles extent, both period and wavelength are reciprocally related to frequency (F), the number of cycles per second,
or Hz:
(1)
F (cycles/sec) =
1
T (sec/cycle)
(2)
F (cycles/sec) =
c (cm/sec)
(cm/cycle)
Note that the numerator in (2) is not 1 but instead c, the speed
of sound.
The spectrum and spectrogram of [i] (Figures 1b, c) show peaks
and horizontal bands, respectively, known as formants, at roughly
300, 2,200, and 2,800 Hz. The corresponding images of [s] (Figures
1e, f) show a broad energy band spanning 4,0007,000 Hz.
Whether a sound is periodic and where in its spectrum energy
is concentrated are nearly sufficient to distinguish all speech
sounds from one another acoustically, and these two properties also reflect the two components of the theoretical model for
transforming articulations into acoustics. All voiced sounds are
periodic, as are trills. Sonorants (vowels, glides, liquids, nasals)
are usually periodic, while obstruents (fricatives, stops, affricates) are aperiodic. Voiced obstruents are both periodic and
aperiodic. Differences in vowel quality and consonantal place of
articulation are both realized acoustically as differences in where
energy is concentrated in their spectra. The remaining property
is duration, which besides conveying short:long contrasts also
contributes to the conveying of tense:lax contrasts between vowels and the voicing and manner contrasts between consonants.
Theory
Speech sounds are the product of the application of a filter that
determines the frequencies in which energy is concentrated
81
Acoustic Phonetics
First resonance
Glottis = closed
Minimum
released or continuously through a fricatives narrow constriction. In strident sounds, the jet breaks up against a baffle
just downstream, increasing turbulence and noise intensity
considerably.
Lips = open
Maximum
2.5
7.5
10
12.5
15
17.5
(a)
Second resonance
2.5
7.5
10
12.5
15
17.5
(b)
Third resonance
2.5
7.5
10
12.5
15
17.5
(c)
Figure 2. The oral cavity as a tube closed at one end and open at the
other: (ac) standing waves corresponding to the first three resonances,
each with a velocity minimum at the closed end and a velocity maximum
at the open end.
RESONANCE. Both periodic and aperiodic sound sources introduce acoustic energy into the oral cavity across a broad enough
range of frequencies to excite any resonance of the air inside
the oral cavity. If the articulators are in their rest positions
and vocal folds are in the voicing position, this cavitys shape
approximates a uniform tube, closed at the glottis and open at
the lips (Figure 2). A resonance is produced by the propagation of acoustic energy away from the source and its reflection
back and forth off the two ends of the tube, which establishes a
standing wave. In a standing wave resonance, the locations of
zero and maximum pressure variation are fixed.To understand
how air resonates, it is easier to consider the change in pressure
level in the standing wave, rather than pressure level itself, that
is, the extent to which the air molecules are being displaced
longitudinally, or equivalently the velocity of the pressure
change, rather than the extent of their instantaneous compression or rarefaction. Air is most freely displaced longitudinally
at the open end, the lips, and least freely at the closed end, the
glottis. As a result, the standing waves that fit best inside the
oral cavity are those whose wavelengths, and thus frequencies,
are such that they have a velocity maximum (antinode) at the
lips, and a velocity minimum (node) at the glottis. Because the
air resonates more robustly at these frequencies than at others, the oral cavitys resonant response filters the sound source,
passing energy in the source at some frequencies and stopping
it at others.
Figures 2ac show the three lowest-frequency standing
waves that fit these boundary conditions. How are their frequencies determined? Figures 2ac show that one-quarter
of the first resonances wavelength spans the distance from
the glottis to the lips (the oral cavitys length, Loc), threequarters of the seconds, and five-quarters of the thirds. More
generally:
(3)
82
Loc =
2n 1
n
4
(4)
Fn =
c
4
Loc
2n 1
Substituting 35,000 cm/sec for c and 17.5 cm for Loc (the average adult males oral cavity length) yields 500, 1,500, and 2,500
Hz as the first three resonances frequencies, values close to
schwas.
Because the variable Loc is in the denominator, resonance
frequencies are lower in adults and mens longer oral cavities than in childrens or womens shorter ones, and likewise when the lips are protruded in a rounded vowel, such
Acoustic Phonetics
Figure 3. Spectra of the vowels (a) [i] as in heed, (b) [u] as in whod,
and (c) [a] as in hod. The individual peaks are the harmonics of the fundamental frequency (F0) of voice sound source, and the formants are the
ranges of amplified harmonics. Peaks corresponding to F1F3 are labeled
at the top of each panel.
Figure 4. Spectrograms of the first 150 ms of the words (a) bad, (b)
dad, and (c) gad. The onsets of F1F3 are labeled.
83
Acoustic Phonetics
constriction
back
8 cm2
1 cm2
front
3 cm
6 cm
8.5 cm
Figure 5. The configuration of the oral cavity with a constriction partway along its length.
(5)
Fn =
c
2
n Lrc
where Lrc is the length of the resonating cavity. The second complication is that the acoustic interaction of the constriction with
the cavity behind it produces a Helmholtz resonance. Its frequency (Fh) is:
(6) Fh =
c
2
Ac
Ab Lb Lc
84
Figure 6. The first three resonance frequencies of the back cavity (filled
symbols) and front cavity (empty symbols) and the Helmholtz resonance
(crosses) produced by incremental movement of the constriction in
Figure 5.
effects of lip rounding, which closes the front cavity at both ends
and introduces another Helmholtz resonance. None of the resonances produced by this front cavity are lower than the back cavity resonances, but the additional Helmholtz resonance is low
enough to constitute the F2 observed in [u] (657 Hz if the labial
constriction has a cross-sectional area of 1 cm2 and a length of 2
cm, and the front cavity is 4.5 cm long).
In the second, perturbation, heuristic, a constrictions proximity to a resonances velocity minimum or maximum determines how it perturbs that resonances frequency away from its
schwa-like value: A constriction near a velocity minimum raises
the formants frequency, while one near a maximum lowers it
instead (expansions have the opposite effects). Figures 4ac show
that minima occur at even quarters of a resonances wavelength
and maxima at odd quarters, and that their locations are at fixed
proportions of the length of the oral cavity. Because constriction locations are also a certain proportion of the distance from
the glottis to the lips, whether they coincide with a minimum or
maximum can be calculated by multiplying both sides of (4) by
the proportion of the oral cavitys length that corresponds to the
constrictions location and rounding the result on the right-hand
side to the nearest quarter (Table I).
The perturbation heuristic successfully predicts the effects of
the bilabial, palatal, velar, and pharyngeal constrictions on all
three formants of [b, i, g, a], and likewise the effects of the alveolar and velar constrictions on F1 and F3 in [d, u], but it fails to
predict F2 raising after [d], and F2 lowering in [u]. The latter can
Adaptation
Table 1. Calculating a constrictions proximity to a resonances velocity minimum or maximum from the constrictions proportional distance from
the glottis to the lips.
Place of
constriction
Segment
Proportion of oral
cavity length
Odd/even
Lower/higher
Labial
1*1/4=1/4
1*32/4=3/4
1*53/4=1/4
Odd
Odd
Odd
F1 lower
F2 lower
F3 lower
Alveolar
7/8
7/8*1/4=7/321/4
7/8*32/4=21/323/4
7/8*53/4=35/324/4
Odd
Odd
Even
F1 lower
F2 lower
F3 higher
Palatal
3/4
3/4*1/4=3/161/4
3/4*32/4=9/162/4
3/4*53/4=15/164/4
Odd
Even
Even
F1 lower
F2 higher
F3 higher
Velar
g, u
2/3
2/3*1/4=2/121/4
2/3*32/4=6/122/4
2/3*53/4=10/123/4
Odd
Even
Odd
F1 lower
F2 higher
F3 lower
Pharyngeal
1/4
1/4*1/4=1/160/4
1/4*32/4=3/161/4
1/4*53/4=5/161/4
Even
Odd
Odd
F1 higher
F2 lower
F3 lower
Summary
Speech sounds articulations produce sound sources by transforming aerodynamic energy into acoustic form, and those
sound sources in turn cause air inside the oral cavity to resonate,
at frequencies determined by the length of the resonating cavities and where they are constricted.
John Kingston
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Fant, C., and M. Gunnar. 1960. Acoustic Theory of Speech Production. The
Hague: Mouton.
Jakobson, Roman, C. Fant, M. Gunnar, and Morris Halle. 1952.
Preliminaries to Speech Analysis, Cambridge, MA: MIT Press.
Ladefoged, Peter, and Ian Maddieson. 1996. Sounds of the Worlds
Languages. Oxford: Blackwell Publishers.
Stevens, Kenneth N. 1998. Acoustic Phonetics. Cambridge, MA: MIT Press.
ADAPTATION
An adaptation is a characteristic in an organism that evolved
because it helped the organism or its relatives to survive and
reproduce. Examples include the vertebrate eye, claws, mammary glands, the immune system, and the brain structures that
underlie the human capacity for language. More completely,
an adaptation is 1) a reliably developing set of characteristics
2) whose genetic basis became established and organized in the
species (or population) over evolutionary time because 3) the
adaptation interacted with recurring features of the body or environment 4) in a way that, across generations, typically caused
this genetic basis to increase its gene frequency.
If a characteristic lacks any of these features, it is not an adaptation. An adaptation is not, therefore, simply anything in an
individual with a good or functional outcome, or that has
useful effects by intuitive standards. Rice cultivation, useful as it
is, is not a biological adaptation because it lacks a specific genetic
basis. Similarly, the English language is not an adaptation,
however useful it might be. In contrast, if a mutation occurred
that modified a neural structure so that the vocal chords could
more reliably produce distinct phonemes, and this gene spread
throughout the species because its bearers prospered due to the
advantages resulting from a lifetime of more efficient communication, then the modified neural structure would qualify as an
adaptation.
Researchers judge whether something is an adaptation by
assessing how likely or unlikely it is that its functional organization was produced by random mutation and spread by genetic
drift. For example, the eye has hundreds of elements that are
arranged with great precision to produce useful visual inputs. It is
astronomically unlikely that they would have arrived at such high
levels of mutual coordination and organization for that function
unless the process of natural selection had differentially retained
them and spread them throughout the species. Consequently,
the eye and the visual system are widely considered to be obvious examples of adaptations. For the same reason, evolutionary
scientists consider it overwhelmingly likely that many neurocognitive mechanisms underlying language are adaptations for
communication (a proposition that Noam Chomsky has disputed; see Lyle Jenkinss essay, Explaining Language, in this
volume). Language competence reliably develops, is believed
to have a species-typical genetic basis, and exhibits immensely
complex internal coordination that is functionally organized to
produce efficient communication, which vastly enhances the
achievement of instrumental goals, plausibly including those
linked to fitness.
85
Ad hoc Categories
Within the evolutionary sciences, the concept of adaptation
plays an indispensable role not only in explaining and understanding how the properties of organisms came to be what they
are, but also in predicting and discovering previously unknown
characteristics in the brains and bodies of species.Evolutionary
psychologists, for example, analyze the adaptive problems our
ancestors were subjected to, predict the properties of previously unknown cognitive mechanisms that are expected to have
evolved to solve these adaptive problems, and then conduct
experimental studies to test for the existence of psychological
adaptations with the predicted design (see evolutionary
psychology). An understanding that organisms embody sets
of adaptations rather than just being accidental agglomerations
of random properties allows organisms to be properly studied as
functional systems. If language is accepted as being the product
of adaptations, then there is a scientific justification for studying
the underlying components as part of a functional system.
The concept of adaptation became more contentious when
human behavior and the human psychological architecture
began to be studied from an adaptationist perspective. Critics
have argued that not every characteristic is an adaptation an
error adaptationists also criticize. More substantively, critics
have argued that it is impossible to know what the past was like
well enough to recognize whether something is an adaptation.
Adaptationists counter that we know many thousands of things
about the past with precision and certainty, such as the threedimensional nature of space and the properties of chemicals, the
existence of predators, genetic relatives, eyes, infants, food and
fertile matings, and the acoustical properties of the atmosphere,
and that these can be used to gain an engineers insight into why
organisms (including humans) are designed as they are.
Julian Lim, John Tooby, and Leda Cosmides
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Gould, S. J., and R. C. Lewontin. 1979. The spandrels of San Marco and
the Panglossian paradigm: A critique of the adaptationist programme.
Proceedings of The Royal Society of London, Series B 205.1161: 58198.
Pinker, Steven. 1994. The Language Instinct. New York: Morrow.
. 2003. Language as an adaptation to the cognitive niche. In
Language Evolution, ed. M. Christiansen and S. Kirby, 1637. New
York: Oxford University Press.
Tooby, John, and I. DeVore. 1987. The reconstruction of hominid
behavioral evolution through strategic modeling. In The Evolution
of Primate Behavior: Primate Models. ed. W. G. Kinsey, 183237. New
York: SUNY Press.
Williams, George C. 1966. Adaptation and Natural Selection: A Critique
of Some Current Evolutionary Thought. Princeton, NJ: Princeton
University Press.
AD HOC CATEGORIES
An ad hoc category is a novel category constructed spontaneously to achieve a goal relevant in the current situation (e.g., constructing tourist activities to perform in Beijing while planning a
vacation). These categories are novel because they typically have
not been entertained previously. They are constructed spontaneously because they do not reside as knowledge structures in
long-term memory waiting to be retrieved. They help achieve a
86
Ad hoc Categories
Adjacency Pair
ADJACENCY PAIR
conversation analysis, an inductive approach to the microanalysis of conversational data pioneered by Harvey Sacks
(1992), attempts to describe the sequential organization of
pieces of talk by examining the mechanics of the turn-taking system. Adjacency pairs reflect one of the basic rules for turn-taking
(Sacks, Schegloff, and Jefferson 1974), in which a speaker allocates the conversational floor to another participant by uttering
the first part of a paired sequence, prompting the latter to provide the second part. Examples are question-answer, greetinggreeting as in (1), and complaint-excuse:
(1) A: Hi there
B: Oh hi
87
Age Groups
(2) A:
B:
A:
B:
(iii) They are organized as a first and a second part, that is,
they are nonreversible. This is the case, incidentally, even in
ostensibly identical first and second parts, such as the greeting-greeting pair in (1), where reversing the order results in
an aberrant sequence.
(iv) They are ordered, so that a particular first part requires a
relevant second part (e.g., greetings do not follow questions).
The fact that the second part is conditionally relevant on the first
part does not mean that only one option is available; in fact, certain first parts typically allow for a range of possible second parts.
If two (or more) options are possible, one will be the more socially
acceptable, preferred response, the other(s) being dispreferred;
this phenomenon is known as preference organization, as in:
(3) A: Have a piece of cake (first part)
B1: Great thanks I will (preferred second)
B2: Ehm actually Ive just eaten but thanks anyway (dispreferred second)
As illustrated in (3), dispreferred second parts tend to be structurally different from preferred seconds (B2 being indirect, including an explanatory account, and containing hesitation markers,
unlike B1).
Another, related phenomenon that merits mention here is
presequencing: Certain adjacency pairs can be introduced or
foreshadowed by a preceding exchange, as in:
(4) A1:
B1:
A2:
B2:
Real-Time Change
This whole exchange forms one unit, in the sense that the occurrence of the question-answer pair A1-B1 is only interpretable
given the subsequent request-compliance adjacency pair A2-B2.
Phenomena such as this are indicative of a level of sequential
organization in conversation beyond two-turn sequencing (see
Schegloff 2007).
Ronald Geluykens
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Sacks, Harvey. 1992. Lectures on Conversation. Oxford: Blackwell.
Sacks, Harvey, Emanuel A. Schegloff, and Gail Jefferson. 1974. A simplest systematics for the organization of turn-taking in conversation.
Language 50: 696735.
Schegloff, Emanuel A. 2007. The Language of Turn and Sequence.
Cambridge: Cambridge University Press.
AGE GROUPS
Age is one of the primary independent variables in
sociolinguistics, along with social class, sex, ethnicity, and
region. Age is the primary social correlate for determining that
language is changing and for estimating the rate of change and
its progress.
88
The study of the progress of change can be carried out in real time
by revisiting survey sites at intervals and observing changes in the
social distribution of variants from one time to the next. Because
changes will not necessarily be completed in the interval but will
be continuing, their progress must be calculated by some kind
of proportional measure, such as the percentage of the variants,
their relative frequency, or their probabilities. The proportional
differences from one visit to the next provide a quantitative measure of the progress of the change.
Studying change in real time has a number of methodological disadvantages. Most obvious is the time required before
attaining a result. Locating subjects on subsequent visits also
poses obvious problems because of mobility, cooperation, or
death. Furthermore, subsequent visits require the addition of
new subjects at the youngest age group each time. This is necessary because of the manner in which linguistic innovations
typically diffuse (see diffusion ) throughout the population.
Instead of spreading outward from the source and affecting
almost everybody in their sphere of influence, as infectious
diseases do in epidemics and technological adoptions do
when, say, spin dryers replace washboards, linguistic innovations tend to be stratified. Under ordinary circumstances,
people acquire their accents and dialects in their formative
Age groups
100
90
80
70
60
50
40
30
20
10
Change in Progress
Correlating linguistic variants with their use by age groups in
the community as the change is taking place is a way of measuring its progress. Inductively, change is evident when a linguistic
variant occurs with greater frequency in the speech of younger
people than in the speech of their elders. There are some exceptions (such as age grading), but the inference of change can be
made with reasonable confidence when the frequency of the
variant is stratified from one age group to the next (Labov 2001).
That is, linguistic changes are almost never bimodal, with one
variant occurring in the speech of younger people and a different one in the speech of older people. Instead, the variants are
typically dispersed along the age continuum in a progressive
gradation.
Figure 1 provides a case study. The variable is the past tense of
the verb sneak, with variants sneaked and snuck. The community
is the Golden Horseshoe, the densely populated region in southern Ontario, Canada. Figure 1 shows the percentage of people
who say snuck, not sneaked, correlated with their age, from octogenarians to teenagers. The correlation shows a progression from
18 percent in the oldest group to 98 percent in the youngest, with
gradation in the intermediate age groups (29 percent of 70s, 42
percent of 60s, and so on).
Other things being equal, it is possible to draw historical
inferences from apparent-time displays like Figure 1. The survey from which the data are drawn took place in 1992. Among
people born 80 or more years prior, that is, before 1913,
sneaked was the standard variant and snuck had very little
1419
2029
3039
4049
5059
6069
7079
0ver 80
89
Age groups
Apparent-Time Hypothesis
Figure 1 draws historical inferences of change based on the
behavior of age groups surveyed at the same time. The replacement of sneaked by snuck is not directly observed as it would be
in real-time studies, in which researchers go to communities at
intervals and track the changes from one time to the next. Instead,
the inference of change is based on the assumption that under
normal circumstances, people retain the accents and dialects
acquired in their formative years. That assumption is known as
the apparent-time hypothesis. Common experience tells us that
the hypothesis is not without exceptions. People are sometimes
aware of expressions that they once used and no longer do, and
sometimes they will have changed their usage after their formative years. If such linguistic capriciousness took place throughout
the entire community, it would invalidate historical inferences
drawn from apparent-time surveys.
However, community-wide changes beyond the formative
years are rare. Real-time evidence, when available, generally
corroborates apparent-time inferences. In the case of sneaked/
snuck, for instance, earlier surveys made in the same region in
1950 and 1972 show proportional distributions of the variants
that approximate the apparent-time results. However, inferring
linguistic changes from the speech of contemporaneous age
groups is not a direct observation of that change. It remains a
hypothesis, and its validity must be tested wherever possible.
Age-Graded Changes
Deeper understanding of linguistic change in progress should
ultimately lead to predictable classes of deviation from the
apparent-time hypothesis. One known deviation is age-graded
change. These are changes that are repeated in each generation,
usually as people reach maturity (Chambers 2009, 200206).
Age-graded changes are usually so gradual as to be almost
imperceptible, so that tracking their progress is challenging.
As an example, in all English-speaking communities, there is a
rule of linguistic etiquette that requires compound subject noun
phrases (NPs) to list the first-person pronoun (the speaker) last.
Adults say Robin and I went shopping, and never say I and
Robin went shopping. There is no linguistic reason for this rule
(that is, the sentences mean the same thing either way), but
putting oneself first is considered impolite (see politeness).
Children, however, do not know this and tend to say Me and
Robin went shopping. At some as-yet-undetermined age, children become aware of the rule and change their usage to conform to adult usage.
Age-graded changes like these violate the apparent-time
hypothesis because the variants used by young people do not
persist throughout their lifetimes. Instead, young people change
as they reach maturity and bring their usage into line with adults.
The occurrence of age-graded changes does not refute the apparent-time hypothesis, but they provide a well-defined exception
to it. Failure to recognize them as age-graded can lead to an erroneous inference that change is taking place.
90
Naming
A common complaint among the healthy aging population is the
increased frequency of word-finding problems in their everyday
speech. A subset of these naming problems is often colloquially described as the tip-of-the-tongue (TOT) phenomenon. In
terms of cognitive models of lexical access, TOTs are described
as a type of word retrieval failure whereby individuals are able to
access the conceptual, semantic, syntactic, and even some
phonological/orthographic information (e.g., number of
syllables or initial sound/letters) of the target word, not enough
information, however, to fully phonologically encode the word
that older adults benefit more from context in noise than do their
younger cohorts.
Lexical comprehension is one aspect of language that is
largely preserved or even enhanced with age. Studies on vocabulary comprehension in older adults using a picture selection task
for auditorily presented words show that they are comparable to
younger adults in this task (Schneider, Daneman, and PichoraFuller 2002). However, lexical-semantic integration at the level of
sentence comprehension may be affected in older adults.
Sentence comprehension in the elderly is known to be poor
in comparison to younger listeners. A number of reasons, both
linguistic and cognitive, have been discussed, including decline
in auditory perceptual skills and lexical-semantic and syntactic
processing capacity, as well as working memory capacity, speed
of processing, and ability to process with competing stimuli and
inhibit noise (Wingfield and Stine-Morrow 2000). Studies on syntactic comprehension of language in older adults demonstrate
that they are slower at judging sentences that are syntactically
complex and semantically improbable (Obler et al. 1991). It has
also been found to be relatively more difficult for this population
than for younger adults to take advantage of constraining context
in a sentence. This leads to difficulties in situations where older
adults need to rely on context but are not able to do so skillfully.
However, in other aspects of sentence comprehension such as
disambiguation, no age-related differences has been reported.
It is apparent that sentence comprehension is largely mediated by the ability to hold these sentences in working memory
while they are being processed. Working-memory decline has
been reported in older adults (Grossman et al. 2002). Moreover,
syntactic-processing difficulties have also been attributed to
reduction in working-memory capacity (Kemptes and Kemper
1997). Executive functions such as inhibition and task switching
have been reported as negatively affected in the elderly.
Comprehension problems eventually affect the older adults
discourse abilities. Often, the elderly find that it gets harder to
follow discourse with advancing age. Studies have shown that
older adults are significantly poorer than younger adults in fully
understanding and inferring complex discourse, and this difficulty is enhanced further with increased perceptual or cognitive
load (Schneider, Daneman, and Pichora-Fuller 2002), such as
noise and length or complexity of the material. A combination of
general cognitive decline and a deterioration of specific linguistic and sensory-perceptual processes contribute to the general
slowing observed in the elderly while they engage in conversation and discourse.
Language Comprehension
It is relatively well known that comprehension in older adults
is compromised in comparison to younger adults. These problems in comprehension may arise from individual or combined
effects of decline in their sensory/perceptual, linguistic, or cognitive domains. Current research is focused on disambiguating
the effects of these processes on phonological, lexical-semantic,
and syntactic aspects of language decline in the elderly.
Research shows that speech perception in older adults is especially affected by noise. This means that even with normal hearing, older adults experience difficulties understanding speech
(e.g., sentences) under noisy conditions such as a cocktail party
or cafeteria. Experiments focusing on sentence perception show
Spoken Discourse
Results from many types of tasks requiring sentence production, such as cartoon description tasks (Marini et al. 2005), have
indicated that older adults tend to produce grammatically simpler, less informative and more fragmented sentences than do
younger adults. Middle-aged and young elderly adults tend to
be better and more efficient story constructors than younger
or older adults. However, older adults usually produce a larger
number of words in their narrative speech, but they can show
difficulties in morpho-syntactic and lexico-semantic processing
by making more paragrammatic errors and semantic paraphasias than younger adults.
91
Dementia
Difficulties in language production, as well as comprehension,
become obvious in dementia. Word-finding difficulties are characteristic of mild cognitive impairment (MCI), Alzheimers disease (AD), and vascular dementia (VaD).
Arto Nordlund and colleagues (2004) indicated that 57.1 percent of the individuals with MCI had significantly lower scores
in different language tasks than did typical aging adults. In AD,
language-processing difficulties are early signs of the disease.
In particular, AD appears to cause a breakdown of the semantic domain of language, which can be reflected in the impaired
comprehension and use of semantic relations between words
(Garrard et al. 2001) and in the reduced semantic noun and verb
fluency performance (Pekkala 2004) in this population.
Comparative studies between AD and VaD have indicated that
cognitive and linguistic performance cannot clearly differentiate
the two types of dementia from each other. Elina Vuorinen, Matti
Laine, and Juha Rinne (2000) indicated that both AD and VaD
involved similar types of semantic deficits early in the disease,
including difficulties in comprehension, naming, and production of semantic topics in narrative speech, while word repetition, oral reading, and fluency of speech output were preserved
in both types of dementia.
Yael Neumann, Seija Pekkala, and Hia Datta
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Brown, Alan S. 1991. A review of the tip-of-the-tongue experience.
Psychological Bulletin 109.2: 20423.
Burke, Deborah M., Donald G. MacKay, Joanna S. Worthley, and
Elizabeth Wade. 1991. On the tip of the tongue: What causes word
finding failures in young and older adults? Journal of Memory and
Language 30: 54279.
92
Agreement
Agreement Maximization
AGREEMENT
( ) person
pl. number
( ) number
m. gender
AGREEMENT MAXIMIZATION
Maximizing, or optimizing, agreement between speaker and
interpreter is part of the principle of charity (see charity, principle of) that, according to philosophers of language in the
tradition of W. V. O. Quine and Donald Davidson, governs the
interpretation of the speech and thought of others and guides
the radical interpretation of a radically foreign language.
According to Davidson, correct interpretation maximizes truth
and coherence across the whole of the beliefs of a speaker. For an
interpreter, maximizing truth across the beliefs he ascribes to a
speaker necessarily amounts to maximizing agreement between
himself and the speaker: He can only go by his own view of what
is true.
Take one of Davidsons own examples: Someone says There
is a hippopotamus in the refrigerator, and he continues: Its
roundish, has a wrinkled skin, does not mind being touched.
It has a pleasant taste, at least the juice, and it costs a dime. I
squeeze two or three for breakfast (1968, 100). The simplest way
of maximizing agreement between this speaker and us will probably be interpreting his expression hippopotamus as meaning
the same as our expression orange. Davidson himself, however,
soon came to consider maximizing agreement as a confused
ideal (1984, xvii) to be substituted with optimizing agreement
(1975, 169). The idea here is that some disagreements are more
destructive of understanding than others (1975, 169). Very
generally, this is a matter of epistemic weight; the more basic a
belief is, and the better reasons we have for holding it, the more
destructive disagreement on it would be: The methodology of
interpretation is, in this respect, nothing but epistemology seen
in the mirror of meaning (1975, 169). According to Davidson, it
is impossible to codify our epistemology in simple and precise
form, but general principles can be given: [A]greement on laws
and regularities usually matters more than agreement on cases;
agreement on what is openly and publicly observable is more
to be favored than agreement on what is hidden, inferred, or ill
observed; evidential relations should be preserved the more they
verge on being constitutive of meaning ([1980] 2004, 157).
Agreement optimization does not exclude the possibility of
error; speakers are to be interpreted as right only when plausibly possible (Davidson 1973, 137). In certain situations this
prevents the interpreter from ascribing beliefs of his own to
the speaker, for instance, perceptual beliefs about objects the
speaker is in no position to perceive. Moreover, if the speaker
has other beliefs that provide him or her with very good reasons
for believing something false, optimizing agreement across all
93
Agreement Maximization
Alliteration
94
ALLITERATION
Linguistically, alliteration, also known as initial or head rhyme,
is defined as the selection of identical syllable onsets within a
specific phonological, morphosyntactic, or metrical domain.
It is usually coupled with stress, as in the three Rs in education: reading, writing, and arithmetic. Etymologically, the term
alliteration (from L. ad- to + littera letter) includes the repetition of the same letters at word beginnings; its dual association
with sounds and letters reflects a common cognitive crisscrossing between spoken and written language in highly literate
(see literacy) societies well illustrated by the famous phrase
apt alliterations artful aid, where the alliteration is primarily
orthographic.
Alliteration based on the sameness of letters is found in visual
poetry, advertising, and any form of playful written language.
Phonologically based alliteration is a frequent mnemonic and
cohesive device in all forms of imaginative language: Examples
from English include idioms (beat about the bush), reduplicative
word-formation (riffraff), binominals (slowly but surely), catch
phrases, refrains, political slogans, proverbs, and clichs. In
verse, alliteration serves both as ornamentation and as a structural device highlighting the metrical organization into feet, cola,
verses, and lines. Along with rhyme, alliteration is a common feature of folk and art verse in languages as diverse as Irish, Shona,
Mongolian, Finnish, and Somali.
The most frequent type of alliteration requires identity of the
onsets of stressed syllables, which makes it a preferred poetic
device in languages with word-initial stress, such as the older
Germanic languages. Within the Germanic tradition, metrically
relevant alliteration occurs on the stressed syllables of the first
foot of each verse (or half-line), where it is obligatory. For Old
English, the language of the richest and most varied surviving alliterative poetry in Germanic, the second foot of the first
half-line may also alliterate. Alliteration is disallowed on the last
stressed syllable in the line.
In Old English verse, alliteration appears with remarkable regularity: Only 0.001% of the verses lack alliteration and
less than 0.05% contain unmetrical alliteration (Hutcheson
1995, 169). Alliteration is, therefore, a reliable criterion used by
Ambiguity
modern editors to determine the boundaries of verses and lines,
though no such divisions exist in the manuscripts. The reinvented alliterative tradition of fourteenth-century England also
uses alliteration structurally, while its ornamental function is
enhanced by excessive verse-internal and run-on alliteration.
As a cohesive device in verse, alliteration refers to the underlying distinctions in the language and relies on identity of phonological categories. The interpretation of onset identity for the
purpose of poetic alliteration varies from tradition to tradition
and can include whole clusters, optionally realized segments,
and even the whole syllable up to the coda. In Germanic, all
consonants alliterated only with one another, the clusters st-,
sp-, sk- could not be split, and all orthographic stressed vowels
alliterated freely among themselves, most likely because their
identity was signaled by the presence of a prevocalic glottal stop
in stressed syllable onsets.
Donka Minkova
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Fabb, Nigel. 1997. Linguistics and Literature: Language in the Verbal Arts
in the World. Oxford: Blackwell.
Hutcheson, Bellenden Rand. 1995. Old English Poetic Metre.
Cambridge: D. S. Brewer.
Minkova, Donka. 2003. Alliteration and Sound Change in Early English.
Cambridge: Cambridge University Press.
AMBIGUITY
Ambiguity refers to the potential of a linguistic expression to have
more than one meaning. Although many expressions (words,
phrases, and even sentences) are ambiguous in isolation, few
remain so when used in a particular context. In fact, people typically resolve all ambiguities without even detecting the potential
for other interpretations. Ambiguity does not imply vagueness;
rather, ambiguity gives rise to competing interpretations, each of
which can be perfectly concrete. Although ambiguity is pervasive
and unavoidable in natural languages, artificial languages developed for mathematics, logic, and computer programming strive
to eliminate it from their expressions.
Ambiguity can be lexical, structural, referential, scopal,
or phonetic. The examples of these phenomena that follow
include well-known classics in English.
Lexical ambiguity refers to the fact that some words, as written or spoken, can be used in different parts of speech (see
word classes) and/or with different meanings. For example,
duck can be used as a noun or a verb and, as a noun, can refer to
a live animal or its meat. Structural ambiguity arises when different syntactic parses give rise to different interpretations. For
example, in addition to being lexically ambiguous, They saw her
duck is also structurally ambiguous:
1. They saw [NP her duck] (the bird or its meat belongs to her)
2. They saw [NP her] [VP duck] (the ducking is an action she carries
out)
A common source of structural ambiguity involves the attachment site for prepositional phrases, which can be at the level
of the nearest noun phrase (NP) or the clause. In the sentence
Danny saw the man with the telescope, either Danny used the
telescope to help him see the man (3) or the man whom Danny
saw had a telescope (4).
3. Danny [VP saw [NP the man] [PP with the telescope]]
4. Danny [VP saw [NP the man [PP with the telescope]]]
95
Amygdala
for sentences to contain upward of 20 words, in which case there
is still the threat of combinatorial explosion.
Marjorie J. McShane
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Cruse, D. A. 1986. Lexical Semantics. Cambridge: Cambridge University
Press. Includes features of and tests for ambiguity.
Small, Steven, Garrison Cottrell, and Michael Tanenhaus, eds. 1988.
Lexical Ambiguity Resolution: Perspective from Psycholinguistics,
Neuropsychology and Artificial Intelligence. San Mateo, CA: Morgan
Kaufmann.
Zwicky, Arnold M., and Jerrold M. Sadock. 1975. Ambiguity Tests and
How to Fail Them. In Syntax and Semantics, ed. J. Kimball, IV: 136.
New York: Academic Press. Discusses tests to distinguish ambiguity
from lack of specification.
AMYGDALA
Studies in animals have established a clear role for the amygdala
in social and emotional behavior, especially as related to fear
and aggression (Le Doux 1996). Human studies, including lesion
studies, electrophysiology, and functional neuroimaging,
have further elucidated the role of the amygdala in the processing of a variety of emotional sensory stimuli, as well as its relationship to behavioral and cognitive responses (Adolphs 2001).
These responses not only guide social behavior but also aid in the
acquisition of social knowledge. The focus of this entry is on the
amygdala and its role in the processing of language, in particular language relevant to social and emotional behavior (see also
emotion and language and emotion words).
The amygdala is an almond-shaped group of neurons located
in the rostral medial temporal region on both left and right sides
of the brain (see left hemisphere language processing and right hemisphere language processing). It
has reciprocal connections to regions, such as the hypothalamus, that are important for coordinating autonomic responses
to complex environmental cues for survival, as well as premotor and prefrontal areas that are necessary for rapid motor and
behavioral responses to perceived threat. Visual, somatosensory, and auditory information is transmitted to the amygdala
by a series of indirect, modality-specific thalamocorticoamygdalar pathways, as well as by direct thalamoamygdalar pathways.
Within the amygdaloid complex, information processing takes
place along numerous highly organized parallel pathways with
extensive intraamygdaloid connections. The convergence of
inputs in the lateral nucleus enables stimulus representations
to be summated. Specific output pathways from the central
nucleus and amygdalohippocampal area mediate complementary aspects of learning and behavioral expressions connected
with various emotional states. The amydgala is thus well positioned to play a role in rapid cross-modal emotional recognition.
It is important for the processing of emotional memory and for
fear conditioning.
In addition, anatomical studies of the primate amygdala demonstrate connections to virtually all levels of visual processing in
the occipital and temporal cortex (Amaral 2003). Therefore,
the amygdala is also critically placed to modulate visual input,
based on affective significance, at a variety of levels along the
96
Language
In order to survive in a changing environment, it is especially
important for the organism to remember events and stimuli
that are linked with emotional consequences. Furthermore, it is
important to be vigilant of emotional stimuli in the environment
in order to allow for rapid evaluation of and response to these
emotional stimuli. In humans, emotional cues are transmitted
linguistically, as well as through body posture, voice, and facial
expression (see gesture).
In the first imaging study to examine language and the
amygdala, a modified Stroop task was utilized, along with a
high-sensitivity neuroimaging technique, to target the neural substrate engaged specifically when processing linguistic
threat (Isenberg et al. 1999). Healthy volunteer subjects were
instructed to name the color of words of either threat or neutral
valence, presented in different color fonts, while neural activity
was measured by using positron emission tomography. Bilateral
amygdalar activation was significantly greater during color naming of threat words than during color naming of neutral words
(see Color Plate 1). Associated activations were also noted in
sensory-evaluative and motor-planning areas of the brain. Thus,
our results demonstrate the amygdalas role in the processing
of danger elicited by language. In addition, the results reinforce
the amygdalas role in the modulation of the perception of, and
response to, emotionally salient stimuli. This initial study further
suggests conservation of phylogenetically older mechanisms of
emotional evaluation in the context of more recently evolved linguistic function. In a more recent study that examines the neural
substrates involved when subjects are exposed to an event that is
verbally linked to an aversive outcome, activation is observed in
the left amygdala (Phelps et al. 2001, 43741).
This activation correlated with the expression of the fear
response as measured by skin conductance response, a peripheral measure of arousal. The laterality of response may relate to
the explicit nature of the fear, as well as to the fact that the stimulus is learned through verbal communication. Fears that are
simply imagined and anticipated nonetheless have a profound
impact on everyday behavior. The previous study suggests that
the left amygdala is involved in the expression of fear when anticipated and conveyed in language. Another study that sought to
examine the role of the amygdala in the processing of positive as
well as negative valence verbal stimuli also demonstrated activity in the left amygdala (Adolphs, Baron-Cohen, and Tranel 2002,
126474). During magnetic resonance (MR) scanning, subjects
viewed high-arousal positive and negative words and neutral
words. In this study, activity was found in the left amygdala while
subjects viewed both negative and positive words in comparison
to neutral words. Taken together, these studies suggest that the
amygdala plays a role in both positive and negative emotional
responses. Furthermore, they suggest that the left amygdala may
be preferentially involved in the processing of emotion as conveyed through language.
Lesion studies have generally suggested that the amygdala
is not essential for recognizing or judging emotional and social
Analogy
information from explicit, lexical stimuli, such as stories (see
ANALOGY
Two situations are analogous if they share a common pattern of
relationships among their constituent elements, even though
the elements are dissimilar. Often one analog, the source, is more
familiar or better understood than the second analog, the target
(see source and target). Typically, a target situation serves
as a retrieval cue for a potentially useful source analog. A mapping, or set of systematic correspondences aligning elements
of the source and target, is then established. On the basis of the
mapping, it is possible to derive new inferences about the target.
Psychological Research
Within psychology, work in the intelligence tradition focused on
four-term or proportional analogies, such as ARM: HAND :: LEG: ?
Charles Spearman (1946) reviewed studies that found high correlations between performance in solving analogy problems and
the g factor (general intelligence). The ability to solve analogylike problems depends on a neural substrate that includes subareas of the prefrontal cortex (Bunge, Wendelken, and Wagner
2005; see frontal lobe). Although there have been reports of
great apes being successfully trained to solve analogy problems,
these results are controversial (Oden, Thompson, and Premack
2001). Complex relational thinking appears to be a capacity that
emerged in homo sapiens along with the evolutionary increase
in size of the frontal cortex. The ability to think relationally
increases with age (Gentner and Rattermann 1991). Greater
sensitivity to relations appears to arise with age due to a combination of incremental accretion of knowledge about relational
concepts (Goswami 1992), increases in working memory
capacity (Halford 1993), and increased ability to inhibit misleading featural information (Richland, Morrison, and Holyoak
2006). Analogy plays a prominent role in teaching mathematics
(Richland, Zur, and Holyoak 2007).
Dedre Gentner (1983) developed the structure-mapping theory of analogy, emphasizing that analogical mapping is guided
by higher-order relations relations between relations. Keith
Holyoak and P. Thagard (1989) proposed a multiconstraint theory, hypothesizing that people find mappings that maximize
similarity of corresponding elements and relations, structural
parallelism, and pragmatic importance for goal achievement.
Several computational models of human analogical thinking
have been developed. Two influential models are SME (Structure
Mapping Engine; Falkenhainer, Forbus, and Gentner 1989),
based on a classical symbolic architecture, and LISA (Learning
and Inference with Schemas and Analogies; Hummel and
Holyoak 2005), based on a neural-network architecture. LISA has
been used to simulate some effects of damage to the frontal and
temporal cortex on analogical reasoning (Morrison et al. 2004).
97
98
b2, b3 (etc.). There is some relation R between a1, a2, and a3,
expressed as R(a1,a2,a3), just as there is another such relation
S between b1, b2, and b3, expressed as S(b1,b2,b3). For A and
B to be analogous, it is required that R and S be exemplifications of the same abstract structure X, as evidenced by a mapping between a1/a2/a3 and b1/b2/b3. This is what is meant by
saying that analogy (e.g., between A and B) is a structural similarity, or a similarity between relations (e.g., R and S). It is not
a material similarity, or a similarity between things (e.g., a1
and b1). More and more abstract analogies are constituted by
similarities between similarities between relations between
things.
In its purely synchronic use (see synchrony and diachrony), analogy is understood to be the centripetal force that
holds the units of a structure together. To simplify an example
given by N. S. Trubetzkoy ([1939] 1958, 6066), in a structure
containing just /p/, /b/, /t/, and /d/, the phoneme /p/ acquires
the distinctive features voiceless and bilabial by being contrasted
with, respectively, (voiced) /b/ and (dental) /t/. The relation
between the pairs /p/ & /b/ and /t/ & /d/ is the same, and so is
the relation between /p/ & /t/ and /b/ & /d/, which means by
definition that there is in both cases an analogy between the
two pairs. This type of analogy-based analysis applies to any wellarticulated structure, linguistic or nonlinguistic: A unit is what
the other units are not (as /p/ is neither /b/ nor /t/ nor, of course,
/d/); and this otherness is based on corresponding oppositions
(like voiceless vs. voiced and bilabial vs. dental).
Synchronic analogy may be characterized as analogy-asstructure. Its counterpart is analogy-as-process,that is, discovery,
manipulation, or invention of structural similarity. Traditionally,
language acquisition was thought to be based on analogy: From
innumerable sentences heard and understood [the child] will
abstract some notion of their structure which is definite enough
to guide him in framing sentences of his own (Jespersen [1924]
1965, 19). After a period of neglect, this traditional view has
again become fashionable in some quarters (Pinker 1994, 417;
Tomasello 2003, 1639).
Only if analogy-as-process leaves a permanent trace that deviates from the current norm is there reason to speak of language
change, the province of diachronic linguistics. Traditionally, the
term analogical change was restricted to morphology, or to cases
where irregularities brought about by sound change are eliminated so as to achieve, or to approach, the goal of one meaning
one form. However, this same goal is involved in such large-scale
changes as have generally been ascribed to a need for harmony
or symmetry. In syntactic change, analogy consists in extending a
reanalyzed structure to new contexts (Anttila 1989, 1024).
Esa Itkonen
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Anttila, Raimo. 1989. Historical and Comparative Linguistics.
Amsterdam: Benjamins.
Jespersen, Otto. [1924] 1965. Philosophy of Grammar. London: Allen and
Unwin.
Pinker, Stephen. 1994. The Language Instinct. New York: Morrow.
Tomasello, Michael. 2003. Constructing a Language. A Usage-Based
Theory of Language-Acquisition. Cambridge: Harvard University
Press.
Analyticity
Trubetzkoy, N. S. [1939] 1958. Grundzge der Phonologie. Gttingen:
Vandenhoeck and Ruprecht.
ANALYTICITY
Analyticity is a property that a statement has when its truth is
in some special way determined by its meaning. Many believe
that no such property exists. Led by W. V. O. Quine (1953), philosophers complain that no one has been able to define the concept of analyticity in a way that is precise and that also fulfills the
purposes to which it is typically put.
To some people, it seems that all uncles are either married or have a sibling is true just because of the meaning of its
constituent words, most prominently because of the meaning
of uncle. But all uncles are less than eight feet tall is true not
because of meanings but because of how the world has turned
out to be. The first sort of statement is said to be analytic, the
second synthetic.
This distinction has far-reaching interest and application.
Empiricists have always had difficulty accounting for the seemingly obvious fact that the truths of logic and mathematics are
both necessary (i.e., they could not be false) and a priori (i.e., they
are known independently of sensory experience). For most of the
twentieth century, it was agreed that analyticity could explain
this obvious fact away as a merely linguistic phenomenon.
The idea that all necessity could be explained away by analyticity fell out of fashion when S. Kripke (1980) convinced most
philosophers that some necessary truths are neither analytic nor
a priori (e.g., water = H2O). But as L. BonJour (1998, 28) points
out, many still think that the unusual modal and epistemic status of logic and mathematics is due to a special relation between
truth and meaning.
However we apply the concept, trouble for analyticity begins
when we remind ourselves that every truth depends, to some
extent, on the meanings of its constituent terms. If the word
uncle had meant oak tree, then both previous examples would
be false. In response, it is said that an analytic truth is one whose
truth depends solely on the meanings of its terms. Our linguistic
conventions alone make the first example true, and the second is
true party because of meaning and partly because of the way the
world is. But how can linguistic convention alone make something true?
We can distinguish the sentence all uncles are either married or have a sibling from the proposition that this sentence
now expresses. Meaning or linguistic convention alone makes
this sentence true in the following way. Given that this sentence
expresses the proposition that it does (i.e., given our current linguistic conventions), it is true. This cannot be said of our other
example. That this sentence means what it does is not sufficient
to determine its truth or falsity (see also sentence meaning).
The world plays a part.
There are three serious problems. First, if this is what it is for
a sentence to be true solely in virtue of its meaning, then it is just
another way of saying that it expresses a necessary truth, and that
tells us nothing about how we know it. Thus, appeal to analyticity
cannot explain the necessity and a priority of logic, mathematics, or anything else. Second, the proposition now expressed
by our first example would be true no matter how we ended up
99
Anaphora
ANAPHORA
Languages have expressions whose interpretation may involve
an entity that has been mentioned before: Subsequent reference to
an entity already introduced in discourse approximates a general
definition of the notion of anaphora (Safir 2004, 4). This works
well for core cases of nominal anaphora as in (1) (Heim 1982):
(1) a. This soldier has a gun. Will he shoot?
b. Every soldier has a gun. ??Will he shoot?
However, (1) versus (2) shows that two different modes of interpretation must be distinguished: i) directly assigning two (or
more) expressions the same discourse entity from the interpretation domain (ID) as a value: co-reference as in (1a), and ii)
interpreting one of the expressions in terms of the other by grammatical means: binding (Reinhart 1983). This contrast is represented in (3).
Coreference in (3a) is restricted in terms of conditions on discourse entities, binding in (3b) in terms of grammatical configuration. Expr1 can only bind expr2 if it c-commands the latter
(Reinhart 1983). This condition is met in (2), but not in (1b),
hence the contrast.
Virtually all languages have words and expressions that are referentially defective they cannot be used deictically (see deixis).
In much of the linguistic literature, these are called anaphors, as
they appear specialized for anaphoric use. Examples vary from
English himself, Dutch zich(zelf), Icelandic sig, Russian sebja,
Chinese (ta) ziji, to Georgian tav tavis, and so on. Such expressions cannot be assigned a discourse value directly. Rather, they
must be bound, often in a local domain approximately the
domain of the nearest subject but subject to variation in terms
of specific anaphor type and language. Furthermore, under
100
Animacy
restrictive conditions, some of these expressions may yet allow a
free logophoric interpretation.
It is an important task to arrive at a detailed understanding
of the ways in which languages encode interpretive relations
between their expressions and of the division of labor between
the components of the language system involved.
Eric Reuland
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Chomsky, Noam. 1981. Lectures on Government and Binding. Dordrecht,
the Netherlands: Foris.
Heim, Irene. 1982. The semantics of definite and indefinite noun phrases.
Ph.D. diss., University of Massachusetts at Amherst.
Reinhart, Tanya. 1983. Anaphora and Semantic Interpretation.
London: Croom Helm.
Reuland, Eric. 2010. Anaphora and Language Design. Cambridge: MIT
Press.
Safir, Ken. 2004. The Syntax of Anaphora. Oxford: Oxford University
Press.
ANIMACY
Languages often treat animate and inanimate nouns differently.
Animacy can affect many aspects of grammar, including word
order, and verbal agreement. For example, in Navajo, the
more animate noun must come first in the sentence (Hale 1973),
and in some Bantu languages, a more animate object must come
before a less animate object. Verbs are more likely to agree with
more animate nouns (Comrie 1989). Animacy can also affect the
choice of case, preposition, verb form, determiner (article), or
possessive marker (Comrie 1989).
What counts as animate differs cross-linguistically. The grammatical category of animates may include certain objectively
inanimate things, such as fire, lightning, or wind. Languages may
make additional distinctions among pronouns, proper nouns,
and common nouns, or between definites and indefinites, and
these are sometimes viewed as part of an animacy hierarchy by
linguists (Comrie 1989). person and number distinctions may
also be included in an animacy hierarchy in this broader sense.
For example, according to Michael Silverstein (1976), subjects
with features at the less animate end of the Animacy Hierarchy in
(1) are more likely to be marked with morphological case, while
the reverse holds of objects.
(1)
Animacy Hierarchy
1pl > 1sing > 2pl > 2sing > 3human.pl > 3human.sing >
3anim.pl > 3anim.sing > 3inan.pl > 3inan.sing
Dyirbal exhibits this pattern in that only third person subjects have morphological case, whereas all human objects do.
Silverstein (1976) postulates that the function of such differential case marking is to flag less animate subjects and more animate objects to avoid ambiguity. It is interesting to note that
the patterns of such differential animacy marking are far more
complex and diverse cross-linguistically for objects than for subjects (Aissen 2003). This may be traced to a relation between animacy and object shift, which produces an associated change in
case or verbal agreement (Woolford 2000, 2007). The less diverse
animacy effects on subject case, which do not affect agreement
101
102
Aphasia
there is no reason to expect it to be accessible to organisms with
a different biological endowment, anymore than humans are
capable of acquiring, say, the echolocation capacities of bats, a
system that is equally grounded in the specific biology of those
animals.
Conclusion
Human language is often considered as simply one more instantiation of the general class of animal communication systems.
Indeed, like others it appears to be highly species specific.
Although relevant experience is required to develop the system
of any particular language, the overall class of languages accessible to the human learner is apparently highly constrained, and
the process of language learning is more like genetically governed maturation than like learning in general. The structural
characteristics of human language are quite different from those
of other communication systems, and it is the freedom of expression subserved by those distinctive properties that gives language
the role it has in human life. (See also primate vocalizations
and grooming, gossip, and language).
Stephen R. Anderson
Aphasia Syndromes
Aphasia has traditionally been categorized into seven subtypes,
including Brocas, Wernickes, global, anomic, conduction,
transcortical sensory, and transcortical motor. These aphasia
variants are characterized by different patterns of speech fluency,
auditory comprehension, repetition, and naming. Although
patients may be classified as having one type of aphasia in the
early period after a brain injury, this classification may change as
language problems resolve with time and treatment.
Brocas aphasia is characterized by a constellation of symptoms, including slow, halting speech with impaired grammar;
disturbed auditory comprehension for grammatically complex
phrases and sentences; and poor repetition. Word-finding problems and difficulty with reading and writing are common. Motor
speech disorders, such as apraxia of speech, a disorder of articulatory planning or coordination, and dysarthria, an impairment
in muscle strength, tone, or coordination, very often co-occur.
Patients with Brocas aphasia often talk in a series of nouns and
verbs, as is the case in the following sample from a patient describing the picnic scene from the Western Aphasia Battery (WAB):
APHASIA
Aphasia is a language impairment caused by brain injury that
affects speech content, auditory comprehension, reading, and
writing to varying degrees. Mild aphasia may result in occasional
word-finding problems, while more severe forms can cause profound deficits in all language domains. Aphasia differs from
developmental disorders in that it occurs after a brain injury to a
person with otherwise normal language skills.
Typically, aphasia results from damage to the left hemisphere of the brain due to stroke, traumatic brain injury, tumor,
or degenerative neurological disease. Nearly all right-handed
individuals and most left-handers are thought to have core linguistic functions semantics, phonology, syntax, and
morphology lateralized to the left hemisphere, while other
aspects of language, specifically prosody and pragmatics, are
associated with the right hemisphere. Approximately one
million people in the United States are afflicted with aphasia,
which is a prevalence rate similar to that of Parkinsons disease.
Roughly 80,000 more acquire aphasia annually.
Recent advances in the study of language have provided greater
insight into aphasia syndromes. Modern neuroimaging technology has helped refine classic models of language localization,
as well as our understanding of aphasia treatment and recovery.
In particular, brain-imaging techniques such as magnetic resonance imaging (MRI), computerized tomography (CT), positron
103
Aphasia
104
Aphasia
105
Art, Languages of
exposure to salient features of a literary work. Implicit learning,
however, is likely to require many exposures, commonly while
immersed in the culture and language of the source tradition. In
isolating literary universals, then, we may take into account the
degree to which a particular property is likely to have been transported from one tradition to another by learning of either sort,
given the degree of contact between the traditions. For example,
the practice of dramatic performance may be transmitted from
one tradition to another through limited interaction, as this may
be learned through a single exposure. The same point does not
hold for background imagery.
Finally, we may wish to expand our study of cross-cultural
patterns to actual borrowings. Here, too, it is crucial to distinguish different types of influence. We may roughly divide influence into two categories hegemonic and nonhegemonic.
Hegemonic influence occurs when the source tradition has
greater economic power (e.g., in the publication and distribution of literary works), more pervasive control of government
or education, or a higher level of prestige (due, for example, to
military strength), or when it is otherwise in a position of cultural domination over the recipient society. Obvious cases are
to be found in colonialism. Common properties that result
from non-hegemonic influences are not universals themselves.
However, they may tell us something about cross-cultural aesthetic or related propensities. Common properties that result
from hegemonic influences, in contrast, may simply reflect the
effects of power.
Patrick Colm Hogan
WORK CITED AND SUGGESTIONS FOR FURTHER READING
Comrie, Bernard. 1981. Language Universals and Linguistic Typology.
Chicago: University of Chicago Press.
Hogan, Patrick Colm. 2005. Literary universals and their cultural traditions: The case of poetic imagery. Consciousness, Literature, and
the Arts 6.2. Available online at: http://www.aber.ac.uk/cla/archive/
hogan.html.
ART, LANGUAGES OF
Languages of Art, a book by Nelson Goodman (190698), was
first published in 1968, with a second edition in 1976. The present entry focuses solely on this book, which raises interesting
questions about language in a general sense and its role in
aesthetic experience. This entry does not attempt to contextualize Goodmans book relative to his philosophy, for which see
Daniel Cohnitz and Marcus Rosenberg (2006) and Catherine Z.
Elgin (1992); Goodmans later and related Ways of Worldmaking
is also recommended (Goodman and Elgin 1978).
By languages (of art), Goodman means more generally symbol systems; natural language is one of the symbol systems, which
include, for example, musical notation or the symbol system of
cubist painting. Certain characteristics of symbol systems, when
used in an artwork, place cognitive demands on its audience,
which make the artwork good to think (to borrow Claude LviStrausss term). The symbol systems from which artworks are
composed enable us to be exploratory, drawing on our cognitive
(including emotional) resources. This is because artworks are
106
made from symbol systems that have one or more of the symptoms of the aesthetic: syntactic density, semantic density, syntactic repletenesss, and exemplification. These notions are defined
later in this entry.
According to Goodman, A symbol system consists of a symbol scheme correlated with a field of reference ([1968] 1976,143).
Goodmans primary interest in defining a symbol system is to
differentiate the notational from the non-notational schemes,
where notation is a technical notion to which his Chapter 4
is devoted. His concern about notations follows from a concern
about forgeries and fakes and with the fact that some types of
art (such as painting) can be faked while others (such as performance of a specific piece of music) cannot. Where a work is
defined by compliance to a score (i.e., it has a notation), it cannot
be faked; such works are called allographic. Where a work is not
defined by compliance to a score, as in the case of a painting, its
authenticity can be established only by tracing the history of its
production back to its origin, and this permits faking; such works
are called autographic.
A symbol system is built on a symbol scheme, which consists of characters (and usually modes of combination for these
characters). For example, for a natural language, a character is a
class of marks, where marks might include anything from single
sounds or letters up to whole spoken or written texts, as in the letter P, a character that is the class whose members are all the writings-down of the letter P. Symbol systems are either notations or
not notations. If the symbol system is a notation, the characters
of which its scheme is comprised must meet two conditions, as
follows:
(1) For a character in a notation, the members can be
interchanged, where different characters can be true copies of one another; this is called the condition of characterindifference and is true, for example, of letters of the English
alphabet.
(2) Characters in a notation must be finitely differentiated
or articulate; for a mark that does not belong to two characters, it must be theoretically possible to determine that it does
not belong to at least one of them (this is explained further
shortly).
(1, 2) are the two syntactic requirements that define a symbol system as notational.
Characters in a scheme are correlated with things outside
the scheme. For example, the marks that make up a score are
correlated with elements in the performance of the score; the
mark that is a written word is correlated with a pronunciation of
that word; and the mark that is a written word is (also and independently) correlated with that words referent. Goodman uses
the term complies and says that the performance complies
with the score, or the referent, or pronunciation, complies with
the written word. The set of things that comply with an inscription (e.g., the set of things named that can be denoted by a
name) is called the compliance class of the inscription. For the
symbol system to be a notation, it must first include a symbol
scheme that is notational (i.e., that satisfies the two syntactic
conditions), and it must also satisfy three semantic conditions,
as follows.
Art, Languages of
(3) Notational systems must be unambiguous; it must be
clear which object complies with each unique element of the
scheme.
(4) In a notation, compliance classes must be disjoint; for
example, a performance cannot comply with two different
scores.
(5) A notational system must be semantically finitely differentiated; for an object that does not comply with two characters, it must be theoretically possible to determine that the
object does not comply with at least one of them.
The notion of finite differentiation is important both in the
syntax and semantics of notational systems; finite differentiation
is articulation, and its lack constitutes density. As we will see,
though articulation is important for a notational system, density
is more generally important in defining works as aesthetic. Finite
differentiation requires gaps between elements in the system
(between characters, or between compliants); if between two
adjacent elements a third can always be inserted, the scheme
lacks finite differentiation. For example, a scheme lacks finite
differentiation if it has two characters, where all marks not longer than one inch belong to one character and all longer marks
belong to the other, and where marks can be of any length.
Between a mark belonging to the character of marks not longer
than one inch and a mark belonging to the character of longer
marks, it is always (theoretically) possible to have a third that
falls between them (this ever-diminishing between-space is a
kind of Derridean mise-en-abme).
A symbol system is called a notation if it meets the five conditions. Goodman asks whether various types of symbol systems
that have been developed in the arts are notations. (A type of
artistic practice may be non-notational just because no notation
has been developed for it; in principle, notations might be developed for all of them, but in practice they have not been.) A traditional musical score is a character in a notational system. The
compliants of the score are the performances, which collectively
constitute the work of music. Similar comments are made for
Labanotation, a scoring system developed for dance. A literary
work is a character in a notational scheme (but not a character
in a notational system): Like the language from which it is composed, it meets the syntactic requirements for a notation, but not
the semantic requirements. A painting is a work that is in a symbol system that is not notational.
Having developed these notions, Goodman uses them as
a way of defining a representation, a problem raised and not
solved in the first part of the book, where, for example, he argues
that we cannot distinguish a representation by criteria such
as resemblance to the represented object. Representation for
Goodman is distinct from description (i.e., the term representation does not correspond to current cognitive science or linguistic
uses, in which propositions or tree structures are representations). A description uses a symbol scheme that is (syntactically)
articulate, whereas a representation uses a symbol system that is
dense (or uses symbols from a dense part of a symbol scheme).
He distinguishes between two types of dense (representational)
schemes, differentiating a diagram from a picture. His example is
a pair of representations that are visually identical, consisting of a
peaking line, one an electrocardiogram and the other a picture
107
Articulatory Phonetics
. 1988. Reconceptions in Philosophy and Other Arts and Sciences.
London: Routledge.
Goodman, Nelson, and Catherine Z. Elgin. 1978. Ways of Worldmaking.
Indianapolis: Hackett.
ARTICULATORY PHONETICS
Articulatory phonetics is that part of phonetics that studies how
speech is produced by the lips, tongue, velum (soft palate), larynx, and lungs to alter the air pressures and airflows and turbulent
noises in the vocal tract and to create the air spaces that yield the
resonances that differentiate speech sounds. The basic vocabulary
of articulatory phonetics is used for the taxonomic, that is, classificatory, description of speech sounds. For example, the initial
sound in the French word pre, father, would be described as
a voiceless bilabial stop, symbolized using the IPA (International
Phonetic Alphabet) symbol [p]. The initial sound in the English
word pear would be described as a voiceless aspirated bilabial
stop and symbolized using the IPA symbol [ph]. However, besides
this essential taxonomic function, articulatory phonetics studies the mechanisms (e.g., muscular, aerodynamic) that produce
speech and especially the how and why of variability in speech
sounds. The following is a brief overview of the subject; for indepth accounts, the readings listed below should be consulted.
History
The study of speech articulation and the development of a
descriptive terminology has an impressive history, with the first
surviving instance being the Adhyy of the Sanskrit grammarian, Pini (ca. 500 b.p.e.), who gave an articulatory account
of the relatively large sound inventory of Sanskrit. Other notable
achievements in the characterization of speech sounds were
given by many Greek grammarians, notably Dionysius Thrax
(first cent. b.p.e.); the Arab and Persian grammarians al Khalil
Ibn Ahmad and Sbawaihi of the eighth century, who described
the Arabic of their times; the First Grammarian of Iceland
(twelfth cent.); and the work commissioned by and credited to
the Korean King Sejong (fifteenth cent.), which provided not only
an articulatory description of Korean as spoken then but also a
transcription, now the official orthography for Korean, hangul,
which is partially iconic in its representation of how the sounds
are produced.
In Europe, the Baroque and modern eras saw dozens of
proposals for the description of speech sounds, for example,
by John Wilkins, Johan Conrad Amman, William Holder,
Francis Lodwick, Alexander J. Ellis, Robert Nares, Ernst Brcke,
Richard Lepsius, Alexander Melville Bell, Henry Sweet, and Otto
Jespersen. Although there is still some variation in the descriptive terms, works such as Catford (1977) and Maddieson (1984)
have helped to standardize the terminology.
The Basics
Speech articulations enable communication between speakers
and listeners because they create sound; it is the sound transmitted to listeners and the perception of these sounds that are
the ultimate goal in speaking. Descriptions of articulation are
intended to capture the gestures that create these distinctive elements in the speech code.
108
Articulatory Phonetics
If an air pocket is trapped between the tongue and palate or
the tongue dorsum and lips, and the tongue is lowered, a large
negative pressure can be generated, which, when released, can
create quite a loud sound. Such sounds, so-called clicks have
velaric ingressive initiation. They are common in cultures all over
the world as interjections, signals to animals, and so on. However,
they are used as speech sounds to differentiate words only in a
few languages of southern and East Africa. They are very common in the Khoisan languages and a few neighboring Bantu languages. For example, Khoi one is [|ui] where the [|] is a dental
affricated click (the sound often symbolized in Western orthographies as tsk or tut). In the Khoisan languages, clicks can be
freely combined with pulmonic and/or glottalic egressives either
simultaneously or in clusters. Velaric egressives, where a positive pressure is created by upward and forward movement of the
tongue, are not found in any languages lexicon but are used in
some cultures as a kind of exaggerated spitting sound, where the
positive pressure creates a brief bilabial trill.
In general, after the mechanism of initiation is specified for
speech sounds, there are three main categories of terms to further characterize them: place, manner, qualifiers. For example,
the Russian [bratj] to take has in word-final position a voiceless
palatalized dental stop. The manner is stop, the place is dental, and voiceless and palatalized are qualifiers.
Manners
There are two fundamental categories of manners, obstruent and
sonorant, each with subcategories. An obstruent is a sound that
substantially impedes the flow of air in the vocal tract to a degree
that turbulent noise is generated either as continuous frication or
as a noise burst. Obstruents may be stops or fricatives. (Ejectives,
implosives, and clicks are inherent stops.) Sonorants, which do
not impede airflow, are subdivided generally into laterals, glides,
approximants, nasals, and vowels.
Stops present a complete blockage of the airflow, for example, the glottal stop in the name of the Hawaiian island Oahu
[oahu]. A special subclass of stops are affricates, which are stops
with a fricative release, as in the initial and final consonants of
English judge []. Another subclass, often categorized in
other ways, is comprised of trills, for example, the initial sound
of Spanish roja red [roxa].
Fricatives present a partial blockage, but with noise generated due to air being forced through a relative narrow constriction, for example, the velar fricative in Dutch groot large [xot].
Fricatives may be further divided into sibilants (s-like fricatives)
made by an apical constriction on or near the alveolar ridge. The
essential characteristic of this class, as opposed to other nonsibilant fricatives, is relatively loud high-frequency noise that exploits
the small downstream resonating cavity, for example, English
schist a category of rock [st] which contains two different
sibilant fricatives. Nonsibilant fricatives either have no downstream resonating cavity (e.g., the bilabial fricative in Japanese
Fuji name of the famous mountain [u i]) or, like the velar
fricative, are made further upstream of the alveolar region and so
have a longer downstream resonating cavity and, thus, lower frequencies. Presence or absence of voicing also affects fricatives
loudness: Even if paired voiced and voiceless fricatives have the
same degree of constriction, voiceless fricatives will have more
Place
The primary places of articulation of speech sounds, proceeding
from the farthest back place to the farthest forward: glottal, pharyngeal, uvular, velar, palatal, alveolar, dental, and labial. Some
of these places have already been illustrated. Finer place distinctions can easily be made if necessary by appending the prefixes
pre- and post-, and multiple simultaneous constrictions can be
differentiated by concatenating these terms as was done with the
labial-velar glide [w]. In most cases, these anatomical landmarks
on the upper side of the vocal tract are sufficient; if necessary, an
indication of the lower (movable) articulator can be specified, for
example, the voiced labial-dental fricative [v] as in French voir
to see [vw] (as opposed to, say, the voiced bilabial fricative
[] in Spanish cerveza beer [sesa]).
Vowels
The descriptors for vowels deviate from those for consonants.
An imaginary quadrilateral space in the mouth (seen sagittally)
is posited, and vowels are said to have the high point of tongue
at regions in this space whose dimensions vertically are high
mid low and horizontally, front central back. (These may
also have further qualifiers. In French the high front unrounded
vowel contrasts with a high front rounded vowel [i] and with a
109
Articulatory Phonetics
high back rounded vowel [u], e.g., dit (s/he) said [di] vs. du of
the [dy] vs. doux soft [du].) Although ostensibly referring to
anatomical features, it is now accepted that these descriptors
actually correspond to acoustic-auditory features of vowels: The
height dimension correlates inversely with their first formant,
the front dimension correlates directly with their second formant, and unrounded-rounded correlates roughly with their
third formant. It is still technically possible to apply the traditional anatomical-physiological descriptors to vowels, in which
case [i] would be a close palatal vowel with spread lips, and [u] a
close velar vowel with lip rounding, and [] a pharyngeal vowel
with a widely open mouth shape. Other vowels would just be
variants of these with either less constriction or intermediate
places of constriction. There is merit in applying the anatomicalphysiological labels, for example, to explain the Danish dialectal
variant pronunciations [bi] ~ [bi] bee. The latter variant with
the voiceless palatal fricative can arise simply from the vowel terminus being devoiced (since, it will be recalled, the degree of turbulence is determined not only by the degree of closure but also
by the velocity of the airflow, which, in a voiceless vowel, is high
to generate turbulence at the point of constriction).
Beyond Taxonomy
The conventional descriptions of speech just reviewed form the
basis for scientific investigations of considerable sophistication
and with applications in fields as diverse as medicine (especially
speech pathology), man-machine communication, first (and
subsequent) language learning, and phonology. These investigations involve study of more than just the anatomical-physiological
character of speech sounds, but also, as was hinted at in the
preceding discussion, speech aerodynamics, speech acoustics,
speech perception, and neurophonetics. Space allows just one
example in the area of phonology: Medial stops emerge seemingly
out of nowhere in words such as glimpse < gleam + s (nominalizing
110
Artificial Languages
suffix), dempster judge < deem + ster, Thompson > Thom + son,
youngster [jkst] < young [j] + ster. One has to ask where the
medial stop came from in these nasal + fricative clusters, neither
element of which is a stop. The answer emerges when one considers that these speech sounds are opposite in the state of the
oral and velic exit valves. The nasal has all oral exit valves closed
and the velic valve open whereas the fricative has an oral valve
open, and the nasal valve closed. If in the transition between the
nasal and fricative the velic valve should close prematurely, then
all exit valves will be closed and thus a brief epiphenomenal stop
will emerge. (For more examples, see Ohala 1997.)
John Ohala
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Browman, C. P., and L. Goldstein. 1986. Towards an articulatory phonology. In Phonology Yearbook 3, ed. C. Ewan and J. Anderson, 21952.
Cambridge: Cambridge University Press.
Catford, J. C. 1977. Fundamental Problems in Phonetics. Bloomington,
IN: Indiana University Press.
Goldstein, L., and C. P. Browman. 1986. Representation of voicing contrasts using articulatory gestures. Journal of Phonetics 14: 33942.
Hardcastle, W. J., and J. Laver. 1999. The Handbook of Phonetic Sciences.
Oxford: Blackwells.
Huffman, M. K. and R. Krakow, eds. 1993. Nasals, Nasalization and the
Velum: Phonetics and Phonology, Vol. 5. San Diego, CA: Academic
Press.
Ladefoged, P. 1964. A Phonetic Study of West African Languages: An
Auditory-Instrumental Survey. Cambridge: Cambridge University
Press.
Ladefoged, P., and I. Maddieson. 1995. The Sounds of the Worlds
Languages. Oxford: Blackwell.
MacNeilage, P. F., ed. 1983. The Production of Speech. New York: SpringerVerlag.
Maddieson, I. 1984. Patterns of Sounds. Cambridge: Cambridge University
Press.
Ohala, J. J. 1990. Respiratory activity in speech. In Speech Production and
Speech Modelling, ed. W. J. Hardcastle A. Marchal, 2353. Dordrecht,
the Netherlands: Kluwer.
. 1997. Emergent stops. Proc. 4th Seoul International Conference
on Linguistics [SICOL] 1115 Aug 1997: 8491.
Rothenberg, M. 1968. The Breath-Stream Dynamics of Simple-ReleasedPlosive Production. Basel: Karger.
Silverman, D. 2006. A Critical Introduction to Phonology: Of Sound, Mind,
and Body. London and New York: Continuum International Publishing
Group.
Sol, M-J. 2002. Aerodynamic characteristics of trills and phonological
patterning. Journal of Phonetics 30: 65588.
ARTIFICIAL LANGUAGES
An artificial language can be defined as a language, or languagelike system, that has not evolved in the usual way that natural
languages such as English have; that is, its creation is due to
conscious human action. However, this definition leaves open
some questions. For one thing, what do we mean by language
or language-like system? Among the systems of communication that could be, and have been, called artificial languages
are systems of logic, for example, predicate calculus, and computer languages, such as BASIC. However, the functions of these
languages are different from the function of natural languages,
Artificial Languages
which is communication among humans. I, therefore, focus on
artificial languages that have this latter function, for example,
Esperanto. Under the heading of artificial languages, one might
also include languages that have been made up in connection
with novels, films, television programs, and so on, for example,
Klingon (fictional or imaginary languages), or as part of some
other imaginary world, or those that have been created for the
enjoyment of their designer (personal languages).
Some languages (philosophical languages) were designed
to reflect the real world better than natural languages. Some of
the earliest known ideas on artificial languages, from the seventeenth century, involve this type. The terms constructed language (or conlang) and planned language are roughly equivalent
to artificial language (although one could point out that some
natural languages have undergone a degree of planning), while
(international) auxiliary language covers only those languages
intended for international communication (of course, some natural languages are also used for this); many, if not most, artificial
languages have been created for this purpose.
Another question concerns our notions of artificial and natural. On the one hand, many (arguably all) natural languages have
been subjected to some human manipulation. Consider, for example, the long line of English prescriptivists who have tried to eliminate some constructions of the language, or organizations such as
the French Academy, which has attempted to keep some English
words out of French. Although many of these manipulations have
not completely succeeded, they have had some effect, and therefore one could argue that English and French are partly artificial.
On the other hand, many consciously created languages were
built from elements of one or several natural languages and could
thus be considered not entirely artificial. Therefore, the boundary
between natural and artificial languages is not entirely clear.
In fact, a common classification of artificial languages is in
terms of whether they are based on natural languages: a posteriori languages are, while a priori languages are not (the philosophical languages belonging to the second group). That is, a
priori languages are (supposedly) built from scratch, not taking anything from natural languages. This is a simplification, as
few, if any, languages are entirely a priori; many contain both a
posteriori and a priori components. Therefore, the distinction
should, rather, be seen as a spectrum, with languages at different
points having differing ratios of a priori and a posteriori components. Artificial languages consisting of substantial proportions
of both types are called mixed languages.
Esperanto stands far above other artificial languages in terms
of success it has vastly more speakers than any others (and
even some native speakers). It has been claimed to have more
than a million speakers, though some would disagree with such
a large number, and of course, the question hinges on how one
defines a speaker. Only a relatively small number of artificial languages have achieved much of a community of speakers. These
include Volapk, Interlingua, and Ido, the latter being a modified
Esperanto. Many artificial languages were not used by anyone
other than their designer and perhaps several other people. In
fact, a large number of artificial languages were never fully developed, with only incomplete descriptions having been published.
Let us now see some examples of sentences in artificial languages. Because the a priori languages do not (intentionally) use
Aspect
elements from any natural languages, on the surface they may
seem rather strange, as shown by the following examples from the
language Oz (which, in spite of its name, was a serious project):
(1)
ap if-blEn-vOs
he HABITUAL-seldom-study
he seldom studies(Elam 1932, 20)
(2)
ep ip-Qks ap
I PAST-see him
I saw him (ibid.)
ASPECT
Situations unfold over time. When we talk about them, we often
specify how they unfold over time (or not). There are many ways
in which language conveys such temporal information. While
tense specifies location of an event in relation to other points
in time (e.g., past, present, future, pluperfect), aspect specifies
111
Aspect
internal temporal structure of a situation (e.g., whether it is
ongoing or completed). This is important information to convey
for our linguistic communication to be successful, and many languages convey it by various means lexical, grammatical, and/
or pragmatic. English grammatically marks tense (-ed, -s, will),
while Chinese does not, relying instead on lexical and pragmatic
means. English also grammatically marks aspect (progressive
be V-ing), while Hebrew does not.
Grammatical marking of aspect, often encoded in auxiliaries and inflections, is known as grammatical aspect, or viewpoint aspect. It is called viewpoint aspect because it signifies
the speakers viewpoint. When one chooses to say He ran a
mile, one is viewing the situation from outside, disregarding
its internal structure (perfective aspect), while if one says He
was running a mile, the beginning and end of this situation
are disregarded and one is focusing on the internal structure of
this situation (imperfective aspect) (Comrie 1976; Smith 1997).
The former is often used to push the narrative storyline forward
(foreground), while the latter is associated with background
information (Hopper 1979).
Equally important in conveying aspectual information is lexical aspect also known as inherent (lexical) aspect, situation
aspect (or situation type), aktionsart, event type, and so on. This
is defined by the temporal semantic characteristic of the verb
(and its associated elements) that refers to a particular situation.
Although there are numerous proposals, the most well known is
the classification proposed by Zeno Vendler (1957):
Achievement: that which takes place instantaneously, and is
reducible to a single point in time (e.g., recognize, die, reach
the summit)
Accomplishment: that which has dynamic duration, but has a
single clear inherent endpoint (e.g., run a mile, make a chair,
walk to the store)
Activity: that which has dynamic duration, but with an arbitrary endpoint, and is homogeneous in its structure (e.g., run,
sing, play, dance)
State: that which has no dynamics, and continues without
additional effort or energy being applied (e.g., see, love, hate,
want)
Lexical aspect has proved to be important in linguistic analysis, acquisition, and processing of aspect. In linguistic analysis,
Carlota S. Smith (1997) proposed the two-component theory, a
system in which the aspectual meaning of a sentence is determined by the interaction between lexical aspect and grammatical aspect. For example, imperfective aspect (e.g., progressive in
English) takes the internal view, and, therefore, it is compatible
with durative predicates of activity and accomplishment and
yields progressive meaning. In contrast, achievement, since it is
nondurative, is not so compatible with imperfective aspect, and
such pairing is often anomalous (e.g., *He is noticing the error) or
results in preliminary stages meaning (e.g., He is dying).
In acquisition, this interaction of lexical and grammatical
aspect has been observed since the 1970s. Cross-linguistically,
when children acquire (perfective) past tense marking, they
show strong association between telic verbs (achievements and
accomplishments), between general imperfective marking (such
112
Auditory Processing
Yap, Foong Ha, Patrick Chun Kau Chu, Emily Sze Man Yiu, Stella Fay
Wong, Stella Wing Man Kwan, Stephen Matthews, Li Hai Tan, Ping Li,
and Yasuhiro Shirai. 2009. Aspectual asymmetries in the mental representation of events: Role of lexical and grammatical aspect. Memory
& Cognition 37: 58795.
AUDITORY PROCESSING
Auditory processing refers to the cognitive processes that enable
a listener to extract a message from the raw material of a speech
signal. The study of auditory processing draws upon a range of
sources within the linguistic sciences: most notably cognitive
psychology, discourse analysis, phonetics, phonology
and neurolinguistics (see brain and language). It is distinct
from theories of empathetic listening (what makes a good listener) in areas such as counseling.
It was not until the 1960s that a significant body of listening research developed with the advent of more sophisticated
recording equipment and the increased availability of spectrograms to display the physical characteristics of the speech
signal (see acoustic phonetics). The many advances in our
understanding of the skill since then include early work on
phoneme perception by the Haskins Laboratories; the recognition that processing occurs on line rather than waiting for
an utterance to be completed; and insights into how listeners
identify word boundaries in connected speech. Evidence of
the extent to which listeners have to build a message for themselves on the basis of inference has resulted in a sharp move
away from a view of listening as a passive skill and the recognition that a listener actively engages in a process of meaning
construction.
Auditory processing falls into two closely linked operations:
Decoding, where acoustic stimuli are translated into linguistic
units;
Meaning construction, which embellishes the bare meaning
of the utterance by reference to knowledge sources outside
the signal. It also requires listener decisions as to the importance of what has been said and how it is linked to the discourse that preceded it.
The listener thus draws upon four information sources. The first
is perceptual, based on the signal reaching the listeners ear. The
second is linguistic, consisting of the listeners stored knowledge
of the phonology, lexis, and syntax of the language being
spoken. The third is external: drawing upon the listeners knowledge of the world, the speaker, the topic, and the type of situation. A final component is the listeners ongoing model of what
has been said so far.
Decoding
Decoding is principally a matching operation in which evidence
from the signal is mapped onto stored representations in the
listeners mind of the phonemes, words, and recurrent chunks
of a language (see mapping). The process was once represented
as taking place in a linear and bottom-up way, with phonemes
shaped into syllables, syllables into words, and words into
clauses. In fact, listeners appear to draw upon several levels of
representation at once, with their knowledge of higher-level units,
113
Auditory Processing
as more of the word is heard. The correct item is identified when
the words uniqueness point is reached and the cohort is reduced
to one possible match.
However, many words do not have a uniqueness point (the
sequence man might be a complete word or the first syllable of
manner or manager), and there are no consistent gaps between
words in connected speech to mark where boundaries fall.
Locating boundaries (lexical segmentation) is unproblematic
when one is listening to languages that bear fixed stress on the
first, penultimate, or last syllable of the word but becomes an
issue when processing languages with variable stress. Research
suggests that listeners exploit certain prosodic features of these
languages in order to establish the most likely points for words to
begin or end. In English, they take advantage of the fact that the
majority of content words in running speech are monosyllabic or
begin with a strong syllable (Cutler 1990).
A further problem for lexical recognition is that many words
in connected speech (particularly function words) occur in
a reduced form. They might be brief, of low saliency, and very
different from their citation forms. Using the gating method,
which presents connected speech in gradually increasing segments, researchers have demonstrated that word identification
in listening is sometimes a retrospective process, with listeners
unable to identify a word correctly and confidently until two or
more syllables after its offset.
There have also been attempts to model lexical recognition
using connectionist computer programs that analyze spoken
input in brief time slices rather than syllables. Early matches
are continually revised as evidence accumulates across slices
(see connectionist models, language structure, and
representation). The most well known of these programs is
TRACE (McClelland and Elman 1986).
The speed with which a word is identified by a listener is
subject to variation. High-frequency words are more rapidly
retrieved than low-frequency ones and are said to be more easily activated. Recognition is also faster when the listener has
recently heard a word that is closely associated with the target.
Thus, encountering a word like doctor facilitates (or primes; see
priming, semantic ) later recognition of nurse, patient, and
hospital. This process, known as spreading activation,
is highly automatic and is distinct from the normal effects of
context.
Syntactic parsing is the third level. Though speech is received
linearly, syllable by syllable, a listener needs to build larger-scale
syntactic structures from it. Listeners appear to retain a verbatim
record of the words heard until a major syntactic boundary. A
wrap-up process then turns the words into an abstract proposition, and they cease to be available to report. Intonation contours frequently coincide with syntactic units and assist listeners
in locating clause and sentence boundaries.
Where there is ambiguity early in an utterance, a listener
has to carry forward parallel hypotheses about how the utterance will end, with one prioritized and the others held in reserve.
Researchers employ garden path sentences (example: The lawyer questioned by the judge apologized) to establish the criteria that influence the preferred interpretation. Listeners appear
to be swayed by multiple factors, including syntactic simplicity,
semantic probability, and argument structure.
114
Meaning Construction
The outcome of decoding is an abstract proposition, which
represents the literal meaning of the utterance independently of
the context. A listener has to build a more complex meaning representation (or mental model), which
(a) adds to and contextualizes the proposition;
(b) links it conceptually to what has gone before. This operation takes place locally as well as at a discourse level. Listeners
need to make local connections between ideas, associating
pronouns with their referents and recognizing logical connectives (in addition, however). But they also have to carry forward
a developing representation of the whole discourse so far.
Meaning construction embraces several different processes.
Many of them are more cognitively demanding in listening than
in reading because the listener is entirely dependent upon the
mental representation that he/she has built up and cannot look
back to check understanding.
ENRICHMENT. The listener adds depth and relevance to the
proposition by drawing upon external information: knowledge
of the world, the topic, the speaker, and the current situation.
Understanding is also deepened by relating the proposition to
the current topic and to the points made so far by the speaker.
INFERENCE. Listeners supply connections that the speaker does
not make explicitly. They might employ scripts to provide
default components for common activities. If a speaker mentions going to a restaurant, the listener takes for granted a waiter,
a menu, and a conventional procedure for ordering.
SELECTION. Listeners do not simply record facts; they select
some, they omit some, and they store some in reduced form. The
same utterance may result in differently constituted messages
in the minds of different listeners. One important factor is the
listeners perception of the intentions of the speaker. Another is
the listeners own purpose for listening. A further consideration
is redundancy: Spoken discourse is often repetitive, with the
speaker reiterating, rephrasing, or revisiting information that has
already been expressed.
INTEGRATION. Listeners integrate the incoming information into
what has been heard so far. Heed is paid to whether it extends an
established topic or whether it initiates a new one.
SELF-MONITORING. Listeners check to see whether incoming
information is consistent with the meaning representation built
differs remarkably from that of age-matched peers. Speech characteristics typical of autism include pronoun reversal (referring
to self as you); unvaried or atypical intonation; neologisms
(Volden and Lord 1991); the use of stereotyped, repetitive, and
idiosyncratic language; and echolalia. Barry Prizant and Judith
Duchan (1981, 246) suggest that echolalia may serve important
communicative and cognitive functions, such as turn-taking for
people with autism.
Significantly, social communication in ASD often fails even
in the presence of apparently intact grammatical skills. This
can be seen in Asperger syndrome, where language skills can
be advanced, vocabulary extensive, and syntax formally correct, even bookish. The speech of individuals with Asperger
syndrome is often pedantic, exhibiting unvaried, stereotyped
phrases and expressions associated with contexts or registers
not presupposed by the immediate situation of talk. The speech
patterns associated with ASD are part of the broader spectrum of
impaired reciprocal social interaction.
Conversation may be the most difficult area of communication for people with ASD. Conventional rules of turn-taking are
often ignored. Speakers may fail to sustain conversation beyond
yes/no answers or speak at length on circumscribed interests,
and they may resist attempts to shift topic. Speakers may also
fail to attend to the conversational needs of listeners and may
have difficulty applying contextual and cultural knowledge in
conversation. They may thus encounter problems interpreting
deictic references, as the following example illustrates: Speaker
1: What did you do on the weekend? Speaker 2: What weekend?
Here, the conventional response to the question posed would be
to interpret the weekend as the one that had just passed. Such
problems with relevance appear to be related to the tendency
in ASD toward an overliteral understanding of communication,
including difficulties interpreting indirect requests and metaphor (Happ 1993).
A number of cognitive theories are currently being explored
to explain the core features of ASD. Executive dysfunction is one
widely accepted cognitive explanation for some behavior difficulties in ASD. This refers to decision-making processes that
are necessary for performing goal-directed activities, which
are thought to originate in the frontal lobes (Russell 1997).
Weak central coherence theory posits a detail-oriented processing style at the expense of global and contextual information
and alludes to poor connectivity between brain regions (Happ
and Frith 2006). Intriguingly, this information-processing style
can often lead to superior performance on certain tasks, such as
the Embedded Figures Task (Witkin et al. 1971), underscoring
the fact that ASD is not merely a set of impairments but involves
unique ways of processing information. The theory most frequently cited to explain communication difficulties in ASD is
theory of mind (ToM) (Baron-Cohen 1995). ToM explains
these difficulties in terms of a cognitive mechanism underlying the ability to recognize others mental states. Many of the
pragmatic impairments that are known to occur in ASD can
be linked to a lack of intuitive mentalizing ability, for example,
difficulties understanding pretense, irony, deception, and
nonliteral language. The ToM hypothesis does not preclude
the presence of assets and islets of ability as suggested by weak
central coherence theory. Cognitive theories and hypothesized
115
Autonomy of Syntax
neural correlates with respect to facial and emotion information processing in the amygdala have so far provided the most
compelling explanations for the communication impairments
seen in ASD. Research into genetic causes appears promising, since some of the strongest genetic effects in autism seem
related to language abilities.
Jessica de Villiers
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
American Psychiatric Association. 1994. Diagnostic and Statistical Manual
of Mental Disorders (DSM-IV). 4th ed. Washington, DC: American
Psychiatric Association.
Baron-Cohen, Simon. 1995. Mindblindness: An Essay on Autism and
Theory of Mind. Cambridge, MA: MIT Press.
Fombonne, Eric, Rita Zakarian, Andrew Bennett, Linyan Meng, and
Diane McLean-Heywood. 2006. Pervasive developmental disorders
in Montreal, Quebec, Canada: Prevalence and links with immunizations. Pediatrics 118.1: 13950.
Frith, Uta. 2003. Autism: Explaining the Enigma. 2d ed. Oxford:
Blackwell.
Happ, Francesca. 1993. Communicative competence and theory of
mind in autism: A test of relevance theory. Cognition 48.2: 10119.
Happ, Francesca, and Uta Frith. 2006. The weak coherence
account: Detail-focused cognitive style in autism spectrum disorders.
Journal of Autism and Developmental Disorders 36.1: 525.
Kana, Rajesh K, Timothy A. Keller, Vladimir L. Cherkassky, Nancy J.
Minshew, and Marcel Adam Just. 2006. Sentence comprehension in
autism: Thinking in pictures with decreased functional connectivity.
Brain 129: 248493.
Kanner, Leo. 1943. Autistic disturbances of affective contact. Nervous
Child 2: 21750.
Prizant, Barry, and Judith Duchan. 1981. The functions of immediate
echolalia in autistic children. Journal of Speech and Hearing Disorders
46: 2419.
Russell, James, ed. 1997. Autism as an Executive Disorder. Oxford: Oxford
University Press.
Volden, Joanne, and Catherine Lord. 1991. Neologisms and idiosyncratic
language in autistic speakers. Journal of Autism and Developmental
Disorders 21.2: 10930.
Witkin, H., P. Oltman, E. Raskin, and S. Karp. 1971. A Manual for the
Embedded Figures Test. Palo Alto, CA: Consulting Psychologists Press.
AUTONOMY OF SYNTAX
Autonomy of syntax refers to what in recent times has been the
dominant assumption concerning the formulation of syntactic
regularities: syntax is determined independently of phonological realization or semantic interpretation. The formal properties of syntax are manipulated purely formally.
Such an assumption is familiar to modern students of linguistics from numerous textbook presentations, such as Andrew
Radfords (1988, 31):
autonomous syntax principle. No syntactic rule can make reference to pragmatic, phonological, or semantic information.
116
Babbling
Langacker 1987,155). Chomsky, however, envisages autonomy
theses of varying degrees of strength (1977, 43), whereby syntax is not necessarily exhausted by the substructure of perfectly
formal rules (1972, 119) of his formulation. Thus, the significant
question with regard to the autonomy thesis may not be a question of yes or no, but rather of more or less, or more correctly,
where and how much (Chomsky 1977, 42). Certainly, provided
again that we have an independent characterization of syntax,
the extent of autonomy and its accommodation are in themselves
interesting empirical questions, with consequences for modularity, universal grammar, and the autonomy of language itself.
And seeking to answer them might be more comprehensible to
opponents of a strong interpretation of autonomy.
Work within the autonomist program(s), whatever the status
of the assumption, has undoubtedly had important results, but
there is room for debate as to how fruitful has been the pursuit of
the autonomy assumption as such. And in addition to the question of how it relates to independent notions of what syntax is, a
major difficulty in evaluating the assumption, and its contribution to these results, is the changing nature of the grammatical
enterprise(s) in which autonomy has been invoked, as well as the
varying degrees of emphasis with which it has been put forward
or denied.
John M. Anderson
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Anderson, John M. 2005. Structuralism and autonomy: From Saussure
to Chomsky. Historiographia Linguistica 32: 11748.
Bloomfield, Leonard. 1926. A set of postulates for the science of language. Language 2: 15364.
Chomsky, Noam. 1957. Syntactic Structures. The Hague: Mouton.
. 1972. Some empirical issues in the theory of transformational
grammar. In Goals of Linguistic Theory, ed. Stanley Peters, 63130.
Englewood Cliffs, NJ: Prentice-Hall.
. 1975. The Logical Structure of Linguistic Theory. New York: Plenum
Press.
. 1977. Essays on Form and Interpretation. Amsterdam: NorthHolland.
. 1995. The Minimalist Program. Cambridge, MA: MIT Press.
Jespersen, Otto. 1924. The Philosophy of Grammar. London: George Allen
and Unwin.
Langacker, Ronald W. 1987. Foundations of Cognitive Grammar,
I: Theoretical Prerequisites. Stanford, CA: Stanford University Press.
Radford, Andrew. 1988. Transformational Grammar: A First Course.
Cambridge: Cambridge University Press.
B
BABBLING
Babbling can be defined as infant vocal production that is broadly
adultlike in phonetic shape but that lacks any identifiable
adult model or intended meaning. The formal criterion broadly
adultlike limits study to the period that follows the childs first
temporally regular or rhythmic C-V (Consonant-Vowel)-syllable
production, also known as canonical babbling (Oller 1980); these
vocal forms include both a complete or nearly complete supraglottal closure and a transition to a recognizable vocalic nucleus,
for example, [dada], [babababa], [aaa]. Prior to that, the
child vocalizes in more primitive ways that are not thought to be
directly related to language.
The lack of an adult model is a question of interpretation
since the first word forms are highly similar in form to concurrent babbling. In fact, there is in most cases a gradual shift from
a predominance of unidentifiable vocalizations, beginning with
the emergence of canonical babbling (at 6 to 10 months), to a
predominance of word use identifiable in the situational context (16 to 22 months). The extent to which the shift is relatively
abrupt or gradual is highly individual.
Finally, there are vocal forms that do not appear to be
based on an adult model but that are nevertheless used with
consistent broad communicative meaning such as request,
rejection, or interest; these transitional forms or protowords
(Vihman and Miller 1988) should thus be distinguished from
babbling, which lacks any apparent communicative goal.
Babbling is a largely self-directed process of exploration
(Elbers 1982, 45).
117
Babbling
signs: Cheek et al. 2001). For example, a French child whose
prelinguistic babbling made considerable use of liquids (mainly
[l]) was found to develop several first words with [l(j)], which is
uncommon in early phonology: allo hello (on the telephone)
[ailo], [hailo], [haljo], [alo]; lolo bottle (babytalk term) [ljoljo];
donne (le) give (it) [d], [dl], [ld], [heldo] (Vihman 1993).
BABBLING DRIFT: THE EFFECT OF PERCEPTION ON PRODUCTION. A
second issue that has aroused interest for half a century is that
of possible drift in babbling toward the sounds of the native language (Brown 1958). The issue has generated considerable heat
and is important since it concerns the extent to which infants can
be taken to be capable of translating their perceptual experience
of the sound patterns of the ambient language into their limited
production repertoire. That is, any identifiable ambient language
influence on prelinguistic vocalizations means that infants have
both perceived the typical sounds of their language and adjusted
their vocal production accordingly.
Many studies, from Atkinson, MacWhinney, and Stoel
(1968) to Engstrand, Williams, and Lacerda (2003), have used
adult perceptual judgments of recorded vocalizations to determine whether infants language of exposure can be identified,
as that would provide evidence of drift; the findings remain
inconclusive, however. Meanwhile, Bndicte de BoyssonBardies and colleagues, using acoustic analyses of vowels
(1989) and tallies of transcribed consonant types (BoyssonBardies and Vihman 1991), established significant prelinguistic adult language influence, although the mechanism for such
an effect remained unclear. More recent work demonstrating
the extent of early implicit or distributional learning (Saffran,
Aslin, and Newport 1996) suggests that infants are capable of
registering dominant patterns of their language within the first
year. Thus, the mechanism needed to account for drift may be
the effect of implicit perceptual learning on production: Those
vocalizations that, as the producing child perceives them, activate perceptual responses already familiar from input patterning would strengthen perceptuomotor connections, leading to
their repeated use.
Theoretical approaches
FRAME AND CONTENT: THE ARTICULATORY BASIS OF BABBLING. The
most widely accepted current model of babbling is that of
Peter F. MacNeilage and Barbara L. Davis (1990 and Davis and
MacNeilage 1990, 1995 (for a review of competing ideas, see Chen
and Kent 2005). The articulatory basis of babbling is claimed to
be frame dominance, meaning that the patterns produced largely
reflect mandibular oscillation without independent control of
lip and tongue movement. The result is strong C-V associations,
such that alveolars are followed by front vowels, labials by central
vowels (the pure frames, requiring no particular tongue setting),
and velars by back vowels. Furthermore, the model predicts that
changes in the mandibular cycle will result in height changes for
vowels and manner changes for consonants in variegated babbling sequences. The work of this team and collaborators investigating a range of other languages (e.g., Dutch, French, Romanian,
Turkish: Davis et al. 2005) have largely supported the predictions
and have demonstrated a tendency for adult languages to show
the C-V associations as well (MacNeilage and Davis 2000) but
118
Babbling
Davis, Barbara L., and Peter F. MacNeilage. 1990. Acquisition of correct
vowel production. Journal of Speech and Hearing Research 33: 1627.
. 1995. The articulatory basis of babbling. Journal of Speech and
Hearing Research 38: 11991211.
Davis, Barbara L., Sophie Kern, Dilara Koba, and Inge Zink. 2005.
Vocalizations in canonical babbling. Paper presented at Symposium,
10th International Congress of the Association for the Study of Child
Language, Berlin.
DePaolis, Rory A. 2006. The influence of production on the perception
of speech. In Proceedings of the 30th Boston University Conference on
Language Development, ed. D. Bamman, T. Magnitskaia, and C. Zaller,
14253. Somerville, MA: Cascadilla Press.
Elbers, L. 1982. Operating principles in repetitive babbling. Cognition
12: 4563.
Engstrand, Olle, Karen Williams, and Francisco Lacerda. 2003. Does
babbling sound native? Phonetica 60: 1744.
Ferguson, Charles A., and Carol B. Farwell. 1975. Words and sounds in
early language acquisition. Language 51: 41939.
Ferguson, Charles A., and Olga K. Garnica. 1975. Theories of phonological development. In Foundations of Language Development, ed. Eric
H. Lenneberg and Elizabeth Lenneberg, 15380. New York: Academic
Press.
Jakobson, Roman. [1941] 1968. Child Language, Aphasia, and
Phonological Universals. The Hague: Mouton. Eng. translation of
Kindersprache, Aphasie und allgemeine Lautgesetze, Uppsala.
Kiparsky, Paul, and Lise Menn. 1977. On the acquisition of phonology.
In Language Learning and Thought, ed. John Macnamara, 4778. New
York: Academic Press.
MacNeilage, Peter F., and Barbara L. Davis. 1990. Acquisition of speech
production: Frames, then content. In Attention and Performance. Vol.
13: Motor Representation and Control. Ed. Marc Jeannerod, 45375.
Hillsdale, NJ: Lawrence Erlbaum.
. 2000. On the origin of internal structure of word forms. Science
288: 52731.
McCune, Lorraine, and Marilyn M. Vihman. 2001. Early phonetic
and lexical development. Journal of Speech, Language and Hearing
Research 44: 67084.
Oller, D. Kimbrough. 1980. The emergence of the sounds of speech
in infancy. In Child Phonology. Vol. 1: Production. Ed. Grace Yenikomshian, James F. Kavanagh, and Charles A. Ferguson, 93112. New
York: Academic Press.
. 2000. The Emergence of the Speech Capacity. Mahwah, NJ: Lawrence
Erlbaum. This book provides a thorough review of babbling studies
conducted with hearing, hearing-impaired, premature, and low SES
(socioeconomic status) infants, as well as provocative ideas about the
evolution of language based on evidence from ontogeny.
Saffran, Jenny R., Richard N. Aslin, and Elissa L. Newport. 1996. Statistical
learning by 8-month-old infants. Science 274: 19268.
Vihman, Marilyn M. 1993. Variable paths to early word production.
Journal of Phonetics 21: 6182.
. 1996. Phonological Development. Oxford: Blackwell. This book
provides an overview of research in infant speech perception and production and their interactions, as well as of theories of phonological
development, early word patterning, and the nature of the transition
into language.
Vihman, Marilyn M., and Sari Kunnari. 2006. The sources of phonological knowledge. In Recherches Linguistiques de Vincennes 35: 13364.
Vihman, Marilyn M., Marlys A. Macken, Ruth Miller, Hazel Simmons,
and James Miller. 1985. From babbling to speech: A re-assessment of
the continuity issue. Language 61: 397445.
Vihman, Marilyn M., and Ruth Miller. 1988. Words and babble at the
threshold of lexical acquisition. In The Emergent Lexicon, ed. Michael
D. Smith and John L. Locke, 15183. New York: Academic Press.
Basal Ganglia
BASAL GANGLIA
Nothing in biology makes sense except in the light of evolution.
Dobzhansky 1973
Neural Circuits
These syndromes follow from the basal ganglia activity in different neural circuits. Neural circuits that link activity in different
parts of the brain appear to be the bases for most, if not all, complex mammalian behaviors. In humans, a class of neural circuits
that links activity in different regions of the cortex through the
119
Basal Ganglia
basal ganglia and other subcortical structures appears to play a
key role in regulating aspects of human linguistic ability, such
as talking and comprehending the meaning of a sentence, as
well as such seemingly unrelated phenomena as decision making, walking, attention, and emotional state. To understand the
nature of neural circuits, we must take account of the distinction
that exists between local operations that are carried out within
some particular part of the brain and an observable behavior
that results from many local operations linked in a neural circuit. Complex brains, including the human brain, perform local
operations involving tactile, visual, or auditory stimuli in particular regions of the brain. Other neural structures perform local
operations that regulate aspects of motor control or hold information in short-term (working) memory, and so on.
The basic computational elements of biological brains are
neurons. Local operations result from activity in an anatomically
segregated, population (a group) of neurons. A given part of the
brain many contain many distinct anatomically segregated neuronal populations that each carry out similar local operations.
But these local operations do not constitute observable behaviors. Each anatomically segregated neuronal population projects
to anatomically distinct neuronal populations in other regions of
the brain, forming a neural circuit. The linked local operations
performed in the circuit constitute the neural basis of an observable aspect of behavior, such as striking the keys of a computer
keyboard.
120
Genetic Findings
Studies of the regulatory gene FOXP2 provide a starting point for
understanding the evolution of the cortical-striatal-cortical circuits that confer human linguistic ability (see genes and language). Other genes undoubtedly are involved and FOXP2 is not
a language gene. FOXP2 governs the embryonic development
of the basal ganglia, other subcortical structures, and lung tissue
and other structures. Its discovery resulted from a long-term study
of an extended family in which many individuals are marked by
a genetic anomaly. A syndrome, a suite of speech and orofacial
movement disorders, and cognitive and linguistic deficits mark
these individuals. They are not able to protrude their tongues
while closing their lips, cannot repeat two word sequences, and
have difficulty comprehending distinctions in meaning conveyed
by syntax (Vargha-Khadem et al. 1998). On standardized intelligence tests, they have significantly lower scores than their nonafflicted siblings. MRI imaging shows that the caudate nucleus (a
basal ganglia structure) is abnormal. fMRI imaging, which provides a measure of neural activity, shows underactivation in the
putamen (the principal basal ganglia input structure), Brocas
Basal Ganglia
area, and its right homolog (Watkins et al. 2002; Liegeois et al.
2003). These structures are connected by neural circuits through
the striatum (Lehericy et al. 2004). The behavioral deficits of
afflicted individuals are similar to those seen in Parkinsons disease and oxygen deprivation (cf. Lieberman 2006 for details).
The role of FOXP2 during early brain development in humans
and of the mouse version ( foxp2) in mice was established by C.
S. Lai and colleagues (2003). The gene governs the expression of
other genes during embryonic development. In both the human
and mouse brain, the gene is active in the interconnected neural structures that constitute the cortical-striatal-cortical circuits
regulating motor control and cognition in humans, including the
caudate nucleus and putamen of the basal ganglia, the thalamus, inferior olives, and cerebellum. Despite the high degree
of similarity, the mouse and human versions are separated by
three mutations. The chimpanzee and human versions are separated by two mutations. W. Enard and colleagues (2002), using
the techniques of molecular genetics, estimate that the human
form appeared somewhere in the last 200,000 years, in the time
frame (Stringer 1998) associated with the emergence of anatomically modern Homo sapiens. The appearance of human speech
anatomy 50,000 years ago presupposes the prior appearance of
this neural substrate (see speech anatomy, evolution of).
In short, the basal ganglia are neural structures that were initially adapted for one function motor control. In the course of
evolution, the human basal ganglia were modified, taking on
additional cognitive and linguistic tasks.
Philip Lieberman
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Dobzhansky, Theodosius. 1973. Nothing in biology makes sense except
in the light of evolution. American Biology Teacher 35: 1259.
Enard, W., M. Przeworski, S. E. Fisher, C. S. Lai, V. Wiebe, T. Kitano, A. P.
Monaco, and S. Paabo. 2002. Molecular evolution of FOXP2, a gene
involved in speech and language. Nature 41: 86972.
Graybiel, Ann M. 1997. The basal ganglia and cognitive pattern generators. Schizoprenia Bulletin 23: 45969.
Grossman, Murray, S. Carvell, S. Gollomp, M. B. Stern, G. Vernon, and
H. I. Hurtig. 1992. Sentence comprehension and praxis deficits in
Parkinsons disease. Neurology 41: 16208.
Hauser, Marc D., N. Chomsky, and W. T. Fitch. 2002. The faculty of
language: What is it, who had it, and how did it evolve? Science
298: 156979.
Kotz, Sonia A., M. Meyer, K. Alter, M. Besson, D. Y. von Cramon, and A. D.
Frederica. 2003. On the lateralization of emotional prosody: An fMRI
investigation. Brain and Language 86: 36676.
Lai, C. S., D. Gerrelli, A. P. Monaco, S. E. Fisher, and A. J. Copp. 2003.
FOXP2 expression during brain development coincides with adult
sites of a pathology in a severe speech and language disorder. Brain
126: 245562.
Lehericy, S. M., M. Ducros, P. F. Van de Moortele, C. Francois, L. Thivard,
C. Poupon, N. Swindale, K. Ugurbil, and D. S. Kim. 2004. Diffusion
tensor tracking shows distinct corticostriatal circuits in humans.
Annals of Neurology 55: 5229.
Lieberman, Philip. 2006. Toward an Evolutionary Biology of Language.
Cambridge: Harvard University Press.
Lieberman, Philip, J. Friedman, and L. S. Feldman. 1990. Syntactic deficits in Parkinsons Disease. Journal of Nervous and Mental Disease
178: 3605.
121
Bilingual Education
for which category members have similar overall shape (cf. car
versus vehicle). Children learn basic level concepts sooner than
other concepts (Brown 1958; Horton and Markman 1980). Basic
level advantages are found in many other domains, including
environmental scenes (Tversky and Hemenway 1983), social
categories (Cantor and Mischel 1979), and actions (Morris and
Murphy 1990).
One explanation for the advantages of basic level categories
over other categories is that they are more differentiated (Rosch
et al. 1976; Murphy and Brownell 1985). Members of basic level
categories have many features in common. These features are
also distinct from those of other categories at this level. In contrast, although members of more specific, subordinate categories (e.g., sports car) have slightly more features in common than
do those of basic level categories, many of these features are not
distinctive. That is, members of a subordinate category share
their features with other subordinates (e.g., members of sports
car share a number of features with other subcategories of car).
In contrast, the members of more general, superordinate categories (e.g., vehicle) have few common features.
Differentiation explains the basic level advantage because
it reflects a compromise between two competing functions of
concepts. Categories should be informative so that one can draw
inferences about an entity on the basis of its category membership. Emphasizing this function leads to the formation of large
numbers of categories with the finest possible discriminations
between categories (Rosch et al. 1976, 384). However, the formation of categories should only preserve important differences
between them that are practical: It is to the organisms advantage
not to differentiate one stimulus from others when that differentiation is irrelevant to the purposes at hand (ibid.). This function
counteracts the tendency to create large numbers of categories
and reflects the principle of cognitive economy (Rosch et al.
1976). Overall, basic level categories have an advantage because
they are relatively general and informative, whereas superordinate categories, though general, are not informative and subordinate categories, though informative, are not general.
The basic level may change with expertise in a way that is
consistent with the differentiation explanation. For example,
James Tanaka and Marjorie Taylor (1991) investigated expertise effects on the basic level in expert dog breeders and birdwatchers. Using a number of tasks, they tested each expert in
both the dog and bird domains. For instance, in a speeded categorization task, experts in their novice domain were fastest at the
basic level and slowest at the subordinate level (as Rosch et al.
1976 found). However, in their area of expertise, categorization
was equally fast at the basic and subordinate levels. For more
detailed reviews of this literature, see Lassaline, Wisniewski,
and Medin 1992; Murphy and Lassaline 1997.)
Edward Wisniewski
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Berlin, Brent. 1992. Ethnobiological Classification: Principles of
Categorization of Plants and Animals in Traditional Societies.
Princeton, NJ: Princeton University Press.
Brown, Roger. 1958. How shall a thing be called? Psychological Review
65: 1421.
122
Cantor, Nancy, and Walter Mischel. 1979. Prototypes in person perception. In Advances in Experimental Social Psychology, ed. L. Berkowitz,
452. New York: Academic Press.
Horton, Marjorie, and Ellen Markman. 1980. Developmental differences in the acquisition of basic and superordinate categories. Child
Development 51: 70815.
Jolicoeur, Pierre, Mark Gluck, and Steven Kosslyn. 1984. Pictures and
names: Making the connection. Cognitive Psychology 16: 24375.
Lassaline, Mary, Edward Wisniewski, and Douglas Medin. 1992.
Basic levels in artificial and natural categories: Are all basic categories created equal? In Percepts, Concepts, and Categories: The
Representation and Processing of Information, ed. B. Burns, 32880.
North Holland: Elsevier.
Morris, Michael, and Gregory Murphy. 1990. Converging operations on
a basic level in event taxonomies. Memory and Cognition 18: 40718.
Murphy, Gregory, and Hiram Brownell. 1985. Category differentiation in
object recognition: Typicality constraints on the basic category advantage. Journal of Experimental Psychology: Learning, Memory and
Cognition 11: 7084.
Murphy, Gregory, and Mary Lassaline. 1997. Hierarchical structure
in concepts and the basic level of categorization. In Knowledge,
Concepts, and Categories, ed. K. Lamberts and D. Shanks, 93131.
London: Psychology Press
Rosch, Eleanor, Carolyn Mervis, Wayne Gray, David Johnson, and Penny
Boyes-Braem. 1976. Basic objects in natural categories. Cognitive
Psychology 8: 382439.
Tanaka, James, and Marjorie Taylor. 1991. Object categories and expertise: Is the basic level in the eye of the beholder? Cognitive Psychology
23: 47282.
Tversky, Barbara, and Kathy Hemenway. 1983. Categories of environmental scenes. Cognitive Psychology 15: 12149.
BILINGUAL EDUCATION
In principle, bilingual education is just the use of two languages
in instruction in a school setting. However, in practice, it covers
a wide array of programs. Bilingual education programs range
from high-status schools promoting international education
through prestige languages, such as English and French, to
highly marginalized schools devoted to the bare-bones schooling of immigrant children.
In its weak form, bilingual education may involve transitional or subtractive bilingualism, leading to monolingualism
(e.g., teaching Spanish-speaking children English to ensure their
assimilation and integration into mainstream America). In its
strong form, bilingual education aims at maintaining the language of a minority child in addition to the learning of a majority language, thus leading to additive bilingualism. Heritage
bilingual schools often practice an ideal version of additive
bilingualism, which stresses bicultural education in addition to
bilingualism.
One of the central concerns of bilingual education is to
address the educational needs/performance of minority children by maintaining their mother tongue. The proponents of
maintaining the mother tongue claim that such maintenance
is critical for linguistic and cognitive growth of the child, school
performance, psychological security, ethnic and cultural identity (see ethnolinguistic identity), self-esteem, and many
other positive personal and intellectual characteristics. The supporters of transitional bilingualism claim that only transitional
Bilingualism, Neurobiology of
bilingualism is capable of saving children from poor academic
performance and allowing for assimilation.
Bilingual education has been steadily gaining strength around
the globe since the era of decolonization. It is further fueled by
the growth of ethnic awareness and the movement to prevent
the extinction of languages of the world. Many countries,
particularly in Europe, that earlier fostered monolithic ideology
have begun to recognize their diversity as a source of social and
economic capital, thus marking a new era of bilingual/multilingual education (e.g., in the United Kingdom, France, and Spain,
among others; see language policy).
Multilingual countries of Asia and Africa continue to nurture
a long tradition of bilingual education. Since 1956, India, for
example, has had as official policy a three-language formula in
education. This formula calls for multilingual education. In addition to learning the two national languages Hindi and English
students are expected to learn a third or a fourth language.
Because of its deep-rooted association with immigrants, bilingual education in the United States is particularly notable for its
turbulent history. On June 2, 1998, the people of California voted
to end a tradition of bilingual education by passing Proposition
227, which gives immigrant children just one year to learn English
before they enroll in regular classes. Many school systems in
other states are waiting either to put in place severe restrictions
on bilingual instruction or to eliminate it completely by passing
English only policies (for details, see Genesee 2006).
While bilingual education is often associated with the education of minority students (e.g., in the United States), the
Canadian immersion programs in French devoted to the
majority Anglophones serve as a model of bilingual education
for majority students (for details, see Genesee 2006).
Tej K. Bhatia
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Crawford, James. 2004. Educating English Learners. Los Angeles: Bilingual
Educational Services.
Genesee, Fred. 2006. What do we know about bilingual education for
majority-language students? In Handbook of Bilingualism, ed. Tej K.
Bhatia and William C. Ritchie, 54776. Oxford: Blackwell.
BILINGUALISM, NEUROBIOLOGY OF
Neurobiology of bilingualism refers to the study of the cerebral organization of multiple languages in the human brain.
From early accounts of selective loss and recovery in bilingual
aphasia (i.e., loss of language due to a brain lesion) to recent
electrophysiological and functional neuroimaging studies, issues
inherent to the bilingual brain have inspired researchers for more
than a century. Investigations into the neural basis of bilingualism focus not only on how two languages (L1 and L2) are represented in the brain (i.e., the anatomical location) but also on how
these languages are processed. Indeed, the main assumption is
that a weaker L2 may be processed through brain mechanisms
that may differ from those underlying L1 processing. After a brief
historical overview, I illustrate findings inherent to the representation of languages, followed by a section focusing on language
processing.
Historical Overview
From a historical standpoint, the first approach to studying
brain organization for bilingualism was the study of bilingual
aphasics. Several clinical aphasia studies have shown that bilingual aphasics do not necessarily manifest the same language
disorders with the same degree of severity in both languages. In
some cases, L1 is recovered better than a L2. In other cases, the
converse obtains. Since the landmark 1895 study of the French
neurologist Albert Pitres, who was the first to draw attention to
the relative frequency of differential language recovery following aphasia in bilinguals, many different recovery patterns have
been described: from selective recovery of a given language
(i.e., one language remains impaired while the other recovers); parallel recovery of both languages, successive recovery
(i.e., after the recovery process of one language, the other language recovers); alternating recovery (i.e., the language that
was first recovered will be lost again due to the recovery of the
language that was not first recovered); and alternating antagonistic recovery (i.e., on one day the patient is able to speak in
one language while on the next day only in the other); to the
pathological mixing of two languages (i.e., the elements of the
two languages are involuntarily mixed during language production). The study of bilingual aphasia is important because
it indicates the cortical regions necessary for performance of a
linguistic task (e.g., speaking in L1).
Clinical case reports indicate a set of relevant factors and
have led to theoretical conjectures. However, at present we lack a
causal account for the various recovery patterns and cannot predict clinical outcomes. Concerning the possible factors involved,
no correlation has been found between the pattern of recovery
and neurological, etiological, experiential, or linguistic parameters: not site, size or origin of lesion, type or severity of aphasia,
type of bilingualism, language structure type, or factors related to
acquisition or habitual use.
Theoretical conjectures arising from the study of bilingual
aphasia developed along two distinct lines, a more traditional
approach and a more dynamic approach. The more traditional
localizationist view argued, for instance, that the specific loss
of one language would occur because the bilinguals languages
are represented in different brain areas or even in different
hemispheres, and hence, a focal brain lesion within a languagespecific area may alter only that specific language, leaving the
other language intact. In contrast, according to the dynamic view
selective recovery arises because of compromise to the language
system, rather than to damage to differential brain representations. A selective loss of a language arises because of increased
inhibition, that is, of a raised activation threshold for the affected
or lost language, or even because of an imbalance in the means
to activate the language due to the lesion. It is worth underlining
that Pitres himself proposed a dynamic explanation of language
recovery in bilingual aphasics: Language recovery could occur
only if the lesion had not entirely destroyed language areas but
temporarily inhibited them through a sort of pathological inertia. In Pitress opinion, the patient generally first recovered the
language to which she/he was premorbidly more exposed (not
necessarily the native language) because the neural elements
subserving the more exposed language were more strongly
associated.
123
Bilingualism, Neurobiology of
The dynamic view not only explains the so-called selective recovery of a language but can also explain many reported
recovery patterns in bilingual aphasia. As outlined by M. Paradis
(1998), a parallel recovery would then occur when both languages are inhibited to the same degree. When inhibition affects
only one language for a period of time and then shifts to the other
language (with disinhibition of the prior inhibited language) a
pattern of alternating antagonistic recovery occurs (see Green
1986). Selective recovery would occur if the lesion permanently
raised the activation threshold for one language, and pathological mixing among languages would occur when languages could
no longer be selectively inhibited.
In general, the aphasia data have provided a rich source of evidence on the range of language disorders and language recovery
patterns in bilinguals. However, there are limitations to the generalizability of such data to neurologically healthy individuals.
Concerns about the lesion-deficit approach include the inability
to determine whether specific language deficits are the result of
damage to a specialized language component at the lesion site,
or if the damaged area is simply part of a larger neural network
that mediates a given component of language. Likewise, aphasia
data do not allow one to separate the effects of injury from those
of neural plasticity or a reallocation of healthy cortical tissue
for the mediation of language functions lost as a result of brain
injury. Nevertheless, studying the effects of brain damage on linguistic function in bilinguals has led to a number of interesting
observations about the nature and course of language impairment and recovery, which in turn has stimulated researchers to
apply functional neuroimaging techniques to the investigation of
bilingual language processing.
124
Bilingualism, Neurobiology of
goal must be maintained in the face of conflicting goals, and the
various actions required to perform the task must be coordinated
(e.g., retrieve or compute the words phonology from its spelling
or retrieve the meaning of the word and select its translation).
Once a given task is established, however (e.g., speaking in L2),
competition with alternative possible tasks (speaking in L1) may
be resolved more automatically. Where individuals wish to alter
their goal (for example, to switch from speaking in one language
to speaking in another), they must disengage from the current
goal and switch to the new goal. Lexical concepts matching the
intended language must be selected and produced, while those
not matching the intended language must be inhibited through
language control mechanisms. For instance, in word production
studies, language control would inhibit potential interferences
from the non-target language. Psycholinguistic evidence points
to the fact that such interference is more common during production in a language that is mastered to a lower degree of proficiency, for example, a weak L2. In that case, for example, when
asked to name a picture in L2, the bilingual speaker has to inhibit
L1 in order to prevent a prepotent interference from L1.
Functional neuroimaging studies using experimental tasks
like picture naming, switching, translating, and so on have elegantly shown that these tasks are paralleled by the activation of
a set of brain areas that are not directly linked to language representation, such as the brain activity within the left prefrontal cortex, the left caudate nucleus, and the anterior cingulate cortex.
The engagement of these areas is even more relevant when subjects have to process a weak L2. The functions generally ascribed
to the prefrontal cortex comprise working memory, response
inhibition, response selection, and decision making, while the
left caudate was reported to be imported for language selection
and set switching. The anterior cingulate cortex is related to such
functions as conflict monitoring, attention, and error detection.
It becomes clear that the engagement of these structures provides a cerebral testimony of the cognitive processes inherent to
bilingual language processing: competition and control between
languages.
Conclusions
Extensive reviews focusing on the bilingual brain as studied with
functional neuroimaging are available in the literature to which
the reader is referred (Abutalebi, Cappa, and Perani 2005; Perani
and Abutalebi 2005; but see also Paradis (2004) for a critical
viewpoint). In broad outlines, functional neuroimaging has shed
new light on the neural basis of L2 processing and on its relationship to native language (L1). First of all, the long-held assumption that L1 and L2 are necessarily represented in different brain
regions or even in different hemispheres in bilinguals has not
been confirmed. On the contrary, functional neuroimaging has
elegantly outlined that L1 and L2 are processed by the same
neural devices. Indeed, the patterns of brain activation associated with tasks that engage specific aspects of linguistic processing are remarkably consistent among different languages, which
share the same brain language system. These relatively fixed
brain patterns, however, are modulated by a number of factors.
Proficiency, age of acquisition, and exposure can affect the cerebral representations of each language, interacting in a complex
way with the modalities of language performance.
125
126
conceptual and linguistic levels (see bilingualism, neurobiology of). These labels and dichotomies demonstrate the complex attributes of bilingualism that make the task of defining and
measuring bilinguals a daunting one. A working definition of
bilingualism is offered by Leonard Bloomfied ([1933] 1984, 53),
who claimed that a bilingual is one who has native-like control
over two languages (i.e., balanced bilingual).
Effects of Bilingualism/Multilingualism
What is the effect of bilingualism/multilingualism on an individual, particularly on a child? The research on this question
is fundamentally driven by two hypotheses: the linguistic deficit hypothesis and the linguistic augmentation hypothesis.
According to the former, bilingual children show serious linguistic and cognitive adverse effects of bilingualism. Exposure to two
languages leads to semilingualism; that is, they become deficient
in both languages, which in turn leads to other disabilities (e.g.,
stuttering) and cognitive impairments (low intelligence, mental
retardation, and even schizophrenia).
Such a hypothesis has become obsolete in light of the findings of the research driven by the linguistic augmentation
hypothesis. Solid on theoretical and methodological grounds,
research by Elizabeth Peal and Wallace E. Lambert (1962) put
to rest such negative and frightening effects of bilingualism.
Their research and the findings of the succeeding research provide ample evidence that the negative conclusions of the earlier
research were premature and misguided due to the theoretical
and methodological flaws. Contrary to the findings of the previous research, bilingual children exhibit more cognitive flexibility
than do monolinguals and perform better on verbal and nonverbal measures. Peal and Lamberts study, which was conducted in
Montreal, revolutionized research on bilingualism and multilingualism by highlighting a positive conception of bilinguals. Their
research has been replicated in many countries, confirming
the augmenting rather than subtracting effect of bilingualism.
Beyond this research, the economic, communicative (intergenerational and cross-cultural), and relational (building relations)
advantages of bilingualism are inarguable.
Conclusion
In short, bilingualism/mulitilingualism is a global phenomenon
that continues to gain further momentum in the age of globalization. It is a by-product of a number of biological, sociopsychological, and linguistic factors. These factors lead to individuals with
varying degree of language competencies. Therefore, it is not surprising that defining and measuring bilingualism/multilingualism continues to be a challenging task. Bilinguals are complex
and colorful in the way they manage and optimize their linguistic
resources. For that reason, they are not a sum of two monolinguals.
Language mixing and shifting are two defining characteristics of
bilinguals. Current socio- and psycholinguistic research attempts
to account for these two salient properties of the bilingual brain.
Tej K. Bhatia
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bhatia, Tej K., and William C. Ritchie, eds. 2006. The Handbook of
Bilingualism. Oxford: Blackwell. This book presents a multidisciplinary
and comprehensive collection of state-of-the-art research on bilingualism and multilingualism. Chapters 7 and 8 deal with bilingual production models, including the discussion of Garrett 1988 and Levelt 1989.
Bhatia, Tej K., and William C. Ritchie. 1996. Bilingual language mixing,
Universal Grammar, and second language acquisition. In Handbook
of Second Language Acquisition, ed. W. C. Ritchie and T. K. Bhatia,
62782. San Diego, CA: Academic Press.
Bloomfield, Leonard. [1933] 1984. Language. Chicago: University of
Chicago Press.
127
Binding
de Bot, Kees, and Judith F. Kroll. 2002. Psycholinguistics. In An
Introduction to Applied Linguistics, ed. Norbert Schmitt, 13349.
London: Arnold.
Edwards, John. 2006. Foundations of bilingualism. In Bhatia and Richie
2006, 731.
Garrett, M. E. 1988. Process in sentence production. In The
Cambridge Linguistic Survey. Vol. 3. Ed. F. Newmeyer, 6996.
Cambridge: Cambridge University Press.
Grosjean, Francois. 1989. Neurolinguists, beware! The bilingual is not
two monolinguals in one person. Brain and Language 36: 315.
Hakuta, Kanji. 1986. Mirror of Language. New York: Basic Books. This
work offers an excellent multidisciplinary account of bilingualism in
general and bilingualism in the United States in particular. Among
other topics, it presents an excellent account of the linguistic deficiency and linguistic augmentation hypotheses.
Levelt, W. 1989. Speaking: From Intention to Articulation. Cambridge,
MA: MIT Press.
MacSwan, Jeff. 2005. Remarks on Jake, Myers-Scotton and Grosss
response: There is no Matrix Language. Language and Cognition
8.3: 27784.
Myers-Scotton, Carol and J. Jake. 1995. Matching lemmas in a bilingual
language competence and production model: Evidence from intrasentential code switching. Linguistics 33: 9811024.
Peal, Elizabeth, and Wallace E. Lambert. 1962. Relation of bilingualism
to intelligence. Psychological Monographs 76: 123.
Ritchie, William C., and Tej K. Bhatia. 2007. Psycholinguistics. In
Handbook of Educational Linguistics, ed. Bernard Spolsky and Francis
Hult, 3852. Oxford: Blackwell.
A systematic class of exceptions to the c-command requirement is found in so-called indirect binding, for example (5), where
the object can be bound from within the subject (sideways):
(5)
(1)
128
BINDING
In quantified logic, binding names the relation between a
quantifier and one or more variables, for example, x and x
in x[P(x) Q(x)]. In linguistics, the term has been used in at
least three domains: first, for the relation between quantified
expressions and pronouns that referentially depend on them,
(1); second for coreference, the relation between two referring
expressions with the same referent, (2), including hypothesized
empty pronouns, (2c); and third, in theories that assume transformations, for the relation between a dislocated phrase and its
trace, (3) (see government and binding theory):
(7)
defended you.
*You
Likewise, reflexive pronouns in English need an antecedent
that is not just within the same finite clause but also c-commands
them:
She
(8)
defended herself.
*Her mother
These conditions on the distribution of reflexive and nonreflexive restrict binding by quantified nominals as well and are
indiscriminately referred to as binding conditions.
While c-command seems relevant in binding conditions
cross-linguistically, other aspects, such as the number of morphological classes (reflexives, nonreflexives, etc.) or the size of
relevant structural domains, vary widely.
Daniel Bring
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bring, Daniel. 2005. Binding Theory. Cambridge Textbooks in Linguistics.
Cambridge: Cambridge University Press.
Chomsky, Noam. 1981. Lectures on Government and Binding. Dordrecht,
the Netherlands: Foris Publications.
Dalrymple, Mary. 1993. The Syntax of Anaphoric Binding. Stanford,
CA: CSLI.
Kuno, Susumo. 1987. Functional Syntax Anaphora, Discourse and
Empathy. Chicago: Chicago University Press.
Biolinguistics
BIOLINGUISTICS
Biolinguistics is the study of the biology of language. The modern biolinguistic program was initiated by Noam Chomsky in
the 1950s (Chomsky 2006), although it has much earlier historical antecedents (see cartesian linguistics). It investigates
the form and function of language, the development (ontogeny)
of language, and the evolution of language (phylogeny), among
other topics. Biolinguists study such questions as the following:
(1) What is knowledge of language?
(2) How does knowledge of language develop in the child?
(3) How does knowledge of language evolve in the species?
To answer the question of what knowledge of language is (1),
biolinguists have proposed various generative grammars,
that is, explicit models of the faculty of language.
The study of generative grammars draws from a variety
of areas, including syntax, semantics, the lexicon, morphology, phonology, and articulatory and acoustic
phonetics.
In addition, the biolinguist investigates the neurological mechanisms underlying the faculty of language (see
where the box represents what the child brings to language learning. We ask how the child maps experience (primary linguistic data)
to a particular language. It is posited that the child moves through a
number of states from an initial state, corresponding to the childs
genetic endowment, to a final state, corresponding to a particular
language. For each subarea discussed, biolinguistics studies the
development or growth of language, often referred to as language
acquisition (e.g., syntax, acquisition of; semantics, acquisition of; and phonology, acquisition of).
The initial state may be characterized by a universal grammar, which is a set of general principles with parameters that
are set by experience, thus accounting for the variation across
languages. For example, there are general principles of word
order that permit some variation; for example, the verb precedes
the object (English) or the verb follows the object (Japanese) (see
x-bar theory). Such a theory is referred to as a principles
and parameters theory. (For specific subareas, see syntax,
of Richard Kaynes (2004) and Charles Yangs (2004) work on competitive theories of language acquisition. (From other perspectives,
see universals, nongenetic; absolute and statistical
universals, implicational universals, and typological
universals.)
In addition to comparative grammar (see also morpholog-
ical typology), universals of language change, syntactic change, semantic change, pidgins, and creoles
provide additional evidence for the nature of universal grammar
and language acquisition.
Moreover, the study of genetic language disorders, as well as
familial and twin studies, has been very fruitful for the study of
language acquisition (see genes and language; specific
language impairment; see also the extensive literature on the
FOXP2 gene [Marcus and Fisher 2003]). Studies of language-isolated children provide information about the critical period
for language learning. The study of sign languages has been
invaluable for investigating language outside the modality of
sound (see also sign language, acquisition of; sign languages, neurobiology of). Finally, the study of linguistic
savants has been quite useful for delineating the modularity
of the language faculty as distinct from other cognitive faculties.
To answer the question of how knowledge of language evolves
in the species, (3), biolinguists integrate data from a variety of
areas, including comparative ethology (see Hauser, Chomsky,
and Fitch 2002; see also animal communication and human
language; speech anatomy, evolution of), comparative
neuroanatomy, and comparative genomics. Since evolution of
language took place in the distant past, mathematical modeling
of populations of speaker-hearers has recently attracted much
interest in work on dynamical systems (see self-organizing
systems). Such studies have proven useful not only for the study
of evolution but also for the study of language acquisition and
change. (For some hypotheses on origins of language, see origins of language; grooming, gossip, and language.)
Questions (1) to (3) might be called the what and how questions of biolinguistics. One can also ask why the principles of
language are what they are, a deeper and more difficult question
to answer. The investigation into why questions is sometimes
referred to as the minimalist program or minimalism. In addition, there is the related question of how the study of language
can be integrated with the other natural sciences, a problem that
Chomsky calls the unification problem (see Jenkins 2000). All
of these questions are certain to continue to fascinate investigators of the biology of language for decades to come.
(For more information on other explicit models of the language faculty, see transformational grammar; standard
129
130
parallels are what one would expect if both species rely on a similar neural substrate for learning and using their communicative
systems.
Relevant genetic evidence is also available. The much-discussed FOXP2 gene is similarly expressed in the basal ganglia of
humans and songbirds (Teramitsu et al. 2004; Vargha-Khadem
et al. 2005). A FOXP2 mutation in humans results in deficits in
language production and comprehension, especially aspects of
(morpho)syntax that involve combining and sequencing linguistic units (Marcus and Hisher 2003; Vargha-Kadham et al. 2005).
One of the neurobiological effects of the mutation is a notable
reduction in the gray matter of the striatum (Vargha-Kadham
et al. 2005). Perhaps, then, the combinatorial aspects of human
language were enabled by the preadaptation of an anterior neural circuit that has been highly conserved over evolutionary time
and across species, and by a genetic mutation in this circuit that
increased its computational space.
Finally, some birdsong, like human language, is compositional; songbirds learn units and rules of combination (Rose et
al. 2004), although the rules of combination are obviously far
less sophisticated than those that characterize human language.
A skeptic might argue that the syntax of human language is too
complex (too highly structured, too recursive, too creative; see
recursion, iteration, and metarepresentation ) to be
modeled as a simple patterned sequence processor that relies
on associative learning mechanisms. In fact, the explanatory
burden placed on rule-based, recursive syntax has diminished
over recent decades. Modern grammars tend to be lexicalist
in nature; that is, much of the knowledge relevant to sentence
structure is stored in the lexicon with individual words, rather
than being computed by abstract phrase structure rules (see
lexical-functional grammar ). Recursion, while clearly
a characteristic of human language, is much more limited in
actual language usage than would be predicted given the standard model. And, because conceptual knowledge (see semantics) has its own structure (Jackendoff 1990), it seems plausible
that some of the burden for structuring the input rests with the
conceptual stream (Jackendoff 2002), rather than entirely with
the syntax.
Birds and humans are fundamentally different in many ways,
as are their systems of communication. Nonetheless, birds and
humans are two of only a handful of vocal learners, and recent
work points to communication-relevant homologies and similarities. It is not unreasonable to think that a comparative approach
might provide important clues to how language evolved and,
perhaps, to the nature of language itself.
Lee Osterhout
WORK CITED AND SUGGESTIONS FOR FURTHER READING
Aldridge, J. Wayne, and Kent C. Berridge. 1998. Coding serial order
by neostriatal neurons: A natural action approach to movement
sequence. Journal of Neuroscience 18: 277787.
Beiser, David G., Sherwin S. Hua, and James C. Houk. 1997. Network
models of the basal ganglia. Current Opinion in Neurobiology
7: 18590.
Brenowitz, Eliot, and Michael D. Beecher. 2005. Song learning in
birds: Diversity and plasticity, opportunities and challenges. Trends
in Neurosciences 28: 12732.
Blended Space
Doupe, Allison J., and Patricia Kuhl. 1999. Birdsong and human
speech: Common themes and mechanisms. Annual Review of
Neuroscience 22: 567631.
Doupe, Allison J., David J. Perkel, Anton Reiner, and Edward A. Stern.
2005. Birdbrains could teach basal ganglia research a new song.
Trends in Neurosciences 28: 35363.
Fuster, Joaquin M. 1995. Memory in the Cerebral Cortex: An Empirical
Approach to Neural Networks in the Human and Nonhuman Primate.
Cambridge, MA: MIT Press.
Jackendoff, Ray. 1990. Semantic Structures. Cambridge, MA: MIT Press.
. 2002. Foundations of Language: Brain, Meaning, Grammar,
Evolution. New York: Oxford University Press.
Lieberman, Philip. 2000. Human Language and Our Reptilian Brain.
Cambridge: Harvard University Press.
Marcus, Gary F., and Simon E. Fisher. 2003. FOXP2 in focus: What can
genes tell us about speech and language? Trends in Cognitive Sciences
7: 25762.
Rose, Gary, Franz Goller, Howard J. Gritton, Stephanie L. Plamondon,
Alexander T. Baugh, and Brendon G. Cooper. 2004. Species-typical
songs in white-crowned sparrows tutored with only phrase pairs.
Nature 432: 7538.
Teramitsu, Ikuku, Lili C. Kudo, Sarah E. London, Daniel H. Geschwind,
and Stephanie A. White. 2004. Parallel FOXP1 and FOXP2 expression in songbirds and human brain predicts functional interaction.
Journal of Neuroscience 24: 315263.
Vargha-Khadem, Faraneh., David G. Gadian, Andrew Copp, and
Mortimer Mishkin. 2005. FOXP2 and the neuroanatomy of speech
and language. Nature Reviews: Neuroscience 6: 1318.
BLENDED SPACE
A blended space is one element of the model of meaning construction proposed by conceptual blending theory. In this
framework, mental representations are organized in small, selfcontained conceptual packets (Fauconnier and Turner 2002,
40) called mental spaces, which interconnect to form complex
conceptual networks. In a conceptual integration network, or
blend, some mental spaces serve as input spaces that contribute elements to a new, blended mental space (Fauconnier and
Turner 1994, 1998, 2002).
The minimal conceptual integration network connects four
mental spaces: two inputs, a generic space that contains all the
structures that the inputs seem to share, and a blended space. The
conventional illustration of this prototypical network, sometimes
called the Basic Diagram (Fauconnier and Turner 2002, 467),
shows four circles marking the points of a diamond, with the
circle representing the generic space at the top and the blended
space at the bottom. However, this four-space model is only the
minimal version of the integration network; in conceptual blending theory, networks can contain any number of input spaces.
Blended spaces can also serve as inputs to new blends, making
elaborate megablends (Fauconnier and Turner 2002, 1513).
What makes a blended space special is that it contains newly
emergent structure that does not come directly from any of the
inputs. For example, understanding This surgeon is a butcher
involves selective projection from inputs of butchery and surgery, but the inference that the surgeon is incompetent arises
only in the blended space.
There is some potential for confusion regarding the terminology used to distinguish blended spaces from other theoretical
Reading by Touch
Vera Tobin
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
131
based languages use the same braille patterns, associated contractions vary. Thus, multilingual reading requires the learning
of language-unique contractions.
During reading, scanning movement across text evokes
intermittent mechanical stimulation from contacting successive braille cells, which activates most low-threshold cutaneous
mechanoreceptors found in the fingertip (Johnson and Lamb
1981). A spatial-temporal transformation of the evoked peripheral activity indicates an isomorphic reproduction of braille
cell shapes across a population of mechanoreceptors (Phillips,
Johansson, and Johnson 1990). Through connecting sensory
pathways, these physiological representations of braille cell
shape are conveyed to primary somatosensory cortex (Phillips,
Johnson, and Hsaio 1988) in the parietal lobe. Despite the
expected isomorphic representation of braille cell shapes in
somatosensory cortex, we do not know whether tactile reading in
fact relies on holistically discriminating shape.
Braille cell patterns also differ in the density of dot-gaps,
which is perceived as variations in texture (Millar 1985). These
texture changes produce a dynamically shifting lateral mechanical shearing across the fingertip as it moves over braille text in
fluent reading. Good braille readers attend to these temporally
extended stimulation patterns, as opposed to global-holistic
spatial shapes (Millar 1997, 337). In addition, top-down linguistic content drives perceptual processing in skillful readers, for
whom the physical attributes of the text are subservient to lexical
content. In other words, they do not puzzle out words, letter by
letter; instead, they recognize them due in part to their physical
properties but also to semantic context, the familiarity of words
stored in their mental lexicon, and so on. Less accomplished
readers trace shape by making more disjointed trapezoidal finger movements over individual cells, a strategy that fluent readers utilize when asked to identify particular letters, which is a
shape-based task (Millar 1997, 337).
132
133
Bounding
(3) *[Handsomei though [S I believe [NP the claim that [S Dick is ti]]],
Im still going to marry Herman.
BOUNDING
For all its modernity and insights into the fundamental workings
of language,Noam Chomskys early writings (1955, 1957) contain a curious gap: They do not contain any explicit discussion of
locality. One does not even find extensive discussion of the fact
that movement appears to be potentially unbounded. This gap is
all the more curious from our current perspective where locality
and long-distance dependencies are arguably the major area of
study in theoretical syntax.
We owe our modern interest in locality to John R. Rosss
([1967] 1986) seminal work in which the concept of island was
introduced. Rosss thesis is full of examples of long-distance
dependencies like (1a and b).
(1) a. Handsome though Dick is, Im still going to marry
Herman.
b. Handsome though everyone expects me to try to force Bill
to make Mom agree that Dick is, Im still going to marry
Herman.
134
Rehabilitation
The next major step forward in neurolinguistics lay in developing
rehabilitation techniques for those with language impairment.
One important group had been identified by the ophthalmologist James Hinshelwood (1902), who described the case of a child
who had particular difficulty learning to read despite normal
135
Lateral Dominance
In the 1950s and 1960s, American psychology was also developing
more rigorous methods of studying behavior, and brain, though
not necessarily linking them yet. Brocas late-nineteenth-century
observation that aphasia tended to arise primarily from damage
to the left hemisphere of the brain, rather than the right, took on
a new life as the techniques of dichotic listening and tachistoscopic presentation evolved to study lateral dominance in nonbrain-damaged individuals. In dichotic listening, series of three
or so pairs of words are presented simultaneously, one to each
ear, and participants are asked to recall as many words as they
can. Because the primary connections between ear and brain are
136
Lenticular nucleus
(putamen and globus pallidus)
BASAL GANGLIA
OF FORK-BRAIN
Figure 2. Schematic representation of the chief ganglionic categories. Adapted from Gray, Henry. Anatomy
of the Human Body. Edited by Warren H. Lewis.
Philadelphia: Lea & Febiger, 1918.
137
138
Brocas Area
BROCAS AREA
Alexander, M. P., and M. A. Naeser. 1988. Cortical-subcortical differences in aphasia. In Language, Communication and the Brain
Research Publications: Association for Research in Nervous and Mental
Disorders. Vol. 66. Ed F. Plum, 21528. New York: Raven Press.
Bates L. and B. Wulfeck. 1989. Comparative aphasiology: A crosslinguistic approach to language breakdown. Aphasiology 3: 11142.
Caplan, D., G. Waters, D. Kennedy, N. Alpert, N. Makris, G. DeDe, J.
Michaud, and A. Reddy. 2007. A study of syntactic processing in aphasia II: Neurological aspects. Brain and Language 101: 15177.
Freud, S. [1891] 1953. Zur Auffassung der Aphasien. Trans. E. Stengel as
On Aphasia. New York: International University Press.
Galaburda, A., and T. Kemper. 1979. Cytoarchitectonic abnormalities in
developmental dyslexia: A case study. Annals of Neurology 6: 94100.
Geschwind, N. 1965. Disconnexion syndromes in animals and man.
Brain 88: 585644.
Goodglass, H., and E. Kaplan. 1972. The Assessment of Aphasia and
Related Disorders. Philadelphia: Lea & Febiger.
Goodglass, H., and Wingfield, A. 1997. Anomia: Neuroanatomical and
Cognitive Correlates. San Diego, CA: Academic Press.
Lecours, A.R., F. Chain, M. Poncet, J.-L. Nespoulous, and Y. Joanette.
1992. Paris 1908: The hot summer of aphasiology or a season in the
life of a chair. Brain and Language 42: 10552.
Menn, L., and L. K. Obler, eds. 1990. Agrammatic Aphasia: A CrossLanguage Narrative Sourcebook. Vol. 3. Amsterdam: John Benjamins.
Obler, L.K., and K. Gjerlow. 1999. Language and the Brain.
Cambridge: Cambridge University Press.
Orton, S. [1937] 1989. Reading, Writing and Speech Problems in Children
and Selected Papers. Repr. Austin, TX: International Dyslexia
Association.
Pitres, A. 1895. Etude sur laphasie des polyglottes. Rev. Md.
15: 87399.
Schuell, H., J. Jenkins, and E. Jimenez-Pabon. 1964. Aphasia in Adults,
New York: Harper & Row.
Sidtis, J. 2007. Some problems for representations of brain organization based on activation in functional imaging. Brain and Language
102: 13040.
139
Brocas Area
During the 1970s, a number of studies reported that comprehension of complex syntactic forms was also disrupted in this
patient group. A seminal study by A. Caramazza and E. B. Zurif
(1976) reported that Brocas aphasics had particular difficulty
understanding semantically reversible versus irreversible sentences (e.g., The cat that the dog is biting is black vs. The apple
that the boy is eating is red). This study concluded that Brocas
area mediated syntactic processes critical to both production
and comprehension. By extension, Brocas area came to be associated with syntactic processing.
Although many subsequent studies supported this general
notion, several others pointed out the need for caution. For example, M. C. Linebarger, M. Schwartz, and E. Saffran (1983) showed
that agrammatic patients could make accurate grammaticality judgments, which would seem to challenge the notion
of Brocas area being broadly involved in syntactic processing.
Second, many patients with fluent aphasia and lesions outside of
Brocas area have been found to exhibit grammatical deficits that
140
Brocas Area
a large left hemisphere lesion that extends beyond Brocas area
to include underlying white matter, adjacent frontal cortex, and
insular cortex (Alexander, Naeser, and Palumbo 1990; Mohr et
al. 1978). Color Plate 3 shows a lesion overlay map of 36 patients
with chronic Brocas aphasia persisting more than one year. As
can be seen, the region of common overlap (shown in dark red)
is not Brocas area but rather medial regions (i.e., more central),
namely, insular cortex and key white matter tracts. Indeed, only
5060 percent of patients with lesions extending into Brocas
area have a persistent Brocas aphasia. Lesions restricted to
Brocas area tend to cause a transient mutism followed by altered
speech output, but not a chronic Brocas aphasia (Mohr et al.
1978; Penfield and Roberts 1959). These findings would suggest
that Brocas area proper might be more involved in later stages of
speech production (e.g., articulation).
Even early studies reported contradictory cases, namely,
patients with Brocas area affected but no Brocas aphasia or
patients with Brocas aphasia but no lesion in Brocas area (e.g.,
Marie 1906; Moutier 1908). Figure 2 shows an example of an
individual with Brocas aphasia whose lesion spares Brocas area
(left) and one with word-finding problems but no Brocas aphasia after a lesion to Brocas area (right).
Although Broca deduced that the critical region for his
patients articulation disturbance was the inferior frontal gyrus,
he realized that his patients lesions most likely extended more
medially. However, he wanted to maintain the brains for posterity and chose not to dissect them (see Figure 3). Recently, N. F.
Dronkers and colleagues (2007) had the opportunity to acquire
three-dimensional MRI images of the brains of Brocas two original patients (Leborgne and Lelong), which are kept in a Paris
museum. They found that the lesions in both patients extended
quite medially, involving underlying white matter, including the
superior longitudinal fasciculus. Moreover, one of the patients
brains (Leborgnes) had additional damage to the insula, basal
ganglia, and internal and external capsules. With respect to
the extent of frontal involvement, Leborgnes lesion affected the
middle third of the inferior frontal gyrus to the greatest extent,
with only some atrophy in the posterior third. In Brocas second
patient, Lelong, the lesion spared the pars triangularis, affecting only the posterior portion of the pars opercularis. Thus, even
what is commonly referred to as Brocas area (BA 44, 45) is not
exactly the region affected in Brocas original patients.
141
Cartesian Linguistics
Caplan, D., N. Hildebrandt, and N. Makris. 1996. Location of lesions in
stroke patients with deficits in syntactic processing in sentence comprehension. Brain 119: 93349.
Caramazza, A., E. Capitani, A. Rey, and R. S. Berndt. 2001. Agrammatic
Brocas aphasia is not associated with a single pattern of comprehension performance. Brain and Language 76: 15884.
Caramazza, A., and E. B. Zurif. 1976. Dissociation of algorithmic and
heuristic processes in language comprehension: Evidence from aphasia. Brain and Language 3: 57282.
Dick, F., E. Bates, B. Wulfeck, M. Gernsbacher, J. A. Utman, and
N. Dronkers. 2001. Language deficits, localization, and grammar: Evidence for a distributive model of language breakdown in
aphasics and normals. Psychological Review 108: 75988.
Dronkers, N. F., O. Plaisant, M. T. Iba-Zizen, and E. A. Cabanis. 2007.
Paul Brocas historic cases: High resolution MR imaging of the brains
of Leborgne and Lelong. Brain 130.5: 143241.
Fiebach, C. J., M. Schlesewsky, G. Lohmann, D. Y. von Cramon, and A.
D. Friederici. 2005. Revisiting the role of Brocas area in sentence
processing: Syntactic integration versus syntactic working memory.
Human Brain Mapping 24: 7991.
Friederici, A. D., M. Meyer, and D. Y. von Cramon. 2000. Auditory
language comprehension: An event-related fMRI study on the processing of syntactic and lexical information. Brain and Language
74: 289300.
Linebarger, M. C., M. Schwartz, and E. Saffran. 1983. Sensitivity to
grammatical structure in so-called agrammatic aphasics. Cognition
13: 36193.
Marie, P. 1906. Revision de la question de laphasia: La troiseme circonvolution frontale gauche ne joue aucun role special dans la function du
langage. Semaine Medicale 26: 2417.
Mohr, J., Pessin, S. Finkelstein, H. H. Funkenstein, G. W. Duncan, and
K. R. Davis. 1978. Broca aphasia: Pathologic and clinical. Neurology
28: 31124.
Moutier, F. 1908. Laphasie de Broca. Paris: Steinhell.
Penfield, W., and L. Roberts. 1959. Speech and Brain Mechanisms.
Princeton, NJ: Princeton University Press.
Stowe, L. A., M. Haverkort, and F. Zwarts. 2005. Rethinking the neurological basis of language. Lingua 115: 9971042.
C
CARTESIAN LINGUISTICS
This term began as the title of a 1966 monograph by Noam
Chomsky. It has become the name of a research strategy for the
scientific study of language and mind that Chomsky in other
works calls rationalist or biolinguistic, which he contrasts
to an empiricist strategy (see biolinguistics). Cartesian
Linguistics illuminates these strategies by focusing on contrasting
assumptions concerning mind and language and their study that
are found in the writings of a selection of philosophers and linguists from the late sixteenth century through 1966. The rationalists include Descartes, the Port-Royal Grammarians, Humboldt,
Cudworth, and clearly Chomsky himself in 1966 and now.
The empiricists include Harris, Herder, the modern linguists
(L. Bloomfield, M. Joos, etc.), and again clearly behaviorists,
connectionists, and others attracted to the idea that children
learn languages rather than growing them.
142
Case
Rationalists adopt a nativist (see innateness and innatism)
and internalist approach to the study of language. Support for
nativism is found in poverty-of-the-stimulus observations. To take
these observations seriously, rationalists believe, someone constructing a theory of language is advised to assume that much of
linguistic structure and content is somehow latent in the infants
mind: Experience serves to trigger or occasion structure and
content, not form and constitute them. Descartes himself was a
rationalist and appealed to poverty facts to support his views of
innate and adventitious (but not made up) concepts/ideas.
Until taken up by his Port-Royal followers, however, there was little attention to the innateness of the structure of language itself.
Descartess greater contribution to the strategy with his name
lies in less discussed but equally important observations concerning the creative aspect of language use (see creativity in
language use). Encapsulated in Descartess Discourse V, these
note that speakers can on occasion produce any of an unbounded
set of expressions (unboundedness), without regard to external
and internal stimulus conditions (stimulus freedom), sentences
that nevertheless are appropriate and coherent with respect to
discourse context (appropriateness and coherence). Taking
these seriously suggests a scientific research strategy that focuses
not on linguistic action/behavior (language use) itself, for that
is in the domain of free human action, but on an internal system,
the language faculty. A science of linguistic action would not
only have to take into account the speakers understanding of reasons for speaking and the job that an utterance is understood to
perform but would also have to say what a person will utter on a
specific occasion. No extant science can do that, and likely none
ever will. Given stimulus freedom, independently specifiable
conditions for utterance are unavailable, and there is no upper
bound on sentences appropriate to a speakers understanding of
discourse circumstance. The scientist of language should focus
on language as an internal system, on competence, not on what
people do by using the tools their systems offer. Science can
say what a language can yield; a theory of competence does that.
But it likely cannot say what will be said, when it will be said, or
whether it is appropriate and, if so, why.
Generally speaking, the rationalist strategy treats languages
as natural (native) systems in the head, not artifacts. The empiricist strategy (not empirical science) treats languages as artifacts
in the head, as socially constituted sets of practices or behaviors,
mastery of which (learning) requires training and negative evidence. Cartesian Linguistics emphasizes that there are few, if
any, linguistic practices and that children grow languages.
James McGilvray
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Chomsky, Noam. 1966. Cartesian Linguistics. New York: Harper and
Row.
. 2003. Cartesian Linguistics. 3rd ed. Ed. and introd. J. McGilvray.
Cambridge: Cambridge University Press.
CASE
This term has been traditionally employed to designate the type of
morphological ending that indicates the syntactic function of the
Categorial Grammar
noun phrase that bears it. In Latin, for instance, the word meaning
girl may surface as puella (nominative case), puellam (accusative
case), or puellae (genitive case), depending on whether it is the subject of the sentence, is the object of a transitive verb, or stands in a
possessor relation with respect to another noun. Languages vary with
respect to the morphological case distinctions they make. Languages
such as Latin have six case distinctions (nominative, genitive, dative,
accusative, vocative, and ablative) whereas languages such as Chinese
have none. One may also find languages like English in which only a
subset of nominal elements display case distinctions as in he (nominative), him (accusative), and his (genitive).
On the basis of a suggestion by Jean-Roger Vergnaud (Rouveret
and Vergnaud 1980), Noam Chomsky (1981) developed a theory
of abstract case (annotated as Case). According to this theory, it is
a property of all languages that noun phrases can only be licensed
in a sentence if associated with Case. Whether or not the abstract
Cases get morphologically realized as specific markings on (some)
noun phrases is a language-specific property. Research in the last
two decades has been devoted to identifying i) which elements
are Case-licensers, ii) which structural configurations allow Case
licensing, and iii) what the precise nature of such licensing is. The
contrast between John/he/*him/*his sang and *John/*he/*him/*his
to sing, for example, has led to the conclusion that in English, the
past tense may license nominative Case, but the infinitival to is not
a Case-licenser. In turn, the contrast between John/he/*him/*his
was greeted and was greeted *John/*he/*him/*his indicates that
nominative is licensed by the past tense if the relevant noun phrase
occupies the subject, but not the object, position.
When a given Case only encodes syntactic information, it is
referred to as structural Case. Nominative and accusative cases
in English are prototypical examples of structural Case. In John
greeted her and she was greeted by John, for example, the pronoun
bearing the thematic role of patient has accusative Case in
the first sentence, but nominative in the second. On the other
hand, when a given Case also encodes thematic information,
it is referred to as inherent Case (Chomsky 1986; Belletti 1988).
The preposition of in English, for example, has been analyzed
as a marker of inherent Case, for it only licenses a noun phrase
that is the complement of the preceding noun: It may license that
country in the invasion of that country but not in *the belief of that
country to be progressive because that country is the complement
of invasion but not of belief.
Jairo Nunes
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Belletti, A. 1988. The Case of unaccusatives. Linguistic Inquiry 19: 134.
Chomsky, N. 1981. Lectures on Government and Binding.
Dordrecht: Foris.
. 1986. Knowledge of Language: Its Nature, Origin and Use.
New York: Praeger.
Rouveret, A., and J. R. Vergnaud. 1980. Specifying reference to the subject: French causatives and conditions on representations. Linguistic
Inquiry 11: 97202.
project word order and build logical forms via general rules of
grammar. First proposed by Kazimierz Ajdukiewicz (1935), it is
arguably the oldest lexicalized formalism. CG is not itself a theory of natural language grammar, but its use has many linguistically relevant ramifications: lexicalization, flexible constituency,
semantic transparency, control over generative capacity, and
computational tractability.
CATEGORIAL GRAMMAR
Categorial grammar (CG) is a family of formalisms that model
Categorial Grammar
Figure 1.
Type Dependency
CG with just function application is weakly equivalent to standard context-free phrase structure grammar. Nonetheless, the
approach is radically different: Syntactic well-formedness in CG
is type dependent rather than structure dependent, and derivation is an artifact rather than a representational level. Also, categories labeling the nodes of categorial derivations are much
more informative than the atomic symbols of constituent
structure produced by PSGs. Subcategorization is directly
encoded in categories like s\np, (s\np)/np, and ((s\np)/np)/
np, rather than with stipulated nonterminal symbols such as
V-intrans, V-trans, and V-ditrans that have no transparent connection to their semantic types. Furthermore, there is a systematic correspondence between notions such as intransitive and
transitive: After the transitive category (s\np)/np consumes its
object argument, the resulting category s\np is that of an intransitive verb.
More importantly, type dependency shifts the perspective
on grammar shared with tree-adjoining grammar and headdriven phrase structure grammar away from a topdown one in which phrase structure rules dictate constituent
structure into a bottom-up one in which lexical items project
structure through the non-language-specific rules (i.e., CGs
universal grammar). Recent developments in the transformational grammar tradition, such as minimalism, have
also incorporated such a lexically driven perspective.
144
CG moves further from PSGs and other frameworks by incorporating other rules that provide new kinds of inference over categories. These rules are responsible for the type-driven flexible
constituency for which CG is well known. The two main branches
of CG can be broadly construed as rule based, exemplified by
combinatory categorial grammar (CCG) (Steedman 2000), and
deductive, exemplified by categorial type logics (CTL) (Moortgat
1997; Oehrle in press). We discuss both of these briefly.
CCG adds a small set of syntactic rules that are linear counterparts of the combinators from combinatory logic. Two combinators, composition (B) and type-raising (T), lead to the following
rules, among others:
Categorial Grammar
Figure 2.
(n\n)/(s/np), which seeks types which are missing objects such
as Olivia likes and Olivia gave Finn, which both have the type s/
np. Extraction is thus handled without appeal to movement or
traces. long-distance dependencies in object extraction are
captured because forward composition allows the unsaturated
argument to be successively passed up until it is revealed to the
relativizer, as in Figure 3. Under the standard assumption that
coordination combines constituents of like types, then right-node
raising is simply constituent coordination:
[Kestrel heard]s/np and [Finn thinks Olivia saw]s/np the plane
flying overhead.
Figure 3.
145
Categorization
Steedman, Mark, and Jason Baldridge. Combinatory categorial grammar. In Non-Transformational Syntax: A Guide to Current Models, ed.
Robert Borsley and Kersti Brjars. Malden, MA: Blackwell. In press.
Vermaat, Willemijn. 2005. The logic of variation. A cross-linguistic
account of wh-question formation. Ph.D. diss., Utrecht University.
Wood, Mary McGee. 1993. Categorial Grammar. London: Routledge.
CATEGORIZATION
William Labov (1973, 342) stated, If linguistics can be said to be
any one thing, it is the study of categories: that is, the study of how
language translates meaning into sound through the categorization of reality into units and sets of units. Labov is here addressing the relation between linguistic expressions and the things
and situations to which the expressions are used to refer. The
circumstances of the world are limitless in their variety; linguistic resources are finite. Since it is not possible to have a unique
name for every entity that we encounter or a special expression
for every event that happens, we need to categorize the world
in order to speak about it. We need to regard some entities, and
some events, as being the same as others.
The relation between a word and a referent is not direct but
is mediated by the words meaning. It is in virtue of its meaning that a word can be used to refer. A words meaning can be
thought of as a concept, and a concept, in turn, can be thought of
as a principle of categorization. To know the word mug (to take
one of Labovs examples) is to have the concept of a mug, which
in turn means being able to use the word appropriately, namely,
for things that are called mugs. This goes not only for names of
concrete things like mugs but also for names of abstract entities
and for words of other syntactic categories. To state that X is on
Y is to categorize the relation between X and Y as an on-relation
rather than an in- or an at-relation. We can make similar claims
for other elements in a language, such as markers of tense and
aspect. To describe an event in the present perfect as opposed to
the past simple or to use progressive as opposed to non-progressive aspect is to categorize the event in a manner consistent with
the concept designated by the morpho-syntactic elements.
On many counts, therefore, linguists need a theory of categorization. The theory must provide answers to two related questions. On what basis are entities assigned to a category? And why
do we categorize the world in just the way that we do?
According to what has come to be known as the classical
or Aristotelian theory, a category is defined in terms of a set of
necessary and sufficient conditions; it follows that
things belong in a category because they exhibit each of the
defining features. There are many problems associated with
this view. First, it often is just not possible to list the defining
features. What, for example, are the defining features of mug as
opposed to cup? Then there is the question of the features themselves. Each feature will itself define a category, which in turn
must be defined in terms of its necessary and sufficient features.
Unless we are prepared to postulate a set of primitive features
out of which all possible categories are constructed, we are
faced with an infinite regress. Finally, the classical theory makes
no predications about why we should have the categories that
we do. Any conceivable combination of features could constitute a valid category.
146
A major landmark in the development of a nonclassical theory of categorization was the work of psychologist Eleanor Rosch
(1978). She argued that categories have a prototype structure, that is, are centered around good examples, and that things
belong to the category in virtue of their exhibiting some similarities with the prototype. The members of a category, therefore,
do not need to share the same set of features. Moreover, some
members can be better or more representative examples of the
category than others.
Rosch addressed not only the internal structure of categories
but also the question of what makes a good category. Good categories the ones that people operate with, and which are likely
to be encoded in human languages are those that deliver maximum information to the user with minimal cognitive effort. We
can approach this matter in terms of the interplay of cue validity
and category validity. Cue validity means that having observed
that an entity exhibits a certain feature, you can assign the entity,
with a fair degree of confidence, to a certain category. Category
validity means that having learned that an entity belongs to a certain category, you have expectations about the likely properties
of the entity. In this way, we can infer quite a lot about the things
that we encounter, on the basis of minimal information about
them.
Categories also need to be studied against broader conceptual and cultural knowledge having to do with human intentions and purposes, presumed causal relations between things
and events, and beliefs about how the world is structured. We
can imagine all kinds of hypothetical categories say, a category comprising things that are yellow, weighing under five kilograms, and manufactured in 1980. Such a category is unlikely to
be lexicalized in any human language. It displays very low cue
and category validity and would, therefore, not be useful to its
users. It would also have no role to play in any broader knowledge system.
As already recognized by Labov, the issue of categorization
applies not only to the categories we use in talking about the
world but also to the analysis of language itself. The very terminology of linguistic description is replete with names of categories, such as phoneme, noun, direct object, word, dialect,
and so on, and practical linguistic description involves assigning
linguistic phenomena to these various categories. Although the
classical approach to categorization is still very strong among
linguistic theoreticians, it is increasingly recognized that the categories of linguistics may have a prototype structure and are to
be understood, in the first instance, in terms of good examples.
As Labov remarked, linguistics is indeed the study of
categories!
John R. Taylor
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Labov, William. 1973. The boundaries of words and their meanings. In
New Ways of Analyzing Variation in English, ed. C. J. Bailey and R. W.
Shuy, 34072. Washington, DC: Georgetown University Press.
Lakoff, George. 1987. Women, Fire, and Dangerous Things: What
Categories Reveal About the Mind. Chicago: University of Chicago
Press.
Murphy, Gregory. 2002. The Big Book of Concepts. Cambridge, MA: MIT
Press.
Causative Constructions
Rosch, Eleanor. 1978. Principles of categorization. In Cognition
and Categorization, ed. E. Rosch and B. Lloyd, 2748. Hillsdale,
NJ: Lawrence Erlbaum.
Taylor, John R. 2003. Linguistic Categorization. Oxford: Oxford University
Press.
CAUSATIVE CONSTRUCTIONS
Causative construction (CC) is defined as form-meaning mapping that encodes a causative situation (CS) in which an entity
(typically human), or causer, acts upon another entity (typically
human), or causee, to induce some action. Cross-linguistically,
this mapping is known to operate using periphrastic syntactic
construction, valence-increasing morphology, or lexical verb.
Three structural types of CCs thus identified are a) the syntactic
causative, which employs a periphrastic causative verb like make
in English (1); b) the morphological causative, which employs a
causative affix like -(s)ase- in Japanese (2); and the lexical causative, wherein a transitive verb with inherently causative meaning is employed (3), e.g. kiseru to put (clothes) on (someone)
in Japanese.
(1)
(2)
(3)
Subject > Direct Object > Indirect Object > Oblique (Whaley
1997, 193)
The accusative case marker o indexes a lesser degree of control retained by the causee than the dative case marker -ni.
This semantic difference between accusative (patient-marking)
case and dative (experiencer-marking) case is captured by the
Otayori yom-as-ase-te
itadaki-masu.
letter read- caus-caus-conj humbly receive-pol:nonpast
Allow me to read this letter.
(Okada 2003, 29)
In this instance, as contrasted with its single causative counterpart yom-ase-ite itadaku (read-caus-humbly receive), the
double causative serves to reinforce the speakers expression of
humbleness and politeness.
CCs have also been productively investigated by more formally oriented linguists (e.g., Kuroda 1993, Miyagawa 1998).
Kaoru Horie
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Cole, Peter. 1983. The grammatical role of the causee in universal grammar. International Journal of American Linguistics 49: 11533.
Comrie, Bernard. 1976. The syntax of causative constructions: Crosslanguage similarities and divergences. In Shibatani 1976b, 261312.
. 1985. Causative verb formation and other verb-deriving morphology. In Language Typology and Syntactic Description. Vol.
3: Grammatical Categories and Lexicon. Ed. Timothy Shopen, 30948.
Cambridge: Cambridge University Press.
Givn, Talmy. 1980. The binding hierarchy and the typology of complements. Studies in Language 4: 33377.
Haiman, John. 1983. Iconic and economic motivation. Language
59: 789811.
147
C-Command
Cerebellum
C-COMMAND
An enduring and fundamental hypothesis within syntactic theory is that the establishment of most, if not all, syntactic relations (agreement, binding, case, control structures,
movement, etc.) requires c-command.
Tanya Reinhart (1979) provides the following definition of
c-command (see also Edward Klimas (1964) in construction
with):
(1)
Sentence
Noun Phrase
Verb Phrase
Noun Phrase
Noun
Verb
Noun Phrase
Marys
mother
criticizes
herself
148
herself, Marys and herself are unable to enter into such a relation
(see anaphora and binding).
Although Reinharts pioneering definition is formally explicit
and strongly supported empirically, questions arise regarding explanatory depth, as with any definition. In this respect,
S. Epstein and colleagues (1998) ask:
(i) Why should this formal relation, and not any definable
other, constrain syntactic relations?
(ii) Why is (first) branching relevant?
(iii) Why must not dominate or equal ?
Epstein and colleagues argue that (iiii) receive natural answers
under a bottom-up derivational approach with recursive application of the binary operation merge, as independently motivated in minimalism (Chomsky 1995). C-command is then
arguably an emergent property of this structure-building process
and is expressible in terms of merge, as in (3):
(3)
CEREBELLUM
The cerebellum is a brain structure located underneath the posterior part of the cerebral hemispheres. Three anatomical loops
Cerebellum
connect the cerebellum to various parts of the nervous system
(Ramnani 2006). The cerebro-cerebellar loop has been of particular interest to language researchers because this pathway supports anatomical connections between the cerebellum and the
cortex, potentially including language-related cortical regions in
the contralateral cerebral hemisphere. Recent advances in neuroimaging techniques, clinical testing, and anatomical methods
provide evidence that strongly implicates the cerebellum in a
broad range of language-related tasks, including those involving speech production and perception, single-word reading, and
higher-level language processing.
Historical Perspectives
Historically, the functions of the cerebellum were thought to be
limited to motor processes, such as motor control, performance,
and skill acquisition. Basic neuroscience research has led to different proposals regarding its role in motor processes such as errordriven learning (Marr 1969) and internal timing (Ivry 1996). In the
late 1980s, H. C. Leiner, A. L. Leiner, and R. S. Dow (1986) proposed
that the cerebellum is not exclusively involved in motor functions
but that it also contributes to cognitive processes. Specifically, they
argued for a putative role in language because the evolutionary
development of the cerebellum paralleled a similar evolution of
cortical areas associated with linguistic functions (e.g., brocas
area) (Leiner, Leiner, and Dow 1993). Based on the homogeneity
of cerebellar cellular organization, a similar role was attributed to
the cerebellum across both motor and non-motor domains.
Empirical work providing support for the claims proposed by
Leiner, Leiner, and Dow began to emerge in the late 1980s and
1990s (Desmond and Fiez 1998). A positron emission tomography (PET) study conducted by S. E. Petersen and J. A. Fiez (1993)
showed increases in cerebellar activity during a verb generation
task. This neuroimaging finding was consistent with a follow-up
case study of a patient with a lateral cerebellum lesion (Fiez et
al. 1992). The patient showed particularly poor performance on
verb generation despite the fact that other neuropsychological
assessments were within the normal range.
Current Perspectives
The cerebellum has been implicated in a broad range of languagerelated tasks. The majority of the work, however, can be related
to one of three domains: 1) speech production and perception,
2) reading, and 3) higher-level word processing. In order to
account for the cerebellums function in language, investigators
have made reference to the timing and error correction functions
that have been attributed to the cerebellum in the motor literature. For a review on how these may be general mechanisms that
contribute to both motor and non-motor processes, see Ivry and
Spencer (2004) and Doya (2000).
SPEECH PRODUCTION AND PERCEPTION. During speech production, the control, coordination, and timing of movements are
essential. Not surprisingly, clinical findings demonstrate profound speech and motor deficits associated with lesions to the
cerebellum. One common speech disorder resulting from damage to the cerebellum is dysarthria, which is characterized by distorted and slurred speech that is often monotonic and of a slower
rate (Duffy 1995). neuroimaging studies provide further
evidence for the involvement of the cerebellum in speech production. In a recent study, participants performed a syllable repetition task in which speech rate was varied. The results showed
increases in cerebellar activity that corresponded with increases
in speech rate (Riecker et al. 2006).
Cerebellar contributions to speech extend to the domain of
perception. Lesions to the cerebellum produce deficits in temporal duration discrimination and impair categorical perception
for consonants that differ in the onset of voicing (Ackermann et
al. 1997). This clinical evidence is consistent with neuroimaging data that show increases in right cerebellar activity during a
duration discrimination task for linguistic items (Mathiak et al.
2002). Other neuroimaging results suggest that the cerebellum
is also involved in learning new perceptual distinctions, such
as the non-native /r/-/l/ phonetic contrast for Japanese speakers (Callan et al. 2003). As in the motor literature, many of these
studies provide evidence that the cerebellum may be important
for coordination and timing in the production as well as the perception of speech (for a discussion, see Ivry and Spencer 2004).
Other research also draws upon knowledge from the motor
literature to emphasize a potential role of the cerebellum in error
correction. The fluency of normal speech has led many models
of speech production to incorporate a mechanism for monitoring and correcting speech errors (Postma 2000). More detailed
computational work has mapped certain processes in speech
production to specific brain regions and defined the cerebellum
as an important component in monitoring (Guenther, Ghosh,
and Tourville 2006). Similar ideas have emerged in models of
verbal working memory. Specifically, J. E. Desmond and colleagues (1997) suggest that a rehearsal process that relies on
inner speech to maintain verbal items in working memory may
also implement an error correction process. In their model,
inputs from frontal and parietal cortex into superior and inferior regions of the cerebellum are used to calculate and correct
discrepancies between phonological and articulatory codes in
order to improve memory performance.
READING. Data from neuroimaging studies consistently show
cerebellar activation during single-word reading tasks (Fiez and
Petersen 1998; Turkeltaub et al. 2002). In addition, individuals with
developmental reading disorders show some of the same symptoms that are often seen in patients with cerebellar damage, such
as poor duration discrimination and impaired gross motor functions (Nicolson, Fawcett, and Dean 1995). These observations led
R. Nicolson, A. Fawcett, and P. Dean (2001) to propose a relationship between cerebellar deficits and developmental reading disorders. Consistent with this idea, anatomical findings have reported
smaller right anterior lobes of the cerebellum in children diagnosed
with developmental dyslexia (Eckert et al. 2003). This work in
developmental reading disorders has focused on the importance of
cerebellar involvement in coordination and timing. Integrating the
neuroanatomical findings with behavioral work on dyslexia will be
key for establishing a specific role for the cerebellum in reading.
Recent neuropsychological research provides mixed findings on the causal relationship between lesions to the cerebellum in adult skilled readers and reading difficulties. One study
found that patients with lesions to the cerebellar vermis had
more errors in single-word reading when compared to controls
149
Cerebellum
(Moretti et al. 2002). On the other hand, a study of native English
speakers with lesions to the lateral cerebellar hemispheres did
not find any reading difficulties at the level of single words or
text (Ben-Yehudah and Fiez 2008). These seemingly inconsistent
findings may be due to differences in the site of the cerebellar
lesions in the two patient groups.
HIGHER-LEVEL LANGUAGE. There is accumulating evidence that
higher-level language processes may also involve the cerebellum,
although this level has received less attention. A meta-analysis of
the neuroimaging literature conducted by P. Indefrey and W. J. M.
Levelt (2004) reveals increased activity in the cerebellum for tasks
that require higher-level word processing; such tasks include picture
naming and verb generation (Indefrey and Levelt 2004) or internal
generation of semantic word associations (Gebhart, Petersen, and
Thach 2002). It is important to note that higher-level language processes seem to recruit more lateral areas, often in the contralateral
right hemisphere of the cerebellum (Indefrey and Levelt 2004).
Neuropsychological studies have observed impairments in higherlevel language processing (Silveri, Leggio, and Molinari 1994; Riva
and Giorgi 2000), including poor performance on a grammaticality
judgment task relative to controls (Justus 2004).
Summary
In summary, these data collectively provide strong support for
cerebellar involvement in many aspects of language, including
speech processing, reading, and higher-level language processing. They also suggest that there may be different regions of the
cerebellum that are involved in different types of language tasks.
This observation is consistent with an emerging concept that
distinct cerebro-cerebellar loops support cerebellar interactions
with cortex, thus potentially enabling the cerebellum to apply
one or more of its suggested functions (e.g., error correction) to
separate input-output loops (Kelly and Strick 2003). According
to this view, language tasks that rely on different cortical regions
would engage distinct cerebro-cerebellar loops that recruit specific cerebellar regions.
Sara Guediche, Gal Ben-Yehudah, and Julie A. Fiez
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Ackermann, H., S. Graber, I. Hertrich, and I. Daum. 1997. Categorical
speech perception in cerebellar disorders. Brain and Language
60: 32331.
Ben-Yehudah, G., and J. Fiez. 2005. Impact of cerebellar lesions on reading and phonological processing. Annals of the New York Academy of
Sciences 1145: 26074.
Callan, D. E., K. Tajima, A. M. Callan, R. Kubo, S. Masaki, and R. AkahaneYamada. 2003. Learning-induced neural plasticity associated with
improved identification performance after training of a difficult second-language phonetic contrast. NeuroImage 19: 11324.
Desmond, J. E., J. Gabrieli, A. Wagner, B. Ginier, and G. Glover. 1997.
Lobular patterns of cerebellar activation in verbal working memory
and finger-tapping tasks as revealed by functional MRI. Journal of
Neuroscience 17.24: 967585.
Desmond, J. E., and J. A. Fiez. 1998. Neuroimaging studies of the cerebellum: Language, learning and memory. Trends in Cognitive Sciences
2.9: 355358. This article reviews neuroimaging evidence suggesting
that the cerebellum is involved in cognitive tasks, including those that
involve learning, memory, and language.
150
Charity, Principle of
Riecker, A., J. Kassubek, K. Groschel, W. Grodd, and H. Ackermann. 2006.
The cerebral control of speech tempo: Opposite relationship between
speaking rate and BOLD signal changes at striatal and cerebellar structures. NeuroImage 29: 4653.
Riva, D., and C. Giorgi. 2000. The cerebellum contributes to higher functions during development: Evidence from a series of children surgically treated for posterior fossa tumours. Brain 123: 105161.
Silveri, M., M. Leggio, and M. Molinari. 1994. The cerebellum contributes to linguistic production: A case of agrammatic speech following a
right cerebellar lesion. Neurology 44.11: 204750.
Turkeltaub, P., G. Eden, K. Jones, and T. Zeffiro. 2002. Meta-analysis of
the functional neuroanatomy of single-word reading: Method and validation. NeuroImage 16.3 (Part 1): 76580.
CHARITY, PRINCIPLE OF
A charity principle is a principle governing the interpretation of
the speech and thought of others. It says that the correct interpretation of certain kinds of expressions, areas of discourse,
or whole languages maximizes truth and rationality across the
(relevant) beliefs of its subject. According to Donald Davidson,
the main defender of a principle of charity, its validity derives
from the essentially rational and veridical nature of belief and
thought.
Principles of charity are of central importance in discussions
of radical interpretation or radical translation. In W. V.
O. Quines version, charity governs the translation of the logical
constants (cf. Quine 1960, 59). According to Donald Davidson,
charity governs the radical interpretation of all expressions of a
language. In an early formulation, it tells the radical interpreter
to optimize agreement by assigning truth conditions to alien
sentences that make native speakers right when plausibly possible (Davidson [1973] 1984, 137). To make native speakers
right is to interpret them as having beliefs that are largely true
and coherent with each other. Later, Davidson distinguished
explicitly between these two aspects of charity:
The Principle of Coherence prompts the interpreter to discover
a degree of logical consistency in the thought of the speaker; the
Principle of Correspondence prompts the interpreter to take
the speaker to be responding to the same features of the world
that he (the interpreter) would be responding to under similar
circumstances. Both principles can be (and have been) called
principles of charity: One principle endows the speaker with a
modicum of logical truth, the other endows him with a degree of
true belief about the world. Successful interpretation necessarily
invests the person interpreted with basic rationality. (Davidson
[1991] 2001, 211)
Coherence restricts belief ascription in terms of the logical relations among the beliefs of a speaker. Correspondence restricts
the ascription of empirical beliefs to a speaker in terms of
their truth. Since this can only be done according to the interpreters own view of what is true, following the principle of
correspondence amounts to agreement maximization
between speaker and interpreter. Here, Davidson more and
more emphasized a causal element; in the most basic perceptual cases, the principle of correspondence calls for the ascription of beliefs shared by speaker and interpreter. The objects of
these beliefs are determined as the shared, external causes of
151
152
grapple with. Howsoever powerful the childs innate mechanisms might be, they do not equate to an attribute (language)
that comes into the world fully formed at birth. Instead, there is a
bridge to be crossed from what Noam Chomsky (1980) has called
the childs initial state (the genetic endowment for language,
present at birth) to the steady state (the mature knowledge of
grammar finally attained). Several explanations are available to
deal with this problem (see innateness and innatism). Of
note here is the simple point that such explanations are required
by nativists and, inevitably, muddy the waters both theoretically
and empirically.
On the nurture side of the nature-nurture fence, speech errors
(grammatical and otherwise) again present a vexing issue that
needs to be addressed. In particular, the behaviorist approach to
language acquisition has been castigated for an excessive reliance
on operant conditioning as a mechanism of language learning.
B. F. Skinner (1957) argued that one of the key processes in language development was the shaping of the childs verbal behavior through reward. On this view, child utterances are rewarded
according to their proximity to the adult models provided. But
this is problematic. In a celebrated demolition of Skinners thesis, Chomsky (1959, 42) pointed out that operant conditioning
cannot tell the whole story since a child will be able to construct
and understand utterances which are quite new, and are, at the
same time, acceptable sentences in his language. Thus, operant
conditioning cannot account for novelty.
Similarly, imitation cannot account for the childs speech,
particularly errors. Although not mentioned by Chomsky (1959),
grammatical errors do, in fact, present the most striking demonstration that language acquisition is not largely based on
imitation. The reason is that children are exposed to very few
grammatical errors in the input they receive from parents.
Hence, there are very few faulty models for children to copy. For
example, Elissa Newport, H. Gleitman, and L. R. Gleitman (1977)
report just one instance of parental ungrammaticality in a corpus of 1,500 utterances directed toward young children. In consequence, one cannot easily blame the parents for child errors.
A further critical point is that the child cannot imitate grammar,
only the products of grammar (sentences). Perhaps not surprisingly, since the demise of behaviorism, several other theories of
language acquisition have been promulgated that do not rely on
operant conditioning or imitation as their mainstay.
Beyond their relevance for the nature-nurture issue, child
errors have been studied because of the insights they furnish
about the processes of language acquisition. For example, an
error like i thought they were all womans reveals that the
child has extracted the regular suffixation rule for forming plurals
of nouns in English. That is, the child knows to add -s to singular
nouns in order to mark plurality. The childs error lies in mistaking woman for a regular noun. This kind of error is commonly
described as an overregularization, since the plural rule for
regular forms (add S) has been applied beyond its conventional
confines to an irregular noun. Thus, errors of this kind illuminate
the childs ability to extract and generalize a morphological rule.
We know that the child is indeed adding -s to make plurals, even in the case of regular plurals like coconuts. This latter fact has been established even though it is conceivable that
the child has simply heard the form coconuts in the input and
153
154
Chirographic Culture
provides a strong stimulus for language scientists to seek explanations for how and why language learners differ from fully competent native speakers.
Matthew Saxton
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Berko Gleason, Jean. 1958. The childs learning of English morphology.
Word 14: 15077.
Brown, Roger, and C. Hanlon. 1970. Derivational complexity and order
of acquisition in child speech. In Cognition and the Development of
Language, ed. J. Hayes, 1153. New York: John Wiley.
Chomsky, Noam. 1959. Review of B. F. Skinners Verbal Behavior.
Language 35: 2658.
. 1965. Aspects of the Theory of Syntax. Cambridge, MA: MIT Press.
. 1980. Rules and Representations. New York: Columbia University
Press.
Chouinard, Michelle M., and E. V. Clark. 2003. Adult reformulations
of child errors as negative evidence. Journal of Child Language
30: 63769.
Hirsh-Pasek, Kathy, R. Treiman, and M. Schneiderman. 1984. Brown
& Hanlon revisited: Mothers sensitivity to ungrammatical forms.
Journal of Child Language 11: 818.
Jaeger, Jeri J. 2005. Kids Slips: What Young Childrens Slips of the
Tongue Reveal about Language Development. Mahwah, NJ: Lawrence
Erlbaum.
Marcus, Gary F., S. Pinker, M. Ullman, M. Hollander, T. J. Rosen, and F.
Xu. 1992. Overregularization in Language Acquisition. Monographs of
the Society for Research in Child Development, serial no. 228.
Maslen, Robert J. C., A. L. Theakston, E. M. V. Lieven, and M. Tomasello.
2004. A dense corpus study of past tense and plural overregularization in English. Journal of Speech, Language, and Hearing Research
47: 131933.
Newport, Elissa, H. Gleitman, and L. R. Gleitman. 1977. Mother, Id
rather do it myself: Some effects and non-effects of maternal speech
style. In Talking to Children: Language Input and Acquisition, ed. C.
Snow and C. Ferguson, 109149. Cambridge: Cambridge University
Press.
Pinker, Steven. 1988. Learnability theory and the acquisition of a
first language. In The Development of Language and Language
Researchers: Essays in Honor of Roger Brown, ed. F. Kessel, 97119.
Hillsdale, NJ: Lawrence Erlbaum.
Saxton, Matthew. 1997. The contrast theory of negative input. Journal of
Child Language 24: 13961.
Saxton, Matthew, P. Backley, and C. Gallaway. 2005. Negative input for
grammatical errors: Effects after a lag of 12 weeks. Journal of Child
Language 32: 64372.
Scholz, Barbara C., and G. K. Pullum. 2006. Irrational nativist exuberance. In Contemporary Debates in Cognitive Science, ed. R. Stainton,
5980. Oxford: Basil Blackwell.
Skinner, Burrhus F. 1957. Verbal Behavior. New York: Appleton-CenturyCrofts.
Smith, Neil. 2004. Chomsky: Ideas and Ideals. 2d ed. Cambridge: Cambridge
University Press.
Strapp, Chehalis M., and A. Federico. 2000. Imitations and repetitions: What do children say following recasts? First Language
20.3: 27390.
CHIROGRAPHIC CULTURE
Writing and script are systems of graphic marks that represent words, syllables, or individual sounds (phonemes)
Chirographic Culture
of a language. Chirography shares the same definition with
the added meaning of writing by hand. The term thus applies
to all the writing systems or scripts that followed the first
invention of writing in Mesopotamia, circa 3200 b.c., and
before Gutenbergs invention of the printing press about
1437. Chirography is generally viewed as the gateway to complex literate societies while leaving behind the archaic oral
cultures.
The nature and extent of the divide between oral and chirographic cultures has long been a matter of debate. In the fifth
century b.c., in Phaedrus and Letter VII, the Greek philosopher
Plato expressed his concerns with the impact of chirography on
human cognition. He warned that writing would weaken human
memory and threaten scholarship by allowing the ignorant to
fake omniscience. As discussed in Khosrow Jahandaries volume
Spoken and Written Discourse (1999), the present consensus is
less critical. Wherever chirography emerged, in Mesopotamia
about 3200 b.c., in China about 1250 b.c. and in Mesoamerica
circa 650 b.c., it is held as a productive supplement of speech.
This is based on the facts that, first, the human voice can be
heard only by a small audience but written documents can be
sent to any destination, and, second, speech disappears instantaneously, leaving no trace, while texts can be preserved over
time. It is, therefore, generally agreed that chirography extends
the network of human communication from culture to culture and makes it possible to trace the roots of cultures and their
evolution in history. Moreover, by reducing to order and clarity
a myriad of details, chirography is credited with revolutionizing
record keeping. Registers allow for administering communities and keeping track of entries and expenditures, profits, and
losses of businesses. Finally, writing is recognized for creating
data banks far beyond the power of human memory, resulting
in turn in the accumulation and articulation of an unlimited
quantity of complex data regarded as instrumental to the formulation of significant syntheses and the creation of new cognitive skills. In other words, literacy is viewed as enhancing the
possibilities for socially distributed cognition, allowing
civilization to grow more complex with the administration of
organizations of greater dimensions and larger political units,
the creation of more extensive economies, the development of
complex sciences and technologies, and a more accurate knowledge of the past.
The major priority of the twentieth century, however, has
been to investigate the impact of chirography on the human
mind. With his famous adage the medium is the message,
Marshall McLuhan emerged as a popular champion of literacy.
In his books The Gutenberg Galaxy (1962) and Understanding
Media (1964), the Canadian media critic advocated that media
were not passive conduits of information but, rather, vortices of
power restructuring human perceptions. McLuhan argued that
by translating sounds into a visual code, writing exchanged an
eye for an ear and that [p]honetic culture endows men with the
means of repressing their feelings and emotions when engaged in
action (1964, 84, 88); he therefore claimed that literate humans
develop the power of acting with detachment from emotional
involvement. He further emphasized the impact of the linear
format of writing, pointing out that civilized societies acquire a
155
Chirographic Culture
Figure 2. Tokens from Uruk, Iraq, ca. 3300 B.C. Courtesy Vorderasiatisches
Museum Berlin, Germany.
156
Chirographic Culture
Figure 3. Pictographic tablet from Uruk, Iraq, ca. 3000 B.C. Courtesy
Deutsches Archaeologisches Institut, Berlin, Germany. The tablet features
a list of goods. In the upper cases, the units of merchandise are indicated by pictographs or images of tokens traced in clay. Numerals are
shown with impressed signs. On the lower case, phonetic signs indicate
the name of the recipient or donor.
Schmandt-Besserat argued that ancient Near Eastern art compositions the way images are organized changed with the
advent of chirography in 31003000 b.c. She showed that preliterate painted or carved compositions consisted of the mere
repetitions of one motif as many times as necessary to cover a
surface. For instance, the same triangle or ibex was replicated
around the body of the vessel. In contrast, by borrowing the
strategies of writing, compositions of the chirographic cultures were able to organize multiple figures into a narrative. To
illustrate this concept, large and small signs of writing denoted
greater or lesser units of goods and similarly, the size of images
indicated status. Gods were represented as larger than kings,
and the images of kings were shown larger than those of their
fellow citizens. Just as signs changed value by being placed to
the right or the left, above or below other signs, the heroes
actions and interactions were indicated by their orientation,
order, and direction. For instance, one figure standing in front
of another was understood as being more important than
one standing behind it. From writing, art also acquired ways
of loading images with information by using symbols akin to
determinatives signs denoting a general class. For instance,
the horned tiara indicated the divine status of a figure in the
same way the star-shaped sign dingir did in Sumerian cuneiform texts. As a result, reading art became akin to reading
a text.
In sum, art, which is, in at least certain respects, a mirror
of culture, signals a conceptual change in design compositions that coincides with the advent of chirography in the
157
158
coincide (as with French object le, which is both unaccented and
distinctively positioned).
An association between them has been noted at least since
the classic work of Jakob Wackernagel (1892), who pointed out
that phonetically weak elements in the ancient Indo-European
languages tended to cluster in second position within the sentence, a position that was quite anomalous from the point of view
of the rest of the syntax. Much later literature has treated the connection between phonological weakness (and especially a lack
of autonomous accent) and unusual positioning as essential,
although the two turn out to be quite separate characteristics,
not only logically but also empirically. The essential distinction
between them was pointed out by Arnold Zwicky (1977) and
further developed in much later literature, including Anderson
(2005).
Conclusion
The analysis of clitics and the principles by which they find their
place in sentence structure and prosody involve an intricate
interplay among all of the major components of grammatical
structure, including syntax, phonology, morphology, and
even semantics. These elements have been invoked by scholars in all of these areas as evidence for fundamental claims about
linguistic structure, and an assessment of those claims is only
possible on the basis of a clearer understanding of the subdivisions among clitics, and the appropriate mechanisms for accommodating their specific properties, than is often found in the
existing literature.
Stephen R. Anderson
Codeswitching
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Anderson, Stephen R. 2005. Aspects of the Theory of Clitics. Oxford: Oxford
University Press.
Selkirk, Elizabeth. 1995. The prosodic structure of function words.
In Papers in Optimality Theory, 43970. University of Massachusetts
Occasional Papers in Linguistics 18.
Wackernagel, Jakob. 1892. ber ein Gesetz der indogermanischen
Wortstellung. Indogermanische Forschungen 1: 333436.
Zwicky, Arnold. 1977. On Clitics. Bloomington: Indiana University
Linguistics Club.
CODESWITCHING
Introduction
Codeswitching (CS) is defined as the use of two or more language
varieties in the same conversation, not counting established borrowed words or phrases. Two general types of structural configurations occur. 1) Intersentential CS, switches for one sentence
or many, is generally studied for its social implications (1). 2)
Intrasentential or intraclausal CS is more studied for its grammatical configurations (24).
(1)
(2)
(3)
(4)
159
Cognitive Architecture
Budzhak-Jones, Svitlana. 1998. Against word-internal codeswitching: Evidence from Ukrainian-English bilingualism. International
Journal of Bilingualism 2: 16182.
Clyne,
Michael.
2003.
Dynamics
of
Language
Contact.
Cambridge: Cambridge University Press.
Gumperz, John J. 1982. Discourse Strategies. Cambridge: Cambridge
University Press.
Jake, Janice, Carol Myers-Scotton, and Steven Gross. 2002. Making a
minimalist approach to codeswitching work: Adding the matrix language. Bilingualism, Language & Cognition 5: 6991.
. 2005. A response to MacSwan (2005): Keeping the matrix language. Bilingualism, Language, and Cognition 8: 2716.
Li, Wei. 2005. How can you tell?: Towards a common sense explanation of conversational code-switching. Journal of Pragmatics
37: 37589.
MacSwan, Jeff. 2000. The architecture of the bilingual language faculty: Evidence from intrasentential code switching. Bilingualism,
Language, and Cognition 3: 3754.
. 2005a. Making a minimalist approach to codeswitching
work: Adding the matrix language. Bilingualism, Language, and
Cognition 5: 6991.
. 2005b. Remarks on Jake, Myers-Scotton and Grosss
response: There is no matrix language. Bilingualism, Language and
Cognition 8: 27784.
Muysken, Pieter. 2000. Bilingual Speech, A Typology of Code-mixing.
Cambridge: Cambridge University Press.
Myers-Scotton, Carol. 1993. Social Motivations for Codeswitching: Evidence
from Africa. Oxford: Oxford University Press.
. 1997. Duelling Languages, Grammatical Structure in
Codeswitching. 2d ed. Oxford: Oxford University Press.
. 2002. Contact Linguistics: Bilingual Encounters and Grammatical
Outcomes. Oxford: Oxford University Press.
. 2006. Multiple Voices: An Introduction to Bilingualism.
Oxford: Blackwell. Chapters 6 and 9 deal with codeswitching for
advanced undergraduates.
Myers-Scotton, Carol, and Janice Jake. 2009. A universal model of codeswitching and bilingual language processing and production. In The
Cambridge Handbook of Linguistic Code-Switching, ed. B. Bullock and
A. Toribio, 33657. Cambridge: Cambridge University Press.
Poplack, Shana.1980. Sometimes Ill start a sentence in English Y
TERMINO EN ESPAOL: Toward a typology of code-switching.
Linguistics 18: 581618.
Treffers-Daller, Jeanine. 1994. Mixing Two Languages: French-Dutch
Contact in a Comparative Perspective. Berlin: Mouton de Gruyter.
Winford, Donald. 2003. An Introduction to Contact Linguistics.
Cambridge: Cambridge University Press. Winford provides a comprehensive overview of codeswitching designed for beginning graduate
students.
As digital computers evolved, so, too, did the notion of computer architecture. Computer designers had to pay attention not
only to the needs of the user but also to additional constraints
that arose with the development of high-level programming
languages and with the invention of new hardware technologies. Brookss more modern definition of architecture reflects
these developments: The architecture of a computer system we
define as the minimal set of properties that determine what programs will run and what results they will produce. The architecture is thus the systems functional appearance to its immediate
user (Blaauw and Brooks 1997, 3). The key element here is that
a computers architecture describes the information-processing
capacities of a device without appealing to its specific hardware
properties. In short, a computers architecture is a description
of its logical and abstract information-processing properties
(Dasgupta 1989).
160
Cognitive Architecture
These two lives are important because the predominant
research methodology used by cognitive scientists is functional
analysis (Cummins 1983). In functional analysis, a researcher
attempts to account for some complex function by decomposing it into an organized system of subfunctions. Each subfunction often becomes the subject of its own functional analysis;
this methodology is intrinsically iterative. However, if this were
all that there was to this methodology, functional analysis would
fall victim to an infinite regress and generate an infinite variety of
unexplained functions (Ryle 1949).
To avoid Ryles regress, functional analysis also attempts to
progressively simplify the proposed subfunctions: The highest
level design breaks the computer down into a committee or army
of intelligent homunculi with purposes, information and strategies. Each homunculus in turn is analyzed into smaller homunculi, but, more important, into less clever homunculi (Dennett
1978, 80). At some point, the less clever homunculi become so
simple that they can be replaced by physical devices that carry
out the desired function. At this point, the functional description is physically subsumed (Cummins 1983) and becomes
explanatory.
From this perspective, the cognitive architecture can be
described as the set of primitive functions that have been subsumed in a functional analysis. Their functional or cognitive role defines how these components mediate complex
information processing. Their physical role defines how such
processing can be physically implemented and explained.
161
Cognitive Grammar
exist in which neural networks have been used to model a variety
of language phenomena (Mammone 1993; Sharkey 1992).
Michael R. W. Dawson
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Blaauw, G. A., and F. P. Brooks. 1997. Computer Architecture: Concepts
and Evolution. Reading, MA: Addison-Wesley.
Brooks, F. P. 1962. Architectural philosophy. In Planning a Computer
System Project Stretch, ed W. Buchholz, 516. New York: McGrawHill.
Chomsky, N. 1965. Aspects of the Theory of Syntax. Cambridge, MA: MIT
Press.
. 1995. The Minimalist Program. Cambridge, MA: MIT Press.
Cummins, R. 1983. The Nature of Psychological Explanation. Cambridge,
MA: MIT Press. Explores the role of the architecture in providing
explanatory power to functional analyses.
Dasgupta, S. 1989. Computer Architecture: A Modern Synthesis. New
York: Wiley.
Dawson, M. R. W. 1998. Understanding Cognitive Science.
Oxford: Blackwell.
Dennett, D. 1978. Brainstorms. Cambridge, MA: MIT Press.
Fodor, J. A. 1983. The Modularity of Mind. Cambridge, MA: MIT Press.
Fodor, J. A., and Z. W. Pylyshyn. 1988. Connectionism and cognitive
architecture. Cognition 28: 371.
Hadley, R. F., and M. B. Hayward. 1997. Strong semantic systematicity from Hebbian connectionist learning. Minds and Machines
7: 137.
Haugeland, J. 1985. Artificial Intelligence: The Very Idea. Cambridge,
MA: MIT Press.
Hauser, M. D., N. Chomsky, and W. T. Fitch. 2002. The faculty of
language: What is it, who has it, and how did it evolve? Science
298: 156979.
Jackendoff, R. 2002. Foundations of Language: Brain, Meaning,
Grammar, Evolution. Oxford: Oxford University Press. Provides an
overview of the state of linguistics that pays special attention to architectural issues.
Mammone, R. J. 1993. Artificial Neural Networks for Speech and Vision.
New York: Chapman and Hall. Contains many examples of artificial
neural networks applied to specific areas of speech and language.
Pylyshyn, Z. W. 1984. Computation and Cognition. Cambridge, MA: MIT
Press. A detailed examination of the role of cognitive architecture in
cognitive science.
. 1991. The role of cognitive architectures in theories of cognition.
In Architectures for Intelligence, ed. K. VanLehn, 189223. Hillsdale,
NJ: Lawrence Erlbaum.
Ryle, G. 1949. The Concept of Mind. London: Hutchinson and Company.
Sharkey, N. E. 1992. Connectionist Natural Language Processing.
Dordrecht: Kluwer Academic Publishers.
Siegelmann, H. T. 1999. Neural Networks and Analog Computation: Beyond
the Turing Limit. Boston: Birkhauser.
COGNITIVE GRAMMAR
Cognitive Grammar (CG) refers to the theory of language articulated most comprehensively in Ronald W. Langacker (1987,
1991), two mutually dependent volumes that are best read
together. Langacker (1988) provides a succinct chapter-length
overview of his theory, while Taylor (2002) and Evans and Green
(2006, 553640) are highly recommended as student-oriented
introductions to the theory. CG is wide ranging in its scope and
provocative in its approach to an understanding of linguistic
162
Cognitive Grammar
The same apparatus is applicable for defining a word class such
as a noun, abbreviated by Langacker as [[thing]/[]], indicating a schematic semantic specification of a thing but without
any specific content phonologically. A morphologically more
complex lexical item such as trees is represented as a composite
structure integrating two symbolic units representing the noun
tree and the plural [z]: [[[tree]/[tri]]-[[pl]/[z]]]. Grammatical
constructions are in principle no different from a lexical item like
trees in terms of the descriptive apparatus required to capture all
the relevant detail, with each of the component structures of a
construction represented by a symbolic unit. Grammatical morphemes appearing in a construction, such as of, are treated as
symbolic units in their own right, with semantic structure (of, for
example, specifies a partwhole relation). The integration of any
two symbolic units goes hand in hand with distinguishing the
dependent and autonomous parts of the composite structure.
As far as semantic structure is concerned, [tree] is autonomous,
while [pl] is dependent, requiring an elaboration by a noun to
complete the structure. In terms of phonological structure, [tri] is
pronounceable as a whole syllable and can be considered autonomous, while the single consonant [z] is dependent.
A striking feature of CG is the detail provided for in the integration of structures into larger composite structures. The analysis of the English passive construction in Langacker ([1991] 2001,
10147) illustrates the theoretical notions relevant to a detailed
grammatical description and is recommended as a prime example of a full-blown CG account of a construction type. Briefly, and
consistent with the foregoing remarks, each morpheme in the
passive (including by and the auxiliary verbs) has its own symbolic representation, giving rise to the overall semantic structure,
just as the active counterpart has its own compositional structure and resulting semantic structure. Passive clauses do not
derive from active clauses in this view, nor do they derive from
some abstract structure underlying actives and passives. Rather,
passive clauses exist in their own right as instantiations of a construction type with its own distinctive way of integrating symbolic units, reflecting a particular construal of the event.
While phonological structure can be fruitfully explored
within CG (see Langacker 1987, 32848, 388400; Taylor 2002,
7895), it is semantic structure that has received most attention
and for which most theoretical apparatus has been developed.
Fundamental to semantic structure is the idea of a network that
is employed to represent polysemy relationships and to provide
motivation for conventional and novel extensions. Each node of
the semantic network, together with the associated phonological
structure, represents a semantic variant of the lexical item.
Two types of relationships figure prominently in these networks: schematicity, whereby one node of the network expresses
a meaning fully contained in another node (see schema), and
extension, understood as a relationship between semantic nodes
of a lexical item involving a conflict in semantic specifications.
The word head, for example, can be assigned a sense [part of a
whole which controls the behavior of the whole] that is
schematic relative to finer-grained elaborations, such as [part
of the human body where thinking is located] and [person
who manages an administrative unit]. In some cases, a highest-level node or superschema can be proposed, encompassing
all lower-level senses in the network, though such superschemas
163
164
165
166
167
168
Cognitive Poetics
one of the most important contributions of cognitive linguistics
to language science.
Patrick Colm Hogan
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Croft, William, and D. Alan Cruse. 2004. Cognitive Linguistics.
Cambridge: Cambridge University Press.
Feyerabend, Paul. 1975. Against Method: Outline of an Anarchistic Theory
of Knowledge. London: Verso.
Frijda, Nico. 1986. The Emotions. Cambridge: Cambridge University
Press.
Lakoff, George, and Mark Johnson. 1980. Metaphors We Live By.
Chicago: University of Chicago Press.
Pulvermller, Friedemann. 2002. The Neuroscience of Language: On Brain
Circuits of Words and Serial Order. Cambridge: Cambridge University
Press.
COGNITIVE POETICS
Cognitive poetics is the study of literary reading that draws on the
principles of cognitive science. In its early phase, the discipline
drew mainly on cognitive linguistics in focusing on the
textual cues for literary reading; in its more recent phase, it has
drawn more readily from cognitive psychology in order to explore
issues of readerly effects and aesthetics. Throughout its short
history, however, practitioners in the field have shown a willingness and propensity for genuinely multidisciplinary study. Work
in cognitive poetics is often characterized by an awareness that
the task of understanding literary reading holistically involves
a serious engagement with several disciplines: Linguistics and
psychology are central, but they are also often enriched from literary scholarship and critical theory, discourse analysis and
social theory, anthropology and historical study, neuroscience
and medical research, aesthetics, ethics, and philosophy. Most
people working in cognitive poetics believe that the systematic
analysis of literary reading is also essential within both linguistics
and psychology.
The field of cognitive poetics coalesced as an identifiable
movement in the mid-1990s, though it is possible to classify retrospectively several areas of work that can be seen as precursors.
The term itself was coined by Reuven Tsur in the 1970s (see Tsur
1992) to refer specifically to his exploration of literary aesthetics
through neuroscience and cognitive psychology. Since then, the
term has been taken up and broadened in scope to include a wide
range of research questions, frameworks for analysis, and areas
for exploration. Various alternative names for the enterprise have
been used during its brief history, each indicating the slightly different emphases of its users. For example, cognitive rhetoric
has been used in North America to point to the connections with
classical rhetoric in uniting form and effect in language study;
cognitive stylistics indicates a focus on detailed and rigorous textual analysis in the European tradition of stylistics, or literary
linguistics; other, more neutral terms, such as cognition and literature, have also been preferred on occasion.
It is apparent, too, that there are cultural contexts underlying
this nomenclature, much of which has to do with the branding
of the new discipline in institutional settings and the intellectual
marketplace. In the United States, the generative paradigm in
169
Cognitive Poetics
linguistics, on the one hand, and poststructuralist critical theory
in literary study, on the other, have meant that cognitive poetics
has been fighting for recognition on two fronts. Though the distinction was important earlier, a West Coast psychological tradition of metaphor study and an East Coast tradition of linguistic
textual analysis have now largely been merged. Similarly elsewhere, a continental European empiricist focus and a British and
East Asian stylistic emphasis have joined together more recently.
If there is any division remaining, it is a tendency for American
scholars to emphasize macrocognitive concerns and for EuroAsian researchers to emphasize the more micrological effects of
stylistic texture, though even this division is fast disappearing as
cognitive poetics develops globally. (See Lakoff and Turner 1989;
Turner 1991; Stockwell 2002; Gavins and Steen 2002; Semino and
Culpeper 2002).
For the majority of writers in cognitive poetics, the basic
common principles of their work involve a rejection of the
distinctions between text and context, form and meaning, abstraction and specification, and literal and metaphorical expression. Fundamental to this position is, most often, the
cognitive linguistic claim that language use is embodied and
that mind and body cannot be separated (see embodiment).
In other words, the linguistic expressions used in all languages
are elaborations of basic physical circumstances of the human
condition. To give an example from early cognitive linguistics,
there are fundamental conceptual metaphors (such as life
is a journey, ideas are objects, theories are buildings, and
so on) that tend to figure abstraction and complexity in familiar
and concrete terms. Such idealized figurative habits condition
our thinking and are manifest in linguistic expressions. These
conceptual metaphors are maintained and exploited in literary
texts just as in any form of language, and a thorough cognitive
poetic analysis is interested in both the conceptual significance
of the underlying scheme and the textual pattern through which
it is expressed (see Johnson 1987).
The principle of linguistic embodiment resolves the key issue
for a theory of interpretation, which is how a single explanation can account for the fact that individual readings are possible
but communal readings are in practice very common. Our basic
human condition creates figurative linguistic commonalities,
and our personal, social, and cultural tracks through these idealizations create individual, group, and cultural distinctiveness.
The literary scholar coming to cognitive poetics thus has a systematic and principled means of exploring individual expression
and sociohistorical patterns in culture.
Cognitive poetics sees reading as a natural evolutionary
process, rather than as an artifice that is distinct from other
human capacities. So, it is particularly interested in the ways
everyday (that is, nonprofessional, nonacademic) readers read,
and it draws connections between the natural creativity of imagination in everyday language (see creativity in language
use) and the particular ways in which linguistic creativity is
manipulated in the literary setting. Meaningfulness is regarded
as a readerly process, rather than a final classification, and so
cognitive poetics researchers have investigated how meanings
are constructed and resolved in the process of literary reading.
The personal experience and social circumstances of the reader
are at least as important as the textual organization of the literary
170
work. Cognitive poetic analyses thus have benefits for the study
of human language processing and the study of mind, as well as
for literary and artistic scholarship.
Within cognitive poetics, several different dimensions of
investigation have emerged. Some of the foremost and earliest
work centers on the human conceptual and linguistic capacity for metaphor or conceptual integration (see conceptual
blending). When cognitive linguists explored the workings of
conceptual metaphors of the sort mentioned earlier, it became
apparent that some expressive, poetic, or innovative metaphors
used in literary settings were causing problems with the basic
theory. Some creative literary language was more concerned with
interesting deviance and unsettling defamiliarization than with
resolving unfamiliar concepts in familiar terms. Furthermore,
some artistic metaphors seemed to affect those very same familiar domains in ways that persisted in the continuing life of the
reader, and most perplexingly of all, some literary metaphors
seemed to take on a life of their own that went beyond the basic
explanatory mappings of their source domains.
The cognitive poetic theory of conceptual integration or
blending has developed as a good account of these and other
features. Briefly, the theory proposes that a set of inputs are generalized to produce a blended space, a mental representation
that the reader uses to develop the emergent logic, texture, and
consequences arising from an engagement with the metaphor.
For example, at a microstylistic level, in Theodore Roethkes
phrase I have known the inexorable sadness of pencils, three
main input spaces (unstoppable motion, human emotion, and
the tool of the writer) are blended to produce a richly integrated
sense of emotional significance that is difficult to express in
any other way. At a more macrological level, the allegorical and
analogical significations of, for example, Margaret Atwoods
novel The Handmaids Tale amount to more than the sum of a
political manifesto, on the one hand, or a dark fantastic narrative, on the other (see Fauconnier and Turner 2002).
Another cognitive poetic dimension with a long history is
comprised of the various frameworks for the consideration of
worlds in literary works. Traditional possible worlds logic
from pragmatics and the philosophy of language has been
augmented as a means of understanding the richness of fictional
projection. (Indeed, the notion of mental spaces in conceptual integration theory owes something to the notion of worlds in
this sense). Such work has been especially fruitful in dealing with
extended prose fiction in which the divergence from reality (the
actual possible world) is most striking or thematized: science
fiction, fantasy, magical realism, dream visions, allegories, and
fairy tales (see Ryan 1991; Ronen 1994; Semino 1997).
Also aiming to integrate text and context in this way is schema
poetics, which draws on the notion of psychological schemas
from both Kantian philosophy and artificial intelligence work.
Here, culturally shared knowledge frames provide a rich context
for linguistic input that always underdetermines the affective
outcome in the reader. In literary research, the notion has been
used to explain mismatches in reader and character knowledge
and as a means of exploring the notion of literariness itself
(see Cook 1994).
Most recently developing out of the worlds tradition is text
world theory, a cognitive poetic approach to discourse processing
Cognitive Poetics
Coherence, Discourse
COHERENCE, DISCOURSE
What is a discourse? What makes a discourse coherent or incoherent? Investigation into these difficult questions has yielded so
many sophisticated proposals that a short, comprehensive survey is well out of reach.
With regard to the first question, it is fair to say that there is
widespread disagreement. Some researchers think of discourses
as texts, which raises questions about how texts are to be identified and individuated. Is sameness of spelling necessary and/
or sufficient to textual identity? Could there be significant variations, such as differences of spelling, among the tokens of a single text type or discourse? Under what conditions are sentence
tokens grouped as parts of a single text? Some investigators
question the very choice of texts as the basic unit of analysis: A
discourse is not a text (type), they say, but a string of spoken or
written sentences in a language. Yet disagreement again resurfaces when we ask how these strings are to be picked out. Some
authors tend to think of a discourse as an utterance (construed
loosely along Gricean lines as anything that is a candidate for
non-natural meaning and produced with communicative
intention), whereas others think that speech-acts, or groups
thereof, are the relevant discursive units. Yet even this is not sufficiently holistic for some discourse analysts, who want to focus
on larger sociocultural patterns, such as the discourse of racism
(see discourse analysis [foucaultian] and discourse
analysis [linguistic]). Some researchers argue for the primacy of face-to-face conversational interactions and contend
that the analysis of discourse coherence should find its point of
departure in specific sequences of immediate communicative
interaction (Schegloff 2001). Another area of divergence concerns the nature and number of the participants or producers of
a single discourse. Is a conversation between two or more parties
one or several discourses?
Investigators who disagree over the very outlines of the
concept of discourse can often still agree that certain kinds of
examples ought to be counted as discourses, and so can meaningfully debate logically independent questions about discourse
coherence. One such question is whether the coherence or
171
Coherence, Discourse
incoherence of a discourse is a matter of degree and, if so, what
sort of vagueness or lack of specification explains this fact. Is
it just ignorance of the real coherence conditions that leads us
to judge that a given discourse is more or less coherent or that
another discourse is a borderline case? Is discourse coherence
by its very nature a genuinely scalar concept? To what extent are
judgments of coherence and incoherence relative to the contexts,
categories, and genres of discourse? One may be tempted to conclude that the coherence of a poem is something quite different
from that of an argumentative essay, but perhaps the subtending
semantic relations are similar or even identical.
The coherence of discourse is not just a matter of logic. The
absence of logical contradictions, or the presence of logical
coherence (see coherence, logical), is hardly sufficient to
establish discursive coherence more generally, as the following
example is designed to show:
(1) The sap dripped. Mike pondered
Battologymeans tiresome repetition.
the
cogito.
(7) I was fed up with the telephone always ringing. Then I fell
ill. (temporal ordering, narration)
172
Coherence, Logical
Clark, Herbert. 1996. Language Use. Cambridge: Cambridge University
Press.
Cohen, Philip, Jerry Morgan, and Martha Pollack, eds. 1990. Intentions in
Communication. Cambridge, MA: MIT Press.
Gernsbacher, Morton Ann, and Talmy Givon. 1995. Coherence in
Spontaneous Text. Amsterdam and Philadelphia: John Benjamins.
Grimes, J. 1975. The Thread of Discourse. The Hague: Mouton.
Grosz, Barbara J., and Candace L. Sidner. 1986. Attention, intentions, and
the structure of discourse. Computational Linguistics 12: 175204.
Halliday, M. A. K., and Ruqaiya Hasan. 1976. Cohesion in English.
London: Longman.
Hobbs, Jerry R. 1985. On the Coherence and Structure of Discourse.
Stanford, CA: Center for the Study of Language and Information.
Hovy, Eduard H., and Elisabeth Maier. N.d. Parsimonious or profligate: How many and which discourse structure relations? Available
online at: http://www.isi.edu/natural-language/people/hovy/papers/
93discproc.pdf (accessed February 7, 2009).
Kehler, Andrew. 2002. Coherence, Reference, and the Theory of Grammar.
Stanford, CA: CSLI Publications.
Mann, William C., and Sandra A. Thompson. 1988. Rhetorical structure theory: Toward a functional theory of text organization. Text
8.3: 24381.
Redeker, Gisela. 2000. Coherence and structure in text and discourse.
In Abduction, Belief and Context in Dialogue: Studies in Computational
Pragmatics, ed. H. Bunt and W. Black, 23363. Philadelphia: John
Benjamins.
Sanders, Ted. 1997. Semantic and pragmatic sources of coherence: On
the categorization of coherence relations in context. Discourse
Processes 24: 11947.
Sanders, Ted, and Wilbert Spooren. 1999. Communicative intentions and
coherence relations. In Coherence in Spoken and Written Discourse,
ed. W. Bublitz, U. Lenk, and E. Ventola, 23550. Amsterdam: John
Benjamins.
Schegloff, Emanuel A. 2001. Discourse as an interactional achievement: III: The omnirelevance of action. In Handbook of Discourse
Analysis, ed. Deborah Schiffrin, D. Tannen, and H. Hamilton, 22949.
Oxford: Blackwell,
Schiffrin, Deborah. 2001. Discourse markers: Language, meaning, and
context. In Handbook of Discourse Analysis, ed. Deborah Schiffrin, D.
Tannen, and H. Hamilton, 5475. Oxford: Blackwell.
COHERENCE, LOGICAL
Logicians generally employ coherence and consistency as
synonyms naming the absence of contradictions in a group of
sentences, propositions, or beliefs, where a contradiction is
the conjunction of a proposition and its negation. In metaphysical terms, logical incoherence or contradiction is the impossible
instantiation of a property and some other, incompatible property, as in the circle was square. Epistemically, a contradiction
is an irrational belief in both a proposition and its denial.
Logical consistency is not a necessary feature of what people
say, write, or think. Nor is the absence of contradictions a sufficient condition on discourse coherence (see coherence,
discourse), as a collection of logically consistent yet unrelated
sentences does not constitute a coherent discourse. In many
contexts, however, logical consistency is a regulative norm for
speakers and interpreters. According to classical logic, a set of
propositions is either coherent or contradictory and trivial (in
the sense that it entails all propositions or explodes). In classical
logic, the ex falso quodlibet argument was held to establish that
173
174
kind and manner of living, than the diversity that is betwixt them
in tongue, language, order and habit, which by the eye deceives
the multitude, and persuades unto them, that they should be as it
were of sundry sorts, or rather of sundry countries. (Statutes 1786,
28 H 8. c.xv.)
Color Classisification
the other hand, Chinua Achebe, another major writer, rejected
this position and opted instead to use English, but a new form
of English, linked to its national home but altered to conform to
African realities.
The range of views on this and related issues and the vehemence with which they are expressed testifies to the ongoing
complexity and significance of the debates surrounding the legacy of linguistic colonialism, many of which are treated as questions of language policy (see language policy).
Tony Crowley
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Achebe, Chinua. 1975. Morning Yet on Creation Day. London:
Heinemann.
Calvet, Louis-Jean. 1998. Language Wars and Linguistic Politics. Oxford:
Oxford University Press.
Crowley, Tony. 2005. Wars of Words: The Politics of Language in Ireland
15372004. Oxford: Oxford University Press.
de Nebrija, Antonio. [1492] 2006. Gramtica Castellana. Barcelona:
Lingua.
Guest, Edwin. 1838. A History of English Rhythms. London: Bell.
Kachru, B. J. 1986. The Alchemy of English: The Spread, Function and
Models of Non-Native Englishes. Oxford: Pergamon.
Ngg Wa Thiongo. 1986. Decolonising the Mind: The Politics of Language
in African Literature. London: James Currey.
Pennycook, Alastair. 1998. English and the Discourses of Colonialism.
London: Routledge.
Spenser, Edmund. [1596] 1633. View of the present state of Ireland. In
The Historie of Ireland Collected by Three Learned Authors, ed Sir James
Ware, 1119. Dublin.
Statutes at Large Passed in the Parliaments Held in Ireland, The. 1786
1801. 20 vols. Dublin.
Watts, T. 1850. On the probable future position of the English language.
Proceedings of the Philological Society 4: 20714.
COLOR CLASSISIFICATION
Color terms label categories of the hue, saturation, and brightness of light reflected from surfaces. Because colors vary from
one another continuously and independently on these three
dimensions, there is no apparent intrinsic structure to the color
space that would prevent speakers of different languages from
cutting up the continuum in different ways. In the absence of
empirical studies, psychologists and anthropologists expected
color classification to be an example of extreme cultural relativism and that the spectrum would be segmented into categories
by different languages in arbitrarily different ways (e.g., Brown
1965, 31516).
B. Berlin and P. Kay (1969) refuted this relativist assumption.
They asked native speakers of 20 different languages to identify
the best examples ( foci) and the boundaries of basic color terms
on a Munsell color chart (a grid of 320 color chips with 40 hues
and 8 levels of brightness, plus a 10-chip gray scale). Although
informants varied enormously in their placement of boundaries
of color categories, they agreed considerably more on the choices
of the foci of the categories. Berlin and Kay found that there were
only 11 basic color categories in their sample of languages, with
foci in black, white, red, yellow, green, blue, brown, gray, pink,
orange, and purple. More surprisingly, they found that the color
175
Color Classisification
W/R/Y
G/B/K
W R/Y
G/B/K
W R Y
G/B/K
W R Y
G B/K
W R/Y
G/B K
W R Y
G/B K
W R K
Y/G/B
W R K
Y/G B
176
W R Y
G B K
Communication
COMMUNICATION
Explicit models of what communication is are not prominent in
cognitive science. Research under the explicit banner of communication does flourish in sociology. There, the emphasis is
most often on phatic rather than ideational communication, a
distinction due to B. Malinowski (1923). The former creates,
maintains, or dissolves communities of communicators. The
latter communicates ideas. Generally, these two aspects of communication are pursued independently, however unjustifiably.
The explicit and general models of language structure that dominate linguistics are sometimes tacitly thought of as providing,
inter alia, models of ideational communication, though the gap
between structure and function is underestimated.
Of course, this avoidance of explicit study might be an analogue of the lack of discussion of life in biology. Life is the subject
matter of biology and, therefore, the word is little heard therein.
Communication is at least a sizable part of what the social sciences are about and, so the argument goes, it is not surprising
that the word is little heard. Unfortunately, this analogy does not
fit unproblematically or absolve us very far.
There is, of course, a huge amount of work on communication phenomena. Linguistics is largely about the structure of
natural languages, and at least one of their functions is communication. psycholinguistics studies the interpretation of
natural language discourses by people. Social psychology has
much to say about both linguistic and nonlinguistic communication, from tone of voice to body language. Sociology provides
extensive studies of communication in all sorts of guises, from
microdialogue to mass media. The humanities likewise. The
issue here is not lack of study, or even lack of study in cognitive
science, but lack of theoretical frameworks for conceptualizing
what human communication is. One might be happy to have lots
of models, better still competing ones, but to have none smacks
of carelessness.
One significant event in the recent cognitive history of the
concept of communication was Noam Chomskys demolition
of the pretense of behaviorist psychologists and linguists to
analyze language (and communication) in terms of finite state
machine models. This computational model is closely related to
C. Shannon and W. Weavers model of communication, one of
the few influential abstract models of communication, in which
a sender issues signals from a finite code book through a channel
to a receiver who decodes the messages from an identical code
book. The amount of information transmitted is a function of
the probability of the occurrence of these signals. Information is
measured by the decrease of uncertainty. The less predictable a
signal, the more we learn from its occurrence. Of course, Shannon
and Weaver were not behaviorists. They assumed that the senders and receivers minds have general capacities for assimilating
messages, though any such assimilation lay outside their model.
But the behaviorists finite state machine can be construed as a
particular application of Shannon and Weavers model, and the
behaviorists claim was that it was a general theory not just of
human communication but of human behavior in general. There
was no mind to assimilate the finite code of messages.
Chomskys demolition of the finite state model as a model of
human language was deservedly famous, thoroughgoing, and
177
Communication
prevalent cause of misalignment. Of course, there is a fine line
between equivocation and enthymeme. If by dog you mean all
members of Canis canis in the current domain, and I mean only
the ones who arent bitches, this could be described as equivocation in the interpretation of a term (here across parties) or enthymatic suppression of a premise by one party (i.e., all relevant dogs
are male). Modern developments of logic have provided formal
theories of this process of reasoning from general knowledge and
contextual specificities to mutual interpretations. But in logics
twentieth-century detour into the foundations of mathematics
and the possibilities of knowledge engineering, this model of
communication has largely been lost.
Discourses come sentence by sentence, and each sentence
has complicated and generally nonmonotonic effects on the
structure of the context in which subsequent sentences will be
interpreted. Entities (people, objects, states, processes, events)
get added to, but also subtracted from, the current model of
the discourse. These effects are functions of both linguistic and
nonlinguistic long-term knowledge, current perceptual circumstance, and much else besides. So the meaning contributed by
each sentence to the discourse is a complex function of more
than its own or any other sentences structure. The general conception of discourse semantics has been developed with great
sophistication by H. Kamp and U. Reyle (1993) and their colleagues under the banner of discourse representation theory and
indeed by other approaches to discourse semantics, though still
with very little explicit connection to communication.
Whereas the classical logic of proof is monotonic in the sense
that adding more assumptions never removes valid conclusions,
defeasible logics of interpretation are nonmonotonic as new
assumptions are added, earlier conclusions may be subtracted.
These defeasible logics for reasoning to interpretations model
the process of discourse interpretation. In a monologue, hearers attempt to construct an interpretation of a speakers utterances that makes the statements true, bringing with them all
available general knowledge and contextual information. When
things go smoothly, this returns a unique minimal model at
every stage (van Lambalgen and Hamm 2004; Stenning and van
Lambalgen 2007). The reason these defeasible logics can yield
unique intended models for discourses against a background of
large bodies of general knowledge is the extensive deployment of
closed-world reasoning. If there is no evidence for the relevance
of facts in the model, then we can conclude that they are not relevant and at least not yet there. There are many subtleties about
how we close the world in constructing the intended model, but
they are variants on this same idea. Whereas in classical logic
we have to search for counterexamples in usually infinite sets of
logically possible models, closed-world reasoning gets us down
to single, small, intended models of the discourse at each point
in its development.
To adapt an example from an experiment designed to invoke
defeasible reasoning to an interpretation, when we are presented
with the following discourse,
She has an essay.
If she has an essay, she is in the library.
178
we may use our general knowledge and may withdraw the inference we made before, constructing a model that can be summarized as:
If she has an essay, and the library is open, she is in the library.
The specification of the models that are the objects of communication on this theory is an achievement of a successful discourse.
Each specification of a situation determines a set of situations differing from it by permutations of the specification: She
does/doesnt have an essay and the library is/isnt open, and
Because of closed-world reasoning, there are no other students
or libraries or essays nor indeed much else in the current model
until we hear about them. It is these models that can be thought
of as Shannon and Weavers code books. Far from getting the
book down from the shelf at the outset of the discourse, it is only
at each stage of development of the discourse that we can see
what code book has been specified. But note that this was never
Chomskys complaint about Shannon and Weavers model. The
complaint was always about the infinity of the code, not about its
on-line construction.
Such defeasible logics formalize the implicit model of communication dominant in psycholinguistics the field that has
most concerned itself with the empirical study of the process of
interpretation. The intended models correspond to what in psycholinguistics is known as the gist of the discourse, or what W.
Kintsch (1988), for example, calls the situation model. It is this
gist that was shown to be rapidly extracted from discourses in
classical studies of text comprehension.
Notice that the most plausible kinds of Shannon and Weaver
signaling within the context of a fully specified code book have to
do with temporal changes of state: She now has an essay to write
and is in the library. She now doesnt have an essay and she is in
the library (shes finished perhaps). She now has no essay and is
not in the library (having left?) what one might call the monologue of the surveillance camera. There are kinds of human communication that are like this (e.g., the stock ticker perhaps), but
one only has to consider such examples to realize what a minor
part they play. Creating local mutual interpretations is not signaling within their possibilities.
Here is the beginnings of a general abstract model of human
communication based on logical theories of discourse processing. In this model, communication is the construction of mutual
interpretations for discourses. It requires an attendant theory
of the structure of a language and of the organization of general
knowledge databases, which might be fully consonant with linguistic theories. But it is not to be confused with such theories.
Its objects of communication are models of discourses (interpretations that make them true), not sentences or meanings.
The contrast between this and other sentence-based theories of
communication can be well illustrated by considering the case of
soliloquy. We do indeed talk to ourselves, either audibly or not,
and intuitively we talk to ourselves for some of the same reasons
we talk to other people, including to help understand what we
believe or want, to formulate a course of action, to persuade ourselves to follow resolutions, to understand what someone said to
us, or to weigh up pros and cons. Needless to say, this process is
Communication
Communication, Prelinguistic
extremely important in learning, as is well testified in the empirical literature (e.g., Chi et al. 1989). To adapt an old saying, talking
to oneself is the first sign of sanity, or at least the search for it.
There is little temptation to understand soliloquy in terms
of the reduction of uncertainty, but we can apply the same logical model of communication as we use for public discourse. We
would not need to talk to ourselves if our knowledge and belief
were a transparent, homogeneous, consistent database of facts
and principles driven by unconflicted motivations. We equivocate and suppress our assumptions in internal argument just
as well as in public, and successful argument with ourselves
can lead to the same kinds of revision in order to gain coherent
interpretations.
Certainly there are differences between soliloquy and public
dialogue, but there are also enormous overlaps. In pursuing our
goals, it may often be a pragmatic matter of convenience whether
we choose to talk to ourselves or a conversational partner. These
functions of soliloquy are functions shared with dialogical communication, and they are functions that have been neglected in
our thinking about communication.
One could, of course, reject the notion that soliloquy is communication and define away these barriers to the consignment
of communication to Shannon and Weaver, but deeper considerations indicate that to do so is to miss much of what is crucial
about public communication. One should observe that this
model is not so incompatible with Chomskys deeper views as at
might at first appear. For example, his objections to functional
linguists who would see language shaped only by public communication is that language evolution may have been driven as
much by the advantages of an internal medium for representation and reasoning as by one for public communication. But
with a more adequate theory of what communication is, and by
dropping the idea that communication is automatically public,
this view is entirely consistent with our claim that communication is about achievement of coherent interpretation, whether by
public utterance or internal soliloquy.
Lastly, the model may help to reconnect the cognitive and
the affective perspectives on phatic and ideational communication alluded to at the outset. One observation is that the fundamental basis of ideational communication is the achievement of
mutually aligned interpretations. The process of getting to these
is, by definition, phatic communication it creates community
through shared interpretation of language. We perhaps forget how disturbing are our rare experiences of complete failure
to achieve this happy state and, in so doing, fail to see that our
cognitive theories of ideational communication contain within
them abstract specifications of just what has to be achieved and
maintained phatically, along with abstract accounts of some of
the processes by which this might be done.
Keith Stenning
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Chi, M., M. Bassok, M. Lewis, P. Reimann, and R. Glaser. 1989. Selfexplanations: How students study and use examples in learning to
solve problems. Cognitive Science 13: 14582.
Halliday, M. A. K. 2004. An Introduction to Functional Grammar.
London: Arnold.
COMMUNICATION, PRELINGUISTIC
Over the last 30 years, the field of developmental psychology
has devoted considerable research to prelinguistic communication, defined, most generally, as the sharing of information
prior to the onset of language. Because language onset is usually identified with the first spoken words, the prelinguistic
period encompasses roughly the first 12 to 18 months. Research
reviews have been framed both in terms of age-related changes
in infants interest in, and behavior during, social interactions
(e.g., Reddy 1999) and in terms of milestones for behaviors specifically related to communication, such as visual regard, turntaking, and gesture (Dromi 1993). These two reviews point to a
general agreement about the infant behaviors that are relevant to
prelinguistic communication (e.g., gesturing and visual regard)
and about the ways in which those behaviors change during the
first year of life. The major controversies in this area concern how
active a role infants play in structuring early episodes of communication and how changes in cognitive functioning relate to
changes in prelinguistic communication.
Shortly after birth, infants recognize familiar people, and
they can recognize their mothers by voice, face, and even by
smell (e.g., DeCasper and Fifer 1980). These perceptual capacities set the stage for infants social interactions to play a special role in prelinguistic communication; infants are interested
in and responsive to their caregivers, who interpret their early
social behaviors as communicative. Even if infants are not yet
aware that others have emotions or ideas to share or that they
themselves might have the same, the fact that the social world
treats them as communicative partners is viewed as a critical feature of social-pragmatic theories of language acquisition (e.g.,
Tomasello 2006).
Social smiling emerges at roughly six to eight weeks of age
and helps to mark the beginning of face-to-face, or en face,
interactions with caregivers, which are characterized by vocal
turn-taking and by the sharing of affect (see review by Adamson
2003). Although the role of the infant in holding up the structure
of these early en face interactions is controversial, it is clear that
infants take an even greater role in initiating social interactions
and maintaining their structure during the middle of the first
179
Communication, Prelinguistic
year. This three-to-eight-month age range has been characterized as a time when infants become increasingly interested in
regularity and surprise. Conventional games such as peek-a-boo
become prominent, and infants begin to take the lead in initiating these games as well as their turns in them.
Infants interest in the attention of others during the last quarter of the first year has been viewed as an important milestone
in prelinguistic communication. For the first time, there is joint
attention, which refers to episodes when infants and partners are
both engaged with the same object-in-the-world. Infants readily
follow the gaze of their partner and attempt to engage them in
attending to the object of interest. Infants use expanded means
to garner the attention of others, including giving objects, showing, and pointing. Gestures, especially pointing, have been the
subjects of intense study in the prelinguistic period because gestures may be used to share ones focus of attention with another
or to direct the attention of another (e.g., Bates 1979).
Also of great interest in late infancy is how new achievements in cognition might relate to changes in communication.
According to Jean Piagets theory of cognitive development (e.g.,
1983), infants begin, in the latter half of their first year, to understand that objects exist when out of sight and that objects exist
independent of our actions on them. This achievement, termed
object permanence, is theorized to be a critical part of the development of symbolic functioning. The words of language are
symbols, that is, arbitrarily spoken or written units that stand for
other objects and events. Thus, the achievement of object permanence and symbolic functioning are important milestones
setting the stage for the onset of formal language.
It is probably important, however, to view the transition
between prelinguistic communication and formal language as
neither abrupt nor all-or-none. Clearly, there continue to be
important relations between cognitive development and language after infants speak their first words; just as clearly, nonsymbolic forms of communication continue throughout the life
span (as, for example, in communication via physical actions or
emotional expressions).
Current research in developmental psychology is focusing on
processes that might be specific to language learning, as well as
on more general cognitive processes (such as categorization)
that might be involved in early word learning (see review by
Hollich, Hirsh-Pasek, and Golinkoff 2000). For example, infants
ability to find patterns in auditory stimuli or to group objects
together if they share similar attributes are general cognitive processes that relate to the problem of learning a language (Hollich,
Hirsch-Pasek, and Golinkoff 2000; Tomasello 2006). Delays in
these milestones of prelinguistic communication, and in the others
noted here, have been the subject of early intervention programs,
and deficits in prelinguistic communication skills have even been
linked to specific developmental disorders, such as autism.
James A. Green and Gwen E. Gustafson
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Adamson, Lauren B. 2003. The still face: A history of a shared experimental paradigm. Infancy 4: 45173.
Bates, Elizabeth. 1979. The emergence of symbols: Cognition and communication in infancy. New York: Academic.
180
Communicative Action
DeCasper, Anthony, and W. Fifer. 1980. Of human bonding: Newborns
prefer their mothers voices. Science 12: 30517.
Dromi, Ester. 1993. The development of prelinguistic communication. In At-Risk Infants: Interventions, Families, and Research, ed. N.
Anastasiow, 1926. Baltimore: Brookes Publishing.
Hollich, George J., K. Hirsh-Pasek, and R. M. Golinkoff. 2000. Breaking
the language barrier: An emergentist coalition model of the origins
of word learning. Monographs of the Society for Research in Child
Development 65.3: 1-137.
Piaget, Jean. 1983. Piagets theory. In Handbook of Child Psychology.
Vol. 1: History, Theory, and Methods, ed. P. Mussen, 10326. New
York: Wiley.
Reddy, Vasudevi. 1999. Prelinguistic communication. In The
Development of Language, ed. M. Barrett, 2550. Hove, East Sussex,
UK: Psychology Press.
Tomasello, Michael. 2006. Acquiring linguistic constructions.
In Handbook of Child Psychology. Vol 2: Cognition, Perception,
and Language, 6th ed. Ed. D. Kuhn and R. Siegler, 25598. New
York: Wiley.
COMMUNICATIVE ACTION
Communicative action is a term introduced by Jrgen Habermas
as part of his attempt to develop a general theory of action for
the social sciences. Among social theorists, it is widely believed
that a purely instrumental (or economic) model of rational action
is unable to account for the orderliness and stability of human
social interaction (Parsons 1968). Classical sociological theorists,
from Max Weber (1978) to Talcott Parsons (1951), tried to remedy
this defect by positing some additional category of value-oriented
or norm-governed action that imposed constraints on the range
of strategically optimizing behavior. Absent from this analysis,
however, was any precise specification of the role that language
played in mediating social interaction. Indeed, in many cases, it
was unclear how speech was supposed to fit into the theory of
action at all (Cicourel 1973, 21).
Habermas took as his point of departure the observation
that not only was a purely instrumental model of rational action
unable to explain the orderliness of social interaction, but it was
also unable to supply an adequate pragmatics for a theory of
meaning. So instead of looking to values or norms for a specification of the structure of noninstrumental rational action, he
turned to speech-act theory. In particular, he looked to the
notion of illocutionary force, as developed by J. L. Austin
and John Searle (Habermas 1984, I: 293; Austin 1975; Searle 1969).
His central intuition was that the limitations of Gricean (Grice
1989), or intentionalist (see communicative intention)
semantics might both reveal the limitations of a strictly instrumental approach to understanding the illocutionary dimension
of speech-acts and provide some indication of the structural
features that a noninstrumental theory of rational action should
exhibit. Once an account of the rationality of speech acts was
developed, his thought was that this could be extended to provide a more general account of the rationality of linguistically
mediated interactions. It is the latter category of action that he
refers to as communicative action.
Although this theory is of primary relevance to social scientists, it is also important to the study of language. Because of the
constraints imposed by the compositionality requirement,
Communicative Action
Speech Acts
any plausible approach to the theory of meaning must incorporate some sort of division of labor between the semantics
and pragmatics (with the former taken to have a compositional
structure, the latter typically not). Theorists of language have,
however, sometimes been naive when it comes to understanding
the constraints that the theory of action imposes upon the pragmatics. For example, it is often simply assumed that individuals
are capable of rule-following at the level of social action; yet
rule-following is, at the level of general action theory, a deeply
contested if not entirely problematic concept (e.g., see Bicchieri
1993). Habermass concept of communicative action is important for showing not only how action theorists might learn from
contemporary developments in the study of language but also
how theorists interested in language might profit from greater
attention to the structure of social action.
Communicative Action
181
Communicative Action
Communicative Intention
Practical Discourse
Finally, it is worth mentioning a further distinction in Habermass
work, between communicative action and practical discourse.
The orderliness of linguistically mediated interactions (i.e., communicative action) is achieved by the binding-bonding effects
of the validity claims raised within speech-acts. It is the rational
acceptance of these claims by listeners that makes it rational, in
turn, for them to accept any constraints on their conduct that
may arise as a consequence. However, this process of acceptance
is usually only tacit and in many cases relies merely upon the
speakers warrant. This means that should the listener suddenly
experience doubts during the course of the subsequent interaction, it is always legitimate for him or her to go back and demand
further justification (i.e., request that the speaker redeem some
validity claim that was associated with the speech-acts). As a
result, the potential for critical scrutiny of social practices is
always present, in every society and culture, even if not explicitly
institutionalized. Such a demand for justification interrupts the
sequence of communicative action and shifts the participants
into discourse, where contested validity claims are reflexively
thematized and debated. Contested rightness claims are discursively tested in a forum that Habermas refers to as practical discourse, which is governed by a set of distinctive inference rules,
in particular a universalization rule that serves as the foundation
for the theory of discourse ethics. The distinction between practical discourse and communicative action is important, in this
regard, because it is only the former that is directly governed by
the universalization constraint.
Joseph Heath
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Austin, J. L. 1975. How to Do Things with Words. 2d ed. Cambridge: Harvard
University Press.
Bicchieri, Cristina. 1993. Rationality and Coordination. Cambridge:
Cambridge University Press.
Cicourel, Aaron. 1973. Cognitive Sociology. Harmondsworth, UK:
Penguin.
Dummett, Michael. 1993. What is a theory of meaning? (II) In The Seas
of Language, 3493. Oxford: Clarendon Press.
182
COMMUNICATIVE INTENTION
Late twentieth-century discussion of the nature of communicative intention was dominated by the theories of British
philosopher Herbert Paul Grice. Grice initially (1957) argued
that the primary intended effect of an indicative utterance
was to get the hearer to believe the proposition expressed; an
essential component of this communicative intention was the
intention to have this effect be achieved through the hearers
recognition of that intention. He eventually acknowledged that
there were counterexamples to this analysis and subsequently
(1968; 1969, 1712) proposed that the primary communicative intention must be that the hearer should at least come to
believe that the utterer has some particular thought or belief.
Grice also allowed that speakers need not intend to change
the attitudes of some specific, actual audience; instead, this
part of the communicative intention concerns what is meant
to happen should there be an audience having such-and-such
characteristics.
Setting aside some of the many refinements (1989, 86116),
Grices characterization of communicative intention runs as
follows:
An utterance, U, is made with a communicative intention if and
only if the utterer, S, utters U with an intention comprised of three
subintentions:
Communicative Intention
It is assumed here that S communicates a belief to some audience, A, just in case As recognition of Ss communicative intention yields R, where R is the formation of the relevant belief in
A. For example, Sally said Congratulations! with communicative intent just in case she meant her saying to congratulate
the person to whom she was speaking and meant for this intention to be recognized by that person; she also had to intend for
that very recognition to be a reason for the recognition of the
congratulation.
Peter F. Strawson (1964) challenged the sufficiency of the
loop or mechanism constituted by subintention (3). Suppose that
Karen thinks that if her tennis racket is lying on the kitchen table,
her friend Laura will think Karen plans to play tennis that day.
Karen knows that Laura is watching her, and she also knows that
Laura does not know that Karen knows Laura is watching. Karen
then puts the racket on the table with the intention of getting
Laura to believe Karen plans to play tennis. Karen also intends
that Lauras recognition of the latter intention will give Laura
reason to believe that Karen in fact means to play tennis. Thus,
all three clauses in Grices definition have been satisfied. Yet in
such a situation, Strawson argues, Karen has not communicated,
at least in Grices sense, to Laura that she plans to play tennis. So
what is missing? Karen must intend not only that Laura recognize her intention to get Laura to think she plans to play tennis
but also that Laura recognize her intention to get Laura to recognize her intention to get Laura to think so.
Grice and other philosophers (e.g., Schiffer 1972; Holdcroft
1978; Recanati 1986) have explored various responses to the
problem raised by Strawson. Grices own preferred response to
the problem was to allow that meaning requires an infinite set of
intentions. Yet this condition is to be understood as defining the
optimal state in relation to which actual communicative states
are measured. He contends, then, that strictly speaking, no
speaker actually means that p in the sense of actually having the
set of infinite intentions required for ideal, non-natural meaning,
but he adds that the speaker is in a situation which is such that
it is legitimate, or perhaps even mandatory, for us to deem him
to satisfy this unfulfillable condition (1989, 302). Grices justification for this move finds its roots in his views concerning the
status of the normative rationality assumptions relied upon in
the entrenched and self-justifying system of both everyday and
philosophical psychology. He evokes the difference between the
titular and factual character of an utterance, where the former is
its idealized, rational character, never actually present in toto,
and where the latter could be a matter of a pre-rational counterpart of meaning (1989, 856). Yet it seems unsatisfactory
to conclude that Karen could never, strictly speaking, actually
communicate to Laura that she wants to play tennis!
Wayne A. Davis (2003) develops an alternative approach to
the relation between communication and semantic intentions.
Not all instances of communication are intentional, and cases
of intentionally communicating something to someone are analyzed as doing something that expresses a mental state, where
this action is the basis of an audiences recognition that the
mental state is expressed. What is expressed depends on what
is intended, but that does not mean that the hearer has to recognize the speakers communicative intention for intentional
communication to take place. The intentional component of
Comparative Method
expression is the performance of an observable action as an
indication of some attitude, where some x indicates some y
whenever x provides some (possibly unreliable) evidence that
y is the case.
Paisley Livingston
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bennett, Jonathan. 1976. Linguistic Behavior. Cambridge: Cambridge
University Press.
Davis, Wayne A. 2003. Meaning, Expression, and Thought.
Cambridge: Cambridge University Press.
Grice, Herbert Paul. 1957. Meaning. Philosophical Review 66: 37788.
. 1968. Intentions and speech acts. Analysis 29: 10912.
. 1969. Utterers meaning and intentions. Philosophical Review
78: 14777.
. 1989. Studies in the Way of Words. Cambridge: Harvard University
Press.
Holdcroft, David. 1978. Words and Deeds: Problems in the Theory of
Speech Acts. Oxford: Clarendon.
Recanati, Franois. 1986. On defining communicative intentions. Mind
and Language 1: 21342.
Schiffer, Stephen. 1972. Meaning. Oxford: Clarendon Press.
Strawson, Peter F. 1964. Intention and convention in speech acts.
Philosophical Review 73: 43960.
COMPARATIVE METHOD
Genetic Relatedness and Common History
Languages sharing a period of common history in a single ancestral language are genetically related. The comparative methods
(henceforth CM) goal is to demonstrate genetic relatedness by
identifying similarities attributable to retention from common
history. Demonstrations of genetic relatedness are presented as
reconstructions (or proto- forms) in the hypothetical ancestral
system. Genetically related languages comprise a language
family, relationships within which are identified through
female kin terms. A form inherited in a daughter is a reflex of the
mothers form. Reflexes of the same form in different languages
are cognates.
Languages sharing a period of common history independent
of the rest of the family constitute a subgroup. Subgrouping is represented as a family tree. (Wave models are an alternative permitting representation of overlapping shared innovations. Textbooks
like Hock [1986, Chap.15] compare the two.) In determining
genetic relatedness, one seeks evidence of shared retention from
a proto-language; in subgrouping, one seeks evidence of shared
innovation after the family began to diverge. In subgrouping, one
must identify unique events common to the history of the subgroup. The features defining the subgroup must demonstrably
not be retentions from an earlier period of common history. (On
subgrouping arguments, see Harrison 2003, 3.2).
Because a language is not an organism passing on genetic
material but behaviors and underlying knowledge, common
history is a property not of languages but the constructions constituting them (see construction grammars). A borrowing
and its source also share common history, and so genetic relatedness privileges common history involving transmission between
individuals speaking the same language. For mixed languages
183
Comparative Method
Table 1.
Trukese
plant
paddle
needlefish
thin
forehead
pandanus
fold
ftuk-i
ftun
taak
mlifi-lif
chaamw
faach
n-num
Mokilese
poadok
padil
doak
manip-nip
soamw
-par
lim
Gilbertese
arok-a
arina
raku
m-manii
ramwa
ara-
num
PMC
*faSok-
*faSla
*[sS]aku
*ma-nifi(nifi)
*camwa
*faca
*lumi
Table 2.
Trukese
ch
mw
aa
Mokilese
mw
oa
oa
*o
*i
Gilbertese
mw
PMC
*f
*S
*k
*l
*c
*m
*mw
184
*a
*u
Comparative Method
regular correspondences proliferate to the point that one must
reconstruct a proto-language phonemic inventory far larger than
that of any of its daughters and possibly larger than one would
consider natural. In such cases, the standard CM fails.
It is vital in comparison that there be some measure of similarity to show that we are comparing likes with likes. The CM
says little about similarity in meaning. Comparativist practice
favors meaning identity. Since semantic similarity remains illdefined, one is guided by experience and common sense. For
phonetic similarity, we can appeal to phonetic theories. What
is seldom appreciated is that the standard CM does not need a
theory of phonetic similarity because the regularity assumption
is a stand-in. We dont need to know that two sounds are similar,
only that there is a regular correspondence between them. Much
of the data for modern theories of phonetic similarity undoubtedly came from regular correspondences identified by the CM.
The empirical validity of the regularity assumption has been
controversial from the outset. Those opposed to the neogrammarian position asserted that each word has its [own] history.
A current manifestation of this opposition is lexical diffusion
(Chen and Wang 1975), the view that sound changes move
through the lexicon, affecting different words at different times.
Words yet unaffected will appear to be exceptions to the change.
For example, the shortening of Early Modern English (ENE) /u:/,
as in good (ENE /gu:d/, English /gd/), has yet to affect food, and
has affected roof only in some dialects.
Examples of nonphonetic conditioning are less often cited. The
Micronesian [aa]-[oa]-[a] correspondence set in Table 2 might be
such a case. In Trukese and some other Micronesian languages,
lengthening affects only the V1 of PMC (C1)V1(C2)V2 nouns. If, as
other evidence suggests, these nouns are the residue of a process
affecting all prosodic phrases, phonetic conditioning is preserved.
Since the regularity assumption is crucial to the CM, it may
be problematic were it proven false. Though William Labov
(1981) argues that the regularity assumption holds for one class
of sound changes while others diffuse through the lexicon, we
need not rely on his assessment to save regularity. Sound change
might begin variably, but given enough time, it moves toward
regularity. (See Durie and Ross 1996 for a range of perspectives
on the regularity assumption.)
SHARED ABERRANCIES. A shared aberrancy is a correspondence
between lexically or morphologically related forms so unusual as
to be unattributable to chance or borrowing. An oft-cited example is the 3s/3p alternations in the present tense of to be in IndoEuropean:
Summary
Any method for determining genetic relatedness must provide a
similarity measure and a means of distinguishing shared retentions from other sources of similarity. None is without flaws and
limitations. The standard CM is unique in defining similarity
through regularity. In some cases it will fail, but less often in principle than the search for shared aberrancies, which depends on
the existence of data of a restricted sort.
The quantitative methods remain to be tested, but there is
reason to doubt that they can replace the standard CM, or even
supplement it where the latter fails. These methods have been
applied to the subgrouping problem, not the genetic relatedness problem, and the two differ crucially. The subgrouping
problem is the search for the best tree for a set of languages
already assumed to be genetically related. These mathematical
techniques can determine genetic relatedness only if they fail to
incorporate unrelated languages into the trees they generate by
failing to identify cognates. And as long as a measure of similarity
is required to identify cognacy, something equivalent to the regularity assumption of the standard CM remains essential.
S. P. Harrison
Sanskrit
sti
snti
Latin
est
sunt
ist
sind
185
Competence
Durie, Mark, and Malcolm Ross, eds. 1996. The Comparative Method
Reviewed: Regularity and Irregularity in Language Change.
Oxford: Oxford University Press
Grace, G. W. 1996. Regularity of change in what? In Durie and Ross
1996, 15779.
Greenberg, Joseph H. 1987. Language in the Americas. Stanford,
CA: Stanford University Press.
Hale, Mark. 2003. Neogrammarian sound change. In Joseph and Janda
2003, 34368.
Harrison, S. P. 2003. On the limits of the comparative method. In Joseph
and Janda 2003, 21343.
Hock, Hans Heinrich. 1986. Principles of Historical Linguistics.
Berlin: Mouton de Gruyter.
Joseph, B., and R. Janda, eds. 2003. The Handbook of Historical Lingusitics.
Oxford: Blackwell.
Labov, William. 1981. Resolving the neogrammarian controversy.
Language 57: 267308.
McMahon, April, and Robert McMahon. 2005. Language Classification by
Numbers. Oxford: Oxford University Press.
Nichols, Johanna. 1996. The comparative method as heuristic. In Durie
and Ross 1996, 3971.
Osthoff, Hermann, and Karl Brugmann. 1967. Preface to morphological
investigations in the sphere of the Indo-European languages I. In A
Reader in Nineteenth Century Historical Indo-European Linguistics, ed.
W. P. Lehmann, 197209. Bloomington: Indiana University Press.
Thomason, Sarah Grey. 2001. An Introduction to Language Contact.
Edinburgh: Edinburgh University Press.
Thomason, Sarah Grey, and Terrance Kaufman. 1988. Language Contact,
Creolization, and Genetic Linguistics. Berkeley: University of California
Press.
COMPETENCE
The competence-performance dichotomy lies at the center of
transformational grammar, the linguistic theory introduced
by Noam Chomsky in the late 1950s (Chomsky 1957). Virtually all
current approaches to grammatical theory that descended from
Chomskys original work take the dichotomy as their starting point.
In brief, competence represents the system of abstract structural
relationships that characterize grammars, and performance the
faculties involved in putting that knowledge to use. It is generally
assumed that performance is determined in part by competence,
but is also a function of physiology, the communicative and social
aspects of language, and general cognitive architecture.
Competence and performance are modern reinterpretations
of the dichotomy between language and speech, which was
bequeathed to the field about a century ago by the great Swiss
linguist Ferdinand de Saussure ([1916] 1966). The French words
that Saussure used for language and speech, langue and
parole respectively, are still encountered today: For Saussure,
langue represents the structural system at the heart of language
a system shared by all members of the speech community; parole
is the individual act of speaking. Saussure compared language
to a symphony. Langue represents the unvarying score, parole
the actual performance, no two of which are identical. Rather
than sticking with langue and parole, Chomsky coined the new
terms competence and performance since he wished to
underscore two important differences between competence and
langue: Competence for Chomsky encompasses syntactic relationships, despite Saussures consignment of much of syntax
to parole; and competence is characterized by a set of generative
186
rules and principles, unlike Saussures langue, which was essentially a taxonomic inventory of grammatical elements.
Chomsky has always considered competence a psychological
construct, defining it as the speaker-hearers knowledge of his
language (1965, 4). Hence, support for the notion tends to be
derived from the apparent disparity between our mental representations of grammatical patterning and the actual use of language in communication. So it is frequently pointed out that
the structural principles that characterize grammars are far from
being in a one-to-one relation with the principles and conventions governing use (Newmeyer 1998). More direct psychological
evidence for competence has been adduced from observations
about child language learning. Experimentation has shown that
even very young children exhibit subtle grammatical knowledge
that points to their possessing a cognitive system encoding strictly
grammatical facts. For example, one-word speakers between 13
and 15 months know that words presented in strings are not isolated units but are part of larger constituents; one-word speakers
between 16 and 19 months recognize the significance of word
order in the sentences that they hear; and 28-month-old children who have productive vocabularies of approximately 315
words and who are speaking in four-word sentences can use a
verbs argument structure to predict verb meaning (Hirsh-Pasek
and Golinkoff 1996). There also appears to be neurological evidence for the competence-performance dichotomy. Numerous
pathological cases have been observed in which grammatical
abilities are lost while other cognitive faculties are preserved,
and vice versa (Pinker 1994).
Some linguists have applied the notion of competence to a far
broader range of abilities than the sort of grammatical knowledge
outlined here. For example, Dell Hymes coined the term communicative competence as the most general term for the speaking and hearing capacities of a person (1971, 16). A broadened
notion of competence was soon applied to such capacities as the
ability of bilinguals to switch languages appropriately (Gumperz
1972), the proper control of stylistic registers (White 1974), the
ability of readers to fathom aspects of literature properly (Culler
1975), and even the use of language by doctors in emergency
wards (Candlin, Leather, and Bruton 1976). The all-too-easy metaphorical extension of the ordinary English word competence has
led Chomsky and others to avoid use of the term in recent years.
Rather, it has become standard to use the term i-language
(short for internalized language). In this usage, I-language contrasts not with performance but with E(xternalized)-language.
Finally, it should be mentioned that some linguists have
questioned the existence of the competence-performance
dichotomy on the basis of the belief that grammatical structure
is an emergent property of language use (see, for example,
Langacker 1987 and Bybee and Hopper 2001).
Frederick J. Newmeyer
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bybee, Joan L., and Paul Hopper, eds. 2001. Frequency and the Emergence
of Linguistic Structure. Vol. 45. of Typological Studies in Language.
Amsterdam: John Benjamins.
Candlin, Christopher N., Jonathan H. Leather, and Clive J. Bruton. 1976.
Doctors in casualty: Applying communicative competence to components of specialist course design. IRAL 14: 24572.
that just as the goal of the analysis of a language is not the description of a corpus of utterances but an explicit account of the linguistic competence of speakers of the language, so ought the goal
of poetics and quite possibly of literary study generally not be the
analysis and interpretation of literary works but an account of the
rules, conventions, and procedures that enable readers to make
sense of literary works as they do (1975, viii, 215, 301, 11330).
His account stresses, for example, the shared knowledge and
processing techniques that enable readers to grasp the plot of a
narrative (a matter on which considerable agreement usually can
be reached) and to construct characters from the implicit and
explicit information scattered through a text, as well as to engage
in the thematic and symbolic interpretation that the institution
of literature encourages (ibid., 189238). He also stresses the distinctive assumptions and operations involved in making sense of
a lyric poem, such as a presumption of significance, the relevance
of sound patterning, and so on (ibid., 13188).
Culler presents literary competence as a revision of the
framework and goals of literary studies, an attempt to integrate
the accomplishments of structuralism and narratology
in literary studies with the program of a generative linguistics,
but others have suggested that taking the concept and the model
of generative grammar seriously would lead to a generative poetics. As a description of competence, a fully adequate
grammar must assign to each of an infinite range of sentences
a structural description indicating how the sentence is understood by the ideal speaker-hearer (Chomksy 1965, 45). Ellen
Schauber and Ellen Spolsky maintain that [a] generative poetics, therefore, will need to describe the derivation of competing
well-formed interpretations and to distinguish them from inadequately derived interpretations (1981, 397). Calling Cullers conception of literary competence focused on literary conventions
and distinctive interpretive operations intolerably restrictive,
Schauber and Spolsky propose that a generative poetics should
integrate three competencies: linguistic competence, communicative competence, and literary competence, on the principle
that literary competence in Cullers sense could never lead to the
derivation of well-formed interpretations (ibid., 398; 1986).
Critiques of the concept of literary competence have suggested that Chomskys specification of the competence of an
ideal speaker-hearer makes the concept of competence inherently elitist. Joseph Dane, while disputing the parallel between
linguistic and literary competence, contrasts a technical sense of
competence as knowledge that makes any literary performance
(including interpretation) possible with the everyday sense
where competence is a matter of qualifications and credentials
(1986, 53, 59). Despite Cullers argument that literary competence
does not involve a supposition that readers will agree upon an
interpretation but only that there are literary conventions that
guide interpretation and make possible some conclusions and
not others, Dane argues that a principle of stability must remain.
Some of us possess this competence; others of us must go to the
university to learn how to be perceptive and competent (ibid.,
60); [c]ompetence is simply that which is possessed by the most
powerful leaders of the literary community (ibid., 72).
The prestige of interpretation in literary studies, where the
task of the critic is to produce a more powerful interpretation,
has blocked the program of the study of literary competence as
187
Compositionality
something shared by readers, though it is implicit in any account
of narratology, for example, or of literary interpretation generally.
The cognitivist turn in literary studies (Turner 1996) provides an
opportunity for returning to aspects of literary competence and
the key question raised by the Chomskian model of the extent
to which such competence involves kinds of knowledge specific
to literature. If our ability to make sense of the world is defined
in terms of perceiving stories, organizing perceptions according
to metaphorical fields, and so on, it may be possible to go on to
identify interpretive moves that are specific to the reading and
appreciation of literary works.
Jonathan Culler
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Chomsky, Noam. 1965. Aspects of the Theory of Syntax. Cambridge,
MA: MIT Press. The classic theorization of a grammar as a description
of linguistic competence.
Culler, Jonathan. 1975. Structuralist Poetics: Structuralism, Linguistics,
and the Study of Literature. London: Routledge. Assesses structuralist
work and attempts to show that description of literary competence is a
fruitful program for literary studies.
Dane, Joseph. 1986. The defense of the incompetent reader.
Comparative Literature 38.1: 5372. A critique of the analogy with linguistic competence and of an implicit elitism.
Schauber, Ellen, and Ellen Spolsky. 1981. Stalking a generative poetics.
New Literary History 12.3: 397413. Starting with literary competence,
lays out a broader program.
. 1986. The Bounds of Interpretation: Linguistic Theory and Literary
Text. Stanford, CA: Stanford University Press. Develops the conception
of a generative poetics at greater length.
Turner, Mark. 1996. The Literary Mind. New York: Oxford University Press.
An important instance of the cognitivist program in literary studies.
COMPOSITIONALITY
The principle of compositionality was first formulated by the
German philosopher Gottlob Frege (1892) and is also referred
as the Frege principle. It states that the meaning of a complex
expression is a function of the meaning of its parts. A mapping
from expressions to meanings that satisfies this principle is
called compositional. Frege identified compositionality as a basic
requirement for an account of the meaning of natural language
(see language, natural and symbolic), and all serious
accounts of sentence meaning are compositional. Therefore,
current research seeks to find more restrictive notions of compositionality that can be used to assign a degree of compositionality
to a semantic analysis, as discussed in this entry. The question
of compositionality has also been asked for nonlinguistic communication systems among humans and other species, which I
mention towards the end.
For a semantics of natural language, compositionality is
a basic requirement because humans can generate infinitely
many sentences (see discrete infinity) and associate them
with one from an infinite set of meanings. Since human memory
is a finite resource, there can only be a finite set of memorized
lexical meanings (see lexical semantics). It follows that
natural language must contain nonlexical expressions and that
the meaning of such nonlexical expressions is determined by a
188
compositional procedure. Therefore, compositionality is a necessary property of any semantics of natural language that claims
complete coverage. The result, however, leaves open what the
lexical expressions of natural language are and how many composition principles there are. Often, words can be assigned a
compositional meaning; for example, the meaning of slept is the
result of sleep combined with past tense. In other cases, however, syntactically complex phrases seem to have a noncompositional meaning. For example, that kick the bucket is synonymous
with die does not follow naturally from the meanings of kick
and the bucket (cf. idioms). In the history of language, complex
expressions often take on a noncompositional meaning over
time (cf. grammaticalization.)
The composition principles are closely tied to a particular
semantic theory. Compositionality plays a central role in formal
semantics and truth conditional semantics of natural
language, while other theories of language meaning have not
addressed compositionality (cf. construction grammars
and cognitive grammar). The textbook by Irene Heim and
Angelika Kratzer (1998) provides one influential account. This
account assumes that humans construct a syntactic representation of a sentence, the logical form, which is then mapped
at the syntax-semantics interface to a meaning. This mapping
is a recursive procedure (see recursion, iteration, and
metarepresentation).
In addition to the meanings of a finite set of lexical items,
general composition rules determine the meaning of complex
phrases. Of the meanings of lexical items, only some aspects
are important for composition. In Heim and Kratzers analysis,
these aspects are captured by the semantic type. For example,
the meaning of proper names like Kai and Berlin have the type
of individuals, and the meanings of both to like and to hate are
of the type of two-place functions. The parts of a complex phrase
can be either lexical items or complex phrases themselves.
Therefore, only one composition rule is required: a rule that
combines the meanings of two subphrases into one. Heim and
Kratzers analysis makes use of three composition rules: function
application, predicate modification, and predicate abstraction.
Which composition rule is applied is determined by the types of
the meanings of the two parts of the complex phrase. The simple
example Kai likes Berlin illustrates only function application. We
assume that the sentence consists of only three lexical items: Kai,
likes, and Berlin, though a full analysis would contain at least
present tense as well. The lexical meanings of Kai and Berlin
are the individual concepts kai and berlin. The lexical entry for
likes is the function like which applied to one individual yields
another function that, when applied to another individual, yields
a sentence meaning. The logical form of the sentence shown
in (1) determines the order in which like is composed with its
arguments.
(1)
like(berlin)(kai)
Kai
like(berlin)
likes
Berlin
Compositionality
Computational Linguistics
COMPUTATIONAL LINGUISTICS
Computational linguists develop working models of various
aspects of languages in the form of computer programs. These
models fall under three main headings: analysis, generation, and
learning. Analysis models take in (usually typewritten) texts and
figure out the details of their linguistic structure, possibly producing a meaning representation. Starting from an abstract representation of a meaning, generation models compose text (e.g.,
a sentence) expressing that meaning in a particular language.
Some systems combine analysis and generation with other tasks.
For example, a database enquiry system analyzes queries in order
to figure out what information is sought, retrieves the requested
information from a database, and uses a generation system to
express that information in natural-language output. Machine
translation systems analyze input in one language and generate
corresponding expressions in a different language. Most systems
rely on grammar rules and resources such as text corpora and
dictionaries. Machine learning researchers build models that
learn the relevant information from training examples to avoid
hand-crafted rules.
189
Computational Linguistics
of grammar in the Chomsky hierarchy. Despite their limitations,
finite-state automata are efficient and adequate for lexical preprocessing, including decomposition of words into morphemes
(e.g., geese goose + Nplu) and normalization of spelling (e.g.,
driver [drive/V + er]Nsing). For text-to-speech conversion,
interpretation of abbreviations and symbols may be necessary,
too, for example, Mr. Mister, 4.36 four pounds thirty-six
(not pound four point three six).
Semantic Analysis
In real-world applications of computational linguistics, syntactic and morphological analysis are merely a means to an end.
In database enquiry systems or machine translation, we may
compute a representation of the meaning of the input. One of
the difficulties facing the computational treatment of meaning
is presented by groups of words with similar or related meanings. For example, in Whats the first class fare? and Whats
the price of a first class ticket?, fare and price have almost the
same meaning (i.e., the answer would be found in the same entry
in the ticket-suppliers database). In order to recognize lexical
relations, many systems use computationally implemented
thesauruses, such as WordNet, Princeton Universitys lexical
database for English.
To represent sentence meanings, computational linguistics often employs formal logic. A question, for example, can be
translated into a logical proposition with some information missing. Experts in formal semantics may express the meaning of
I would like the cheapest flight from Washington to Atlanta as
the predicate calculus formula:
(1)
Machine Translation
Conversion from natural languages to logical formulae and vice
versa suggests one method of machine translation:
(2)
190
pairs) than to develop translation rules and bilingual dictionaries for all 210 pairs. One technical problem with this approach
is that logical formulae are not unique representations of meaning: for example, if A then B is equivalent to not (A and not
B). But an overseas booking clerk might be confused if your
statement If meals arent served in economy class, I want a first
class ticket were translated as I dont want no meals served
in economy class and a first class ticket, even though this is
logically correct. Consequently, real machine translation systems combine transfer methods, which map structures of one
language to the other, direct methods that use word and phrase
correspondences with as little linguistic manipulation as possible, and some statistics to help choose the most likely ways of
expressing the output.
Machine translation and information retrieval requires
the generation of linguistic output, a task with its own particular challenges. When there are many equivalent ways of saying
the same thing (e.g., John drove the car, the car was driven by
John), the most appropriate variant must be chosen, observing
pragmatic conventions, such as putting given information
before new, the conventional order of words (big, red bus,
not red, big bus), and the time sequence they suggest: The
accused broke his leg and fell out of the window does not mean
the same as the accused fell out of the window and broke his
leg. Sentence generation often uses a slot-filling technique: The
agent of an action is placed in subject position; the undergoer
is the object, and so on. focus might prompt a particular sentence pattern, for example, it was the policeman who broke his
leg. Outputs are also generated in dialogue systems and in text
summarization. Dialogue systems collect information from the
user and provide information that the user requires according
to the accepted conventions of dialogue sequence. The dialogue
may be managed via a script that successively prompts the user
for gobbets of information. This is akin to form filling, as when
purchasing products on the Internet. The order in which the
user gives the information may not matter so long as all required
fields are eventually filled in. To navigate its script, the system
takes the lead in the conversation.
In text summarization, documents are analyzed in order to
extract the most important pieces of information according to
various criteria, such as discourse structure and word frequency.
This information is then used to generate a summary of the original, to a required length. Often, the summary simply consists of
the most relevant extracts of the original.
Probabilistic Methods
Computational linguists employ a wide range of probabilistic
methods that are helpful in various problems, especially sentence
and word-sense disambiguation. For example, the girl saw the
dog with the telescope has at least two structures and meanings,
depending on whether the girl or the dog has the telescope. Both
these structures and meanings are legitimate, but in real-world
applications such as machine translation, we may need to determine which structure and meaning is intended. Parsers quickly
reveal that average-length sentences have many possible structures, some quite implausible and unwanted. It is impractical to
model a speakers world knowledge, such as the fact that dogs
cannot have telescopes. But it is feasible to use the statistics of
Concepts
word combinations, such as the fact that telescope occurs with see
more often than dog, to select more likely analyses.
Probabilities can also help with word-sense disambiguation: For example, in she joined the club, it is not hard to work
out that club is more likely to be association of persons than
heavy staff of wood or suit of cards, simply on the basis of the
frequency of the collocation join club. Consequently, lexical
semantic analysis often combines probabilistic methods and
symbolic resources such as thesauruses.
Learning
In order to determine probabilities of rules, word senses, and so
on, a system must be trained on a corpus of language data, in
effect learning them. We would like computers to do more of the
hard work of finding the best grammar for a language. Consider
the pairs of rules that, according to X-BAR THEORY, define the
structure of noun phrases in all languages:
(3)
(4)
(5)
CONCEPTS
Concepts are the central constructs in most modern theories of
the mind. Humans (and arguably other organisms) are seen as living in a conceptually categorized world (see categorization).
191
Concepts
assumptions of the classical view. phonemes were analyzed as
sets of universal, abstract, binary features (Chomsky and Halle
1968). word meaning, the province of semantics, was likewise represented by a componential analysis of features (see
feature analysis); for example, bachelor was rendered as the
features +human, +male, +adult, +never married (Katz and Postal
1964). A complex concept such as bald bachelor was considered
the unproblematic intersection of the features of bachelor with
those of bald. Synonymy, contradiction, and other relational
aspects of word meaning were accounted for in a similar fashion.
Syntax was analyzed by formal systems such as transformational grammar (Chomsky 1965) that also relied on decomposition into features (see Taylor 2003). Such an understanding
of language was adopted with enthusiasm by computer science
because meaning could be divorced from world knowledge and
readily represented by the substitutable strings of symbols on
which computers work.
192
Theories
The theories approach to concepts takes advantage of peoples
intuitions that life activities and the concepts that map them take
place in a context larger than is offered by either formal description or laboratory experiments. The basic claim is that concepts
get their meaning through mental theories. There are actually
two groups of theory theorists: cognitivist-oriented cognitive
psychologists, who primarily address categorization issues, and
developmental psychologists of the theory theory school, who
address conceptual change. The first group (Medin 1989; Medin
and Wattenmaker 1987; Murphy and Medin 1985) has used the
idea of theories primarily as criticism of previous categorization
research. These theorists point out that previous accounts of
concepts cannot properly define or explain either attributes or
similarity and that previous experiments on conceptual categories are all subject to context effects; for example, judgments of
the prototypicality of animals changes if a zoo context is specified. They do not, however, themselves give an account of attributes or similarity, nor do they specify what a theory is or give
any concrete examples of a theory defining a concept.
In contrast, the theory-theory school of developmental psychology (see Gopnik and Meltzoff 1997) explicitly defines theory
as analogous to scientific theories, much like Kuhnian paradigms,
and argues that cognitive development should be viewed as the
successive replacement of one paradigm theory held by the child
by another. Interest in concepts tends to be from the point of
view of change in the childs (rather than the researchers) theory
of what a concept is. When specific concepts are studied (such
as biological types Carey 1985; Keil 1979), the thrust is to show
them as parts of larger theoretical units.
In linguistics, the discussion tends to be formulated in terms
of the relation of word meaning to general knowledge in a variety of specific contexts. Such contexts have been characterized as
schemas, frames, scripts, image schemas, domains, and
perspectivization. For example, Ronald Langacker (1990) talks
of the seven-day week as the semantic domain within which
Monday is understood, and George Lakoff (1987a) points to five
frames needed to explain our use of the word mother (genetic,
birth, nurturance, genealogical, and marital). Computer science
has worked on similar formulations in the design of the type of
program known as story understanders. Such work tends to be
classified under theories despite its lack of general explanatory
hypotheses because, lacking specification of what is to count
as a theory, virtually any demonstration of the embedding of
Concepts
individual concepts in larger semantic complexes or in world
knowledge has been argued as support for the theories view a
diffuseness that has also been used as a critique of that view.
Conceptual Atomism
All of the previous approaches are subject to philosophical criticisms, even the classical views (for reviews, see Fodor 1998 and
Laurence and Margolis 1999). Jerry Fodors (1990, 1998) conceptual atomism attempts to sidestep such issues by arguing that
the concept BIRD (Fodors notation) simply expresses the single
atomic property bird. The concept derives that meaning from its
causal history (as in Kripke [1972] 1980 and Putnam 1975 see
also essentialism and meaning). Concepts have no structure and are not decomposable into any kind of properties, internal or external. This view has not been adopted by psychologists
who presume (despite Fodors denial) that atomic concepts
would need to be innate (how could they be learned?) and that
it seems highly unreasonable to propose biologically innate concepts for everything in the universe, including televisions, penicillin, and so on. More broadly, conceptual atomism has not so
far appeared generative of psychological empirical research.
Conclusion
Concepts occur in use only in particular moments in particular
situations. From the perspective of their use, one can see the
aspects of mental and interpersonal activities in which each of the
seven accounts of concepts offers insight. For example, one might
be actively seeking to find the attributes for a classical definition
by means of thoughts that use mental prototypes. This would be
done against a background of loosely organized frames, scripts,
and so on the sort of knowledge structures toward which the
theories view points. The concepts being used could be atomic
and inherent to that moment, given that recognition of items had
already been performed and was now inherited from previous
moments. In short, each of the views maps a particular intuition
that humans seem to hold about concepts: The classical view
offers essences of a sort; the prototype view highlights concrete,
holistic mental representations; the theories view points toward
a background of conceptually structured world knowledge; connectionism points to a subsymbolic neuronal substrate; atomism
brings in history (and simplicity); combination models attempt
to bring it all together; and the ecological approach attempts to
bring it all together in terms of the ways in which concepts participate in real-world uses.
Eleanor Rosch
193
Concepts
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Armstrong, Sharon, Lila Gleitman, and Henry Gleitman. 1983. What
some concepts might not be. Cognition 13: 263308.
Barsalou, Lawrence. 1987. The instability of graded structure: Implications
for the nature of concepts. In Neisser 1987, 10140.
Bruner, Jerome, Jacqueline Goodnow, and George Austin. 1956. A Study
of Thinking. New York: Wiley.
Cantor, Nancy, Walter Mischel, and J. C. Schwartz. 1982. A prototype
analysis of psychological situations. Cognitive Psychology 14: 4577
Carey, Susan. 1985. Conceptual Change in Childhood. Cambridge,
MA: MIT Press.
Chomsky, Noam. 1965. Aspects of a Theory of Syntax. Cambridge, MA: MIT
Press.
Chomsky, Noam, and Morris Halle.1968. The Sound Pattern of English.
New York: Harper and Row.
Fodor, Jerry. 1990. Information and representation. In
Information, Language, and Cognition, ed. Philip Hanson, 17590.
Vancouver: University of British Columbia Press.
. 1998. Concepts: Where Cognitive Science Went Wrong. New
York: Oxford University Press.
Gabora, Liane, Eleanor Rosch, and Diederik Aerts. 2007. Toward an
Ecological Theory of Concepts. Ecological Psychology 20.1: 84116.
Gopnik, Allison, and Andrew Meltzoff. 1997. Words, Thoughts, and
Theories. Cambridge, MA: MIT Press.
Jackendoff, Ray. 1983. Semantics and Cognition. Cambridge MA: MIT
Press.
Kahneman, Daniel, Paul Slovic, and Amos Tversky, eds. 1982. Judgment
under Uncertainty: Heuristics and Biases. New York: Cambridge
University Press.
Katz, Jerrold, and Paul Postal. 1964. An Integrated Theory of Linguistic
Descriptions. Cambridge, MA: MIT Press.
Keil, Frank. 1979. Semantic and Conceptual Development: An Ontological
Perspective. Cambridge: Harvard University Press.
Kripke, Saul. [1972] 1980. Naming and Necessity. Cambridge: Harvard
University Press.
Lakoff, George. 1987a. Cognitive models and prototype theory. In
Neisser 1987, 63100.
. 1987b. Women, Fire, and Dangerous Things: What Categories
Reveal about the Mind. Chicago: University of Chicago Press.
Langacker, Ronald. 1990. Concept, Image, and Symbol: The Cognitive
Basis of Grammar. Berlin: Mouton de Gruyter.
Laurence, Stephen, and Eric Margolis. 1999. Concepts and cognitive
science. In Concepts: Core readings, ed. Eric Margolis and Stephen
Laurence, 381. Cambridge, MA: MIT Press.
Medin, Douglas. 1989. Concepts and conceptual structure. American
Psychologist 44: 146981.
Medin, Douglas, and William Wattenmaker. 1987. Cognitive cohesiveness, theories, and cognitive archeology. In Neisser 1987, 2562.
Murphy, George, and Douglas Medin. 1985. The role of theories in conceptual coherence. Psychological Review 92: 289316.
Neisser, Ulric, ed. 1987. Concepts and Conceptual Development: Ecological
and Intellectual Factors in Categorization. Cambridge: Cambridge
University Press.
Osherson, Daniel, and Edward Smith. 1981. On the adequacy of prototype theory as a theory of concepts. Cognition 9: 3558.
Pinker, Steven. 1989. Learnability and Cognition: The Acquisition of
Argument Structure. Cambridge, MA: MIT Press.
Putnam, Hillary. 1975. Mind, Language and Reality. New York: Cambridge
University Press.
Rogers, Timothy, and James McClelland. 2004. Semantic Cognition: A
Parallel Distributed Processing Approach. Cambridge, MA: MIT Press.
Rosch, Eleanor. 1973. Natural categories. Cognitive Psychology
4: 32850.
194
Conceptual Blending
. 1978. Principles of categorization. In Cognition and
Categorization, ed. Eleanor Rosch and Barbara Lloyd, 2748. Hillsdale,
NJ: Lawrence Erlbaum.
. 1996. The environment of minds: Towards a noetic and hedonic
ecology. In Cognitive Ecology (Handbook of Perception and Cognition).
2d ed. Ed. Morton Friedman and Edward Carterette, 524. San Diego,
CA: Academic Press.
. 1999. Reclaiming concepts. Journal of Consciousness Studies
6.11/12: 6177.
Rosch, Eleanor, and Barbara Lloyd, eds. 1978. Cognition and
Categorization. Hillsdale, NJ: Erlbaum.
Rosch, Eleanor, and Carolyn Mervis. 1975. Family resemblances: Studies
in the internal structure of categories. Cognitive Psychology
7: 573605.
Rosch, Eleanor, Carolyn Mervis, Wayne Gray, David Johnson, and
Penelope Boyes-Braem. 1976. Basic objects in natural categories.
Cognitive Psychology 8: 382439.
Smith, Edward, and Douglas Medin. 1981. Categories and Concepts.
Cambridge: Harvard University Press.
Smith, Edward, Edward Shoben, and Lance Rips. 1974. Structure and
process in semantic memory: A featural model for semantic decisions. Psychological Review 81: 21441.
Taylor, John. 2003. Linguistic Categorization. Oxford: Oxford University
Press.
Wittgenstein, Ludwig. 1953. Philosophical Investigations. New
York: Macmillan.
CONCEPTUAL BLENDING
Conceptual blending is a basic mental operation that has been
explored as a central mechanism indispensable to grammar and
language. Conceptual blending leads to new meaning, global
insight, and conceptual compressions useful for memory and
manipulation of otherwise diffuse ranges of meaning. It plays a
fundamental role in the construction of meaning in everyday life,
in the arts and mathematics, in the natural sciences, and in the
social and behavioral sciences. The essence of the operation of
conceptual blending is to construct a partial match between two
inputs and to project (see projection [blending theory])
selectively from those inputs into a blended mental space,
which dynamically develops emergent structure.
It has been suggested that the capacity for complex conceptual blending (double-scope integration) is the crucial capacity
needed for language and for higher-order cognition of the sort
that characterizes cognitively modern human beings.
A systematic study of conceptual blending was initiated in
1993 by Gilles Fauconnier and Mark Turner, who discovered the
structural uniformity and wide applications. The central introductory statement of the field is Fauconnier and Turner 2002.
(See also Turner 1996 and 2001; Fauconnier 1997; Fauconnier
and Turner 1996 and 1998; and Turner and Fauconnier 1999.)
The blending Web site at http://blending.stanford.edu presents
an extensive body of work done since then by many researchers in various fields on the theory of conceptual blending and its
empirical manifestations in language and grammar, mathematics, art, natural and social science, literature, social pragmatics,
and music. Additional research considers mathematical and
computational modeling of conceptual blending and experimental investigation in the cognitive neuroscience of neural and
cognitive processes.
Conceptual Blending
Informally, there are two distinct events in this story, the run
by the clipper in 1853 and the run by the catamaran in 1993 on
(approximately) the same course. In the magazine quotation, the
two runs are merged into a single event, a race between the catamaran and the clippers ghost. The two distinct events correspond to two input mental spaces, which reflect salient aspects
of each event: the voyage, the departure and arrival points, the
period and time of travel, the boat, and its positions at various
times. The two events share a more schematic frame of sailing from San Francisco to Boston; this is a generic space, which
connects them. Blending consists in partially matching the two
inputs and projecting selectively from these two input spaces into
a fourth mental space, the blended space, as shown in Figure 1.
In the blended space, we have two boats on the same course
that left the starting point, San Francisco, on the same day.
Pattern completion allows us to construe this situation as a race
(by importing the familiar background frame of racing and the
emotions that go with it). This construal is emergent in the blend.
The motion of the boats is structurally constrained by the mappings. Language signals the blend explicitly in this case by using
the expression ghost-ship. By running the blend imaginatively
and dynamically by unfolding the race through time we have
the relative positions of the boats and their dynamics.
Crucially, the blended space remains connected to the inputs
by the mappings, so that real inferences can be computed in the
inputs from the imaginary situation in the blended space. For
example, we can deduce that the catamaran is going faster overall
in 1993 than the clipper did in 1853, and, more precisely, we have
some idea (four and a half days) of their relative performances.
We can also interpret the emotions of the catamaran crew in
terms of the familiar emotions linked to the frame of racing.
Generic space
cross-space mapping
Input space 1
Input space 2
selective projection
Blended space
Figure 1.
Computer Interfaces
A nice example of conceptual blending in action and design is
the desktop interface, in which the computer user moves icons
around on a simulated desktop, gives alphanumeric commands, and makes selections by pointing at options on menus.
Users recruit from their knowledge of office work, interpersonal
commands, pointing, and choosing from lists. All of these are
inputs to the imaginative invention of a blended scenario that
serves as the basis for integrated performance. Once this blend
is achieved, it delivers an amazing number of multiple bindings across quite different elements, bindings that seem, in retrospect, entirely obvious. A configuration of continuous pixels
on the screen is bound to the concept folder, no matter where
that configuration occurs on the screen. Folders have identities,
which are preserved. The label at the bottom of the folder in one
view of the desktop corresponds to a set of words in a menu in
another view. Pushing a button twice corresponds to opening.
195
Conceptual Blending
Pushing a button once when an arrow on the screen is superimposed on a folder corresponds to lifting into view. Of course, in
the technological device that makes the blend possible, namely,
the computer interface, there is no ordinary lifting, moving, or
opening happening at all, only variations in the illumination of a
finite and arranged number of pixels on the screen. The blend is
not the screen; the blend is an imaginative mental creation that
lets us use the computer hardware and software effectively. In
the blend, there is lifting, moving, opening, and so on happening, imported not from the technological device at hand, which
is only a medium, but from another input, namely, our mental
conception of work we do on a real desktop.
196
Emergent structure arises routinely, as in This surgeon is a lumberjack, which suggests that the surgeon is incompetent, though
incompetence is a feature of neither surgeon nor lumberjack.
There are opposing pressures within an integration network
to maximize topology matching, integration, unpacking of the
blend, Web connections, compression, and intentionality. More
complex integration networks (multiple blends) allow multiple
input spaces, and successive blending in which blends at one
level can be inputs at another.
Compression
Blending is a remarkable tool of compression over vital relations
like time, space, causeeffect, identity, and change. In the newborn ritual, time is compressed: An entire lifetime becomes, in
the blend, the short time it takes to climb the stairs; in the desktop
interface, the complex sequence of events that move the mouse
horizontally and cause an apparent vertical motion of the arrow
(and other objects) on the screen is compressed and integrated
into a single action moving the arrow. This is a compression of
space, causeeffect, and change.
Language Science
The role of conceptual blending in language has been investigated in many areas.
FICTIVE MOTION. Languages have means of describing static
scenes in terms of fictive motion:
The fence runs all the way down to the river.
This works by having an imaginary trajector move along the relevant dimension of an object, in this case the fence or along some
imaginary path linking two objects. This is a remarkable mode
of expression: It conveys motion and immobility at the same time.
Objective immobility is expressed along with perceptual or conceptual motion. This apparent contradiction is a consequence of
conceptual blending, which allows several connected, but heterogeneous, mental spaces to be maintained simultaneously within
a single mental construction. An input space containing a static
scene of a fence and a river is blended with an input space that
contributes a moving trajector on a path with a reference point.
COUNTERFACTUALS. Human thought depends heavily on the
capacity for counterfactual thought, and counterfactuals are
complex blends. Most of us can effortlessly understand statements like In France, Watergate would not have hurt Nixon. This
counterfactual is intended to highlight some differences between
the American and French cultural and political systems. It is a
blend that brings in aspects of the French system from one input
and the Watergate scandal and President Richard Nixon from the
other. In the blend, we have a Watergate-like situation in France.
Running this blend delivers attitudes quite different from those
in the American input, and so in the blend, the president is not
harmed. Counterfactuals can blend frames and identities in powerful ways (If I were you ). Such blends have been shown to
play a major role for reasoning in the natural and social sciences.
THE ORIGIN OF LANGUAGE. The central problem in the origins
of language is that conceptual structure is vast relative to
expressive structure. The central problem of expression is that we
Conceptual Blending
197
198
Imposters
A major problem in the study of conceptual change concerns
cases where other patterns of cognitive developmental change
may give an appearance of conceptual change when none is
actually happening. Two common cases involve increasing
access and shifting relevance.
Increasing access refers to cases where cognitive limitations
having nothing to do with the concept per se limit its use (Rozin
1976). Thus, younger children might differ from older ones in
terms of memorial or attentional capacities that make them
unable to access a concept in a certain set of tasks. For example,
children might fail to engage in transitive reasoning in a wide
range of tasks, not because the children lack the concept of transitivity but because of the memory burdens imposed by having to
keep several inequalities in mind at the same time. When those
memory burdens are reduced by intensive practice with the
inequalities, the concept of transitivity is easily accessed, even
as the learning of the inequalities might be quite difficult (Bryant
and Trabasso 1971). One way of thinking about increasing access
can be seen in the metaphor of a young child learning to use a
Conceptual Metaphor
heavy hammer. We might note at first that the child cannot use
the hammer at all and think that he or she has a hammer deficit, only to find out that with a much lighter hammer, the child
reveals a full understanding of hammers. In other cases, there
may be a real deficit in the form of the missing concept.
Shifting relevance refers to changes in which several possible conceptual interpretations first come to mind in a task. For
example, when young children are asked if worms eat, they
might initially say that worms do not because they interpret eat
in a psychological manner involving feelings of satiation, hunger,
and pleasant tastes. Older children who interpret eat in a biological sense of providing nutrition might judge that worms do it.
Such changes have been interpreted as evidence for the emergence of the new conceptual domain of biology (Carey 1985). Yet
younger children may also be able to access the biological sense
of eat when shown that such an interpretation is appropriate
(Gutheil, Vera, and Keil 1998).
In summary, there are several distinct varieties of conceptual
change as well as other patterns of cognitive change that can
masquerade as conceptual change. It is critical in discussions of
conceptual change and of the relations of conceptual change to
other topics, such as word meaning, to know which senses are
in play.
Frank Keil
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bryant, Peter E., and Thomas Trabasso. 1971. Transitive inferences and
memory in young children. Nature 232: 4568
Carey, Susan. 1985. Conceptual Change in Childhood. Cambridge,
MA: MIT Press.
. 1988. Conceptual differences between children and adults.
Mind and Language 3: 16781.
. 1991. Knowledge acquisition: Enrichment or conceptual
change? In The Epigenesis of Mind: Essays on Biology and Cognition,
ed. S. Carey and R. Gelman, 25791. Hillsdale, NJ: Erlbaum.
Chinn, C., and W. Brewer. 1993. The role of anomalous data in knowledge acquisition: A theoretical framework and implications for science
instruction. Review of Educational Research 63: 149.
Fodor, Jerry. 1998. Concepts: Where Cognitive Science Went Wrong. The
1996 John Locke Lectures. Oxford: Oxford University Press.
Gopnik, Allison, and Henry Wellman. 1994. The theory-theory. In
Mapping the Mind: Domain Specificity in Cognition and Culture, ed.
L. Hirschfeld and S. Gelman, 25793. New York: Cambridge University
Press.
Gutheil, G., A. Vera, and F. Keil. 1998. Houseflies dont think: Patterns of
induction and biological beliefs in development. Cognition 66: 3349.
Inagaki, Kayoko, and Giyoo Hatano. 2002. Young Childrens Naive
Thinking about the Biological World. New York: Psychology Press.
Keil, Frank. C. 1986. Conceptual domains and the acquisition of metaphor. Cognitive Development 1: 7396.
. 1989. Concepts, Kinds, and Cognitive Development. Cambridge,
MA: MIT Press.
. 2003. Thats life: Coming to understand biology. Human
Development 46: 36977.
Keil, Frank C., and Nancy Batterman. 1984. A characteristic-to-defining
shift in the acquisition of word meaning. Journal of Verbal Learning
and Verbal Behavior 23: 22136.
Kemler, Deborah G., and Linda B. Smith. 1978. Is there a developmental trend from integrality to separability in perception? Journal of
Experimental Child Psychology 26: 498507.
CONCEPTUAL METAPHOR
According to proponents of conceptual metaphor theory, conceptual metaphors are metaphors that we have in our minds
that allow us to produce and understand abstract concepts. The
theory was first expounded by George Lakoff and Mark Johnson
(1980), who argued that conceptual metaphors structure how
people perceive, how they think, and what they do. According
to Lakoff (1993), conceptual metaphors represent habitual ways
of thinking, in which people metaphorically construe abstract
concepts such as time, emotions, and feelings, in terms of more
concrete entities.
199
Conceptual Metaphor
Conceptual Metaphors
e.g., ARGUMENT IS WARFARE
Linguistic Metaphors
e.g., Mr. Marshall had the knives out
for Mr. Manning
conduit metaphor in which communication is seen as transfer from one person to another, allowing us to talk, for example,
about conveying information, and getting the message across.
Another conceptual metaphor, PROGRESS THROUGH TIME IS
FORWARD MOTION, results in expressions such as plan ahead,
200
back in the 60s and to move on. In the same way, ARGUMENT
is often thought of in terms of WARFARE, UNDERSTANDING is
often expressed in terms of SEEING, LOVE is often thought of in
terms of a PHYSICAL FORCE, and IDEAS are often thought of in
terms of OBJECTS. Conceptual metaphors are thought to exist
for every abstract concept that we have, although there is no oneto-one mapping: A single abstract concept can be understood
through several conceptual metaphors, and a single conceptual
metaphor can be used to explain several abstract concepts. Some
conceptual metaphors are universal, whereas others vary from
language to language (cf. metaphor, universals of).
Conceptual metaphors are often very complex, and one
conceptual metaphor will frequently give rise to a series of
mappings. For example, the conceptual metaphor THINKING
IS PERCEIVING gives rise to mappings such as IDEAS ARE
THINGS PERCEIVED (its quite clear to me); ATTEMPTING TO
GAIN KNOWLEDGE IS SEARCHING (Im still looking for a solution); and BEING IGNORANT IS BEING UNABLE TO SEE (you
have allowed yourself to be blinded to the truth) (Gibbs 2006).
Conceptual metaphors differ from conceptual metonymies in
that they involve mappings between different domains, whereas
in conceptual metonymies, one part of the single domain is used
to refer to another, related part of that domain.
Conceptual Metaphor
At times, the ability to understand linguistic metaphors
(when they are first encountered) may rely on the successful identification of a relevant conceptual metaphor; at other
times, it may not. However, the ability to identify an appropriate conceptual metaphor in itself is rarely sufficient to allow a
complete understanding of a linguistic metaphor. Additional
metaphoric thinking is usually required, which takes into
account the context in which the metaphor appears and the
function that it is intended to perform. For example, in order
to understand the metaphor slavery was well on the road to
extinction, it may be helpful (but not necessary) to think in
terms of the conceptual metaphor PROGRESS IS FORWARD
MOTION. However, further metaphoric thinking is required to
understand that considerable progress has already been made
and that there is likely to be no turning back. Thus, conceptual metaphors sometimes help us to understand linguistic
metaphors, but they are not always a necessary prerequisite
nor a sufficient condition (see necessary and sufficient
conditions ).
201
Conduit Metaphor
in Comparison and Contrast, ed. Rene Dirven and Ralf Prings, 53354.
Berlin: Mouton de Gruyter.
Kovecses, Zoltan. 2002. Metaphor: A Practical Introduction. Oxford: Oxford
University Press.
Lakoff, George. 1987. Women, Fire and Dangerous Things: What
Categories Reveal About the Mind. Chicago and London: University of
Chicago Press.
. 1993. The contemporary theory of metaphor. In Metaphor and
Thought 2d ed. Ed. Andrew Ortony, 20251. Cambridge: Cambridge
University Press.
Lakoff, George, and Mark Johnson. 1980. Metaphors We Live By.
Chicago: University of Chicago Press.
Littlemore, Jeannette, and Graham Low. 2006. Figurative Thinking and
Foreign Language Learning. Basingstoke, UK: Palgrave MacMillan.
Low, Graham. 1999. Validating metaphor research projects. In
Researching and Applying Metaphor, ed. Lynne Cameron and Graham
Low, 4865. Cambridge: Cambridge University Press.
. 2003. Validating models in applied linguistics. Metaphor and
Symbol 18.4: 23954.
Ramachandran, V. S., and E. M. Hubbard. 2001. Synaesthesia a window into perception, thought and language. Journal of Consciousness
Studies 8.12: 334.
Stefanowitsch, Anatol, and Stefan Gries, eds. 2006. Corpus-based
Approaches to Metaphor and Metonymy. Berlin: Mouton de Gruyter.
CONDUIT METAPHOR
The conduit metaphor (Reddy [1979] 1993) models communication as a process in which the speaker puts information into
words and gets it across to a receiver, who tries to find the meaning in the words. Words are understood as containers, meanings as objects that can be put into words. Reddy was concerned
with the biasing influence this model has on our thinking about
communication.
Jrg Zinken
WORK CITED
Reddy, Michael J. [1979] 1993. The conduit metaphor: A case of frame
conflict in our language about language. In Metaphor and Thought,
ed. A. Ortony, 164201. Cambridge: Cambridge University Press.
tionist models, language structure, and representation). Although a few connectionist models have been used to
directly implement traditional types of grammar (e.g., Fanty 1986),
most aim to offer new ways of capturing key properties of grammar,
such as constituent structure and recursion (see recursion, iteration, and metarepresentation). In particular,
the latter models seek to demonstrate how important aspects of
grammar may emerge through learning, rather than being built
into the language system. This entry, therefore, focuses on the
radical connectionist models as they promise to provide new ways
of thinking about grammar and, as such, potentially could provide
the most substantial contribution to the language sciences.
Words in sentences are not merely strung together as beads
on a string but are combined in a hierarchical fashion. Grammars
202
203
204
otherwise. Second, there is a parallel element in symbolic architectures, at least when they are fully elaborated. For example,
in minimalism, the production of a sentence involves at least
some phonological, logical, and syntactic parallelism. Moreover,
connectionist networks clearly involve rules (even rules that
constitute operations over variables, the key set of rules stressed
by Gary Marcus [2001; see particularly 3583]). The rules are
embodied in the ways activation operates (e.g., through summation of inputs to thresholds).
Finally, it is not entirely clear that distributed versus unified
is really an opposition. For example, lexical items have components in symbolic accounts (e.g., semantic features). Writers in
the tradition of symbolic architecture do seem to envision those
components as occurring in one place, much like items in a dictionary entry occur on a single page. However, in terms of the
theories themselves, that only means that the components are
conceptually related, accessed together, and so on. A symbolic
architecture need not be committed to localism in the neural
substrate. The meaning of a given symbol may be realized in a
pattern of activation across different areas of the brain.
This is not to say that there are no differences between connectionist and symbolic accounts. But many of these are just the
general sort of differences that arise when one moves through
distinct levels of structure. For example, connectionist models
stress dispersal. Symbolic accounts stress unities. But this need
not constitute a contradiction anymore than differences between
physical and social accounts or macroscopic and microscopic
accounts need constitute a contradiction. We dont stop speaking
of trees and discussing their biology or their gross physical properties just because we discover that they are composed of atoms.
Nor do we say that a railway system is not reasonably treated as a
single thing just because it is dispersed in space. A similar point
holds for rules. It is indeed the case that symbolic systems tend to
involve many more rules and much more specific rules than connectionist architectures. But these rules are reasonably thought
of as emerging from neural networks. The existence of a neural
substrate without, for example, specific grammatical rules does
not invalidate a linguistic discussion of grammatical principles
anymore than the existence of a particle substrate invalidates an
engineers discussion of macroscopic causal laws.
Consider, for example, the head-directionality parameter
in principles and parameters theory. As Mark Baker
explains in his entry on parameters, Roughly put, when a
word-level category X [the head] merges with a phrase Y to create
a phrase of type X, there are two ways that the elements can be
ordered: The order can be X-Y within XP [X phrase], or it can be
Y-X. In languages throughout the world, heads tend to be added
in the same position, either first or last (not half in each, as one
might expect from random distribution). Grossly oversimplifying, we could imagine a connectionist model in which the networks for a range of processes develop some alternation between
initial position and final position. For example, it is easy to set up
a model in which some set of items triggers a directionality node,
which in turn activates either a beginning or an end node. Suppose
it activates the beginning node. Once that node is activated in
the context of a task say, identifying a determiner it will lead
to a behavioral output of checking the beginning. The absence
of the determiner at the beginning will initiate the correction
v
o
205
T
c
o
v
o
v
ov
F
Figure 2. A simplified model of the psychologically operative truth conditions for If x is a farm animal, then x is oviparous or viviparous. The letter
c stands for the entire conditional. T stands for true and F for false. The
letters o and v stand for oviparous and viviparous; - stands for not;
and -ov stands for neither oviparous nor viviparous. Arrows indicate
excitatory relations. Lines with solid circles indicate inhibitory connections.
Solid lines indicate a connection strength of 1. Broken lines indicate a connection strength of .5. Activation of a node occurs at 1. Thus, for example,
both -o and -v would have to be activated for -ov to be activated. This is
because the .5 connection strengths reduce the activation they communicate to .5. Note that this is not a model of logical/empirical truth conditions but of psychologically operative truth conditions. Thus, it seeks to
capture an initial presumption about the truth of the conditional based,
in this case, on authority. This is presumably more psychologically realistic
than a model that yields T only in cases of inductive validity.
connectionist models, language structure, and representation and connectionism and grammar). Nonetheless,
this model may partially illustrate some of the relations between
connectionist models, symbolic architectures, and neurobiology
outlined previously. For example, we do not seem to gain anything
206
Quasi Regularity
A key, but underrecognized, aspect of all of natural languages is
the fact that they exhibit quasi regularity. That is, linguistic expressions generally have properties that are shared with many other
forms, while also having properties that are more idiosyncratic.
In general, they are neither completely regular nor completely
irregular, but instead are best understood as lying somewhere
along a continuum of regularity, with most forms somewhere
between the extremes. Languages do have a tendency to pull
novel and infrequent forms into conformity with other forms, but
they also have a tendency to maintain (even promote) forms with
item-specific idiosyncrasies and clusters of similar items that
share such idiosyncrasies. These idiosyncrasies coexist within
items that also exhibit sensitivity to more general regularities.
Lets consider a range of examples from several different subdomains of language:
Inflectional morphology
keep-kept
tell-told
say-said
have-had
All these cases exhibit the correct regular past tense inflection together with a vowel change or consonant deletion in the
stem. The pattern in keep is shared with a number of other verbs,
including sleep, creep, sweep, and somewhat more distantly
dream, kneel, and mean.
Derivational morphology
predict, prefer
dirty, rosy
idolize, replicate,
As many have noted, and Joan L. Bybee (1985) and Luigi Burzio
(2002) have considered extensively, there are many derived
inflectional forms that preserve semantic characteristics of
their constituents but bring in altered or additional elements of
meaning. Treating these items as either fully compositional or
as fully opaque misses out on capturing one or the other aspects
of their meaning.
Idioms, constructions, and collocations
I want to see a doctor.
She felt the baby kick.
Hour by hour I grow more and more lonely.
Such expressions nearly always participate in general patterns
characteristic of other expressions with overlapping constituents
(I saw a doctor, I saw a movie, I want to see a baseball game) but
carry idiosyncratic meaning.
Spelling sound correspondences
pint
clown
great
Consider PINT. This item is an exception, but three of the letters
(P, N, and T) take their regular correspondences, and there are
related cases, such as MIND, FIND, MILD, and so on, that overlap with PINT and share the atypical vowel correspondence seen
in this item. The other cases have similar properties.
The point, in all of these cases, is that an inherent feature
of language is its quasi regularity: irregular forms are nearly
always partially regular. Approaches to language that relegate
all quasi-regular forms to the status of exceptions are, in general,
eschewing an account of an important and productive aspect of
language.
207
208
neighbors items like BOPE with a high degree of consistency with conventional spelling-sound rules, but shows considerable inconsistency in productions of forms such as:
GROOK, PREAD, MAVE
Models that work from raw speech input (Keidel et al. 2003).
Rules or Connections?
S. Pinker reacted strongly against the connectionist model of D.
E. Rumelhart and McClelland (Pinker and Prince 1988), taking
strong exception to the notion that language exhibits the graded,
similarity-based properties characteristic of connectionist models. He subsequently (Pinker 1991, 1999) accepted that language
does indeed have some gradient-like properties similar to those
of neural networks, but maintained that such properties are
restricted to operating within the lexicon. He made the case that
there is a separate, pure, rule-based system, operating according to principles quite distinct from those characteristic of connectionist networks. He made his case on the basis of a series
of empirical claims, arguing for the special status of categorical, structure- but not content-sensitive rules in performance,
acquisition, and breakdown under brain damage, also claiming support from cross-linguistic evidence. McClelland and K.
Patterson (2002) evaluated all of these claims and found instead
that properties of performance, acquisition, breakdown, and
cross-language variation do not support the view that language is
based on categorical rules; rather, the evidence supports the idea
that language knowledge is graded, semantic- and phonologicalcontent sensitive, and that there is no separate system for regular
as opposed to irregular aspects of language.
The fact that connectionist models can capture both systematic and idiosyncratic properties of linguistic forms leads to the
following suggestions:
209
McClelland, J. L., and K. Patterson. 2002. Rules or connections in pasttense inflections: What does the evidence rule out? Trends in Cognitive
Sciences 6.11: 46572.
McClelland, J. L., M. St. John, and R. Taraban. 1989. Sentence comprehension: A parallel distributed processing approach. Language and
Cognitive Processes 4: 287335.
Pinker, S. 1991. Rules of language. Science 253: 5305.
. 1999. Words and Rules. New York: Basic Books.
Pinker, S., and A. Prince. 1988. On language and connectionism: Analysis
of a parallel distributed processing model of language acquisition.
Cognition 28: 73193.
Plaut, D. C., and L. M. Gonnerman. 2000. Are non-semantic morphological effects incompatible with a distributed connectionist approach
to lexical processing? Language and Cognitive Processes 15: 44585.
Plaut, D. C., J. L. McClelland, M. S. Seidenberg, and K. Patterson. 1996.
Understanding normal and impaired word reading: Computational
principles in quasi-regular domains. Psychological Review
103: 56115.
Plunkett, K., and V. Marchman. 1991. U-shaped learning and frequency
effects in a multi-layered perceptron: Implications for child language
acquisition. Cognition 38: 43102.
Rumelhart D. E., and J. L. McClelland. 1986. On learning past tenses of
English verbs. In Parallel Distributed Processing. Vol 2: Psychological
and Biological Models. Ed. D. E. Rummelhart and J. L. McClelland.
Cambridge, MA: MIT Press.
210
211
The object varies, but the relation is always the same. Moores
point now is that there is no reason to think that the object must
always be some psychological state. He writes, I am as directly
aware of the existence of material things in space as of my own
sensations (1903, 453). If we can indeed appeal to this generic
relation of consciousness to external phenomena, doesnt this
give us our account of what is needed to provide a semantic
understanding of the terms we use?
For this approach to work, there has to be some response to
the lines of argument noted earlier. First, if consciousness does
do any work in our understanding of language, then it must do
some causal work. It must make some difference to what happens. But if we think of causality as a matter fundamentally of
the mechanistic interactions of physical particles, then how can
consciousness be causally significant? At best, it will seem that
an appeal to consciousness is an appeal to some ghostly quasi
mechanism. However, it is arguable that the trouble here is not
the appeal to consciousness but the mechanistic conception of
causation. The mechanistic conception is independently objectionable (cf, e.g., Woodward 2003). Arguably, we should think of
X causing Y as a matter of what would have happened to Y had
things been different with X. And we can certainly make sense of
the idea that things would have gone differently had one not been
conscious of this or that external phenomenon. So the idea that
the grasp of semantics is provided by experience of the things we
talk about does not seem to face any intractable difficulty over
the possibility of a causal role for consciousness. The idea that
consciousness as such must be a mere epiphenomenon has, in
any case, little intuitive appeal.
There were two classical objections to the idea that consciousness plays a role in our understanding of language. The first was
the point about causality. The second objection was that this
idea would make communication impossible; we would end up
each talking about only our own sensations. But this objection
disappears when we leave behind Russells view that we can be
212
Paracomplete Language?
The principle of excluded middle (PEM, sometimes LEM for
law) has it that, for any (declarative) sentence A, the disjunction of A and its negation is logically true (or valid). This is usually put by saying that, where V is disjunction and ~ negation, all
instances of AV~A are valid (in the given language). A paracomplete language is one in which PEM fails, that is, one in which
AV~A is not valid.
A popular response to the Liar paradox (and related semantic
paradoxes) is that it indicates the failure of PEM in English. After
all, premise (1) of the previous argument toward paradox relies
on PEM. Without PEM, the conclusion that some sentences are
true and not true fails to find a sound argument.
Paracomplete responses to the paradox are the most common approaches today. One problem with them is similar to
the meaningless response: They have trouble expressing their
position. After all, if (say) the Liar-instance of PEM fails, then,
presumably, neither the Liar nor its negation is true, in which
case, the Liar (e.g., the italicized sentence) is not true. But the
Liar says that its not true, and so, if it really is not true, it seems
to speak truly.
Paraconsistent Language?
(3) If the italicized sentence in CTP is not true, then the italicized sentence in CTP is true.
213
Constituent Structure
(6)
Maliki-rli ka
kurtu kartirti-rli paji-rni wita-ngku
dog-ERG PRES child tooth-ERG bite-NPST small-ERG
The small dog is biting the child with its tooth.
(Simpson 1991, 261)
J. C. Beall
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Beall, J. C., ed. 2008. Revenge of the Liar. Oxford: Oxford University Press.
This is a collection of very recent papers on truth and paradox.
Martin, Robert., ed. 1984. Recent Essays on Truth and the Liar Paradox.
Oxford: Oxford University Press. This collects a variety of contemporary approaches to truth, including so-called revision theory and contextual theories not mentioned here.
Tarski, A. 1983. Logic, Semantics, Metamathematics: Papers from 1923 to
1938. Ed. John Corcoran. Indianapolis: Hackett Publishing. Tarskis
classic approach toward defining truth for a language remains influential, but it is highly implausible as an account of truth for natural
languages (as Tarski himself thought).
CONSTITUENT STRUCTURE
words in sentences cluster into groups called constituents (or
phrases). For example, in (1), the dog, barked at a cat, at a cat,
and a cat are constituents:
(1)
This is the rat that ate the cheese that lay in the house that Jack
built.
214
215
216
Construction Grammars
is inclusive-or (including iii). For most sentences that children
experience, however, one or the other of the expressions surrounding or (its disjuncts) is false (excluding iii): Eat your veggies, or youll have to go to bed, Is his name Ted, or Fred? This
leads to the expectation that children should initially attribute
the exclusive-or meaning to the word or. In fact, there is evidence
that despite the input, children initially interpret or as inclusiveor as in classical logic. One possibility is that classical logic (or a
universal grammar) imposes constraints on childrens initial interpretations of logical words. On this scenario, children
initially assign or the truth conditions associated with inclusiveor, and later learn to limit these truth conditions to those associated with exclusive-or, based on principles of conversation. This
could be an example of an innate constraint but one that may or
may not be domain specific.
To summarize, constraints typically direct children to begin with
narrow hypotheses (i.e., to start small) and only broaden these
hypotheses if the input demands it. Constraints function quite differently in different parts of the linguistic system. Some constraints
are viable candidates for being domain specific and innately specified, but others may be domain general and learned.
Stephen Crain, Rosalind Thornton
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bloom, Paul. 2000. How Children Learn the Meaning of Words.
Cambridge, MA: MIT Press.
Braine, Martin, and Patricia Brooks. 1995. Verb argument structure and
the problem of avoiding an overgeneral grammar. In Beyond Names
for Things: Young Childrens Acquisition of Verbs, ed. M. Tomasello and
W. E. Merriman, 35376. Hillsdale, NJ: Lawrence Erlbaum.
Brown, Roger, and Camille Hanlon. 1970. Derivational complexity and
order of acquisition in child speech. In Cognition and the Development
of Language, ed. J. Hayes, 153. New York: Wiley.
Chomsky, Noam. 1971. Problems of Knowledge and Freedom. New
York: Pantheon.
. 1981. Lectures in Government and Binding. Dordrecht, the
Netherlands: Foris.
Crain, Stephen, and Rosalind Thornton. 1998. Investigations in Universal
Grammar. Cambridge, MA: MIT Press.
Marcus, Gary, Steven Pinker, Michael Ullman, Michelle Hollander, T. John
Rosen, and Fei Xu. 1992. Overregularization in Language Acquisition.
Monographs of the Society for Research in Child Development, no. 57.
Chicago: University of Chicago Press.
Markman, Ellen, and G. Wachtel. 1988. Childrens use of mutual exclusivity to constrain the meanings of words. Cognitive Psychology
20: 12157.
Morgan, James, and Lisa Travis. 1989. Limits on negative information in
language input. Journal of Child Language 16: 53152.
Pinker, Steven. 1984. Language Learnability and Language Development.
Cambridge: Harvard University Press.
. 1999. Words and Rules. New York: Basic Books.
Quine, W. V. O. 1960. Word and Object. Cambridge, MA: MIT Press.
CONSTRUCTION GRAMMARS
There has been a broad convergence in many quarters in recent
years toward a view of grammar in which constructions play a central role; approaches that share this view are here referred to as
construction grammars. Construction grammars view linguistic
Construction Grammars
(3) An emphasis is placed on psychological validity. A linguistic theory must interface naturally with what we know
about acquisition, processing, and historical change.
Complex word
Idiom (filled)
Passive
(7) Language-specific generalizations across constructions are captured via inheritance networks much like those
that have long been posited to capture our nonlinguistic
knowledge.
217
Construction Grammars
Him, a doctor?!
N P N construction
Whats X doingY?!
Covariational conditional
construction
218
Contact, Language
Fillmore, C. J., P. Kay, and M. C. OConnor. 1988. Regularity and idiomaticity in grammatical constructions: The case of let alone. Language
64: 50138.
Goldberg, A. E. 1995. Constructions: A Construction Grammar Approach
to Argument Structure. Chicago: Chicago University Press.
. 2002. Surface generalizations: An alternative to alternations.
Cognitive Linguistics 13.4: 32756.
. 2006. Constructions at Work: The nature of generalization in language. Oxford: Oxford University Press.
Haiman, John. 1985. Iconicity in Syntax. Cambridge: Cambridge
University Press.
Hale, K., and J. Keyser. 1997. On the complex nature of simple predicators. In Complex Predicates, ed. A. Alsina, J. Bresnan, and P. Sells,
2965. Stanford, CA: CSLI.
Jackendoff, R. 2002. Foundations of Language. Oxford: Oxford University
Press.
Kay, P. 2002. English subjectless tagged sentences. Language
78.3: 45381.
Kay, P., and C. J. Fillmore. 1999. Grammatical constructions and linguistic generalizations: The whats X doing Y? construction. Language
75.1: 134.
Lakoff, G. 1987. Women, Fire, and Dangerous Things: What Categories
Reveal about the Mind. Chicago: University of Chicago Press.
Lambrecht, K. 1990. What, me worry? Mad Magazine sentences revisited. Proceedings of the 16th Annual Meeting of the Berkeley Linguistics
Society. Berkeley, CA: Berkeley Linguistics Society, 21528.
Langacker, Ronald. 1987a. Foundations of Cognitive Grammar. Vol 1.
Stanford, CA: Stanford University Press.
. 1987b. Nouns and verbs. Language 63: 5394.
. 1988. A usage-based model. In Topics in Cognitive Linguistics,
ed B. Rudzka-Ostyn, 12761. Philadelphia: John Benjamins.
. 1991. Foundations of Cognitive Grammar. Vol. 2. Stanford,
CA: Stanford University Press.
. 1992. Reference point constructions. Cognitive Linguistics 4: 139.
. 2003. Construction grammars: Cognitive, radical and less so.
Paper presented at the International Cognitive Linguistics Conference,
Logrono, Spain.
Marantz, A. 1997. No escape from syntax: Dont try morphological analysis in the privacy of your own lexicon. In University of Pennsylvania
Working Papers in Linguistics. Vol. 4.2. Ed. A. Dimitriadis and L. Siegel,
20125. Philadelphia: University of Pennsylvania Press.
Michaelis, L. 2004. Implicit and explicit type-shifting in construction
grammar. Cognitive Linguistics 15: 167.
Michaelis, L. A., and K. Lambrecht. 1996. The exclamative sentence type
in English. In Conceptual Structure, Discourse and Language, ed A. E.
Goldberg, 37598. Stanford, CA: CSLI.
Sag, I. A., T. Wasow, and E. M. Bender. 2003. Syntactic Theory: A Formal
Introduction. Stanford, CA: Center for the Study of Language and
Information.
Saussure, F. de. [1916] 1959. Course in General Linguistics. Trans W.
Baskin. New York: Philosophical Library.
Steels, L., and J. DeBeule. 2006. A (very) brief introduction to fluid
construction grammar. Paper presented at the Third International
Workshop on Scalable Natural Language Understanding, New York
City.
Williams, E. 1994. Remarks on lexical knowledge. Lingua 92: 734.
CONTACT, LANGUAGE
Language contact occurs when individuals who regularly use
different language varieties communicate with each other.
Language variety should be understood in a very broad sense,
including varieties that are traditionally considered to be different
languages, as well as varieties of the same language, since the linguistic effects of either sort of contact are similar.
The social contexts in which language contact occurs are varied and have been common throughout human history. Most of
us, even if we are ourselves monolingual, interact with people
who are bilingual or bidialectal and, thus, are participants in
language contact situations whether we are aware of it or not.
Language contact can result in a wide variety of possible
outcomes for the language varieties involved, ranging from no
discernible effects to the borrowing of a few vocabulary items to
profound structural change. It may even result in the creation of
entirely new language varieties.
219
Contact, Language
Whether varieties are autonomous or heteronomous, linguistic elements of various sorts may be exchanged between them.
This exchange is referred to as borrowing. Virtually any linguistic
feature can be borrowed: vocabulary, grammatical morphemes,
grammatical constructions, semantic relations, sounds, and so
on. Vocabulary borrowing is the most common, but its easy to
find instances of borrowing of virtually any grammatical element.
Borrowing, particularly of vocabulary, can occur even when contact between speakers of the source and target languages is fairly
casual; when languages coexist in bilingual situations, however,
really intensive borrowing can take place.
bilingualism implies communicative skill in two autonomous varieties. (The term bidialectalism refers to communicative skill in two heteronomous varieties.) When bilingual
situations persist over a long period, convergence may take
place: The two varieties may converge toward each other, but
more usually one variety converges toward another, reflecting
local political or social dominance. If the varieties are sufficiently
different to begin with, and the convergence is extensive, the
result may be metatypy, a complete change in language type.
Cases of metatypy have not been commonly attested, but they
certainly exist: Amharic, a Semitic language spoken in Ethiopia,
has a grammar that is rather different from its Semitic kin due
to metatypic convergence with the Cushitic languages spoken in
the region prior to the arrival of Semitic languages.
At the beginning, bilingual situations are usually characterized by varying degrees of imperfect learning: situations where
in learning variety B, speakers of variety A learn B imperfectly,
making various sorts of grammatical errors in B, speaking B with
an accent, and so on. These linguistic features may become permanent components of the bilingual communitys command of
language B and may ultimately affect the speech of others speaking language B. If the bilingual community undergoes language
shift if speakers of A cease to speak A and adopt B instead the
results of imperfect learning are referred to as substratic influences on the B variety they speak. The particular sort of English
spoken in Ireland, for example, is often said to be the result, at
least in part, of this sort of substratic influence, in this case from
the Irish language to the English now spoken in the country.
Superstratic influence is also possible. A superstratic language
is one with high prestige, either in all formal contexts or in some
specified domains. Classical languages (Latin, classical Greek,
Sanskrit, Koranic Arabic, classical Tamil, etc.) can serve as superstrata, but so can living languages when these languages have
sufficient prestige. French has served as a superstratic language
for Europe, a role now filled largely by English. Chinese served
as a superstratum for Japanese, Arabic for much of the Moslem
world, Russian for other languages within the old Soviet Union,
and so on. Superstratic languages do not require many speakers
within communities for their influence to be strong: In the last
few centuries, only a minority of people in Europe learned Latin
to any significant degree, yet Latin influence on the languages
of Europe remains profound and extends beyond the extensive
borrowing of vocabulary to the borrowing of syntactic constructions, rhetorical strategies, and so on.
The discipline of second language acquisition refers
to linguistic systems that arise in the course of learning another
language as an adult as an interlanguage. In the ideal case, the
220
Contact, Language
221
Control Structures
CONTROL STRUCTURES
In many languages, when the subject of an embedded clause is
identical in reference (coreferential) with some noun phrase in
the main clause, the former may (or must) be left syntactically
unexpressed. Thus, sentence (1a) can be paraphrased as (1b),
where the understood subject of the embedded clause corresponds to the pronoun he in (1a). This unexpressed subject is
standardly notated by PRO, as represented in (1c). The referential dependence of PRO on George is expressed by the sharing of
an index (here, subscript i). This dependence is called control
(originally equi-NP deletion, see Rosenbaum 1967).
(1) a. George hoped that he would meet the Pope.
b. George hoped to meet the Pope.
c. Georgei hoped [PROi to meet the Pope].
222
Conversational Implicature
(3) a. Janei admitted that it was likely that [PROi perjuring herself]
was a mistake.
b. [PROarb to blame everything on fate] is all too common.
CONVERSATIONAL IMPLICATURE
The British philosopher Herbert Paul Grice observed that the total
significance of an utterance embraces not only what is said but
what is implied. His term of art for the latter was implicature,
and he identified conversational implicature as an important
type of implicit meaning or signification.
Grice used the following example to introduce this type of
implicit meaning in his 1967 William James lectures at Harvard: A
and B discuss a mutual friend, C, who has recently started working at a bank. When A asks B how C is getting on, B replies, Oh
quite well, I think; he likes his colleagues, and he hasnt been to
prison yet. Given knowledge of English and of contextual factors, A can readily grasp that B has said that C is getting on well,
likes his colleagues, and hasnt been to prison yet. A may also
understand that B has implied something else with the remark
about prison, and, according to Grice, a rational reconstruction
of the bases of this understanding reveals a complex inferential
process a process based on principles that the persons involved
probably would not be able to articulate unless they had studied
the Gricean literature.
At the top of Grices list is the cooperative principle,
roughly, the idea that it is rational for participants in conversations to advance the accepted purpose or direction of the talk
exchange to which they contribute. As it is routinely recapitulated, Grices theory specifies that in addition to this basic presumption about rational cooperation, hearers should act and
think in terms of conversational maxims or imperatives, which
include the following:
Maxim of Quality. Make your contribution true; so do not
convey what you believe false or unjustified.
Maxim of Quantity. Be as informative as required.
223
Conversational Implicature
linked to larger philosophical themes, including his defense of a
causal theory of perception and contentions about ambiguity
and presupposition in ordinary language use. Grice argues,
for example, that the word or is not ambiguous in English since
the exclusive interpretation (according to which or means that
either but not both disjuncts is true) can be understood as a conversational implicature and not as a second meaning of or.
According to the thesis known as Grices Razor, it is better to
posit conversational implicatures than ambiguity (and in some
cases, presupposition) because implicatures can be derived
more economically from the independently motivated principles
of cooperative rationality.
Although it is widely acknowledged that implicature is an
important phenomenon, questions are raised about the explanatory and descriptive value of Grices theory. Wayne A. Davis
(1998) argues that Grices maxims and cooperative principle
predict a range of implicatures that do not actually occur, while
other implicatures that do occur cannot be derived from them.
He also argues that the theory has no genuine explanatory payoffs. The proximal causes of a speakers conversational moves
are that persons attitudes, not general tendencies to cooperate or an audiences presumptions about the latter. According
to Davis, Grice wrongly assumes that the production and recognition of implicature are processes explicable in terms of the
same principles and maxims. This premise is misleading if what
a speaker implicates or means is not caused by what others presume or know about that speaker. To implicate a meaning that
extends beyond what one says is to say something with certain
intentions, and the speakers intentions do not directly depend
on what others know or presume. Having and expressing intentions is one thing, whereas communicating them to others is
something else. Grice appears to assume that audience uptake of
a certain kind is necessary or even sufficient to the realization of
communicative intentions.
Grices explicit analysis of conversational implicature can
be read as indicating that the very existence of implicature (as
opposed to its successful uptake or understanding by some audience) requires the presumption, on the part of a hearer, that the
speaker has observed or acted in accordance with the cooperative principle. Thus, Grice writes that S implicates q only if he
is presumed to be observing the conversational maxims (1989,
30). On an alternative reading, the actual hearers presumptions
and other beliefs are not necessary, since some implicatures
are made by a speaker but remain unrecognized by the target
audience. What Grice has proposed is an account of successfully communicated implicatures, but not implicature tout court.
Jennifer Saul (2002, 241) suggests that what matters for Grice
is not what particular hearers actually think, but what they are
required to think. Grice indeed stressed that his focus was on
the rationality or irrationality of conversational conduct (1989,
369). The thesis that implicature must be calculable or capable
of being worked out can, then, be taken as belonging to a normative theory of the conditions under which speakers can successfully realize the rational intention to implicate rather than
to state some thought. Yet it is unclear why the norms of communicative rationality should apply to both noncommunicative
and communicative linguistic behavior. Is it persuasive to argue
that it is simply impossible for a speaker to have implicated some
224
Conversational Repair
. 1989. Studies in the Way of Words. Cambridge: Harvard University
Press.
Kasher, Asa, ed. 1998. Pragmatics: Critical Concepts. London: Routledge.
Levinson, Stephen C. 2000. Presumptive Meanings. Cambridge, MA: MIT
Press.
Saul, Jennifer. 2002. Speaker meaning, what is said, and what is implicated. Nos 36: 22848.
Sperber, Dan, and Deirdre Wilson. 1986. Relevance: Communication and
Cognition. Cambridge: Harvard University Press.
Sterelny, Kim. 1982. Against conversational implicature. Journal of
Semantics 1: 18794.
Wierzbicka, Anna. 1991. Cross-Cultural Pragmatics: The Semantics of
Human Interaction. New York: Mouton de Gruyter.
CONVERSATIONAL REPAIR
Conversational repair (hereafter, repair) refers to a common
practice in the interactive social organization of conversations
in which speakers suspend the smooth progressivity of the talk
to deal with some ostensible problem in speaking, hearing, or
understanding the talk. Repair does not always involve hearable
errors or mistakes that require correction. Therefore, the term
repair, rather than correction, is used to capture the more general
domain of such occurrence in conversation analysis.
The organization of conversation is a turn-taking system in
which speakers take turns to converse. A repair may be done
by the speaker of the trouble source in the same turn (sameturn self-repair). Or it may be done by anyone but the speaker.
Furthermore, a repair may be initiated by the speaker of the
trouble source or by others. Repair is often carried out with repetition/recycling, replacement, or restructuring of the utterance,
although not all repair attempts may be successful. Studies find
that self-repair prevails even when a repair is initiated by others.
Next is a brief discussion of same-turn self-repair, for which
an emerging utterance may be stopped, aborted, recast, continued, or redone. Such repair often involves self-initiation with
some nonlexical initiators, such as cutoffs, sound stretches, uhs,
and so on, followed by repair. Following is an example containing two instances of repair with cutoffs, replacements, insertion,
and repetition/recycling (the asterisk indicates where repair
initiates).
(1) And tshe-* this girls fixed up onna da- * a blind da:te.
In the first instance of repair, the speaker cuts off the pronoun
tshe- (i.e., the repairable; the - indicates glottalized cutoff) and
replaces it with a full noun phrase, this girl. The second instance
is where date is cut off to introduce a modifier by recycling the
entire noun phrase: a blind date.
Repair is highly patterned, with some basic mechanisms
occurring cross-linguistically. But specifics of the mechanisms
differ. For instance, recycling often occurs at a turn beginning
when the utterance overlaps with the ending of the previous
speakers turn. However, the syntactic unit that is recycled differs
from one language to another: Some allow repetition of single
words, whereas some require larger syntactic units to be recycled (e.g., in the example, the entire noun phrase is recycled).
Furthermore, speakers of a tone language such as Mandarin
make tone-related recycles. Therefore, repair mechanisms are
constrained by the grammar of individual languages.
Conversation Analysis
Conversation analysis finds that the grammar of repair is
vital for syntax-for-conversation. Repair is closely related to
syntax because it affects the shape and/or components of
a sentence. Syntax organizes elements through which talk
is constructed. Syntax-for-conversation cannot exist without
repair because speakers constantly search for the next item
due for the interactive needs of the conversation. The study of
repair, therefore, demonstrates how interaction and grammar
shape each other.
Liang Tao
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Fox, Barbara, Markoto Hayashi, and Robert Jasperson. 1996. Resources
and repair: A cross-linguistic study of the syntactic organization
of repair. In Interaction and Grammar, ed. Elino Ochs, Emanuel
Schegloff, and Sandra Thompson, 185237. Cambridge: Cambridge
University Press.
Schegloff, Emanuel, Gail Jefferson, and Harvey Sacks. 1977. The preference for self-correction in the organization of repair in conversation.
Language 53: 36182.
CONVERSATION ANALYSIS
Conversation analysis (CA) is the study of talk (and other conduct) in human interaction that began with the pioneering work
of Harvey Sacks (1995) and his collaborators Emanuel Schegloff
and Gail Jefferson (e.g., Sacks, Schegloff, and Jefferson 1974;
Schegloff, Jefferson and Sacks 1977). CA seeks to establish technical specifications of the practices people use to co-construct
orderly and mutually understandable courses of action. These
specifications constitute a cumulative, empirically derived body
of knowledge that is foundational to CA as a discipline. Since its
beginnings within sociology in the late 1960s and early 1970s,
CA has become hugely influential, both as an emerging discipline in its own right and across the fields of sociology, psychology, anthropology, linguistics, and education. It is increasingly
applied in studies of institutional and organizational interaction
(including news interviews, court proceedings, emergency and
help-line calls, and doctorpatient interaction; see Drew and
Heritage 1992; Heritage and Maynard 2006) and in sociological studies of the operation of social norms and the reproduction of culture (especially related to gender [see gender and
language] and sexualities; Kitzinger 2000, 2005). Although
it originated in the analysis of talk from American-English speakers, CAs basic findings have now been replicated across many
other languages.
The intellectual roots of CA lie in a synthesis of the sociological
traditions established by Erving Goffman and Harold Garfinkel
traditions that, like other broadly social constructionist theoretical frameworks, offer models of people as agents and of a
social order grounded in contingent, ongoing, interpretive work
(see Heritage 1984). CA aims to build a science of social action,
rather than to contribute to the study of language per se. It relies
on analysis of recordings of naturally occurring human interaction (i.e., not invented or hypothetical data and not data generated by researchers via interviews or in laboratories). Recordings
are transcribed according to a distinctive transcription notation system (Jefferson 2004), but it is the recordings themselves
225
Conversation Analysis
(and not transcripts of them) that are the primary data. Sound
files are increasingly being made available on the World Wide
Web (see http://www.sscnet.ucla.edu/soc/faculty/schegloff/
sound-clips.html for sound clips from Schegloffs publications),
enabling readers of published work to access the original data.
Much early CA was based only on audio recordings (since the
technology for video recording was not yet available, but see
Goodwin 1981), which precluded analysis of such interactional
features as gesture, body deployment, and gaze. Video recordings of face-to-face interactions are now the norm. Although new
data are continually being collected, several core data sets have
been shared within the CA community since the 1970s (e.g., the
telephone conversation known as Two Girls [TG], which can
be accessed at http://www.cambridge.org/9780521532792,
Appendix 2). These shared data are widely used in teaching,
frequently reanalyzed for new phenomena, and appear in publications by a range of different authors. Analysis of these kinds
of ordinary conversations is the point of departure for studying
more specialized communicative contexts (the legal process, the
medical encounter) in which social institutions are talked into
being (Heritage 1984).
Conversation analysis has produced few theoretical manifestos but has, rather, concentrated on fine-grained empirical
studies of interaction. These studies rest upon three fundamental theoretical assumptions (Heritage 1984): i) Talk is a form of
action; that is, people use it to do things like complaining, complimenting, disagreeing, inviting, telling, and so on; ii) action
is structurally organized; that is, turns at talk are systematically
related to one another, such as (for example) when an acceptance follows an invitation or a self-deprecation follows a compliment (see adjacency pair); and iii) talk creates and maintains
intersubjectivity; that is, a first speaker understands, by what a
second speaker does, how that second speaker heard his or her
first turn as when a second speaker produces a turn hearable as
an answer, thereby showing herself /himself to have heard the
prior turn as a question.
The focus in CA research is on identifying generic orders of
organization in talk-in-interaction that are demonstrably salient
to the participants analyses of one anothers turns at talk in
the progressively unfolding interaction. Data are rarely coded
or quantified since manifest similarities in talk may turn out to
have very different interactional meanings. Key discoveries of CA
include turn-taking, action formation, sequence organization,
repair, word selection, and overall structural organization, each
of which is now sketched out.
Turn Taking
The classic paper by Sacks, Schegloff, and Jefferson (1974) presents a model to describe the practices whereby people (mostly)
speak one at a time. Summarized very simply, the model proposes that the building blocks out of which turns are composed
(turn constructional units or TCUs) can be whole sentences,
phrases, sometimes just single words, or even nonlexical
items which, in context, are recognizable to a co-participant
as possibly constituting a complete turn. Each speaker is initially entitled to just one TCU, after which another speaker has
the right (and sometimes the obligation) to speak next. As a
speaker approaches the possible completion of a first TCU in a
226
Action Formation
Researchers have focused on how speakers deploy talk (and other
conduct) in order to fashion a turn designed to be recognizable to
their recipients as doing a particular action, that is, how people
do complaining, or inviting, or declining, and so on (Atkinson
and Heritage 1984). Since CA (unlike speech-act theory) starts
from the analysis of singular episodes of human interaction and
undertakes to understand action as the co-participants understand it, one outcome of this kind of analysis is a very detailed
understanding of how (for example) complaining or inviting
are done that often departs from vernacular understandings.
Another outcome is the discovery of actions that have no vernacular name (e.g., confirming an allusion; Schegloff 1996a).
Sequence Organization
The most basic type of sequence involves two turns at talk by different speakers, the first constituting an initiating action (first
pair part) and the second an action responsive to it (second pair
part): for example, an invitation and an acceptance or declination of it; a news announcement and a news receipt (see adjacency pair). Most initiating actions can be followed by a range
of sequentially relevant (i.e., appropriately fitted) next actions,
some of which further the action of the prior turn (e.g., accepting
an invitation) and are termed preferred responses, and others of
which do not (e.g., rejecting an invitation) and are termed dispreferred. The basic two-turn adjacency pair sequence can be and
frequently is expanded. Pre-expansions are turns that come
before and are recognizably preliminary to some other action;
for example, a turn such as What are you doing tonight? can be
recognizable in context as preliminary to an invitation (hence, a
pre-invitation); a turn such as Guess what is virtually dedicated
to preannouncement. Insert expansions come between the first
and second pair parts, for example, between an invitation and
the acceptance or declination of it (Do you wanna come round
tonight? / What time? / About six. / Okay where the invitation and its acceptance are separated by an insert sequence).
Postexpansions come after the second pair part and may accept
or assess it. For example, You want me to bring you anything?
(offer: first pair part) / No, no nothing (declination: second
pair part) / Okay (acceptance of the declination, expanding the
sequence to a third turn). The authoritative work on adjacency
pairs and expansions of them, the organization of preference
and dispreference, and other types of sequence organization is
Schegloffs (2007) primer.
Repair
Interactional co-participants must manage troubles in speaking, hearing, and/or understanding talk if the interaction is not
Conversation Analysis
to founder when trouble arises (see conversational repair).
Repair is a method for fine-tuning a turn in the course of its production and for maintaining intersubjectivity. Researchers have
shown some of the practices that speakers use across a range of
different positions in talk, both in repairing their own talk (e.g.,
by deleting, inserting, or replacing a word; Schegloff, Jefferson,
and Sacks 1977) and in initiating repair on the talk of others (e.g.,
with open-class repair initiations like huh?; Drew 1997). Most
repairs are completed by the speaker of the trouble source in the
same turn (more accurately, the same TCU) as the trouble source
but can be delayed to third turn or third position, or even later
(Schegloff 1992).
Word Selection
Turns at talk are composed of lexical items selected from among
alternatives. For example, when English-language speakers refer
to themselves, they can often select between I or we (the latter
choice sometimes being used, for example, to index that they are
speaking on behalf of an organization or a couple). Alternatively,
they can self-reference in distinctive (marked) ways (e.g., selfnaming or self-description) that perform analyzable actions.
Likewise, explicit self-reference in (so-called zero-anaphora)
languages in which this is not required has been shown to be
interactionally meaningful (see the Lerner and Kitzinger 2007
collection on selection issues in self-reference). Category-based
reference to nonpresent persons also involves choices between
alternatives (Schegloff 1996b); for example, law enforcement
officers can be referred to as police or cops, and speakers selection of one or the other may be responsive to whether the speaker
is appearing in court (Jefferson 1974) or talking with adolescent
peers (Sacks 1995). CA explores how word selection is done as
part of turn design and how it informs and shapes the understanding achieved by the turns recipient.
Conclusion
Although much research remains to be done, there is, for each of
these orders of organization, an established set of core findings,
foundational to the discipline of CA. An outstanding bibliographical source of information about CA is available on the Ethno/CA
Web site maintained by Paul ten Have at http://www2.fmg.uva.
nl/emca/resource.htm.
Celia Kitzinger
227
Cooperative Principle
COOPERATIVE PRINCIPLE
Introduced by the British philosopher Herbert Paul Grice
(191388), the cooperative principle and related maxims are part
of his theory of conversational implicature.
Grice formulates the principle as an imperative: Make your
contribution such as required, at the stage at which it occurs, by
the accepted purpose or direction of the talk exchange in which
you are engaged (1975, 45). He observes that it is a well-recognized empirical fact that this ceteris paribus principle applies
to all talk exchanges that do not consist of wholly disconnected
remarks, and he adds that he would like to be able to argue
that the principle is grounded in rationality. To that end, he suggests that persons participating in conversational exchanges
do so with certain shared purposes, such as exchanging information and influencing and being influenced by others. These
shared purposes, Grice suggests, are in general only realized if
the exchanges are conducted in accordance with the cooperative principle. For those who know this, it is rational to behave in
accordance with the cooperative principle and to expect others
to do so as well. Thus, this principle and presumption are rational, given assumptions about shared conversational ends, effective means to those ends, and rationality.
Many researchers (e.g., Brown and Levinson 1987; Clark
1996) describe the cooperative principles applications and take
up thorny questions about its relation to the associated maxims
of quantity, quality, relation, manner, and politeness. There is
disagreement as to whether norms of conversational etiquette
derive from, are complementary to, or are in tension with the
cooperative principle. Asa Kasher (1976, 1982) argues against
the assumption of shared conversational purposes and contends
that the cooperative principle is superfluous since the needed
maxims can be derived from a more fundamental principle of
rational behavior: Given a goal, adopt the most effective and
least costly means to its realization. Wayne A. Davis (1998) contends that the cooperative principle lacks explanatory value and
is hopelessly ambiguous among normative, motivational, behavioral, and cognitive readings. As it is only the speakers motives
and beliefs that are causally involved in the intentional production of implicit meanings, what the speaker implicated or implicitly expressed does not depend on the thoughts or presumptions
of the audience. In other words, Grice erred when he made the
hearers presumption that the speaker observes the cooperative
principle a condition on the speakers expressing one thing by
saying something else.
Paisley Livingston
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Brown, Penelope, and Stephen C. Levinson. 1987. Politeness: Some
Universals in Language Usage. Cambridge: Cambridge University
Press.
Clark, Herbert H. 1996. Using Language. Cambridge: Cambridge
University Press.
Davis, Wayne A. 1998. Implicature: Intention, Convention, and Principle
in the Failure of Gricean Theory. Cambridge: Cambridge University
Press.
Grice, Herbert Paul. 1975. Logic and conversation. In Syntax and Semantics.
Vol. 3. Ed. Cole and Morgan, 4158. New York: Academic Press.
228
Grice, Herbert Paul.. 1989. Studies in the Way of Words. Cambridge: Harvard
University Press.
Kasher, Asa. 1976. Conversational maxims and rationality. In Language
in Focus: Foundations, Methods, and Systems, ed. A. Kasher, 197211.
Dordrecht, the Netherlands: Reidel.
Kasher, Asa. 1982. Gricean inference revisited. Philosophica
29: 2544.
(1a) is a case of sluice-stranding, where an isolated whphrase stands in place of an understood indirect question.
It contains not only the wh-phrase but also a preposition
from whose complement the wh-phrase has apparently been
moved. It is technically possible to derive this construction
through some combination of wh-movement and deletion.
The difficulty is that sluice-stranding is both more productive
and more restricted than a derivational account would suggest.
Sluicing in general is possible where the purported extraction
site normally forbids extraction (Ross 1969). (2a) illustrates for
ordinary sluicing of a prepositional phrase; (2b) illustrates for
sluice-stranding.
(2)
(4)
d. where to/from/*near.
e. * which (book) with/to/from/next to/about/beside. There are
Conclusions
Syntactic constructions appear to be ranged on a continuum
from words through idioms through truly idiosyncratic constructions through more general but still specialized constructions to the most general corelike structures and principles of
universal grammar. It is likely that certain peripheral constructions may be related to the core in systematic ways, say, by
relaxing certain conditions of core grammar (Chomsky 1986,
147). But C/P per se, however valuable heuristically, may not
merit genuinely theoretical status. (Cf. head-driven phrase
structure grammar and construction grammar.)
The implication for learning is that the learner stores current
analyses of novel utterances in the lexicon, with idiosyncratic
and general properties (see lexical acquisition). The learning procedure attempts to construct more general lexical entries
on the basis of positive experience, where common parts of existing lexical entries are retained and differing parts are replaced by
a variable. The resulting lexical entry functions as a schema or
rule that encompasses existing entries and permits construction
of new utterances. In turn, this schema, along with others, may
be further abstracted into a still more general schema by replacing further dimensions of variation with variables (Tomasello
2003), producing in the limit grammatical rules of full generality
where warranted (see also syntax, acquisition of).
Peter W. Culicover
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Chomsky, Noam. 1965. Aspects of the Theory of Syntax. Cambridge, MA:
MIT Press.
. 1981. Lectures on Government and Binding. Dordrecht, the
Netherlands: Foris.
. 1986. Knowledge of Language. New York: Praeger.
. 1993. A minimalist program for linguistic theory. In The View
from Building Twenty, ed. Kenneth Hale and Samuel J. Keyser, 152.
Cambridge, MA: MIT Press.
229
Corpus Callosum
Culicover, Peter W. 1999. Syntactic Nuts. Oxford: Oxford University
Press.
Culicover, Peter W., and Ray Jackendoff. 2005. Simpler Syntax.
Oxford: Oxford University Press.
Ross, John R. 1969. Guess who. In Proceedings of the Fifth Annual
Meeting of CLS, ed. Robert I. Binnick et al., 25286. Chicago: Chicago
Linguistics Society,
Tomasello, Michael. 2003. Constructing a Language. Cambridge: Harvard
University Press.
CORPUS CALLOSUM
Although language processing relies predominantly on left
hemisphere networks, certain functional units are also localized in the right hemisphere. As a result, there is a strong
need for an interaction between the two hemispheres during
most language-related processes. The neuronal basis for this
interaction is provided by a brain structure located between the
two hemispheres, the so-called corpus callosum (CC), which is
the major interhemispheric fiber tract. The more than 200 million
axons forming the CC originate from nearly all cortical regions,
including the language areas, and they primarily link homologue
regions of the hemispheres. The fibers cross the interhemispheric
gap ordered by their cortical origin. Due to functional specialization of the cerebral cortex, this anatomical organization also
establishes a functional topography within the CC. Thus, different subregions of the tract are related to specific functional
networks.
Viewed in a midsagittal section of the brain (see Figure 1),
two subregions of the CC seem to be particularly relevant for language processing: Fibers passing through portions of the anterior
CC connect the language production network situated in the left
inferior frontal cortex (see frontal lobe and brocas area)
with its contralateral homologue, while axons in the posterior
CC interconnect the cortical areas in the temporal lobes (see
also wernickes area) which are responsible for language
perception.
230
Corpus Linguistics
temporal lobes seems to prevent the interhemispheric exchange
required to integrate prosodic and syntactic information.
One often-quoted modulatory role of the CC is the functional
inhibition of the contralateral hemisphere while the ipsilateral
hemisphere is engaged in a task for which it is specialized. The
advantage of such an inhibitory mechanism might be the reduction of interfering influence coming from the opposite hemisphere. A finding recently published by Alexander Thiel and
coworkers (2006) could be interpreted in this vein. The authors
measured the activation of the left and right inferior frontal gyrus
(IFG) in a verb generation task using positron emission tomography (see neuroimaging). In some of the administered trials,
repetitive transcranial-magnetic stimulation, a method to induce
a temporary disruption of ongoing neuronal activity, was simultaneously applied over the left IFG. Besides a reduction of the
activation in the stimulated left IFG, this virtual brain lesion also
induced a relative increase in the response measured in the right,
nonstimulated IFG. Thus, the suppression of the left IFG area
seems to result in a disinhibition of its contralateral homologue.
The studies cited here illustrate that not only the exchange of
information but also the coordination of bihemispheric processing
is supported by transcallosal connections. Furthermore, a recent
functional imaging study has shown that interindividual differences, which can be found in size and micro-architecture of the
CC, have consequences for language processing (Westerhausen
et al. 2006, 80). Here, the degree of activation differences between
left and right inferior frontal language areas (in a word production task) appeared to be directly related to differences in the fiber
architecture of the callosal connection. Whether structural CC
differences between individuals also trigger differences in performance or are even associated with language disorders (as was
hypothesized for dyslexia) has still to be confirmed.
Ren Westerhausen
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Friederici, Angela D., D. Yves von Cramon, and Sonja A. Kotz. 2007. Role
of the corpus callosum in speech comprehension: Interfacing syntax
and prosody. Neuron 53: 13545.
Gazzaniga, Michael S. 2000. Cerebral specialization and interhemispheric communication Does the corpus callosum enable the human
condition? Brain 123.7: 12931326.
Thiel, Alexander, Birgit Schumacher, Klaus Wienhard, Stefanie Gairing,
Lutz W. Kracht, Rainer Wagner, Walter F. Haupt, and Wolf-Dieter
Heiss. 2006. Direct demonstration of transcallosal disinhibition in
language networks. Journal of Cerebral Blood Flow and Metabolism
26.9: 11227.
Westerhausen, Ren, Frank Kreuder, Sarah Dos Santos Sequeira, Christof
Walter, Wolfgang Woerner, Ralf A. Wittling, Elisabeth Schweiger, and
Werner Wittling. 2006. The association of macro- and microstructure of the corpus callosum and language lateralisation. Brain and
Language 97: 8090.
CORPUS LINGUISTICS
This term refers to linguistic research that uses corpus data as
the primary object of study. The term, therefore, describes a
methodology rather than a field of linguistics; corpus research
has been carried out in most areas of formal and applied
linguistics, including phonetics, phonology, morphology, syntax, semantics, pragmatics, discourse analysis (linguistic), sociolinguistics, language acquisition,
PSYCHOLINGUISTICS, HISTORICAL LINGUISTICS, dialectology, and lexicography.
231
Corpus Linguistics
232
Creoles
As that example illustrates, the tacit knowledge of language
(syntax, semantics, phonetics, morphology, categorization, dialect, orthography) and also of the world
that is needed to use language imaginatively is richly detailed
and widely various. Both hearer and speaker must possess this
knowledge if the creative usage is to be understood.
Margaret A. Boden
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Binsted, Kim, and G. D. Ritchie. 1997. Computational rules for punning
riddles. Humor: International Journal of Humor Research 10: 2576.
Boden, M. A. 2004. The Creative Mind: Myths and Mechanisms. 2d ed.
London: Routledge.
Boden, M. A.. 2006. Mind as Machine: A History of Cognitive Science.
Oxford: Oxford University Press.
Fauconnier, G. R., and Mark Turner. 2002. The Way We Think: Conceptual
Blending and the Minds Hidden Complexities. New York: Basic Books.
Hofstadter, D. R., and Melanie Mitchell. 1993. The copycat project: A model of mental fluidity and analogy-making. In Advances
in Connectionist and Neural Computation Theory. Vol. 2: Analogical
Connections. Ed. Keith Holyoak and John Barnden, 31112. Norwood,
NJ: Ablex.
Humboldt, Wilhelm von. [1836] 1988. On Language: The Diversity
of Human Language-Structure and Its Influence on the Mental
Development of Mankind. Trans. Peter Heath. Cambridge: Cambridge
University Press.
CREOLES
SocioHistorical, Terminological, and Epistemological
Background
The concept Creole has not been operationalized with rigorous
and reliable criteria in linguistic theory. At best, it is a sociohistorically and politically motivated concept, often misidentified
as linguistic (DeGraff 2005b, 2009; Mufwene 2008).
Etymologically, the word Creole derives from the Portuguese
crioulo and/or Spanish criollo raised in the home (from criar
to raise, to breed). In Caribbean history, the labeling of biological species, including humans, as Creole seems to have
preceded the labeling as Creole of certain speech varieties.
Both uses referred to nonindigenous varieties that developed
locally, in contrast to their counterparts from Europe and Africa.
The original uses of the word were thus devoid of any specific
structural correlates (Mufwene 2001, 311; Chaudenson and
Mufwene 2001, Chap. 1).
In keeping with this original usage and to avoid circularity and the sort of controversial linguistic assumptions that
are noted in Mufwene 2008, I here ostensively use Creole as a
label for certain speech varieties that became emblematic of the
newly created communities the Creole communities on and
around colonial Caribbean plantations. These are the classic
Creole languages.
Caribbean Creole languages developed mostly among
Europeans and Africans via language acquisition by adults and
children in a complex mix of language-contact settings. The
complex sociohistorical factors therein included a continuum
of social divides and power asymmetries (Chaudenson and
Mufwene 2001). One end of this continuum was marked by
Creole Exceptionalism
The term Creole exceptionalism (DeGraff 2003) covers a subset of long-standing hypotheses whereby Creole languages
constitute a sui generis class on phylogenetic and/or structural
grounds. Here is a sample:
(i) Creoles are degenerate offshoots of their European
ancestors;
(ii) Creoles
genealogy;
are
special
hybrids
with
exceptional
(iii) Creoles are the only contemporary languages with a history of abnormal transmission that deprives them of any
structurally full-fledged ancestors.
(iv) The Pidgin-to-Creole transition recapitulates the transition from pre-human protolanguage to human language.
(For a fuller development of these arguments, see DeGraff 2005a,
2009.)
Creoles
Bloomfield (1933, 472), The creolized language has the status of
an inferior dialect of the masters speech.
Even in the latter half of the twentieth century, certain linguists claimed that structural linguistic factors, related to (e.g.)
morphological simplicity and a vocabulary [that] is extremely
poor, are among the greatest obstacles to the blossoming of
Creoles (Valdman 1978, 345; cf. Whinnom 1971, 110; Samarin
1980, 221; Seuren and Wekker 1986; and Quint 1997, 58). Pieter
Seuren (1998, 292) has elevated the alleged extraordinary simplicity of Creole languages to historical universal.
There is no reliable empirical or theoretical basis for the claim
that Creole languages are uniformly less complex than their
European ancestors. For example, certain aspects of my native
Haitian Creole (HC) signal an increase in complexity to the extent
that these properties of HC have no counterpart in French, HCs
European ancestor (DeGraff 2001b, 284). Furthermore, HC, like
any other language, expands its vocabulary as needed, via productive affixation, neologisms, borrowings, and so on (DeGraff
2001a; Fattier 1998).
CREOLES AS SPECIAL HYBRIDS? Lucien Adams (1883) hybridologie linguistique hypothesis posited different linguistic
templates for different races. The latter belong to distinct evolutionary rungs, with their respective linguistic templates ranked
in a corresponding hierarchy of complexity. Upon language
contact, these templates will cross-fertilize (i.e., hybridize),
and the most primitive grammar (in this scenario, the grammar of the lower race of speakers, i.e., the non-European
speakers) imposes an upper bound of complexity on the hybrid
grammar. In such scenario, the European contribution to the
hybridization of European and non-European languages is limited to superficial traits, such as the phonetic shapes of words
only these shapes, and not the complex grammars of European
languages, can be acquired by the allegedly inferior minds of
the non-Europeans.
Claire Lefebvres (1998) relexification hypothesis is far
removed from Adams race-theoretical postulates. For Lefebvre,
it is because the Africans in Haiti had very limited access to
French that they were virtually unable to learn any aspect of
French grammar. Thus, they could only overlay French-derived
phonetic strings on their native substrate grammars, with the latter being kept nearly intact in the original Creole languages.
Consider again HC. Current results from L2A research predicts that HC structure would have indeed evolved under some
influence from the substrate languages. L2A research also documents that adult learners at every stage acquire more than phonetic strings from their target. Unsurprisingly, HC instantiates,
alongside substrate-influenced patterns, a wide range of superstrate-derived properties that apparently have no analogues in
the substrate languages (DeGraff 2002).
Adams and Lefebvres proposals share one non-uniformitarian assumption, namely, that Creole creators, unlike L2 learners elsewhere, were unable to learn anything abstract about
their target language. Yet the lexicon and morphology of HC
demonstrate that Creole creators were able to segment and
parse target speech (here, French), including affixes. Such segmentation and parsing contradict the claim that the creators of
HC could not access any abstract property of French grammar.
234
Creoles
The broken transmission and linguistic fossils doctrines
are further undermined by a vast range of comparative data and
empirical and theoretical observations. As mentioned earlier,
there is ample evidence for systematic lexical and morpho-syntactic correspondences between radical Creoles and their European
lexifiers from the onset of Creole formation onward (Fattier 1998;
DeGraff 2001a, 2005b, 2009; Mufwene 2008). There is also ample
evidence for transfer from the African substrate languages into
Creole grammars. This is as expected given the aforementioned
facts of Caribbean history and the results from L2A research. The
sort of structureless pidgin that is an essential ingredient in the
traditional Pidgin-to-Creole scenario renders mysterious any systematic set of structural correspondences between Creoles and
their ancestor languages. Besides, the magnitude of structural
gaps in the history of non-Creole languages seems comparable
to, and sometimes even greater than, that of their counterparts
in Creole diachrony (DeGraff 2005b, 2009), pace Thomason and
Kaufman (1988, 812, 206) and Thomason (2002, 105).
If the rigorous criteria of the Comparative Method [CM]
include the establishment of recurring phonological correspondences in morphemes of identical or similar meanings, including
much basic vocabulary the establishment of systematic morphosyntactic correspondences (Thomason 2002, 103), then the
available evidence puts Caribbean Creoles squarely in the scope
of the CM (DeGraff 2005b, 2009; Mufwene 2008, pace Thomason).
Such evidence militates against the postulation, in Creole formation, of an exceptional and abnormal break in transmission
with subsequent creation of all new linguistic structure from the
hypothetical scraps of a Pidgin.
235
236
The label critical linguistics has given way to critical discourse analysis, as the field developed to include wider areas
of social concern and more social theory. In particular, in addition to the writings of Marx and Marxian writers (see marxism
and language ), CDA practitioners have often made appeal
to Antonio Gramsci, Michel Foucault, and the British sociologist Anthony Giddens. The label CDA became associated with
the work of, in particular, Norman Fairclough (e.g. 1989, 1992),
Teun van Dijk (e.g. 1993, 1998, 2005) and Ruth Wodak (1996;
Wodak et al. 1999). Faircloughs work, mainly cast within a neoMarxian mold, has developed detailed concepts and models of
discourse, intertextuality, and genre, while leaving the
linguistic dimension of discourse comparatively undeveloped.
His work can be characterized as social theory that gives full
recognition to the constitutive role of language in society, in the
form of interlinked discursive practices. Van Dijks work is
rooted in formal discourse analysis of the 1970s with a cognitive
tendency, and has sought to provide deeper discourse-based
understanding of major pragmatic notions such as context,
as well as of the social-theoretic notion of ideology. Wodaks
work has developed the discourse historical method (Wodak
and Meyer 2001), a methodology for empirical investigation
that advocates the study of intersecting texts and samples of
talk collected from various milieus and representing various
genres, for which analysis of the historical context is regarded
as crucial. Numerous other scholars of the same period, whose
accomplishments it is not possible to describe here, produced
work that was overtly or implicitly critical, particularly in
France, Germany, the Netherlands and Belgium, the Hispanic
world, and Australia.
The CDA literature has introduced a number of key concepts and claims. The principle that discourse is constitutive of
social processes and structures has been extended to include
other semiotic systems, notably pictorial ones (Kress and van
Leeuwen 2001). Discourses, in the plural, are understood as relatively stable uses of language serving the structuring of social
life, organizations, and political systems. Such discourses consist of interlinked discursive practices or genres. Discourses
may be of various kinds and operate in different ways. Thus,
political ideologies, scientific worldviews, ethical systems, and
the like are said to represent the world in particular ways. Genres
regulate interaction and thus control social behavior, examples
being interviews, news broadcasts, medical consultations, educational examinations, and so forth. When these discursive
practices are viewed as interlinked, they involve intertextuality and interdiscursivity, leading to the colonization of one
discourse by another and to hybrid genres a process regarded
as integral to and an index of social change. Discursive practices viewed as a complex network are referred to as an order of
discourse (Fairclough 1992). A crucial claim of CDA is that the
details of particular instances of text and talk are related in complex ways to these structures, structures regarded as embodying
power.
CDA has not produced a theory of language that explains how
meanings produced by utterances have the tight connection to
social structures, which is often claimed. There are nonetheless
numerous examples of the description, analysis, and interpretation of utterances in the CDA literature, and these are dependent
special focus in many CDA works on issues such as racism, militarism, media bias, marketization, and gender.
This survey necessarily neglects much work that has a critical
stance without subscribing to the CDA label. Such work would
include the writing on language and gender that has arguably had
an impact on modern social behavior (for an overview see Eckert
and McConnell-Ginet 2003). It would also include work on society and on controversial political matters in North America (G.
Lakoff 1996); R. Lakoff 1990; Lemke 1995). Further, it should also
include the emerging scholarship on discourse in its sociopolitical context in China (cf. Gu 2001) and on a smaller scale in the
Middle East and Africa. It may be the case that something like the
CDA approach emerges in periods of significant socioeconomic
or political change: In some respects, CDA may be considered to
have the character of a social movement.
Paul Chilton
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bourdieu, Pierre. 1982. Ce que parler veut dire: Lconomie des changes
linguistiques. Paris: Fayard.
Chilton, Paul. 1996. Security Metaphors: Cold War Discourse from
Containment to Common House. New York: Peter Lang.
. 2004. Analysing Political Discourse: Theory and Practice.
London: Routledge.
. 2005. Missing links in mainstream CDA: Modules, blends
and the critical instinct. In A New Agenda in (Critical)l Discourse
Analysis: Theory and Interdisciplinarity, ed. Ruth Wodak and Paul
Chilton 1950. Amsterdam: Benjamins.
Eckert, Penelope, and Sally McConnell-Ginet. 2003. Language and
Gender. Cambridge: Cambridge University Press.
Fairclough, Norman. 1989. Language and Power. London: Longman.
. 1992. Discourse and Social Change. Cambridge, UK: Polity.
Fowler, Roger, Gunther Kress, Robert Hodge, and Tony Trew, eds. 1979.
Language and Control. London: Routledge.
Goatly, A. 2007. Washing the Brain: Metaphor and Hidden Ideology.
Amsterdam: Benjamins.
Gu, Yueguo. 2001. The changing orders of discourse in a changing
China. In Studies in Chinese Linguistics. Vol 2. Ed. Haihua Pan, 3158.
Hong Kong: Linguistic Society of Hong Kong.
Hodge, Robert, and Kam Louie. 1998. The Politics of Chinese Language
and Culture. London and New York: Routledge.
Kress, Gunther, and Robert Hodge. [1979] 1993. Language as Ideology.
London: Routledge.
Kress, Gunther, and Theo van Leeuwen. 2001. Multimodal Discourse: The
Modes and Media of Contemporary Communication. London: Arnold.
Lakoff, George. 1996. Moral Politics: What Conservatives Know That
Liberals Dont. Chicago: University of Chicago Press.
Lakoff, Robin. 1990. Talking Power: The Politics of Language in Our Lives.
New York: Basic Books.
Lemke, Jay. 1995. Textual Politics: Discourse and Social Dynamics.
London: Taylor & Francis.
Martin, Jim R. 2004. Positive discourse analysis: Power, solidarity and
change, Revista Canaria de Estudios Ingleses. 49: 179200.
OHalloran, Kieran. 2003. Critical Discourse Analysis and Language
Cognition. Edinburgh: Edinburgh University Press.
Saussure, Louis de, and Peter Schulz, eds. 2005. Manipulation and
Ideologies in the Twentieth Century: Discourse, Language, Mind.
Amsterdam: Benjamins.
Van Dijk, Teun. 1993. Elite Discourse and Racism. Newbury Park,
CA: Sage.
. 1998. Ideology. London: Sage.
237
Critical Periods
. 2005. Contextual knowledge management in discourse production: A CDA perspective. In A New Agenda in (Critical) Discourse
Analysis: Theory and Interdisciplinarity, ed. Ruth Wodak and Paul
Chilton, 71100. Amsterdam: Benjamins
Widdowson H. G. 2005. Text, Context, Pretext: Critical Issues in Discourse
Analysis. Oxford: Blackwell.
Wodak, Ruth. 1996. Disorders of Discourse. London: Longman.
Wodak, Ruth, and Michael Meyer. 2001. Methods of Discourse Analysis.
London: Sage.
Wodak, Ruth, Rudolf de Cillia, Martin Reisigl, and Karen Liebhart. 1999.
The Discursive Construction of National Identity. Edinburgh: Edinburgh
University Press.
CRITICAL PERIODS
The idea of a critical period for language acquisition is one of the
most debated issues in language acquisition theory. One reason
for controversy in the empirically based discussion of the critical
period hypothesis (CPH) has been the different understandings
of what a critical period actually means and what effects it might
have on language. Another more basic philosophical cause for
lack of unanimity is the notions central symbolic role in the
naturenurture divide and the concomitant ideological preferences among researchers to stress biological or environmental
aspects of language development. It is probably fair to say that
the rhetorical tone that is sometimes noticeable in debates on
the CPH originates to a considerable extent from the various
epistemological commitments of the protagonists, nativist or
constructivist, cognitive or social constructivist, general cognitive or modular, and so on.
The CPH is relevant for all kinds of language learning. When
much more of the discussion is about second language learning (see second language acquisition) than about first
language learning, it is a reflection of the fact that first language
learning, except in cases of isolation from language input, starts
from age zero. This means that data from delayed first language
acquisition are rare. On the contrary, for second language acquisition (SLA), massive data from language learning at different
phases of the life span are available. Most of this entry, therefore,
addresses second language acquisition rather than first language
acquisition.
The notion of critical period has been used by ethologists to
explain the fact that the development of several aspects of species-specific behavior are dependent on early stimulus exposure
or experiences. A critical period can generally be defined, therefore, as a time span in early life during which the organism is
responsive to those stimuli in the external environment that are
crucial or relevant for a behavior (or capacity) to develop eventually in keeping with a species-specific standard. If the organism
does not encounter or experience the particular stimuli during
the time span for sensitivity, that behavior will either not develop
at all or eventually reach an end state that differs from the species-specific ultimate standard. In addition to maturationally
constrained learning in humans and other species, there are two
other kinds of learning, namely, learning that occurs with equal
success at any time over the lifespan and learning that becomes
effective only at later phases of cognitive development.
Examples of maturationally constrained behaviors range
from imprinting in geese and song learning in songbirds (see
238
Critical Periods
that could explain an end point for the critical period at puberty.
However, such a role for lateralization was soon demonstrated
not to be correct: The widely accepted theory in the 1960s that
lateral specialization is progressive, increasing with age from
infancy to adolescence has long been abandoned in the face of
accumulated evidence that indicates that cerebral lateral specialization is established from early infancy or even during
fetal development (Paradis 2004, 107).
Lennebergs proposals were based mainly on general informal
observations, but subsequent research has provided theoretical
frameworks and substantial empirical data that, taken together,
can be given an interpretation that is compatible with the predictions of the hypothesis. The first point, on the distinction between
(implicit) acquisition and (explicit) learning, has been addressed
in a series of theoretical discussions about language learning differences between children and adults (see DeKeyser 2003), with
some of its most well known exponents in S. Krashens (1988)
acquisition-learning hypothesis, R. Bley-Vromans (1989) fundamental difference hypothesis, and S. Felixs (1985) competition
hypothesis. Also, the various perspectives on access to universal grammar (UG) (full/direct, partial/indirect, or no access)
in UG-framed SLA theories (see White 2003) can be translated
into a CPH framework (cf. Pinkers 1994 use-it-then-loose-it
hypothesis).
As for the second point, a multitude of studies have singled
out puberty or early adolescence as an end point for high or
nativelike proficiency levels in a second language. Several other
studies have pointed to a discontinuity at earlier ages, in particular around age six, especially for phonology and grammar (Long
2005; see, especially, Johnson and Newport 1989).
In relation to point three on obtained nativelikeness in
younger and older starters, empirical observations are by and
large compatible with the CPH; that is, child learners frequently
have results in the range of native controls, whereas this is rare,
or even claimed never to have been demonstrated, in adult
learners (Hyltenstam and Abrahamsson 2003; Long 2005). In
many studies, however, nativelikeness is not focused upon. It is,
rather, the level of ultimate attainment, nativelike or not, among
younger and older learners that is correlated with age of onset
(AO). One of the most robust results in CPH-related research is a
strong negative correlation between AO and ultimate attainment
in a second language.
In addition, in relation to Lennebergs fourth point, the
hypothesis correctly seems to predict the fact that delayed exposure to an L1 more severely affects the level of ultimate attainment than delayed exposure to an L2. Case studies of abused
children who have not been exposed to any language before
puberty (Curtiss 1977) show severe limitations in the development of grammar and pronunciation, whereas learning a second
language from the same age allows high levels of proficiency in
that language. It has frequently been pointed out that deprivation data are difficult to interpret in terms of a critical period for
language, but similar results have been obtained from studies
of delayed first language exposure to American Sign Language
(ASL) (e.g., Mayberry 1993).
A number of counterarguments to the CPH have also been
put forward over the decades. There is a current consensus
that some of these are, in fact, not valid arguments, as they deal
239
Critical Periods
and claimed that studies show up to 15 percent of postpuberty-investigated samples that reach this level. At the other end,
many studies have shown that nativelike ultimate attainment is
far from always obtained among learners with prepuberty AOs
(Abrahamsson and Hyltenstam 2009).
Research related to the idea of a critical period for language
is central to language acquisition theory, in particular for the
understanding of defining differences between first and second
language acquisition. Intensive efforts at understanding the factors behind age differences in language acquisition outcomes
have been made over the last 15 years, but the field is far from
approaching a consensus about the role of maturational constraints or critical periods. However, current claims for reconsidering definitions of concepts, analytical instruments, and
research methodologies (see, e.g., Birdsong 2005; Hyltenstam
and Abrahamsson 2003; Long 2005) are promising for decisive
steps forward in the near future.
Kenneth Hyltenstam
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Abrahamsson, N., and K. Hyltenstam. 2009. Age of onset and nativelikeness in a second language: Listener perception versus linguistic scrutiny. Language 59: 249306.
Bialystok, E., and B. Miller. 1999. The problem of age in second-language acquisition: Influences from language, structure, and task.
Bilingualism: Language and Cognition 2: 12745.
Birdsong, D. 2005. Interpreting age effects in second language acquisition. In Handbook of Bilingualism: Psycholinguistic Perspectives, ed. J.
Kroll and A. M. B. De Groot, 10927. Cambridge: Cambridge University
Press.
Birdsong, D., ed. 1999. Second Language Acquisition and the Critical
Period Hypothesis. Mahwah, NJ: Lawrence Erlbaum.
Birdsong, D., and M. Molis. 2001. On the evidence for maturational
constraints in second-language acquisition. Journal of Memory and
Language 44: 23549.
Bley-Vroman, R. 1989. What is the logical problem of foreign language
learning? In Linguistic Perspectives on Second Language Acquisition,
ed. S. Gass and J. Schachter, 4168. Cambridge: Cambridge University
Press.
Boothe, R. G. 1997. A neonatal visual deprivation syndrome. Perception
26: 766.
Curtiss, S. 1977. Genie: A Psycholinguistic Study of a Modern-day Wild
Child. New York: Academic Press.
DeKeyser, R. M. 2003. Implicit and explicit learning. In Doughty and
Long 2003, 31348.
DeKeyser, R. M., and J. Larson- Hall. 2005. What does the critical
period really mean? In Handbook of Bilingualism: Psycholinguistic
Approaches, ed. J. F. Kroll and A. M. B. De Groot, 88108. Oxford: Oxford
University Press.
Doughty, C., and M. Long, ed. 2003. Handbook of Second Language
Acquisition. Oxford: Blackwell.
Felix, S. 1985. More evidence on competing cognitive systems. Second
Language Research 1: 4772.
Flege, J. E., E. M. Frieda, and T. Nozawa. 1997. Amount of native-language (L1) use affects the pronunciation of an L2. Journal of Phonetics
25: 16986.
Hyltenstam, K., and N. Abrahamsson. 2003. Maturational constraints in
SLA. In Doughty and Long 2003, 53988.
Johnson, J. S., and E. L. Newport. 1989. Critical period effects in second language learning: The influence of maturational state on the
240
and Johnson 1980). When meaning making is based on such elementary human experiences, the result may be (near-)universal
meaning (content) though under a particular interpretation
(construal), that is, conceived of in a certain manner, to use
Hoyt Alversons phrase (1991, 97).
Language, on this view, consists of a set of linguistic signs, that
is, pairings of form and meaning (which can range from simple
morphemes to complex syntactic constructions). Learning
a language means the learning of such linguistic signs. Thus,
language can be regarded as a repository of meanings stored in
the form of linguistic signs shared by members of a culture. This
lends language a historical role in stabilizing and preserving a
culture. This function becomes especially important in the case
of endangered languages (see extinction of languages),
and it often explains why minorities insist on their language
rights (see language policy).
Members of a culture interact with one another for particular purposes. To achieve their goals, they produce particular discourses. Such discourses are assemblies of meanings that relate
to particular subject matters. When such discourses provide a
conceptual framework within which significant subject matters are discussed in a culture, and when they function as latent
norms of conduct, the discourses can be regarded as ideologies (see, e.g., Charteris-Black 2004; Musolff 2004; Goatly 2007).
Discourse in this sense is another source of making meaning in
cultures. A large part of socialization involves the learning of how
to make meaning in a culture.
241
242
Cycle, the
Conclusion
Culture and language are connected in many ways, and the
interconnections can be studied from a variety of different
perspectives. Following Geertz, I tried to develop a view of the
relationship that is based on how we make sense of our experiences linguistic or otherwise. Recent cognitive science and cognitive linguistics provide us with new ideas and methodological
tools with which we can approach the issue of meaning making
in cultures, both in its universal aspects and in its infinite crosscultural variety.
Zoltn Kvecses
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Alverson, Hoyt. 1991. Metaphor and experience: Looking over the
notion of image schema. In Beyond Metaphor: The Theory of Tropes
in Anthropology, ed. J. Fernandez, 94117. Stanford, CA: Stanford
University Press.
Charteris-Black, Jonathan. 2004. Corpus Approaches to Critical Metaphor
Analysis. Houndsmill, UK: Palgrave Macmillan.
Encyclopedia Britannica Ready Reference. 2003. Chicago: Encyclopedia
Britannica. Electronic version.
Fillmore, Charles. 1982. Frame semantics. In Linguistics in the Morning
Calm, 111137. Hanshin: The Linguistic Society of Korea.
Foley, William A. 1997. Anthropological Linguistics: An Introduction.
Oxford and Malden, MA: Blackwell.
Geertz, Clifford. 1973. The Interpretation of Cultures. New York: Basic
Books.
Gibbs, Raymond W. 2006. Embodiment and Cognitive Science. New
York: Cambridge University Press.
Glick, J. 1975. Cognitive development in cross-cultural perspective. In
Review of Child Development Research. Vol. 4. Ed. F. Horowitz, 595
654. Chicago: University of Chicago Press.
Goatly, Andrew. 2007. Washing the Brain: Metaphor and Hidden Ideology.
Amsterdam: John Benjamins.
Johnson, Mark. 1987. The Body in the Mind. Chicago: University of
Chicago Press.
Kimmel, Michael. 2001. Metaphor, Imagery, and Culture: Spatialized
Ontologies, Mental Tools, and Multimedia in the Making. Ph.D. diss.,
University of Vienna.
Kvecses, Zoltn. 2005. Metaphor in Culture: Universality and Variation.
Cambridge: Cambridge University Press.
CYCLE, THE
The syntactic cycle was originally formulated in Chomsky (1965)
as a general principle of grammar that constrains the way transformational rules can apply in the derivation of sentences
containing embedded clauses. The term cycle refers to the property of syntactic derivations whereby transformations apply
within syntactic subdomains before they apply to larger syntactic domains that contain them. Thus, in (1), where A, B, and C
denote syntactic domains to which transformational rules can
apply, the set of transformations must apply to C before B, and
then to B before A (see Figure 1). If a rule X applies to domains
AC in a single derivation, it applies successive cyclically to C,
then B, and finally A that is, starting with the smallest cyclic
domain and proceeding stepwise to the largest (see Boeckx 2007
for discussion).
Noam Chomsky sharpens his original formulation as the
strict cycle condition (SCC):
No rule can apply to a domain dominated by a cyclic node A in
such a way as to affect solely a proper subdomain of A dominated
by a node B which is also a cyclic node. (1973, 243)
243
Cycle, the
(1)
Deconstruction
A
Figure 1.
The SCC includes the further restriction that transformations
may not revisit a subdomain after they have applied to the larger
domain that contains it. For example, once a transformation
applies to domain B in Figure 1, no transformation can apply
solely within the subdomain C. Since the earliest formulations,
clauses (complementizer phrase (CP)/inflection phrase (IP)
in current analysis) have been designated as cyclic domains.
Nominal phrases (i.e., noun phrase [NP] and determiner phrase
[DP]) and more recently light verb phrase (vP) have also been
proposed as additional cyclic domains.
Empirical motivation for the SCC involves deviant sentences
whose derivation violates the SCC, for example (2) under the
partial derivation given in (3).
(2)*Who did you wonder what bought?
(3) a.
b.
c.
d.
244
D
DECONSTRUCTION
Deconstruction is a practice of exceptionally close and vigilant
critical reading that aims to reveal the various contradictions
or moments of aporia (of paradox or strictly irresolvable doubt)
endemic to the texts of Western philosophy, literature, and other
kinds of writing. It is best approached through the work of Jacques
Derrida (19302004), the most vigorous exponent of deconstruction and a thinker centrally concerned with issues in semantics,
hermeneutics (see philology and hermeneutics),
speech-act theory, and philosophy of language and logic (see
especially Derrida 1973, 1978, 1982, 1989).
Perhaps the most striking example is Derridas lengthy
treatment of Jean-Jacques Rousseau in Of Grammatology
(Derrida 1976). Here, he shows how Rousseaus ideas about a
vast range of topics nature, culture, language, society, ethics,
politics, history, sexual relations, personal identity, literature,
Deconstruction
and music are affected by a curious logic of supplementarity
that constantly twists his argument back against itself and thereby
subverts its manifest intent. Rousseau wants to say and does
quite explicitly state that in each case, there is (or once was)
an original, authentic, natural, uncorrupted state that then gives
way to a decadent, artificial, and degraded state where human
beings are condemned to live at a distance from their true nature
and enter into all kinds of intrinsically bad (since by very definition unnatural) relationship with themselves, each other, and
the world around them. Yet in each case, his argument comes up
against a stubbornly insistent counterlogic that throws its claims
into doubt by reversing the conceptual order of priorities upon
which that argument relies.
Thus, Rousseau sees language and music as having their
common point of origin in a mode of passionate speech-song
that expresses human feelings directly and without, as yet, any
need for those various bad supplements of syntactic and
lexical structure in language or harmony and counterpoint in
music that had since come to work their insidious, corrupting
effects. Indeed, as Derrida notes, the key word supplement
occurs with remarkable frequency and weight of semantic
implication throughout Rousseaus writing on these topics. Yet
when he treats them in a more sustained and reflective way,
Rousseau has perforce to concede that the melodious aspect of
music is always dependent on a background sense of its harmonic implications, just as the expressive aspect of language
depends on the existence of grammatical and lexical structures in the absence of which it could communicate nothing
whatsoever.
The same applies to his thinking about matters of history, politics, and civil society. Here, Rousseau purports to trace a process
of epochal decline from the close-knit, organic, natural communities that once existed before the advent of all the corrupting
forces of power, class, education, authority, political influence,
acquired expertise, and so forth which are falsely considered
progress or civilization by those same decadent standards.
And again, what passes for culture among the denizens of
modern society is, in truth, just another melancholy sign of the
falling away from nature or the ever more false and artificially
cultivated manners, practices, and modes of expression that figure as mere supplements to an otherwise perfectly self-sufficient
natural state. Yet here also Derrida shows that the term supplement is subject to a kind of dislocating force, or logico-semantic
torsion, that twists the operative sense of Rousseaus argument
against his avowed intent. In each case, the supplementary item
turns out to be not so much a mere supplement ( = add-on,
accessory, optional extra) but a supplement in the opposite,
palliative sense: that which is required in order to complete or
make good an otherwise defective, non-self-sufficient, or inadequate mode of being. Thus, it is strictly impossible a downright
contradiction in terms to posit the existence of a social state
of nature that would somehow precede and contrast with those
subsequent states whose hallmark, according to Rousseau, was
their basis in various, increasingly complex forms of societal and
cultural distinction. This attempt to describe what can never in
truth have existed a society unmarked by any of those structures (however primitive) that constitute the very conditions
of possibility for social life can in fact be seen to define the
245
Deconstruction
(philosophical) order of priorities by treating language as the
condition of possibility for thought, or in narrowly professional
terms linguistics as the discipline now poised to occupy the high
academic ground. Rather, what emerges from a critical reading
of Benvenistes texts is the absolute necessity that any such argument should turn out to controvert its own leading premise by
taking for granted a whole range of indispensable distinctions
that derive from a prior philosophic discourse, in this case one
that finds its first clear statement in Aristotles doctrine of the
categories.
There is a similar twist of argument, and again one that
is ignored by most commentators, in Derridas essay White
Mythology: Metaphor in the Text of Philosophy (1982, 20771).
On the usual account, what he here purports to show taking a
lead from Nietzsche is the saturation of philosophic discourse
by various (predominantly visual or tactile) types of metaphor
that cannot be expunged, as some philosophers would wish, or
even brought within the bounds of rational acceptability through
a systematic treatment or method of classification. That is, they
are so pervasive and go so far toward defining the very nature,
self-image, and operative scope of that discourse that it is strictly
impossible for philosophy either to manage without them or
come up with some rigorously theorized account that would
finally reduce them to order on its own methodological terms. To
suppose that philosophy has managed to resolve this problem
to achieve a clear demarcation between concept and metaphor
or literal and figural language is to take it for granted that the
sense aimed at through these figures is an essence rigorously
independent of that which transports it, which is an already philosophical thesis, one might even say philosophys unique thesis,
the thesis which constitutes the concept of metaphor (Derrida
1982, 229). However, this confidence may appear ill-placed if
one considers the extent to which philosophy depends upon a
range of metaphorical terms and distinctions like that between
metaphor (etymologically a means of transport or carrying
away) and concept (grasping, comprehending, holding together
in thought) which make up its very element.
What I have said so far about White Mythology fits in well
enough with the received view among mainstream analytic philosophers: that Derridas approach has more in common with
literary criticism than with philosophy properly so called, that
is, the practice of rigorous conceptual analysis. However, this is
a partial and highly prejudicial reading, as soon becomes clear
if one looks beyond the opening section where his approach
might plausibly be construed along these echt-Nietzschean
lines to later passages where Derrida goes out of his way to forestall or disqualify that interpretation. His counterargument (as
with the response to Benveniste) is that philosophy has provided
all the terms and categorical distinctions that must be seen as
absolutely prerequisite to any discussion of these issues, among
them most crucially the distinctions between concept and metaphor, reason and rhetoric, or philosophy and literature.
Hence, Derridas cardinal point: that this will require not only
the highest degree of conceptual precision but also a detailed
knowledge of their history and various stages of elaboration and
refinement to date.
White Mythology makes good this claim by examining a
great range of texts by philosophers, linguists, rhetoricians, and
246
historians of science and showing how they take for granted not
only the existence of certain prior philosophical concepts and
categories but also the necessity of bringing them to bear in the
process of deconstructing the kinds of uncritical or prejudicial
thinking that often go along with them. For if the concept of
metaphor is itself what Coleridge dubbed a philosopheme a
distinctively philosophic notion then we can have no means
of questioning the supposed priority of concept over metaphor
except by way of the discourse wherein that topic has received
its most decisive statements and elaborations. Thus, it makes no
sense to proclaim, with postphilosophical adepts like Rorty,
that we should give up the old deluded quest for truth, clear and
distinct ideas, conceptual precision, and so on and henceforth
embrace the Derridean ideal of philosophy as just another kind
of writing that at best offers new and adventurous modes of creative self-description (Rorty 1982). This involves not only a snippety reading of Derrida but also a failure to grasp his point that
such gestures can amount to no more than a kind of rhetorical
hand waving or a claim to have come out on the far side of philosophy, while in fact regressing to a prephilosophical stage of
unreflective immersion in language.
Such is indeed the charge that Habermas brings against
Derrida, namely, his having leveled or annulled the crucial genre
distinction between philosophy and literature or language in its
constative (i.e., truth-based or logical) and its performative (suasive and rhetorical) modes (Habermas 1987; see also performative and constative). It seems to me, on the contrary,
that Derridas most significant achievement will be seen to lie in
his contributions to philosophy of language and logic and, above
all, his remarkably inventive and original yet none the less rigorous rethinking of the relationship between these disciplines. (For
some early indications, see Norris and Roden 2002.)
Christopher Norris
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Aristotle. 1990. Categories and De Interpretatione. Trans. J. L. Ackrill.
Oxford: Clarendon.
Benveniste, Emile. 1971. Problems in General Linguistics. Trans. Mary
Meek. Coral Gables, FL: University of Miami Press.
Dasenbrock, Reed Way, ed. 1989. Re-Drawing the Lines: Analytic
Philosophy, Deconstruction, and Literary Theory. Minneapolis:
University of Minnesota Press.
Derrida, Jacques. 1973. Speech and Phenomena and Other Essays
on Husserls Theory of Signs. Trans. David B. Allison. Evanston,
IL: Northwestern University Press.
. 1976. Of Grammatology. Trans. Gayatri C. Spivak. Baltimore: Johns
Hopkins University Press.
. 1978. Writing and Difference. Trans. Alan Bass. London: Routledge
and Kegan Paul.
. 1982. Margins of Philosophy. Trans. Alan Bass. Chicago: University
of Chicago Press.
. 1982. Margins of Philosophy. Trans. Alan Bass. Chicago: University
of Chicago Press.
. 1989. Afterword: Toward an ethic of conversation. In Limited Inc,
ed. Gerald Graff, 11154. Evanston, IL: Northwestern University Press.
Habermas, Jrgen. 1987. On levelling the genre-distinction between
philosophy and literature. In The Philosophical Discourse of
Modernity: Twelve Lectures, trans. Frederick Lawrence, 185210.
Cambridge: Polity Press.
Definite Descriptions
Deixis
Norris, Christopher. 1989. Philosophy as not just a kind of writing: Derrida and the claim of reason. In Dasenbrock 1989, 189203.
. 1990. Deconstruction, postmodernism and philosophy: Habermas
on Derrida. In Whats Wrong with Postmodernism: Critical Theory and
the Ends of Philosophy, 4976. Hemel Hempstead: Harvester.
. 2000. Deconstruction and the Unfinished Project of Modernity.
London: Athlone.
. 2002. Derrida on Rousseau: Deconstruction as philosophy of
logic. In Norris and Roden 2002, II: 70124.
Norris, Christopher, and David Roden, eds. 2002. Jacques Derrida. 4 vols.
London: Sage.
Rorty, Richard. 1982. Philosophy as a kind of writing. In Consequences
of Pragmatism, 89109. Brighton: Harvester.
. 1989. Two versions of logocentrism: A reply to Norris. In
Dasenbrock 1989, 20416.
Rorty, Richard, ed. 1967. The Linguistic Turn. Chicago: University of
Chicago Press.
Searle, John R. 1977. Reiterating the differences: A reply to Derrida.
Glyph 1: 198208. Baltimore: Johns Hopkins University Press.
DEFINITE DESCRIPTIONS
1.
2a.
I
tepoti rua inanahi.
LOC.past Dunedin 3d yesterday
They were in Dunedin yesterday.
2b.
2c.
Hei
te ata
tua haere ai.
LOC.fut ART morning 1d go
PART
We will go in the morning.
DEIXIS
By deixis (from the Greek to point) we mean here all
cues provided by a language that localize a speech event and its
participants in space and time. By contrast, reference is based on
the privative distinction related to the deictic center (origo) /
not related to the deictic center. There are two reference systems to localize a referent. In a positional system, the speaker is,
or both speaker and hearer are, the deictic center whose position
is used to localize an entity. A dimensional system relies on the
speech participants orientation and understanding of the environment, in which case the deictic center may be something else
instead of the speech participants. The choice for one or both of
these systems evokes major differences among languages.
There are six categories of deixis: person, space, time, discourse, emphathy, and social status (Levinson 1983). Person
deixis usually distinguishes the speaker (first person) from
the hearer (second person) and the non-speech or narrated
participant (third person). This can be encoded by inflection
In Algonquian languages (North America), third person pronouns signal whether their referent is more or less topical
in the narration (for example, proximal bi versus obviative
yi in Navajo). Similarly, a languages deictic system may signal the speakers empathy toward the referent and its status
within society. In Javanese (Indonesia), for example, social
247
Deixis
Table 1. Pronoun systems in four languages
English
Quechua (Peru)
Tamil (India)
Biak (Indonesia)
nuqa
nn
ai
2d person singular
you
qan
au
3d person singular
he
she
it
pay
avan (masculine)
ava (feminine)
atu (neuter)
we
nuqanchis
nm
u (dualis)
o (trialis)
we
nuqayku
nka
nu (dualis)
mo (plural)
2d person plural
you
qankuna
nka
mu (dualis)
mo (plural)
3d person plural
they
paykuna
avar
su (dualis)
so (trialis)
si (animate plural)
na (inanimate plural)
Proximal
Bounded
Unbounded
Distance-neutral
Medial
Distal
Bounded
Unbounded
Visibility
Number
Visible
Singular
Ito ~ ity
Itsy
Iry
Io
Iny
Plural
Ireto
Iretsy
Irery
Ireo
ireny
Singular
Izato
Izaty
Izary
Izao
izany
Plural
Izatero
Invisible
248
249
Recent Developments
Chomsky (1981) introduced the principles and parameters
(P&P) model. In this model, the language faculty is considered as
a set of genetically determined principles that are operative in all
languages. In order to account for the differences between languages, principles are assumed to have parameters. A parameter
is a variable in the principle with a predetermined set of values.
Language acquisition is then analyzed as finding the right values
for the parameters. The P&P model solves the tension between
the demands of descriptive and explanatory adequacy because it
provides a basis for explaining language acquisition while describing the range of attested languages. It does not provide an obvious
constraint on the proliferation of parameters, however. This means
that the indeterminacy problem arises again, but at a higher level.
250
The higher-level indeterminacy problem raised by the proliferation of parameters cannot be solved within the framework of
Figure 1. Chomskys (1995) minimalist program (MP) (see
minimalism) addresses this problem by considering the questions of the unification of linguistics with biology and the emergence of the language faculty in evolution (see biolinguistics).
The latter aspect is elaborated by Marc D. Hauser, Chomsky,
and W. Tecumseh Fitch (2002). In the analysis proposed by ten
Hacken (2007), the MP adds a new pair of a real-world entity and
a theoretical entity on top of the three represented in Figure 1.
These entities determine and explain evolution, respectively.
They are not part of linguistics proper but operate more generally. It is not necessary to specify their exact nature in order to
use them as constraints on the way the language faculty can have
emerged and should be shaped.
If the levels of adequacy are considered only in the way they
are defined by Chomsky (1964), they have lost their relevance in
the MP, as stated by Chomsky (2002, 12933). If they are considered as reflecting a wider-ranging concern in empirical science,
they have to be reformulated in a more general way. Ten Hacken
(2006) considers two alternative ways of adapting the levels of
adequacy to the expanded framework arising from the MP. One
is to add a new level of adequacy, directly connected with the
new pair of entities added on top of Figure 1. The other approach
is to relativize descriptive and explanatory adequacy with respect
to the level of the entities to which they are applied.
Pius ten Hacken
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Chao, Yuen-Ren. [1934] 1957. The Non-uniqueness of phonemic solutions of phonetic systems. Bulletin of the Institute of History and
Philology 4: 36397. Repr. in Readings in Linguistics: The Development
of Descriptive Linguistics in America 19251956, ed. Martin Joos, 3854.
Chicago: University of Chicago Press.
Chomsky, Noam. 1964. Current Issues in Linguistic Theory. Den
Haag: Mouton. This book is the original source for the terms descriptive, observational, and explanatory adequacy.
. 1981. Lectures on Government and Binding. Dordrecht: Foris.
. 1995. The Minimalist Program. Cambridge, MA: MIT Press.
. 2002. On Nature and Language. Cambridge: Cambridge University
Press.
Gazdar, Gerald, Ewan Klein, Geoffrey Pullum, and Ivan Sag. 1985.
Generalized Phrase Structure Grammar. Cambridge: Harvard
University Press.
Harris, Zellig S. [1951] 1960. Methods in Structural Linguistics.
Chicago: University of Chicago Press. Repr. as Structural Linguistics, 1960.
Hauser, Marc D., Noam Chomsky, and W. Tecumseh Fitch. 2002. The
faculty of language: What is it, who has it, and how did it evolve?
Science 298: 156979.
Hockett, Charles F. 1954. Two models of grammatical description.
Word 10: 21031.
Householder, Fred W. 1965. On some recent claims in phonological
theory. Journal of Linguistics 1: 1334.
Quine, W. V. 1972. Methodological reflections on current linguistic
theory. In Semantics of Natural Language, ed. Donald Davidson and
Gilbert Harman, 44254. Dordrecht: Reidel. This article offers an alternative conception of adequacy for grammars.
ten Hacken, Pius. 2006. The nature, use and origin of explanatory adequacy. In Optimality Theory and Minimalism: A Possible Convergence?
ed. Hans Broekhuis and Ralf Vogel, 932. Linguistics in Potsdam 25.
251
252
253
Dialect
Patajali. 1971. Yogasutra. Ed. and trans. J. R. Ballantyne and Govind
Shastri Deva. Varanasi, India: Indological Book House.
Webster, John. Duchess of Malfi. [1623] 1961. In Elizabethan Drama, ed.
Leonard Dean, 271360. 2d ed. Englewood Cliffs, NJ: Prentice-Hall.
DIALECT
In pursuit of a useful definition of dialect, one may begin by
distinguishing language from dialect. Think of a language as a
set. A dialect is a set member of that set. Just as a set does not
exist in the absence of its set members, so a language does not
exist in the absence of its dialects. A dialect is often identified
as a variety of a language spoken by a group of people. Thus, a
dialect involves the shared linguistic behavior and knowledge
of a group, not of the idiosyncratic individual. This variety may
be understood, to varying degrees, by people who speak other
contemporary varieties of the same language. This criterion of
mutual intelligibility is problematic as different languages can
also show degrees of mutual intelligibility. (Consider Swedish
versus Norwegian.) Consequently, to speak of a dialect presupposes a prior identification of the language to which the variety
belongs. Within this language, each variety will differ from others
in terms of pronunciation, vocabulary, and/or grammar. More
precisely, each variety may differ in its phonetic implementation
rules, phonology, lexical items or semantics, morphology,
syntax, and the pragmatic functions of discourse markers or
syntactic structures. Although differences of speech-act realizations, narrative types and topics, or conversational routines
can also exist, less research has explored these issues.
Generally, differences between varieties of the same language
may be quantitative or qualitative in nature. What is a quantitative difference? Consider words like the or that. Most speakers of
English will sometimes pronounce the initial consonant as a fricative [] and sometimes as a stop [d]. This alternation between
[] and [d] can be quantified. In turn, the frequencies with which
individuals say [] versus [d] may correlate to categories of identity, such as age, gender, class, or ethnicity. Alternating forms,
like [] ~ [d], are called variants of a sociolinguistic variable. This
variable is represented as (dh) in the literature. Sociolinguistic
variables like (dh) show quantitative differences across social
classes (upper favors [] more than lower), genders (women
favor [] more than men), and age (working adults favor [] more
than young children). Such frequency differences have also been
termed group-preferential differences.
What is a qualitative difference? Consider the Spanish la casa
(house) and la caza (hunt). In Puerto Rico, they are both pronounced [kasa]. In Spain, casa is [kasa] but caza is [kaa]. Thus,
in one dialect we find a feature that is absent from another. This
is a qualitative difference. Some researchers identify this as a
group-exclusive difference.
Qualitative and quantitative dialect differences may map onto
groups defined by geographical regions or by social characteristics. Thus, researchers speak of regional or social dialects. This
differs from the common belief that dialects are only regional.
Because differences among the genders or age groups occur, we
may ask if varieties associated with gender or age are dialects as
well. In practice, researchers reserve social dialects for class and
ethnic varieties, even though gender and age differences in a
254
by continual mainstreaming of its message, but also by continual and only partially successful marginalization of its residual
polyphony and heteroglossia.
255
Diffusion
256
DIFFUSION
The study of linguistic diffusion is concerned with describing and
explaining how languages or language features spread over time
and space. On a macrolevel, diffusion refers to the dispersion of
languages from a common point of origin. Through migration
and subsequent isolation, thousands of languages have developed from a highly limited set of protolanguages. The physical,
demographic, and social constraints on language dispersion cannot be reduced to a simple algorithm, and the reconstruction of
protolanguages, language family relationships, and patterns
of spatial dispersion remain a primary challenge in historical
linguistics.
On a language-specific level, diffusion is concerned with the
spread of particular linguistic innovations across the varieties of
a language or, in some cases, across languages. Linguists, particularly sociolinguists, dialectologists, and historical linguists,
seek to identify the mechanisms of transmission and the factors
that promote or inhibit the spread of language traits. A linguistic
change is initiated in a particular locale at a given point in time
and spreads outward from that point in progressive stages so that
earlier changes reach the outlying areas later. The wave model
assumes that a change spreads in concentric layers, as waves
radiate outward from a central point of contact when a pebble is
dropped into a pool of water. Forms that follow this straightforward time and distance relation follow the pattern of contagious
diffusion (Bailey et al. 1993).
Because of physical, social, and psychological factors, a model
that considers only time and distance is too simplistic to account
for the spread of linguistic forms. Diffusion researchers cite at
Diffusion
least five factors that influence the dispersion of customs, ideas,
and practices: 1) the phenomenon itself, 2) communication networks, 3) distance, 4) time, and 5) social structure. Although linguistic structures are inherently quite different from phenomena
such as technological innovations, they are subject to many of
the same social and physical factors that influence the nature of
diffusion in general.
A gravity model or hierarchical model of language (Trudgill
1974) often provides a better profile of the diffusion of linguistic
forms than a simple wave model. In the gravity model, which is
borrowed from the physical sciences, diffusion is a function not
only of the distance from one point to another, as with the wave
model, but of the population density of areas that stand to be
affected by a nearby change. The interplay between the population density of two areas and the distance that separates them
parallels the effects of density and distance on gravitational pull.
Changes are most likely to begin in heavily populated cities that
serve as cultural hearths. From there they radiate outward, but
not in a simple wavelike pattern; rather, innovations first reach
moderate-size cities that fall under the area of influence of some
large, focal city, leaving nearby, sparsely populated areas unaffected. Gradually, innovations filter down from more populous,
denser areas to less densely populated areas, affecting rural
areas last, even if such areas are quite close to the original focal
area of the change. The spread of change is thus like skipping a
stone across a pond, rather than dropping a stone into a pond, as
in the wave model. The model of change following this pattern is
referred to as cascade diffusion.
One of the noteworthy examples of cascade diffusion is a
vowel shift currently taking place in the northern cities of the
United States. Part of this elaborate rotation involves the shift
of the vowel of thought so that it sounds more like the vowel of
lot. The lot vowel, in turn, sounds more like the vowel of trap,
which moves closer to the pronunciation of the vowel of dress.
This vowel shift proceeds from larger cities in the North to successively smaller ones, leaving in-between rural areas relatively
unaffected until the later stages of the change.
Gravity models of change include factors of distance and
communication networks as a function of population density,
but they do not recognize the role of other social and psychological factors. For example, changes do not spread evenly across
all segments of the population. Members of upwardly mobile
social classes usually adopt linguistic innovations more quickly
than do members of other classes, and women and younger
people are often leaders in certain kinds of language change. It
is therefore essential to track changes not only across geographical space and population density but also across different age,
ethnic, gender, and social status groups (cf. age groups and
gender and language).
In terms of social networks, the first people to adopt changes
are those with loose ties to many social groups but strong ties to
none due to the face that strong ties inhibit the spread of change.
In order for the changes to make their way into more close-knit
groups, they need to be picked up by people who are central figures in these groups but who are willing to adopt change nonetheless, perhaps for reasons of prestige. Because these early
adopters are well regarded in their social groups, the changes
they adopt are likely to be picked up by other members of these
Digital Media
groups, thereby diffusing throughout the group and to other
groups or communities within a population (Labov 2001).
One important study of language diffusion in the southern
United States (Bailey et al. 1993) shows that although many
linguistic innovations follow the more common hierarchical
pattern of cascade diffusion, some features may display the
opposite diffusion pattern. For example, the use of the special
intentional modal fixin to, as in Theyre fixin to go now, once
heavily concentrated in rural areas of the American South, has
now been adopted in some larger, urban population centers.
The explanation for this contrahierarchical diffusion pattern is
tied to its symbolic marking of traditional southern American
speech. In the face of a large influx of outsiders into the region,
native urban residents may seek to assert their southern identity by adopting selected structures strongly associated with the
regional South, showing that the social meaning attached to
linguistic forms has to be considered along with geographical,
demographic, and interactional factors in explaining linguistic
diffusion.
Walt Wolfram
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bailey, Guy, Tom Wikle, Jan Tillery, and Lori Sand. 1993. Some patterns
of linguistic diffusion. Language Variation and Change 5: 35990.
Labov, William. 2001. Principles of Linguistic Change. Vol. 2. Social
Factors. Malden, MA, and Oxford: Blackwell.
Rogers, Everett M. 1995. Diffusion of Innovations. 5th ed. New York: Free
Press.
Trudgill, Peter. 1974. Linguistic change and diffusion: Description and
explanation in sociolinguistic dialect geography. Language in Society
3: 21546.
DIGITAL MEDIA
Digital media are those media whose means of production and
distribution are digitized via computers; the term is commonly
used in contrast to older forms of media, such as print (for text)
or analog devices (for sound and images). In language studies,
digital media most commonly refer to the Internet, the World
Wide Web, mobile telephony, and other networked and wireless
technologies that support human communication known as
computer-mediated communication (CMC) and the transmission of information. Digital media can also refer to digital storage
devices for data, sound, video, and graphics. Here, we are concerned primarily with the former sense, especially the impact of
digital communication technologies on peoples individual and
collective use of and relation to language.
History
Communication via digital media can be traced to the invention of packet switching technology in the 1960s, which enabled
messages to be exchanged among networked computers. The
ARPANET, the predecessor of the Internet, was implemented as
a United States defense department project in 1969; by the mid1970s, it had become popular for human communication via
e-mail and mailing lists. In 1979, the USENET was created as an
alternative, grassroots network; USENET newsgroups, along with
various BBS (bulletin board systems) and networks hosted on
257
Digital Media
private servers during the 1980s, were eventually integrated into
the Internet, the term used after 1983 for the collection of networks that had grown around the ARPANET. By the late 1980s,
the Internet offered public real-time chat via Internet Relay Chat
and MUDs (Multi-User Dimensions), along with email, mailing
lists, and newsgroups. Around the same time, Internet service
providers (ISPs) were starting to make the Internet accessible
to people in their homes, rather than just from businesses and
universities.
The introduction of the World Wide Web in 1991 and the first
graphical browser in 1993 transformed the Internet by enabling
networked multimedia. By the mid-1990s, Internet telephony
and videoconferencing were available, along with graphical virtual worlds. Despite the increasing availability of bandwidth to
support multimedia, however, text retained its popularity. The
late 1990s saw the emergence of several text-based applications: instant messaging, weblogs (blogs), and text messaging on
mobile phones (especially in Europe and Asia).
A more recent trend has been toward mobile media and flexible access. Starting with external hard drives for external data
storage and continuing with laptops, personal digital assistants
(PDAs), iPods, and smartphones, digital media have moved
away from desktop computing toward more distributed, lightweight, faster devices.
Language-Related Issues
The rapid rise in popularity of digitally mediated communication
over the past two decades has attracted considerable interest
from language scholars. The central debates have focused on
how to classify such communication relative to speech and writing, the effects of technology on language and language use, the
purported anonymity of text-based CMC and its social and linguistic consequences, and the long-term effects of digital media
on individual languages and the global language ecology.
Computer-mediated communication is sometimes claimed
to constitute a third modality of language, alongside speech and
writing. Text-based CMC, by far the most common manifestation
of digital communication, blends the production and reception
features of writing (typing on a keyboard or otherwise entering
characters into an alphanumeric interface; reading messages on
a screen) with the structural and interactional features of spoken
conversation (e.g., informality, phatic content, relatively rapid
exchange of messages), making it a hybrid modality with distinctive characteristics (Crystal [2001] 2006). Moreover, the personal
accessibility and wide public reach of the Internet have led some
to characterize it as fundamentally transformative of human
communication, a revolution as profound as that triggered by
the printing press.
At the same time, the novelty of digital language should not
be overstated. It is often possible to trace the roots of so-called
emergent or digitally native CMC genres (Crowston and Williams
2001) to older written and oral genres. An example is the blog,
which, while arguably a historically unprecedented hybrid of
personal, interpersonal, and mass communication, manifests
continuities with handwritten diaries, phone calls to friends and
family, project logs, and letters to the editor. Ultimately, what may
be most unique about digital media is their tendency to support
a convergence of language features, genres of communication,
258
and communication technologies that were previously considered distinct. The incorporation of text chat into multiplayer
online games and the ability to send text messages from mobile
phones to interactive television (iTV) programs illustrate the latter trend.
Theoretical debate has also centered around the effects of
digital technology on human communication. A strong technological determinism position holds that production and reception constraints on CMC inevitably shape digitally mediated
language and language use. Such a position finds support in
research findings that technical constraints on message exchange
disrupt and reshape turn-taking patterns across a range of digital
genres (Herring 1999). A weaker version of technological determinism holds that features of specific technologies predispose
users to communicate in certain ways, but that users may override those predispositions. For example, the synchronicity of
CMC systems tends to affect message length, complexity, and
formality (with messages in asynchronous modes being generally longer, more syntactically complex, and more formal than
in synchronous modes), although both formal and informal language can be found, for example, in email (asynchronous) and
chat (synchronous), depending on the topic and purpose of the
communication.
The social construction of technology theory goes further to
assert that users shape technologies through their use as much or
more than their use is shaped by those technologies (Bijker and
Law 1992). This view receives support from computer-mediated
cooperative work and online education, where the nature of
the tasks structures communication in often predictable ways.
Further, many face-to-face social and interactional dynamics,
including gendered patterns of communication, are reproduced
in digital discourse, albeit differently in academic discussion
forums than in chat. In an effort to account for such variation,
a fourth position holds that there is no single way in which technology influences mediated language; rather, it depends on the
particular constellation of technical and social variables that
characterizes a given sample of mediated discourse (Herring
2007). A desideratum for future research is a coherent theory
that can predict when specific types of media will have particular
communicative effects.
Another nexus of debate concerns the purported anonymity of digitally mediated communication. Because social cues
conveyed through prosody, facial expression, and physical
appearance of message senders are filtered out in text-based
CMC, many early scholars believed that digitally mediated communication was depersonalized and that users identities were
masked or irrelevant. This was thought to give rise to flaming or
hostile language (and antisocial behavior, in general); play with
identity and liberatory (or inauthentic, depending on ones perspective) online self-presentations; and compensatory linguistic
strategies, such as creative spellings and emoticons (faces made
out of ascii characters), in order to enhance ones social presence and signal ones intentions. These linguistic strategies have
been referred to as textspeak by David Crystal ([2001] 2006; for
examples, see Figure 1).
Alternative perspectives have also been advanced on these
phenomena, however. True anonymity is infrequent, since most
people who communicate digitally use consistent identifiers, and
Digital Media
HK English: Hee hee . . . dunno why I always like to send u mails ar! Part is
becoz I wanna keep contact with u la!
French:
Ca sera donc tjs 1 plaisir 2te revoir! :-) [So it will always be a pleasure to
259
Diglossia
accepted into mainstream disciplinary approaches. In language
studies, new media currently provide application domains (e.g.,
for language learning) and sources of data for empirical analysis
and, increasingly, for theorizing about language from cognitive,
social, and evolutionary perspectives.
Susan C. Herring
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bijker, Wiebe, and John Law, eds. 1992. Shaping Technology/Building
Society: Studies in Sociotechnical Change. Cambridge, MA: MIT Press.
Crowston, Kevin, and Marie Williams. 2001. Reproduced and emergent
genres of communication on the World-Wide Web. The Information
Society 16.3: 20116.
Crystal, David. [2001] 2006. Language and the Internet.
Cambridge: Cambridge University Press.
Danet, Brenda, and Susan C. Herring, eds. 2007. The Multilingual
Internet: Language, Culture, and Communication Online. New
York: Oxford University Press.
Herring, Susan C. 1999. Interactional coherence in CMC. Journal of
Computer-Mediated Communication 4.4. Available online at: http:
//jcmc.indiana.edu/vol4/issue4/herring.html.
. 2004. Slouching toward the ordinary: Current trends in computermediated communication. New Media & Society 6.1: 2636.
. 2007. A faceted classification scheme for computer-mediated
discourse. Language@Internet. Available online at: http://www.
languageatinternet.de/articles/761.
Thurlow, Crispin. 2006. From statistical panic to moral panic: The
metadiscursive construction and popular exaggeration of new
media language in the print media. Journal of Computer-Mediated
Communication 11.3: article 1. Available online at: http://jcmc.indiana.edu/vol11/issue3/thurlow.html.
DIGLOSSIA
In his seminal article, Charles Ferguson (1959, 435) defined
diglossia as
a relatively stable language situation in which, in addition to the
primary dialects of the language (which may include a standard
or regional standards), there is a very divergent, highly codified
(often grammatically more complex) superposed variety, the
vehicle of a large and respected body of written literature, either
of an earlier period or in another speech community, which is
learned largely by formal education and is used for most written
and formal spoken purposes but is not used by any section of the
community for ordinary conversation.
260
261
262
In applying the approaches to the language of this short discourse, we focus on some aspects of text (the information in,
structure of, and relationships between successive clauses) and
context (social identities, relationships, and institutions). We
progress from a focus on knowledge (how to communicate information and take social action) to the use of the linguistic code
to convey a variety of situated meanings by people in particular roles that are embedded within (and sustain) larger cultural
practices and social structures.
pragmatics analyzes how we communicate more than
the semantic content of language by depending upon our ability to draw inferences based on such general principles as the
cooperative principle (appropriate quantity, quality,
263
264
continuation, then, Mr. Clarks statement could have been followed by others actions. Indeed, Lets get started and the utterance that follows are parts of a particular pair of sequentially
related actions: They sequentially implicate another action, the
students response to the summons (see adjacency pair).
Mr. Clarks ability to maintain his turn and develop the
sequence of classroom activities is one way that his role in the
situation is established and reinforced. Interactional sociolinguistics reveals how numerous features of language provide
clues to (or indices of) the social situation, activities, participant
identities, and relationships that may actually have a role in creating the context of interaction. The use of lets, for example, suggests that the speaker has authority over the hearer, thus evoking
a situation in which participants have an asymmetrical power
relationship (e.g., doctor/patient, parent/child). We can narrow
down the nature of the authority and the situation by noting that
the activity being started by Mr. Clark (review), and the object of
the review (problems), indicates a learning environment or one
in which an expert is instructing a novice. That the problems
were from last night reveals a cyclical pattern, a structured routine often found in formal institutions.
Language choices inferences about meaning, actions, roles,
relationships, and participation are all embedded in broader
cultural matrices of recurrent practices, knowledge, and meanings, which include beliefs about who should do what and how
they should do so, as well as the evaluations based on larger
values and ideologies (see ideology and language) of
particular outcomes of what is said and done. Ethnography of
communication elucidates these connections. For example,
the collaborative review of the problems portrays a cultural
belief system in which learning and attaining information arises
when novices work on their own (problems from last night)
and then, at a given point in time, present and review their solutions with an expert.
Just as language, inferences, actions, roles, relationships,
and participation are all embedded in culture, so too are they
intertwined with social processes and structures that sustain (or
restrict) power and privilege. critical discourse analysis
(see also discourse analysis [foucaultian]) explores how
ways of speaking can put those processes into place and reinforce
(or challenge) received means of authority. Mr. Clarks ability to
manage the use of time, select the activity in which to engage,
and organize the way in which information becomes distributed
as knowledge is consistent with a school setting in which his role
is not challenged. The power created by the institutional setting
links his discourse to broader social, cultural, and civic agendas.
What Mr. Clark says and more fundamentally, his ability to do
so thus positions him as one who can reinforce social structural
norms (who teaches whom? how? when?) and as an arbiter of the
official set of values, beliefs, and ideologies that are sanctioned
means of maintaining a stock of received knowledge. And the
fact that he is speaking on a special day that happens only once
a year when parents are permitted to visit en masse to observe
firsthand their childrens education highlights the public and
civic function of his role.
OPENING AN ORAL HISTORY INTERVIEW. The discourse analyzed
in (2) is from the beginning of an oral history interview. After
Conclusion
Discourse analysis provides a range of methodologies that are
applicable to different facets of language in text and context.
Although we have been able to consider only some components
of discourse analysis, our discussion and sample analyses help
us extract several general principles (Schiffrin 1994):
(1) Analysis of discourse is empirical: Data are based on people using language, not linguists thinking about how people
use language.
(2) Analyses are accountable to the data: They have to explain
the data in both sequential and distributional terms.
(4) Discourse is not just a sequence of linguistic units: Its coherence (see coherence, discourse) cannot be understood if
attention is limited just to linguistic form and meaning.
(5) Resources for coherence jointly contribute to participant
achievement and understanding of what is said, meant, and
265
Discrete Infinity
done. In other words, linguistic forms and meanings work
together with social and cultural meanings, and interpretive
frameworks, to create discourse.
(6) The structures, meanings, and actions of everyday spoken discourse are interactively achieved.
(7) What is said, meant, and done is sequentially situated;
that is, utterances are produced and interpreted in the local
contexts of other utterances.
(8) How something is said, meant, and done speakers
selection among different linguistic devices as alternative
ways of speaking is guided by relationships among the
following:
(a) speaker intentions;
(b) conventionalized strategies for making intentions
recognizable;
(c) the meanings and functions of linguistic forms in relation to the text and context in which they appear;
(d) the sequential context of other utterances;
(e) properties of the textual mode, for example, narrative,
description, exposition;
(f) the social context, for example, participant identities
and relationships, structure of the situation, the setting;
(g) a cultural framework of beliefs and actions.
When brought together, this set of heuristic tools leads to the
following principle: Our uses of language, and the functions
that it accomplishes, are interactively constructed by people
using language together (e.g., taking turns at talk, drawing inferences about communicative intentionality) and drawing
upon properties of language and its ability to join smaller units
(clauses, sentences) into larger units (texts) that both reflect and
create the social contexts in which they emerge.
Deborah Schiffrin
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Halliday, Michael, and R. Hasan. 1976. Cohesion in English.
London: Longman.
Harris, Zellig. 1952. Discourse analysis. Language 28: 130.
Jaworski, Adam, and N. Coupland, eds. 1999. The Discourse Reader.
London: Routledge.
Johnstone, Barbara. 2006. Discourse Analysis. Oxford: Blackwell.
Labov, William. 1972. The transformation of experience in narrative
syntax. In Language in the Inner City. Philadelphia: University of
Pennsylvania Press, 35496.
Labov, William, and J. Waletzky. 1967. Narrative analysis. In Essays
on the Verbal and Visual Arts, ed. June Helm. Seattle: University of
Washington Press, 1244.
Schiffrin, Deborah. 1994. Approaches to Discourse. Oxford; Blackwell.
Schiffrin, Deborah, D. Tannen, and H. Hamilton, eds. 2001. Handbook of
Discourse Analysis. Oxford: Blackwell.
DISCRETE INFINITY
This locution was brought into linguistics by Noam Chomsky
(for instance, Chomsky 1988, 170) to characterize the fact that
human languages are built up from discrete units (morphemes
266
267
268
269
Dyslexia
270
DYSLEXIA
Introduction: What Is Dyslexia?
Dyslexia is a specific learning difficulty affecting literacy
development. Children and adults with developmental dyslexia
show difficulties in reading and spelling that are not explicable in terms of their age, intelligence, or educational experience.
Children with dyslexia typically have marked difficulties in learning to read and spell words, though their understanding of what
they read may be good. These difficulties are often accompanied
by difficulties in short-term memory and organization. In adulthood, the word-reading difficulties may resolve, but spelling and
other underlying difficulties remain.
Theories of Dyslexia
Current theories of dyslexia are cast at either the biological or cognitive levels of explanation. The predominant cognitive account
of dyslexia views the primary cause as a phonological processing impairment (Vellutino et al. 2004). According to this hypothesis, children with dyslexia have phonological deficits that cause
Dyslexia
a wide range of symptoms, not all of which are directly related
causally to the reading deficits (e.g., verbal short-term memory
problems and word-finding difficulties). As far as is known, such
symptoms are equally common among children learning to read
in all languages.
Many other theories of dyslexia accept phonological difficulties as a proximal cause of reading problems but cite more lowlevel deficits as their distal cause. For example, the automization
deficit hypothesis (Nicolson and Fawcett 1990) proposes that
difficulties in the cerebellum in dyslexic children place similar constraints on learning of all skills, including phonology,
naming abilities and basic motor skills. The proposal of William
Lovegrove, Frances H. Martin, and Walter L. Slaghuis (1986) that
people with dyslexia have impairments of the magnocellular
system (the division of the visual system that responds to rapid
changes) has also generated much research. Findings are mixed,
with some studies reporting no evidence of abnormal sensitivity and others suggesting that group differences between people
with dyslexia and normal readers may be related to uncontrolled
differences in IQ. Research investigating visual attention problems in dyslexia is also inconclusive.
An influential hypothesis is that dyslexia stems from a deficit in basic auditory processing. Specifically, a rapid auditory processing deficit found with both speech and nonspeech
sounds would affect the perception of consonants distinguished
by rapid changes in the speech signal, and further, poor speech
perception would affect the development of phonological
processing skills (Tallal 2004). Investigation of auditory deficits
in dyslexia has extended to such tasks as frequency discrimination, frequency modulation, binaural processing, and backward
masking. However, as with findings on visual impairments, the
literature is replete with conflicting results, and an alternative
suggestion is that the deficit is not a general auditory impairment
but is specific to the processing of speech sounds. Investigations
of speech perception in dyslexia have highlighted subtle impairments, although again there are conflicting results. The lack of
consensus in the field regarding sensory impairments has led to
the proposal that they frequently occur in dyslexia but are not
causally linked to it (Ramus 2004). Further investigation of this
complex issue is needed.
Etiology of Dyslexia
GENETIC FACTORS. It has long been known that dyslexia runs in
families; however, because families share genes as well as environments, it is important to attempt to disentangle genetic and
environmental influences. Twin studies have been helpful in this
regard (Pennington and Olson 2005). Most twin studies of reading and reading disability report that both reading and phonological awareness are heritable skills, and thus it can be inferred
that dyslexia has a genetic basis. Furthermore, molecular genetic
studies have found gene markers of dyslexia as well as some candidate genes, though it is far from clear what the genetic mechanisms are (Fisher and Francks 2006).
It is important to note that the genes implicated in dyslexia
indicate a susceptibility to reading difficulties but not that reading problems are fully genetically determined. The interaction
of different skills in determining reading outcomes can be seen
in studies of children at family risk of dyslexia followed from
Comorbidity
Dyslexia shows some similarities with specific language
impairment, and there is some debate as to whether they
should be characterized as the same disorder (Bishop and
Snowling 2004). There is also evidence of comorbidity between
dyslexia and various emotional and behavioral problems. Most
strikingly, dyslexia is highly comorbid with attention-deficit
hyperactivity disorder (ADHD) and, in particular, attention difficulties (Willcutt and Pennington 2000). Recent research suggests shared genetic risk factors as a possible cause. Children
with dyslexia also show an increased risk of developing clinically
271
Dyslexia
Ellipsis
Reading Intervention
Theoretical knowledge of the relationship between phonological
skills and learning to read has led to the development of effective
reading intervention programs that promote phonological skills
in the context of reading (National Reading Panel, 2000) (see
teaching reading). Such interventions are effective both for
diagnosed dyslexics and for children who are at risk of reading
problems. An underresearched issue is the problem of children
who, despite high quality intervention, do not respond to teaching and continue to have reading impairments. These children
are often socially disadvantaged and may show additional emotional and behavioral difficulties.
Conclusions
Dyslexia is a highly researched developmental disorder. There
is now clear evidence that difficulties in phonological skills are
a major proximal cause of reading difficulties across languages.
There is also evidence that reading is a complex skill influenced
both by genetics and by the environment. However, outstanding issues remain. Notably, models of the disorder are moving
toward a multiple deficit model, and it is unknown which is the
most appropriate support for children who do not respond to
standard phonics-based reading intervention.
Margaret J. Snowling and Julia M. Carroll
272
E
ELLIPSIS
Ellipsis is the nonexpression of some lexical material specifically, a word or words forming a syntactic constituent that
is needed for the full interpretation of a sentence but is not
expressed because it can be recovered from the linguistic or realworld context. Under a traditional syntactic definition of ellipsis, elliptical gaps must be able to be filled with overt material,
thus distinguishing them from other types of gaps, like traces
of moved constituents. All natural languages permit ellipsis,
but they differ with respect to which constituents can be elided
in which configurations. Ellipsis falls within the larger field of
reference resolution.
Most studies of ellipsis concentrate on formalizing the licensing and recoverability conditions for elided constituents. The
former must account for what makes ellipsis grammatical in
given configurations whereas the latter concerns the ways in
which the meaning of the elided material can be understood
from the context. When the meaning of an elided constituent is
understood by coreference with a previously introduced linguistic constituent, that constituent is called the antecedent.
A cross-linguistic sampling of the many types of constituents
that are subject to ellipsis includes arguments of a verb (1); head
nouns in noun phrases with an overt quantifier, modifier, and so
on (in [2], laps); main verbs in so-called gapping constructions
Embodiment
(in [2], swam); verb phrases selected by an overt auxiliary (3);
and main verbs in sentences containing two or more overt arguments or adjuncts (4). The elided categories in the examples are
indicated by []. Textual antecedents, when present, are shown
in boldface.
1.
2.
3.
4.
EMBODIMENT
Embodiment refers to the ways in which persons bodies and
bodily interactions with the world shape their minds, actions,
and personal, cultural identities. Embodied accounts of mind
and language embrace the idea that human symbols are
grounded in recurring patterns of bodily experience, and therefore reject traditional dualistic, disembodied views of human
cognition and linguistic meaning. The study of embodiment
demands recognition that thought and language arise from the
continuous dynamic interactions among brains, bodies, and
the world. There are, in fact, three levels of embodiment that
together shape the embodied mind (Lakoff and Johnson 1999).
Neural embodiment concerns the structures that characterize
concepts and cognitive operations at the neurophysiological
level. The cognitive unconscious consists of the rapid, evolutionarily given mental operations that structure and make possible conscious experience, including the understanding and
use of language. The phenomenological level is conscious and
273
Emergentism
understood as things that can be grasped. Studies also show
that people understand metaphorical fictive motion sentences,
such as The road runs along the coast, in terms of implicit,
imaginary sensations of movement implicit in these sentences
(Matlock 2004). People are not aware of these simulations, and
so language processing is not dependent on deliberate thought
about motion. In general, psycholinguistic studies provide additional support for the broad claim, also now made in computational modeling research, known as simulation semantics
(Feldman and Narayanan 2004) that language use is closely tied
to embodied imagination.
The empirical work in cognitive science on embodiment in
language and thought (see Gibbs 2006) mirrors other debates in
philosophy and literary studies on the role of embodied imagination in literary and aesthetic experience. Readers emotional
involvement with fiction, for instance, may arise from their simulations of themselves as the characters they read about and their
fictional actions (Nichols 2006). In this manner, reading may not
be an abstract, purely mental process with little engagement of
the bodily imagination but is fundamentally tied to our powers
to recreate what it must be like to be and move like the people
we are reading about. Debate about this, and about other issues
related to embodiment in thinking and language, is central in
much contemporary scholarship in the humanities and cognitive sciences.
Raymond W. Gibbs, Jr.
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Feldman, J., and S. Narayanan. 2004. Embodied meaning in a neural
theory of language. Brain and Language 89: 38592.
Gibbs, R. 2006. Embodiment and Cognitive Science. New York: Cambridge
University Press.
Hauk, O., I. Johnsrude, and F. Pulvermuller. 2004. Somatotopic representation of action words in human motor and premotor cortex.
Neuron 41: 3017.
Klatzky, R. L., J. W. Pellegrino, B. P. McCloskey, and S. Doherty. 1989. Can
you squeeze a tomato? The role of motor representations in semantic
sensibility judgments. Journal of Memory and Language, 28: 5677.
Lakoff, G., and M. Johnson. 1999. Philosophy in the Flesh: The Embodied
Mind and Its Challenge to Western Thought. New York: Basic Books.
Matlock, T. 2004. Fictive motion as simulation. Memory & Cognition
32: 13891400.
Nichols, S., ed. 2006. The Architecture of The Imagination: New Essays on
Pretense, Possibility, And Fiction. Oxford: Oxford University Press.
Wilson, N., and R. Gibbs. 2007. Real and imagined body movement
primes metaphor comprehension. Cognitive Science, 31: 72131.
EMERGENTISM
A significant body of linguistic research can be situated in the
philosophical and scientific tradition known as emergentism.
This entry offers a brief overview of this work, with a focus on its
guiding principles and on the proposals it makes concerning the
nature of human language.
274
properties that amount to more than the sum of its parts. The
physical world offers many examples of this, as Mill observes
(p. 243):
The chemical combination of two substances produces, as is well
known, a third substance with properties different from those of
either of the two substances separately, or both of them taken
together. Not a trace of the properties of hydrogen or oxygen is
observable in those of their compound, water.
Mills insight is relevant to the study of so-called complex systems ranging from atoms to the weather whose dynamic
nonlinear behavior involves many interacting and interconnected parts. (A system is dynamic if it is constantly in flux; it
is nonlinear if effects are out of proportion to causes, as when
a neglected candle causes a fire that destroys an entire city.
See self-organizing systems.) However, the question of
whether and to what extent language is an emergent phenomenon remains controversial.
Linguistic Emergentism
Although it is widely agreed that emergentist approaches to language necessarily stand in opposition to theories of the language
faculty that posit an innate universal grammar, other tenets
of linguistic emergentism are less well defined, and there is no
consensus within the field as to how precisely the standard problems of linguistic analysis should be confronted. Nonetheless,
the starting point for a substantial portion of emergentist work
seems to involve a commitment to the emergentist thesis for
language:
The phenomena of language are best explained by reference to
more basic nonlinguistic (i.e., nongrammatical) factors and
their interaction.
Emergentism
change, and evolution, the eliminativist position is far from universally accepted within emergentism. Symbolic representations
of one form or another are evident in the work of many emergentists (e.g., Goldberg 1999; Tomasello 2003; OGrady 2001, 2005),
who nonetheless reject the view that the properties of those representations should be attributed to innate grammatical principles (see innateness and innatism).
Language Acquisition
To date, emergentist work within linguistics has focused most
strongly on the question of how language is acquired (see, e.g.,
the many papers in MacWhinney 1999). The impetus for this
focus stems from opposition to the central claim of grammatical nativism, which is that the principles underlying a good deal
of linguistic knowledge are underdetermined by experience and
must therefore be innate.
Emergentism is not opposed to nativism per se the fact that
the brain is innately structured in various ways is not a matter of
dispute. However, there is opposition to representational nativism, the view that there is direct innate structuring of particular
grammatical principles and constraints (Elman et al. 1996, 369 ff;
Bates et al. 1998), as implied by many of the proposals associated
with universal grammar.
Contemporary emergentism often includes a commitment
to explaining linguistic development by reference to the operation of simple learning mechanisms (essentially, inductive generalization) that extract statistical regularities from experience.
It is interesting that there is as yet no consensus as to what form
the resulting knowledge might take local associations and
memorized chunks (Ellis 2002), constructions (Goldberg 1999;
Tomasello 2003), or computational routines (OGrady 2001,
2005). In addition, there is variation with respect to the exact
relationship that is assumed to hold between learning and relative frequency in the input. Some work implies a quite direct
relationship (e.g., Ellis 2002), but other work suggests something
less direct (e.g., Elman 2002).
Emergentist work on language acquisition often makes use
of computer modeling to test hypotheses about development.
Jeffrey Elman and his colleagues (e.g., Elman 2002) have been
able to show that a simple recurrent network (SRN) can achieve
at least some of the milestones associated with language acquisition in children, including the identification of category-like
classes of words, the formation of patterns not observed in the
input, retreat from overgeneralizations, and the mastery of subjectverb agreement. (An SRN learns to produce output of its own
by processing sentences in its input; it is specifically designed to
take note of local co-occurrence relationships or transitional
probabilities given the word X, whats the likelihood that the
next word will be Y?)
Emergentist modeling has yielded impressive results, but it
raises the question of why the particular statistical regularities
exploited by the SRN are in the input in the first place. In other
words, why does language have the particular properties that it
does? Why, for example, are there languages (such as English) in
which verbs agree only with subjects, but no language in which
verbs agree only with direct objects?
Networks provide no answer to this sort of question. In fact,
if presented with data in which verbs agree with direct objects
275
Emergentism
that of its base. The words inadequate and inaudible are a case in
point. Because adequate is more frequent than the affixed form
inadequate, its presence in the derived word is relatively salient,
leading to a high native-speaker rating for structural complexity. In contrast, inaudible, which is more frequent (and therefore
more salient) than audible, receives a low rating for structural
complexity.
If this is right, then morphological structure exists but not
in the categorical form commonly assumed. Rather, what we
think of as morpheme boundaries emerge to varying degrees
of strength from the interaction of more basic factors, such as
frequency, semantic transparency, and even phonotactics.
(The low-probability sequence in inhumane creates a sharper
morpheme boundary than the high-probability sequence in
insincere.)
THE LEXICON. There have been various attempts to develop an
emergentist approach to the lexicon, which is traditionally seen
as the repository of information about morphemes and words.
One possibility, suggested by Joan Bybee (1998), among others,
is that the lexicon emerges from the way in which (by hypothesis)
the brain responds to and stores linguistic experience by creating units whose strength and productivity is determined largely
by frequency of occurrence. Some of these units correspond to
words, as in a traditional lexicon, but many are phrases and other
larger units of organization, including possibly abstract constructions (see usage-based theory).
Elman (2005) also argues against a pre-structured lexicon,
proposing instead that lexical knowledge is implicit in the effects
that words have on the minds internal states, as represented
in the activation patterns created by an SRN. Because an SRN
focuses on co-occurrence relationships (see above), these effects
are modulated by context a words meaning, like its syntactic
category, emerges from the contexts in which it is used rather
than from an a priori vocabulary of linguistic primitives.
PHONOLOGY. Pioneering work on emergentist phonology was
carried out by Patricia Donegan (1985), who noted the unhelpfulness to language learners of classic distributional analysis. As
she observed, it is implausible to suppose that children record
sets of phonetic representations in memory and then compare
them in the hope of determining which phonetic contrasts are
distinctive and which are predictable from context (see speech
perception in infants and speech production). Instead,
Donegan suggests, children begin with a set of processes (nasalization, devoicing, and so forth) that emerge as responses to the
physical limitations of the human vocal tract and the auditory
apparatus. (These limitations are inborn, of course, but are not
inherently linguistic in character, despite their linguistic consequences.) A languages phonemic inventory and allophonic
patterns then emerge as specific processes are suppressed in
response to experience.
A simple example involves the process that palatalizes /s/ in
front of a high front vowel, giving the pronunciation [i] for /si/
in many languages (e.g., Japanese). A child learning English is
forced to suppress this process upon exposure to words such
as see, which is pronounced [si], without palatalization. This, in
turn, results in the admission of // to the phonemic inventory
276
of English: Because the palatalization process has been suppressed, the [] in words such as [i] she must be interpreted
as a sound in its own right, rather than as a process-induced
variant of /s/.
Crucially, this conclusion is drawn without the need for comparison of minimal pairs or similar distributional analysis; the
phonemic inventory emerges in response to a much simpler
and more basic phenomenon the suppression of processes
based on exposure to particular individual words. Boersma
(1998) and Hayes, Kirchner, and Steriade (2004) discuss a broad
range of other phonological phenomena from an emergentist
perspective.
Concluding Remarks
There is currently no comprehensive emergentist theory of language or its acquisition, but there are various emergentist-inspired
research programs devoted to the construction of such a theory.
For the most part, this work is based on the simple thesis that the
core properties of language are best understood by reference to
the properties of quite general cognitive mechanisms and their
interaction with one another and with experience. The viability
of this idea can and must be measured against its success in confronting the classic empirical challenges of linguistic analysis
figuring out how language works and how it is acquired.
William OGrady
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bates, Elizabeth, Jeffrey Elman, Mark Johnson, Annette Karmiloff-Smith,
Domenico Parisi, Kim Plunkett. 1998. Innateness and emergentism.
In A Companion to Cognitive Science, ed. W. Bechtel and G. Graham,
590601. Oxford: Blackwell.
Bates, Elizabeth, Judith Goodman. 1999. On the emergence of grammar
from the lexicon. In The Emergence of Language, ed. B. MacWhinney,
2979. Mahwah, NJ: Erlbaum.
Bates, Elizabeth, Brian MacWhinney. 1988. What is functionalism?
Papers and Reports on Child Language Development 27: 13752.
Boersma, Paul. 1998. Functional Phonology: Formalizing the Interactions
between Articulatory and Perceptual Drives. The Hague: Holland
Academic Graphics.
Bybee, Joan. 1998. The emergent lexicon. In Proceedings of the 34th
Regional Meeting of the Chicago Linguistic Society: The Panels, 42135.
Chicago: Chicago Linguistic Society. An excellent example of emergentist thinking about the lexicon.
Donegan, Patricia. 1985. How learnable is phonology? In Papers on
Natural Phonology from Eisenstadt, ed. W. Dressler and L. Tonelli,
1931. Padua: Cooperativa Libraria Editoriale Studentesca Patavina.
Ellis, Nick. 2002. Frequency effects in language processing. Studies in
Second Language Acquisition 24: 14388.
Elman, Jeffrey. 1993. Learning and development in neural networks: The
importance of starting small. Cognition 48: 7199. A much-cited and
widely admired illustration of the value of computational modeling in
the study of language acquisition.
. 2002. Generalization from sparse input. In Proceedings of the
38th Regional Meeting of the Chicago Linguistic Society, 175200.
Chicago: Chicago Linguistic Society. A highly readable summary of
several important SRN-based studies of language acquisition.
. 2004. An alternative view of the mental lexicon. Trends in
Cognitive Science 8: 3016.
Elman, Jeffrey, Elizabeth Bates, Mark Johnson, Annette Karmiloff-Smith,
Domenico Parisi, Kim Plunkett. 1996. In Rethinking Innateness: A
Emergent Structure
Connectionist Perspective on Development. Cambridge, MA: MIT
Press.
Goldberg, Adele. 1999. The emergence of the semantics of argument structure constructions. In The Emergence of Language, ed. B.
MacWhinney, 197212. Mahwah, NJ: Erlbaum.
Goldberg, Adele, and Ray Jackendoff. 2004. The English resultative as a
family of constructions. Language 80: 53268.
Gregg, Kevin R. 2003. The state of emergentism in second language
acquisition. Second Language Research 19: 4275.
Hay, Jennifer.and 2003. Causes and Consequences of Word Structure. New
York: Routledge.
Hay, Jennifer, and R. Harald Baayen. 2005. Shifting paradigms: Gradient
structure in morphology. Trends in Cognitive Science 9: 3428. An
excellent survey of work on emergentist morphology.
Hayes, Bruce, Robert Kirchner, and Donca Steriade, eds. 2004. Phonetically Based Phonology. Cambridge: Cambridge University Press.
MacWhinney, Brian. 1998. Models of the emergence of language.
Annual Review of Psychology 49: 199227.
. 2002. Language emergence. In An Integrated View of Language
Development: Papers in Honor of Henning Wode, ed. P. Burmeister,
T. Piske, and A. Rohde, 1742. Trier, Germany: Wissenshaftliche Verlag.
. 2004. A multiple process solution to the logical problem of language acquisition. Journal of Child Language 31: 883914.
. 2005. The emergence of grammar from perspective. In Grounding
Cognition: The Role of Perception and Action in Memory, Language and
Thinking, ed D. Pecher and R. Zwaan, 198223. Cambridge: Cambridge
University Press.
MacWhinney, Brian, ed. 1999. The Emergence of Language. Mahwah,
NJ: Lawrence Erlbaum.
Marcus, Gary. 1998. Rethinking eliminative connectionism. Cognitive
Psychology 37: 24382.
. 2001. The Algebraic Mind. Cambridge, MA: MIT Press.
McClelland, James, and Karalyn Patterson. 2002. Rules or connections
in past-tense inflection: What does the evidence rule out? Trends in
Cognitive Science 6: 46572. An update and survey of connectionist
work on inflection.
Mill, John Stuart. [1843] 1930. A System of Logic Ratiocinative and
Inductive. London: Longmans, Green, and Co.
OGrady, William. 2001. An emergentist approach to syntax. Available
online at: http://www.ling.hawaii.edu/faculty/ogrady/. A summary of
the detailed arguments for an emergentist theory of syntax found in
OGrady (2005).
. 2005. Syntactic Carpentry: An Emergentist Approach to Syntax.
Mahwah, NJ: Erlbaum. An emergentist approach to syntax that seeks
an understanding of many of the classic problems of syntactic theory
in terms of processing.
Palmer-Brown, Dominic, and Jonathan Tepper, Heather Powell. 2002.
Connectionist natural language parsing. Trends in Cognitive Science
6: 43742.
Smolensky, Paul. 1999. Grammarbased connectionist approaches to
language. Cognitive Science 23: 589613.
Steedman, Mark. 1999. Connectionist sentence processing in perspective. Cognitive Science 23: 61534.
Tomasello, Michael. 2003. Constructing a Language: A Usage-Based
Theory of Language Acquisition. Cambridge: Harvard University
Press. A widely cited example of a usage-based approach to language
acquisition.
EMERGENT STRUCTURE
This term is used in many fields, including scientific and social
disciplines, where it has been applied to a variety of adaptive
self-organizing systems from termite mounds to grocery
checkout lines. The term emergent refers especially to an openended process in which systematicity is partial and incomplete
and in which a system is in a constant course of (re)formation.
In the study of language, the expression emergent grammar
was coined by Paul Hopper (1987) as a methodological proposal for approaching the relationship between grammar and
the local structure of natural discourse. Logically prior, fixed
grammar was, Hopper argued, inconsistent with the kinds of ad
hoc linguistic decisions made by speakers. The notion of emergent grammar inverts the standard picture of grammar, as well
as the generally accepted logical priority of structure over text.
Linguistic structure is thus to be seen as a product of, rather than
a prerequisite to, discourse. Since discourse is ongoing, structure
is emergent, that is, continually in a process of formation according to the current needs of the interaction. (See Weber 1997 for
further discussion.)
The database for the study of language from this perspective is a corpus of transcribed texts, usually oral, and, recently,
conversational (Ochs, Schegloff, and Thompson 1996). In this
respect, too, emergent grammar differs from structural and cognitive grammar, in which conclusions are normally based on
isolated constructed sentences. The explanation for grammar,
according to this theory, lies in frequency (Bybee and Hopper
2001) and the associated routinization of forms (Haiman 1994).
High-frequency forms tend to become phonetically reduced
and to be restructured (Bybee 2001). Typical examples are the
English pronoun+modal sequences like Ill, youre, weve, and so
on. Emergent grammar is thus relevant to the more general study
of grammaticalization.
Incipient structure that is, looking backward at the historical origins of a structured system or forward to the predicted
course of events leading to a structured system, as in the study
of first language acquisition and of most varieties of cognitive linguistics is more properly described as emerging
than as emergent. The noun emergence is ambiguous in this
respect.
Paul J. Hopper
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bybee, Joan. 1998. The emergent lexicon. Papers of the Annual Meeting
of the Chicago Linguistic Society 34: 42135.
. 2001. Phonology and Language Use. Cambridge: Cambridge
University Press.
Bybee, Joan, and Paul Hopper, eds. 2001. Frequency and the Emergence of
Linguistic Structure. Amsterdam: Benjamins.
Haiman, John. 1994. Ritualization and the development of language.
In Perspectives on Grammaticalization, ed. William Pagliuca, 328.
Amsterdam: John Benjamins.
Hopper, Paul. 1987. Emergent grammar. Papers of the Annual Meeting
of the Berkeley Linguistic Society 13: 13957.
Lewin, Roger. 1992. Complexity: Life at the Edge of Chaos. New
York: Macmillan. A readable account of chaos theory and the emergence of structured systems.
Ochs, Elinor, Emanuel Schegloff, and Sandra Thompson, eds.1996.
Interaction and Grammar. Cambridge: Cambridge University
Press.
Weber, Thilo. 1997. The emergence of linguistic structure: Paul
Hoppers emergent grammar hypothesis revisited. Language Sciences
19.2: 17796.
277
Historical Perspectives
Historical perspectives illustrate the multiple domains and methodologies of research on emotional words. In the mid-1800s,
neuropsychological case studies of Hughlings Jackson (1958)
and others helped shape current ideas concerning automatic
or uncontrollable production of emotional words (see, e.g., Van
Lancker 1987). Carl Jungs (1910) work with emotional words
in free association tasks also shaped procedures for diagnosing clinical disorders such as schizophrenia (see also psychoanalysis and language). From 1950 to 1975, experimental
psychologists used classical conditioning concepts to analyze
the learning of emotional words (e.g., Staats 1968) and adopted
perceptual defense paradigms to determine whether ego-protective processes shield threatening stimuli (taboo words) from
awareness (e.g., Dixon 1971). However, both lines of research
were largely abandoned: perceptual defense because of methodological flaws and the learning of emotional words because computer metaphors dominated the study of language and cognition
and downplayed emotion during the period of 1975 to 1990
(Jay 2003).
Methodological Perspectives
RATING STUDIES. Rating studies provide a method for determining the emotional qualities of words. A classic example is
the semantic differential (Osgood, Suci, and Tanenbaum 1957),
where ratings of words on bipolar connotative scales reflect three
underlying dimensions: evaluation (the valence component,
e.g., negativepositive); activity (e.g., fastslow); and potency
(e.g., strongweak). L. H. Wurm and D. A. Vakoch (1996) argued
that evolutionary considerations and relations between processing time data and the evaluation, activity, and potency ratings for
words indicate an affective lexicon (for avoiding threats) that differs from the general lexicon (for obtaining valuable resources).
Other rating studies involving the affective lexicon include
Bellezza, Greenwald, and Banaji (1986), Bradley and Lang
(1999), and Jay (1992). Unrepresented in current rating studies
278
Theoretical Perspectives
Current research on emotional words illustrates a gamut of theoretical perspectives that differ in their scope and goals and in
the nature and specificity of the predictions they make. Jays
(2000) neuro-psychosocial theory of cursing summarizes likelihood estimates of various forms of cursing, based on neurological (e.g., conscious state, brain damage), psychological (e.g.,
personality, age, history), and social context (e.g., culture, class)
279
Conclusion
Both historical and contemporary research on emotional
words reflects a wide variety of theoretical and methodological
approaches in fields ranging from neuroscience to psycholinguistics to cognitive and clinical psychology. Further research is
required to piece together these multiple domains and to develop
a general understanding of emotional words and their relation to
other cognitive processes. However, emotional words currently
seem poised to resume their central position in the language sciences and related disciplines.
Kristin L. Janschewitz and Donald G. MacKay
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Adolphs, R., J. A. Russell, and D. Tranel. 1999. A role for the human
amygdala in recognizing emotional arousal. Psychological Science
10: 16771.
Bellezza, F. S., A. G. Greenwald, and M. R. Banaji 1986. Words high and
low in pleasantness as rated by male and female college students.
Behavior Research Methods, Instruments, and Computers 18: 299303.
Blake, M. L. 2003. Affective language and humor appreciation after
right hemisphere brain damage. Seminars in Speech and Language
24:10719.
Borod, J. C., R. L. Bloom, and C. S. Haywood. 1998. Verbal aspects
in emotional communication. In Right Hemisphere Language
Comprehension: Perspectives from Cognitive Neuroscience, ed. M.
Beeman and C. Chiarello, 285307. Mahwah, NJ: Erlbaum.
Bradley, M. M., and P. J. Lang. 1999. Affective Norms for English Words
(ANEW): Instruction Manual and Affective Ratings. Technical Report
C-1, The Center for Research in Psychophysiology, University of Florida.
Brown, W. S., L. K. Paul, M. Symington, and R. Dietrich. 2005.
Comprehension of humor in primary agenesis of the corpus callosum. Neuropsychologia 43: 90616.
Bucci, W. 1997. Symptoms and symbols: A multiple code theory of somatization. Psychoanalytic Inquiry 17: 15172.
280
In her critical essay, J. T. Irvine (1990) wrote that many linguists tend to get cold feet when it comes to considering how
emotions are expressed verbally. Accordingly, if we use the terminology of Ferdinand de Saussure, we can say that emotion
is accepted as integral to the parole, which is linguistically less
meaningful than langue language in its broadest sense. Thus,
examination of emotion is pushed to the periphery of linguistics. Irvine also remarks that though there are languages with
phonological and morphological units that indicate
emotional states in speech, linguists frequently tend to combine
such elements with general descriptions of grammar, rather than
emphasizing such verbal expressions of emotion. However, she
notes two important linguistic texts that also deal with emotion
in language, namely, Edward Sapirs lexicon of emotions as mirroring culture and Roman Jakobsons work relating to the emotive function of language.
C. Caffi and R. W. Janney (1994) examined the rhetorical
strategies for expressing emotion by comparing psychological
categories of emotions with linguistic categories. They define
emotional communication as directed strategies for imparting
emotional information in speech or writing, insisting that such
expressions must be analyzed linguistically, because language
spoken or written is the means for conveying emotion. Their
model comprises linguistic markers, including specific emotion
words, obligatory words, syntax markers, and spoken language
mechanisms (i.e., tones and intonation, prosody, length of
syllables, etc.).
Caffi and Janneys writings infer that there are significant
connections between textual linguistic usage and emotions,
as evidenced also in diaries, letters, and other autobiographical writings. For instance, language was used to measure emotion in a study by G. Collier, D. Kuiken, and M.E. Enzle (1982).
The researchers noted that when describing negative emotions,
people use more complex constructions than in describing positive emotions, and this also applies to expressions of negative
as opposed to positive personal qualities. The positive is always
more clearly expressed. Assessing descriptions of levels of positivity or negativity of emotions or traits indicated that the feelings or traits described in more complex detail tended to be more
negative, that is, the more complex the descriptions, the more
likely they were to be negative.
An earlier study by C. E. Osgood (1958) dealt with the connection between emotion and language, establishing the link
between the lingual characteristics of a text and the motivation
level of the author when writing it. The research studied suicide
notes, written under the influence of very strong emotion the
last letter as compared with ordinary correspondence with
family or close friends.
In her chapter How and why is emotion communicated?
S. Planalp (1999) writes that verbal expression of feelings is
endemic to the process of communication, even though people do not always use words. They do not, as a rule, announce
that Im angry or Im feeling depressed at the moment, but
there are other verbal indications, like swearing or extravagant
outbursts such as I could kill him!
J. W. Pennebaker and M. E. Francis (1996) analyzed personal texts describing thoughts about commencing higher studies by first-year students at college. Linguistic and cognitive
281
282
Emotion Words
Ortony, A., G. L. Clore, and M. A. Foss. 1987.The referential structure of
the affective lexicon. Cognitive Science 11: 34164.
Osgood, C. E. 1958. Some effects of motivation on style of encoding.
In Style in Language, ed. T. A. Sebeok, 293306. Cambridge, MA: MIT
Press.
Pennebaker J. W. 2007, The world of words. Available online at: http://
www.liwc.net/.
Pennebaker, J. W., and M. E. Francis. 1996. Cognitive, emotional and
language processes in disclosure. Cognition and Emotion 10: 60126.
Pennebaker, J. W., R. B. Slatcher, and C. K. Chung. 2005. Linguistic
markers of psychological state through media Interviews: John Kerry
and John Edwards in 2004, Al Gore in 2000. Analyses of Social Issues
and Public Policy 5: 197204.
Planalp, S. 1999. Communicating Emotion: Social, Moral and Cultural
Processes. Cambridge: Cambridge University Press.
EMOTION WORDS
What Counts as an Emotion Word?
Languages differ in the size and range of their emotion vocabularies. There are, for example, more than 500 words in English,
750 in Taiwanese Chinese (Russell 1991), and 256 in Filipino
(Church, Katigbak, and Reyes 1996). In addition, translation
equivalents often cover overlapping but not identical semantic
space (Wierzbicka 1999). Clearly, the investigation of the emotion lexicon requires the careful delimitation of what counts as
an emotion word.
Empirical approaches to this question are driven by
prototype theory (Fehr and Russell 1984; Rosch 1978), according to which semantic categories are recognized not by lists of
necessary and sufficient features but in terms of a gestalt
or configurational whole. This approach suggests that emotion
is a fuzzy category, and emotion words fit the category in a
graded manner.
A number of taxonomies have been proposed. G. L. Clore,
A. Ortony, and M. A. Foss (1987) distinguished eight categories
in English: 1) pure affective states (e.g., happy), 2) affective-behavioral states (e.g., cheerful), 3) affective-cognitive states (e.g.,
encouraged), 4) cognitive states (e.g., certain), 5) cognitivebehavioral states (e.g., cautious), 6) bodily states (e.g. sleepy),
7) subjective evaluations of character (e.g., attractive), and 8)
objective conditions (e.g., abandoned). Analyses of prototypicality ratings of 585 candidate emotion words confirmed the
empirical discriminability of the eight categories, and words in
the first three (affective) categories had the highest typicality
ratings.
Phillip Shaver and colleagues (1987) used cluster analysis of
prototypicality ratings of English emotion words to display a prototype hierarchy, with two superordinate categories encompassing positive versus negative terms and five basic-level terms: love,
joy, anger, sadness, and fear. The rest of the terms are subordinates under these basic terms (Shaver et al. 1987; Storm and
Storm 1987). It is interesting to note that negative emotion words
generally outnumber positive emotion words, perhaps explained
by the greater cognitive processing required by negative events
in comparison with positive events (Schrauf and Sanchez 2004).
The Indonesian emotion lexicon has the same overall structure
(Shaver, Murdaya, and Fraley 2001), but in the Chinese lexicon,
a love category does not emerge separate from happiness-
283
Emplotment
cross-linguistic study. Journal of Multilingual and Multicultural
Development 25.2/3: 26684.
Schwanenflugel, P. J., C. Akin, W. Luh. 1992. Context availability and
the recall of abstract and concrete words. Memory and Cognition
20: 96104.
Shaver, P., U. Murdaya, R. C. Fraley. 2001. Structure of the Indonesian
emotion lexicon. Asian Journal of Social Psychology 4: 20124.
Shaver, P. R., S. Wu, J. C. Schwartz. 1992. Cross-cultural similarities and
differences in emotion and its representation: A prototype approach.
In Review of Personality and Social Psychology. Vol. 13: Emotion. Ed.
M. S. Clark, 175212. Newbury Park, CA: Sage.
Shaver, Phillip, Judith Schwartz, Donald Kirson, Cary OConnor. 1987.
Emotion knowledge: Further exploration of a protoytpe approach.
Journal of Personality and Social Psychology 52: 106186.
Storm, C., T. Storm. 1987. A taxonomic study of the vocabulary of emotions. Journal of Personality and Social Psychology 53: 80516.
Wierzbicka, A. 1999. Emotions Across Languages and Across
Cultures: Diversity and Universals. Cambridge: Cambridge University
Press.
Zammuner, V. L. 1998. Concepts of emotion: Emotionness and dimensional ratings of Italian emotion words. Cognition and Emotion
12: 24372.
EMPLOTMENT
Emplotment is the organization of events into a narrative. The
concept was developed most influentially by Hayden White in
his treatment of historiography. White distinguishes five levels
of conceptualization in the writing of history (1973, 5). The first
is the chronicle, a simple listing of events in the order of their
occurrence. The second is the formation of these events into a
basic causal sequence or story. The third, emplotment proper,
is their further elaboration into a narrative with a point. (The
fourth and fifth levels, mode of argument and mode of ideological implication, go beyond emplotment and thus the main
concerns of this entry.) According to White, different historians
commonly organize even the same sequence of events into divergent histories, reflecting different strategies of emplotment. The
same point applies beyond writers on history. Everyone emplots
events, from political figures shaping public policy to ordinary
people in conversational storytelling.
We might consider the events of September 11, 2001, by way
of illustration. A chronicle would simply list the events of the
day. A basic story would set out the causal relations the organization of the conspirators, their practice, their final execution
of their plans, and so on. It should be clear that, even here, there
are different ways in which events may be selected and grouped
together and different ways in which causal links may be posited. For example, in the lead-up to the invasion of Iraq, some
commentators suggested that various actions of the Iraqi government were part of the September 11 causal sequence; others
denied the connection, arguing that this was not even a plausible part of the basic story. The third level, emplotment, embeds
the causal sequence in a more elaborated structure. In the case
of the Bush administration, that structure was a war narrative in
which the events of September 11 constituted an act of war. For
many others, that structure was a crime narrative in which the
events were a (massive) criminal violation. As these cases suggest, differences at the level of causal interpretation and differences at the level of emplotment not only manifest intellectual
284
Emplotment
sorts of prototype as well, including prototypes for categories of
event sequences thus, narrative prototypes.
We may broadly distinguish, then, between two types of causal
understanding. The first sort, found in what might be called the
general sciences, is experimental, based on the isolation of causal
features by the controlled manipulation of variables. The second sort, found in ordinary life and in what might be called the
particularistic sciences, is prototypical, based on the formation
of distinctive, (loosely) statistical structures. Of course, there
are intermediate cases. Moreover, there are different degrees to
which statistical derivations may be made explicit and rigorous.
For example, there are areas of economic history where we might
achieve relatively high levels of explicitness and rigor. In other
cases, however, it is very difficult to make the statistical process at all scientific, for the selection of a comparison set (which
guides causal inferences) is already so thoroughly imbued with
the implicit prototypes of the researcher.
This division in types of causal understanding is, of course,
connected with the orientation of the particularistic sciences to
the explanation of particulars. But there are many particulars.
Just how does our interest in certain particulars arise? In both
particularistic and general sciences, we attend to the explanation of individual objects or events when we care about them.
We care about something when it has an emotional impact,
which is to say, when it engages some emotion system. Thus, an
understanding of emotion systems is crucial for understanding
our explanatory aims in particularistic study. As it turns out, an
understanding of emotion systems also gives us a way of understanding Whites first and second levels of conceptualization.
Specifically, both a chronicle and a basic story involve three
fundamental cognitive operations selection, segmentation,
and structuration. (On these processes, see Hogan 2003, 3840.)
There are countless aspects of any given sequence of events and
countless construals of those events and their components.
Even a chronicle selects certain aspects while ignoring others,
clusters those aspects together into the events that compose
the chronology, and gives those aspects at least some degree of
internal structure. For example, in a chronology of the events of
September 11, we might include a statement that the hijackers
took over one airplane at such and such a time. That statement
selects various aspects of the situation and organizes them into
a brief causal moment. Moreover, in going beyond a chronicle
and telling the basic story, we might begin with the first plane
crash, or we might begin with the conspiracy of the hijackers, or
we might begin with various aspects of U.S. policy in the Muslim
world (seen by the hijackers as justification for the September 11
attacks). The questions that arise here concern just why we select
certain aspects and construals over others and how we come to
decide that events begin and end at certain points.
The simple answer to these questions is that our initial selection and construal (as manifest in a chronicle) are the product of
our emotion systems. (For a more technical discussion of these
issues, and for research supporting this analysis, see Hogan
2008 and Chapter 4 of Hogan 2009.) We are emotionally sensitive to certain sorts of properties, conditions, alterations, and so
on. These draw our attentional focus. Our sense of a beginning
and an ending (thus, our fashioning of a basic story) are equally
guided by our emotional responses. The beginning of the story
is the point at which our emotion systems are engaged. The end
of the story is the point at which our emotion systems return to
their normal state. This is why Americans tend to view the conspiracy of the hijackers as the beginning of the story. With limited
exceptions, they lack emotional interest in what preceded and
motivated the attacks.
Of course, things do not end with this level of selection, and so
on. Whenever we isolate aspects of a particular event sequence
due to their emotional force, we simultaneously activate cognitive
structures for understanding and responding to that sequence.
These structures crucially involve prototypes, which supplement
our emotional responses in elaborating interpretations, explanations, expectations, directing attentional focus, and so on. In the
case of event sequences, these are crucially narrative prototypes,
including subprototypes bearing on actions and on agents. This,
then, leads us to the level of emplotment proper.
The narrative prototypes that guide emplotments undoubtedly include the broad structures of narrative universals.
For example, the hijackers may have emplotted their actions in
terms of a sacrificial narrative in which the suffering of the home
society will be relieved by God due to the voluntary death of a
member of that society. In contrast, the U.S. government emplotted these same actions as the foreign aggression component of a
heroic plot. Beyond these cross-cultural patterns, emplotments
also derive from more culturally specific narrative structures,
including structures related to culturally defined practices, such
as those of legal systems (as in the emplotment of the September
11 attacks as criminal acts).
Both universal and culturally specific narratives are bound up
not only with emotions but also with values related to those emotions. Thus, it is unsurprising that different emplotments tend to
import different social agendas and different political attitudes
into the interpretation of the event sequence. (In Whites system, this appears in the fifth level of conceptualization, mode
of ideological implication [1973, 5].) It is also unsurprising that
they tend to be points of consequential political contestation.
285
Encoding
in discourse but to more narrowly grammatical concerns as well.
The most radical view of this relation would be that emplotment
is cognitively fundamental to certain aspects of grammar (a point
suggested by authors such as Mark Turner [1996]). Alternatively,
it may simply be that certain features recur in grammar and narrative, due to shared cognitive sources (an account that may be
suggested by certain aspects of frame semantics) or due to the
effects of grammar on emplotment.
Patrick Colm Hogan
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Barsalou, Lawrence. 1983. Ad hoc categories. Memory and Cognition
11.3: 21127.
Frye, Northrop. 1957. Anatomy of Criticism: Four Essays. Princeton,
NJ: Princeton University Press.
Hogan, Patrick Colm. 2003. Cognitive Science, Literature, and the Arts: A
Guide for Humanists. New York: Routledge.
. 2008. Stories, wars, and emotions: The absoluteness of narrative
beginnings. In Narrative Beginnings, ed. Brian Richardson, 4462.
Lincoln: University of Nebraska Press.
. 2009. Understanding Nationalism: Narrative, Identity, and
Cognitive Science. Columbus: Ohio State University Press.
Kahneman, Daniel, and Dale Miller. 1986. Norm theory: Comparing
reality to its alternatives. Psychological Review. 93.2: 13653.
Kroeger, Paul R. 2004. Analyzing Syntax: A Lexical-Functional Approach.
Cambridge: Cambridge University Press.
Turner, Mark. 1996. The Literary Mind. Oxford: Oxford University Press.
Tversky, Amos. 1977. Features of similarity. Psychological Review
84: 32752.
White, Hayden. 1973. Metahistory: The Historical Imagination in
Nineteenth-Century Europe. Baltimore: Johns Hopkins University
Press.
ENCODING
Encoding refers both to the process of laying down information in memory as the result of exposure to certain stimuli and
to the organization of that information once it has been laid
down. From the viewpoint of language, this encompasses both
the encoding of linguistic knowledge and the encoding of verbal
experience (Francis 1999). The former refers to what we think of
as knowing a language and the latter refers to knowing things
in language (e.g., recalling a conversation). In either case, most
of what we know about encoding comes from studying its interaction with retrieval.
286
nonc/Statement (Foucault)
What Is Essentialism?
The idea of essences has been important in Western philosophy
since at least the time of Plato. Discussion of essences flourished
in classical and medieval philosophy and has been revived in
recent decades by philosophers such as Saul Kripke, due primarily to work in modal logic, which is to say, the formalization of
necessity and possibility. In the humanities, essentialism is often
used to refer to any view that is not historicist. In philosophy
including the semantics of formal logic the term is used
more narrowly for the belief that objects have definitive features
and incidental features. The incidental features may change without the identity of the object changing. However, if a definitive
feature changes, then the objects identity changes. For example,
in sufficient quantities and with the right light, water appears to
be blue, but if we take a glass of water from a larger (blue) body,
it appears colorless. In scooping the glass out of the water, we
have altered the color of the water, yet we would not say that it
is a different thing than it was before. An incidental property of
appearance has changed, but the stuff itself remains the same.
In contrast, suppose we take that glass of water and induce a
chemical change so that it continues to appear clear but is no
287
288
water that looks, tastes, and quenches thirst like H2O, even if it
has a different chemical composition? This brings us to our second topic.
289
a term with an object. Water names H2O because its use leads
back to a link with a certain substance having a certain essence.
Al Gore names Al Gore because its use leads back to a certain individual, tracing a chain of transmission in reverse. (The
chain of transmission, presumably went from Gores parents to
their friends [Bob, Trudy this is baby Al], and so on.) This is
called the causal theory of reference. (This intentional or semantic causality is, of course, different from the physical causality
of putatively essential properties, such as chemical composition.
In order to keep the two distinct, I refer to the latter as physical
causality in the remainder of this entry.)
As with natural kinds, however, there has to be some limit on
just what we can vary about Al Gore in order to guarantee that
he is the same person across possible worlds. It is not clear that
there is any property that has the sort of physical causal force
that chemical composition has for water. Relying on intuitions
about possible worlds, Kripke concludes that the earliest physical causal factors are the crucial ones here. It seems to me, he
writes, that anything coming from a different origin would not
be this object ([1972] 1980, 113). But here again we run into the
issue of just what intuition tells us. Put differently, do we learn
anything about identity here, or do we only learn something
about the general importance of physical causality for the way
in which we think about the world, and thus the way we draw
intuitive inferences about identity? In other words, does this tell
us something epistemological and ontological, or simply something psychological and evolutionary?
Conclusion
If they are valid, the preceding arguments may suggest that,
ultimately, there is no real issue of essences (or essential identity) in semantics. (For further discussion of these issues, see
Hogan [1996] 2008, Preface, Chapter 2, and Chapter 3, particularly ixxi and 6370.) However, there are important issues of
physical causal analysis (issues to which the active development of essentialist theories has helped to draw our attention).
More exactly, there is no issue of metaphysics and the foundations of epistemology (requiring guidance by possible worlds
theorization). Rather, there is only 1) an issue of empirical
science regarding the physical causal properties of objects (or
substances) and 2) the related, also empirical, issue of how our
brains organize the world in terms of physical causality, identity relations, and so on. These forms of empirical study, then,
may impact our descriptive account of the way in which meaning operates. They should also have consequences, of a more
limited sort, for our normative definitions of terms, particularly
in scientific contexts where precise physical causal analysis is
paramount.
On the other hand, not everyone agrees with these arguments far from it, in fact. Essentialism is an important and
highly influential position, and advocates of essentialist semantics have responses to the preceding claims. For example, some
writers would insist that we should not take causality as a primitive notion. Rather, we need to explain causality in terms of possible worlds (see Lewis 1986, 157269), leading us back to modality
and, presumably, essences.
Patrick Colm Hogan
290
Logical Positivism
By calling ethical assertions nonsense, Rudolf Carnap meant
not only that they lack truth value but that they are outside
the sphere of rational argument altogether. The real world
influence of this doctrine was enormous. For example, Lionel
Robbins (1932), one of the most influential economists of the
1930s, enthusiastically endorsed it, as did Milton Friedman and
Paul Samuelson. And the idea of value-free science obviously
influenced other social sciences as well
Even a critic of logical positivism must grant that it had one
enormous virtue, and that was its capacity for self-criticism. If
logical positivists seemed to the economists just mentioned to be
logicians of science who had discovered how to demarcate the
cognitively meaningful from nonsense, the positivists themselves were dissatisfied with their formulations of the supposed
demarcation principle and constantly revised it (Hempel 1963;
see also Putnam 2002, 727).
Besides criticisms faced from within the movement itself,
the attempts to formulate a criterion of cognitive meaningfulness encountered criticisms from W. V. Quine, a lifelong friend
of Carnaps who shared his admiration for science and symbolic
logic and his noncognitivism with respect to ethics.
Basically, all the logical positivist formulations presupposed 1)
that all meaningful language except for pure mathematics and
logic could be reduced to observation terms (it was supposed to
be clear which these are) and 2) that mathematics and logic are
analytic or tautologous. Quine ([1951] 1961) famously demolished these two dogmas of empiricism, as he called them, and
urged to the satisfaction of almost all philosophers of science
and philosophers of mathematics that neither the program of
reducing all meaningful language to the positivists observation
vocabulary nor the idea that mathematics consists of tautologies (or, alternatively, truths by convention) is defensible. The
pillars on which the positivist criterion of demarcation rested fell
in 1950.
Just as it was a friend of Carnap who rebutted many of Carnaps
claims, so it was a friend of Quine, Morton White, who pointed
out that with the positivists criterion of cognitive significance
demolished, the whole basis for the positivists claim that ethical
sentences are nonsensical was also gone. First of all, the notion
of an observation is extremely unclear: Why isnt I saw X steal
Ys wallet? an observation sentence? (Incidentally it would
seem that stealing is a fairly clear notion by comparison to being
an observable predicate [White 1956, 109].) Secondly, after
pointing out that once we accept the holistic view of confirmation originally proposed by Pierre Duhem, according to which
scientific explanation and prediction puts to the test a whole
body of beliefs, rather than the one which is ostensibly under test
291
Ethnolinguistic Identity
292
ETHNOLINGUISTIC IDENTITY
According to Epicurus (Letter to Herodotus), the different languages of the world arose historically from the differences in feelings and sensory perception among peoples: [M]ens natures
according to their different nationalities [ethn] had their own
peculiar feelings and received their peculiar impressions, and
so each in their own way emitted air formed into shape by each
of these feelings and impressions, according to the differences
made in the different nations by the places of their abode as well
(Bailey 1926, 756). The idea of a strict linkage between a people and its language is also found in the Book of Genesis (10:5),
when, after the Flood, Noahs descendants spread out over the
earth, every one after his tongue, after their families, in their
nations (King James Version).
The Bible is the most direct source of the modern (post-Renaissance) conception of the nation as a people linked by birth,
language, and culture and belonging to a particular place. This
had not been the general European way of thinking prior to the
Renaissance, when religious belonging provided a first division
among peoples and dynastic rule a second. Language meant
Latin, pan-European, and largely insulated from the vernacular
dialects spoken by most people going about their daily lives.
These vernaculars were not thought of as language or as having
any importance beyond the practical needs of communication, whereas Latin was the sacred vehicle of divine rites and
divine knowledge.
As the Reformation increased access to the Bible, the sense
of national belonging and the nation-language nexus spread
(see nationalism and language). Concern arose for vernaculars to be raised to the status of the language, making
them eloquent, able to fulfill some of the functions previously
Typology of Events
Event types are often distinguished based on Z. Vendlers (1957)
classic typology of lexical aspect or Aktionsarten, a German term
for kinds of action, which groups verbs into subclasses based
on their temporal features. These subclasses or event types are
processes (e.g., activities such as walk and run), accomplishments (events that culminate by the use of temporal adverbials,
e.g., build or cook in an hour), and achievements (instantaneous
events that finish in a short time period, e.g., win and find). Event
itself is part of a larger notion called situations, divided into
two categories: events and states (e.g., know and love) (Mani,
Pustejovsky, and Gaizauskas 2005). Hence, the term situation
aspect is often used in preference to the term lexical aspect.
293
Evidentiality
Future Trends
Beyond the current state, two main trends may be noted. First,
event structure is currently studied with reference to words and
sentences, mostly in isolation. In the future, we need to study it
in context, such as in speech and texts. D. Townsend and colleagues (2003) lead this trend. Second, current research mostly
studies how different types of objects and other functions influence telicity. But we also ought to look at how event structure in
complex verbal constructions, such as serial verbs, is computed.
lexical-functional grammar analyses like Bodomo (1993,
1997) and Alsina, Bresnan, and Sells (1997) lead this trend.
Tenny and Pustejovsky (2000), Mani, Pustejovsky, and
Gaizauskas (2005), and Dolling, Heyde-Zybatow, and Schafer
(2007) are further recent book-length readings that put most of
these issues in perspective.
Adams Bodomo
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Alsina, A. Joan Bresnan, and Peter Sells. 1997. Complex Predicates.
Stanford, CA: CSLI.
Bodomo, A. 1993. Complex predicates and event structure: An integrated analysis of serial verb constructions in the Mabia languages of
West Africa. Working Papers in Linguistics 20, Trondheim, Norway.
. 1997. A conceptual mapping theory for serial verbs. In On-line
Proceedings of LFG97, ed. M. Butt and T. King, CSLI, Stanford
University. Available online at: http://www-csli.stanford.edu/publications/LFG2/bodomo-lfg97.html.
Borer, H. 2005. Structuring Sense. Vol. 2. The Normal Course of Events.
Oxford: Oxford University Press.
Butt, Miriam, and Gillian Ramchand. 2001. Complex aspectual structure in Hindi/Urdu. Oxford Working Papers in Linguistics, Philosophy
and Phonetics 6: 130. M. Liakata, B. Jensen, and D. Maillat edited this
volume.
Dolling, Johannes, Tatyana Heyde-Zybatow, and Martin Schafer. 2007.
Event Structures in Linguistic Form and Interpretation. Berlin: Walter
de Gruyter.
294
EVIDENTIALITY
This is a grammatical category that has source of information as
its primary meaning whether the narrator actually saw what is
being described, made inferences about it based on some evidence, or was told about it, and so on. Tariana, an Arawak language from Brazil, has five evidentials marked on the verb. If I
saw Jos play football, I will say Jos is playing-naka, using the
visual evidential. If I heard the noise of the play (but didnt see it),
I will say Jos is playing-mahka, using the nonvisual. If all I see
is that Joss football boots are gone and so is the ball, I will say
Jos is playing-nihka, using the inferential. If it is Sunday and
Jos is not home, the thing to say is Jos is playing-sika since
my statement is based on the assumption and general knowledge that Jos usually plays football on Sundays. And if the information was reported to me by someone else, I will say Jos is
playing-pidaka, using the reported marker. Omitting an evidential results in an ungrammatical and highly unnatural sentence.
About a quarter of the worlds languages have some grammatical marking of information source. The systems vary in their
complexity. Some distinguish just two terms. An eyewitness versus non-eyewitness distinction is found in Turkic and Iranian
languages. Other languages mark only the nonfirsthand information, for example, Abkhaz, a northwestern Caucasian language. Numerous languages express only reported, or hearsay,
Evidentiality
information, for example, Estonian. Quechua languages have
three evidentiality specifications: direct evidence, conjectural,
and reported.
Systems with more than four terms have just two sensory
evidentials and a number of evidentials based on inference and
assumption of different kinds; these include Nambiquara languages, from Brazil, and Foe and Fasu, of the Kutubuan family
spoken in the Southern Highlands of Papua New Guinea.
The terms verificational and validational are sometimes used
in place of evidential. French linguists employ the term mdiatif
(Guentchva 1996). A summary of work on recognizing this
category, and naming it, is in Jacobsen (1986) and Aikhenvald
(2004).
Evidentiality does not bear any straightforward relationship
to truth, the validity of a statement, or the speakers responsibility. The truth value of an evidential may be different from
that of the verb in its clause. Evidentials can be manipulated to
tell a lie: One can give a correct information source and wrong
information, as in saying He is dead-reported when you were
told that he is alive, or correct information and wrong information source, as in saying He is alive-visual when, in fact, you
were told that he is alive but did not see this. The ways in which
semantic extensions of evidentials overlap with modalities
and such meanings as probability or possibility depend on the
system and on the semantics of each individual evidential
term. In many languages (e.g., Quechua, Shipibo-Konibo, or
Tariana, all from South America), markers of hypothetical and
irrealis modality can occur in conjunction with evidentials on
one verb or in one clause. This further corroborates their status
as distinct categories.
Nonvisual and reported evidentials used with the first person
often refer to uncontrolled spontaneous action or have overtones
of surprise, known as mirative.
Every language has some lexical way of referring to information source, for example, English reportedly or allegedly. Such
lexical expressions may become grammaticalized as evidential
markers. Nonevidential categories may acquire a secondary
meaning relating to information source. Conditionals and other
nondeclarative moods may acquire overtones of uncertain information obtained from some other source for which the speaker
does not take any responsibility; the best-known example is the
French conditional. Past tense and perfect aspect acquire
nuances of nonfirsthand information in many Iranian and
Turkic languages, and so do resultative nominalizations and
passives. The choice of a complementizer, or a type of complement clause, may serve to express meanings related to the way
in which one knows a particular fact. In English, different complement clauses distinguish an auditory and a hearsay meaning
of the verb hear: Saying I heard Brazil beating France implies
actual listening, whereas I heard that Brazil beat France implies
a verbal report of the result. These evidential-like extensions are
known as evidentiality strategies. Historically, they may give rise
to grammatical evidentials.
The maximal number of evidentials is distinguished in statements. The only evidential possible in commands is the reported,
to express command on behalf of someone else: eat-reported!
means eat following someones command! Evidentials often
come from grammaticalized verbs. The verb of saying is
Evolutionary Psychology
a frequent source for reported and quotative evidentials, and
the verbs feel, think, hear can give rise to a nonvisual evidential.
Closed-word classes deictics (see deixis) and locatives may
give rise to evidentials, both in small and in large systems.
Evidentials vary in their semantic extensions, depending on
the system. Reported information often has overtones of probability or unreliability, while visual evidentials may develop
meanings of certainty. They can be extended to denote the
direct participation, control, and volitionality of the speaker.
morphemes marking tense, aspect, mood, modality, and evidentiality may occur in the same slot in the structure of a highly
synthetic language.
Evidentiality is a property of a significant number of linguistic
areas, including the Balkans, the Baltic area, India, and a variety
of locations in Amazonia. Evidentials may make their way into
contact languages, as they have into Andean Spanish. The texts
genre may determine the choice of an evidential. Traditional
stories are typically cast in reported evidential. Evidentials can
be manipulated in discourse as a stylistic device. Switching
from a reported to a direct (or visual) evidential creates the effect
of the speakers participation and confidence. Switching to a
nonfirsthand evidential often implies a backgrounded aside.
Evidentiality is interlinked with conventionalized attitudes to
information and precision in stating its source.
Alexandra Aikhenvald
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Aikhenvald, Alexandra Y. 2004. Evidentiality. Oxford: Oxford University
Press.
Aikhenvald, Alexandra Y., and R. M. W. Dixon, eds. 2003. Studies in
Evidentiality. Amsterdam: John Benjamins.
Barnes, J. 1984. Evidentials in the Tuyuca verb. International Journal of
American Linguistics 50: 25571.
Guentchva, Z., ed. 1996. Lnonciation mdiatise. LouvainParis: ditions Peeters.
Jacobsen, William H., Jr. 1986. The heterogeneity of evidentials in
Makah. In Evidentiality: The Linguistic Coding of Epistemology, ed.
Wallace L. Chafe and Johanna Nichols, 328. Norwood, NJ: Ablex. See
other papers therein.
EVOLUTIONARY PSYCHOLOGY
This term is used in several different, related senses. Among
behavioral, social, and cognitive scientists, it properly refers to
a new scientific paradigm or framework, together with the discipline that has grown up around this framework, and the body
of knowledge produced by the researchers working within that
framework. Some scholars outside the field, as well as many
journalists and lay people, use it more loosely to refer to any finding, speculation, or discussion that links evolution and behavior,
whether well informed or not. Evolutionary psychology as both
a research framework and a discipline is organized around the
proposition that the design features of the mechanisms comprising a species psychology reflect the character of the adaptive
problems they evolved to solve. This proposition was uncontroversial when applied by biologists to other species (e.g., Williams
1966). However, it generated significant debate and opposition
once it began to be applied to humans, who because of culture,
295
Evolutionary Psychology
intelligence, language and complexly variable social systems
appear notably different from other species (Sahlins 1977).
The field shares some tenets with early Chomskyan proposals that the human mind contains numerous mental organs
specialized for carrying out different cognitive tasks, such as a
language acquisition device (Chomsky 1965). The antifunctionalist strain in Chomskys thinking led him to largely set
aside natural selection for communicative functions in his discussions of language (Chomsky 1972). In contrast, evolutionary
psychologists such as the psycholinguist Steven Pinker (1994),
argue that the existence of mental organs can only be explained
as the consequence of natural selection. This is because selection
is the only process known to science that builds complex functional systems into the designs of organisms (Williams 1966). By
this standard, the intricate functional interdependence of the
various cognitive mechanisms underlying language provides
very strong evidence for the organizing role of natural selection
in constructing such mechanisms (Pinker and Bloom 1992).
Evolutionary psychology began to emerge in the 1970s and
1980s when a small number of researchers tried to synthesize
several distinct research orientations in a mutually consistent
way (Tooby and Cosmides 1992). The most important of these
orientations were cognitive science, with its commitment to
information-processing descriptions of psychological mechanisms; modern primatology, huntergatherer studies, and
paleoanthropology, which together offered the prospect of characterizing the conditions in which humans evolved; evolutionary
biology (including behavioral ecology, sociobiology, ethology,
and evolutionary game theory); and neuroscience, with its
prospect of discovering the physical implementation of cognitive mechanisms. Evolutionary psychologists argued that cognitive mechanisms were, ipso facto, biological adaptations,
a proposition that inevitably connected cognitive science to
evolutionary biology. If cognitive mechanisms are adaptations,
they then must exhibit an evolved organization, have an evolutionary history, and have been naturally engineered to carry out
evolved functions. Most importantly, the identification of cognitive mechanisms with adaptations allowed the entire technical
apparatus developed within biology concerning adaptations to
be imported and validly applied to cognitive science.
Evolutionary psychologists start from the premise that
the brain, like our other organs, is the product of evolution.
Specifically, the brain is viewed as an information-processing
organ that evolved over evolutionary time in order to regulate
behavior in an adaptively successful way. In a world filled with
the disordering force of entropy, biologists and physicists recognize that natural selection is the only known natural physical process that can push the designs of organisms uphill into
functionally organized systems. It follows that whatever functional organization there is to be found in the design of the brain
reflects the history of selection that acted ancestrally on the
species. Evolutionary psychologists use the cause-and-effect
relationships between ancestral selection pressures and the
resulting functional architectures of the brains mechanisms as
one powerful new tool to guide scientific discovery. On this view,
the structure of each psychological mechanism should reflect
the actions of the selection pressures that built it. Consequently,
by considering ancestral adaptive problems, evolutionary
296
Evolutionary Psychology
psychologists to derive hypotheses about the design of human
information-processing mechanisms from the large preexisting body of theories already developed and empirically tested
within modern evolutionary biology. For example, evolutionary biologists know that for organisms like humans, mating with
close relatives causes genetic defects to express themselves at far
higher rates in the incestuously produced children. This has led
evolutionary psychologists a) to the general prediction that natural selection had built a program in humans designed to identify close genetic relatives; b) to detailed predictions about the
cues that the program would use to identify genetic relatives; and
c) to detailed predictions about how this kin detection program
would be coupled to increased sexual aversion to individuals
it identified as genetic relatives (as well as increased altruism,
as predicted by kin selection theory). The analysis of ancestral
selection pressures and huntergatherer conditions made it possible to design studies that could test (and did confirm) these
propositions. These studies, in turn, mapped the informationprocessing architecture of these functionally specialized programs (Lieberman, Tooby, and Cosmides 2007). In contrast, the
disregard by sociocultural anthropologists (and Freudians) of the
selection pressures that select strongly against incest prevented
them from discovering the existence of these evolved mechanisms. Once a mechanism is mapped, its population-level social
and cultural expressions can also be analyzed such as moral
attitudes about incest in the case of kin detection and human linguistic variation in the case of language.
Evolutionary psychology originally emerged among anthropologists, cognitive scientists, biologists, and psychologists,
although it has subsequently diffused into many other disciplines. Evolutionary psychology is not a subfield of psychology,
and it is not devoted to the study of a specific class of phenomena. Rather, it is an approach to the behavioral, social, cognitive,
and neural sciences that can be applied to any of the topics they
deal with. Originally reacting against the mutually contradictory
claims about the mind and human nature advanced in different
disciplines, evolutionary psychologists constructed what they
argue is a logically integrated scientific framework that attempts
to reconcile into a single body of knowledge the results drawn
from all relevant fields. Its advocates view it as an interdisciplinary nucleus around which a single unified theoretical and
empirical behavioral science is being crystallized. Of course, not
everyone in behavioral science agrees, with disagreements ranging from disputes over specific analyses to broader rejection of
the program, often in favor of culturalist and social constructionist views.
A second feature that distinguishes evolutionary psychology
is the importance it places on achieving information-processing
descriptions of the designs of evolved mechanisms, rather than
stopping at behavioral or neuroscience descriptions. Along with
most cognitive scientists, evolutionary psychologists believe that
the brain, like any other computational system, can usefully be
mapped both in physical terms (which, for the brain, means in
neurophysiological and neuroanatomical terms) and also complementarily in information-processing terms. Evolutionary
psychologists go on to stress that the brain and its subsystems
evolved as an organ (or set of organs) of computation: The brain
evolved in order to regulate behavior and physiology adaptively
based on information it is exposed to. Because the evolved function of a neural (or psychological) mechanism is inherently
computational (i.e., as a program mapping informational inputs
to outputs), the only form of description that can accurately
characterize how its organization solves its adaptive problem
is an information-processing description. Physical descriptions of brain subsystems cannot, by their nature, fully capture
the information-processing interrelationships that embody the
function of an evolved program (mechanism, adaptation, etc.).
So, for example, however interesting it is to identify the brain
regions implicated in various aspects of language processing, it
is still important to develop a parallel account in terms of computational steps (data structures, operations, etc.). Similarly,
simply observing that humans behaviorally tend to avoid incest
inside the nuclear family is very different from having mapped
the information-processing steps in the evolved programs that
take prespecified cues to kinship as input, compute from them
magnitudes that capture estimated genetic relatedness, and then
pass these magnitudes into the sexual-choice motivational subsystem, where they generate sexual disgust at mating with those
it identifies as genetic relatives.
A third difference in perspective between evolutionary psychologists and most other behavioral scientists is in how numerous and functionally specialized they expect the psychological
mechanisms of a species to be. For most of the last century, the
majority view among learning theorists, cognitive scientists, and
neuroscientists has been that the psychological mechanisms
that operate on experience to produce knowledge are likely to
be small in number, and to be primarily content independent
and general purpose (Pinker 2002; Tooby and Cosmides 1992).
Content independence means that a cognitive procedure (such
as association formation in connectionism) operates in the
same way regardless of the content it is processing. Hence, on
this view, the same cognitive procedures are expected to operate on all contents uniformly, whether language, fighting, eating,
sex, family interactions, or intergroup conflict. This blank slate or
environmentalist view can be expressed by comparing the operation of learning mechanisms or cognitive mechanisms to the
operation of a tape recorder that processes all sounds uniformly,
regardless of their meaning: The content that ends up on the tape
reflects only the content present in the environment, and nothing in the tape-recording machinery itself introduces content of
its own that was not present in the environment.
From a selectionist perspective, however, such a blank-slate
viewpoint seems extremely implausible, as well as inconsistent
with what is known about the cognitive architectures of nonhumans (Gallistel 1990). Mutations for specialized design features
that exploit the rich recurrent structure of particular problem
domains should spread by natural selection whenever they costeffectively improve the organisms propensity to solve important
adaptive problems in a fitness-promoting way. That is, if there is
a particular set of cues that solves the problem of kin detection,
then the mind could evolve a specialization that is designed to
take only those cues as input. For a problem-solving strategy to
be applied generally across contents, it cannot employ problemsolving shortcuts that work only on particular problem subsets,
such as grammar acquisition, depth perception, kin detection,
or mate selection. Hence, evolutionary psychologists consider it
297
Evolutionary Psychology
likely that the mind solves the diverse computational problems
posed by stereopsis, color vision, echolocation, face recognition, object mechanics, navigation, and reasoning about social
exchange by using at least some principles and operations that
are particular to each respective domain. Evolutionary psychologists argue that evolved specializations that are activated only by
certain content domains or adaptive problems seem virtually
inevitable, rather than implausible or exceptional outcomes of
the evolutionary process. This is because selection inherently
favors efficiency and puts no weight per se on uniformity or simplicity (Tooby and Cosmides 1992).
Moreover, unlike a tape recorder, the designs of such evolved
psychological mechanisms might be expected to regularly introduce particular contents, motivations, interpretations, and
conceptual primitives into the human mind that are not simply
derived from the environment. From an engineering perspective, it is easy to see how such reliably developing contents could
enhance adaptive performance. For example, the environmental
regularity of venomous snakes posed an evolutionarily long-enduring adaptive problem. This regularity appears to have selected
for an evolved computational device implemented in the brains
of African primates (including humans). This adaptation contains a psychophysical specification of snakes linked to a system
that motivates snake avoidance. Additionally, this avoidance is
up-regulated to the extent that the individual is exposed to conspecifics who display fear toward snakes (hman and Mineka
2001). This depends on mental content about snakes being built
into the mechanism. The human mind is suspected to contain
neurocomputational versions of what philosophers would once
have called innate ideas, such as snake, spider, mother, predator,
food, word, verb, agency, object, and patient (Tooby, Cosmides,
and Barrett 2005). By augmenting the cognitive architecture in
such a fashion, natural selection could supercharge perceiving,
learning, reasoning, and decision making in evolutionarily consequential domains.
At a minimum, evolutionary psychologists expect that in addition to whatever general-purpose cognitive machinery humans
have, we should also be expected to have a wide array of domainspecific mechanisms, including specialized learning mechanisms. So, for example, although the snake phobia system, the
kin detection mechanism, and the language acquisition system
are all learning mechanisms, they are each specialized only for
their particular type of content (snakes linked to fear intensity,
kinship cues linked to incest aversion and altruistic motivation,
and language inputs linked to linguistic competence). For this
reason, evolutionary psychologists do not regard learning as constituting an alternative explanation for the claim that a particular kind of behavioral output was shaped by evolution. Evidence
that something is learned is not in the least inconsistent with the
claim that much of the knowledge produced was supplied by
specialized learning mechanisms permeated with evolved content. Critics of evolutionary psychology view its multiplication
of hypothesized cognitive mechanisms (e.g., specializations for
language acquisition, kin detection, mate selection, and so on)
to be unparsimonious. Evolutionary psychologists respond that
although parsimony may have been a useful principle in physics,
evolutionarily engineered systems are not designed to be simple
but, rather, to be adaptively effective.
298
Evolutionary Psychology
neo-associationistic mechanisms, or other general-purpose
alternatives (Pinker 1994). First, computationally intricate linguistic capacities develop precocially far earlier than comparable cognitive achievements in other domains. Second, genetic
and developmental conditions can doubly dissociate language
and general intelligence (i.e., one can speak well with low intelligence and be unable to speak but have otherwise unimpaired
intelligence). Third, underneath linguistic variability are design
features like linear order, constituency (see constituent
structure), predicate-argument structure, case markers,
morphophonemic rules, and phonological rules that are a) universal and b) well designed to communicate propositional information, such as who did what to whom, but poorly designed for
many other cognitive tasks, such as statistical induction, imagery, face recognition, and so on (see phonology, universals
Exemplar
Generation of Culture, ed. J. Barkow, L. Cosmides, and J. Tooby, 19136.
New York: Oxford University Press.
Tooby, J., L. Cosmides, and H. C. Barrett. 2005. Resolving the debate
on innate ideas: Learnability constraints and the evolved interpenetration of motivational and conceptual functions. In The Innate
Mind: Structure and Content, ed. P. Carruthers, S. Laurence, and S.
Stich, 30537. New York: Oxford University Press.
Tooby, John, and I. DeVore. 1987. The reconstruction of hominid
behavioral evolution through strategic modeling. In The Evolution of
Primate Behavior: Primate Models, ed. Warren Kinsey, 183237. New
York: SUNY Press.
Williams, George C. 1966. Adaptation and Natural Selection: A Critique
of Some Current Evolutionary Thought. Princeton, NJ: Princeton
University Press.
EXEMPLAR
This term occurs importantly in research and theorization in category identification, recognition, categorization, and learning. It
is used interchangeably with the terms instance or item across
various strands of research, including psychology, religion, and
history.
Within the context of category learning, for instance, the term
exemplar refers to a specific instance, such as a specific cat to
which a parent points when teaching a child the concept and
name of cat. Alternatively, during remediation of language skills
in children with severe disabilities, researchers have utilized various exemplars of graphical symbols to improve communication
(Schlosser 2003). In studies examining category relearning in
individuals who have suffered brain damage, training in naming
of a subset of exemplars results in improved naming of untrained
exemplars within the category (Kiran 2007).
Within the topic of categorization of semantic concepts,
the terms specific usage comes in the context of exemplar
theory. Briefly, this theory suggests that a category is represented by a collection of members (exemplars) that have been
previously encountered, experienced, and stored as unique
and individual memory traces. A new object/item is judged as a
member of a given category provided that it is sufficiently similar
to the stored exemplars (Komatsu 1992). This specific interpretation of exemplar is at odds with an alternate view of categorization, namely, the prototype theory, which suggests that a
category is represented in terms of a single summary representation (i.e., a prototype).
Not all theorists agree that exemplar and prototype models
are competitors; there is yet another class of models according
to which categorization decisions are made using exemplars,
although the effect of using exemplars necessitates the creation
of abstractions that can be later applied to novel exemplars (Ross
and Makin 1999). Similarly, some connectionist networks
assume that a category is represented by summary information
across the entire network and, depending upon the input provided, specific connection strengths in the network have greater
influence on the overall activation (Knapp and Anderson 1984).
Finally, the interpretation of the term exemplar can also be
influenced by the level of category structure. As Edward Smith
and Douglas Medin (1999) argue, the term can refer to a specific instance of the concept (e.g., your favorite blue jeans
in the category clothing) or to a subset of the concept (blue
299
Exemplar Theory
jeans). Further, whereas experimental investigations of exemplars typically refer to them as a basic level concepts (e.g.,
apple), exemplars can also refer to an individual entity such as
Macintosh apple, which is a subordinate concept.
Swathi Kiran
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Knapp, A., and J. Anderson. 1984. Theory of categorization based
on distributed memory strorage. Journal of Experimental
Psychology: Learning, Memory and Cognition 10: 61637.
Kiran, Swathi. 2007. Semantic complexity in the treatment of naming
deficits. American Journal of Speech Language Pathology 16: 112
Komatsu, Lloyd. 1992. Recent views of conceptual structure.
Psychological Bulletin 112: 50026.
Ross, Brian, and Valarie Makin. 1999. Prototype versus exemplar models in cognition. In The Nature of Cognition, ed. R. Sternberg, 20542.
Cambridge, MA: MIT Press.
Schlosser, Ralf. 2003. The Efficacy of Augmentative and Alternative
Communication: Toward Evidence-Based Practice. Amsterdam and
Boston: Elsevier.
Smith, Edward, and Douglas Medin. 1999. The exemplar view. In
Concepts: Core Readings, ed. E Margolis and S. Laurence, 20721.
Cambridge, MA: MIT Press.
EXEMPLAR THEORY
An important goal of linguistic theory has been to develop explicit
approaches to describing, modeling, and explaining linguistic
behavior. Certainly, most familiar to linguists are the rule-based
models that derive abstract rules from linguistic exemplars, then
use those rules to predict linguistic behavior. Such rule systems
are readily shown to be empirically inadequate, both diachronically, as in the shift from digged to dug as the past tense of dig,
and synchronically (see synchrony and diachrony), as in
the overgeneralization (see overregularizations) of glew
as the past tense of glow in place of glowed. Thus, rule-based
models of language behavior must also incorporate some sort of
component that can account for analogical interactions among
linguistic items.
Generally speaking, there are two broad categories of analogical models under investigation in current linguistic research.
One group consists of approaches that use linguistic exemplars
to derive an analogical system but which then do not consult
those individual exemplars of linguistic experience further in
predicting linguistic behavior. Best known among this category of analogical models are the connectionist approaches
to language. Connectionist models typically pool their training
input into schematic, prototype-like representations of a category that do not retain individualizing information about the
exemplars used to train the models. Such representations, however, make the models empirically inadequate in two respects.
First, connectionist models incorrectly predict behaviors such
as categorization and response times in terms of similarity to
the prototype encoded in the network, rather than in terms of
similarity to individual exemplars. Second, there is abundant,
and growing, evidence that memories for individual linguistic experiences do influence subsequent linguistic behaviors.
Research has derived clear evidence of specific exemplars
300
subsequently influencing phonetic and phonological output (Pierrehumbert 2001), lexical and morphological output
(Goldinger 1997 and Bybee 2002), and childrens manipulations of syntactic structures (Tomasello 2003; see also syntax,
acquisition of).
Exemplars are not prototypes. They are individual instances of
linguistic usage retained in memory. Given the empirical necessity for incorporating memory for exemplars into models of language behavior, the crucial research question in exemplar-based
approaches becomes whether they alone can account for the
spectrum of linguistic behaviors or whether there remains independent empirical justification for the rule-based components.
Exemplar-based models of language are founded on the simple notion that in language use, speakers will compare a current
linguistic expression and its context (linguistic and nonlinguistic)
with their personal collections of memories for similar expressions and then choose at least one of the tokens in memory an
exemplar as the basis for deciding how to interpret or otherwise operate on that expression. Usually, the token(s) selected
will be similar to the input currently being considered and its
context. Such models imply that the brain stores vast inventories of memories for individual episodes of linguistic experience
and that it employs some procedure for comparing the features
of the new input or current context to the features of those
remembered exemplars, and then has some basis for choosing
one of those exemplars as the model for an analogical response
(or interpretation).
Given the empirical evidence that memories for individual
exemplars of previous linguistic behavior influence current
linguistic behavior, we restrict our discussion in this entry to
explicitly defined exemplar-based approaches that use actual
exemplars (typically gleaned from linguistic corpora) to predict
linguistic behavior. The approaches discussed here are all computationally based and have actually been tested against real
linguistic behavior. The algorithms are also publicly available to
researchers.
In the three approaches that follow, the exemplars are
retained and directly used to predict linguistic behavior. A data
set of relevant exemplars is constructed from actual linguistic
corpora; then an algorithm is applied that compares the new
input to the exemplars in the data set and selects certain of those
exemplars while lessening or even zeroing out the chances of
other exemplars in the data set being used. Typically, the exemplars in the data sets are composed of outcomes associated with
various linguistic variables or features. A prediction is then made
for the outcome defined by an input set of variables (the given
context). Normally, exemplars that are in some sense closer to
the given context have a higher chance of being selected, but
sometimes the algorithm may select more distant exemplars. In
other words, these approaches sometimes allow exemplars that
are not nearest neighbors that is, most similar to be used.
Exemplar Theory
nouns (Nakisa and Hahn 1996). The GCM determines the conditional probability of assigning a given linguistic form say, the
base form of a noun to a particular form class, for example, a
particular plural form. It does so by comparing the features of the
test form with the weighted sum of those features in all the exemplars of one response category, divided by the weighted sum of
those features across the exemplars of all the possible response
categories. The model also factors in a response bias value for the
different categories. Thus, it arrives at a conditional probability
for choosing any one response over the alternatives.
In the application of the GCM to nonlinguistic data, the
accuracy of the models predictions depends crucially upon
the weightings assigned to the different stimulus features and
the response biases for the alternative categories. Typically,
both are determined ahead of time by constructing a confusion
matrix for the exemplars to be used. The resulting weightings
(said to account for the effects of selective attention during training) and response biases then apply only to the given data set
of exemplars. The feature weightings determined ahead of time
for a given data set are equivalent to the information gain values
described for the memory based learning model that follows and
are subject, therefore, to the same theoretical criticisms in that
they must be calculated ahead of time for a given data set and do
not generalize to a new data set. In applying the GCM to natural
language data, Ramin Nakisa and Ulrike Hahn (1996), of course,
were not able to obtain feature weightings for German nouns in
native speakers of the language, and the model therefore did not
perform as well as a competing connectionist model.
the best results for a particular task and a particular data set, and
although the differences among them are usually not great, there
appears to be no principled basis for choosing one measure over
another.
One important contribution of the MBL studies to exemplarbased modeling theory is that reducing the size of the data set by
omitting very low frequency exemplars, redundant exemplars,
and very rare but exceptional exemplars actually reduces the
level of correct predictability. Thus, MBL researchers now recognize the need to construct large, complete data sets in order to
maximize overall predictability.
301
Extinction of Languages
to assume that access to exemplars is not always available and, in
general, can be considered a random phenomenon.
One important result from AM is that one cannot assume in
advance which variables are significant and thus ignore the others. Often, the potential value of a variable remains latent until
the model is required to predict the outcome for an appropriate given context. This kind of result can occur when gangs of
non-neighbors are called upon to predict the behavior of a given
context.
Royal Skousen and Steve Chandler
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bybee, Joan. 2002. Phonological evidence for exemplar storage of multiword sequences. Studies in Second Language Acquisition 24: 21521.
Daelemans, Walter, and Antal van den Bosch. 2005. Memory-Based
Language Processing. Cambridge: Cambridge University Press.
Goldinger, Stephen D. 1997. Words and voices: Perception and production in an episodic lexicon. In Talker Variability in Speech Processing,
ed. K. Johnson and J. W. Mullennix, 3365. San Diego, CA: Academic
Press.
Nakisa, Ramin, and Ulrike Hahn. 1996. Where defaults dont help: The
case of the German plural system. In The Proceedings of the 18th
Annual Conference of the Cognitive Science Society, 17782. Hillsdale,
NJ: Lawrence Erlbaum.
Nosofsky, Robert M. 1992. Exemplar-based approach to relating categorization, identification, and recognition. In Multidimensional
Models of Perception and Cognition, ed. F. G. Ashby, 36393. Hillsdale,
NJ: Lawrence Erlbaum.
Pierrehumbert, Janet B. 2001. Exemplar dynamics: Word frequency, lenition and contrast. In Frequency Effects and Emergent Grammar, ed.
J. Bybee and P. Hopper, 119. Amsterdam: John Benjamins.
Skousen, Royal. 1989. Analogical Modeling of Language. Dordrecht:
Kluwer.
Skousen, Royal, Deryle Lonsdale, and Dilworth S. Parkinson. 2002.
Analogical Modeling: An Exemplar-Based Approach to Language.
Amsterdam: John Benjamins.
Tomasello, Michael. 2003. Constructing a Language: A Usage-Based
Theory of Language Acquisition. Cambridge: Harvard University Press.
EXTINCTION OF LANGUAGES
An increasing number of books, scholarly articles, and media
reports have predicted that 5090 percent of the worlds approximately 6,900 some languages may be at risk of extinction within
the next hundred years (see, for example, Krauss 1992; Nettle and
Romaine 2000; Crystal 2000; Abley 2003). This alarming figure does
not include dialects because no one knows exactly how many
languages and dialects there are, and there are no clear criteria for
distinguishing between language and dialect (see Wolfram and
Schilling-Estes 1998 for discussion of dialect endangerment).
Estimates of the number of languages in danger of extinction
vary depending on the criteria used to assess risk. UNESCOs
World Atlas of the Worlds Languages in Danger of Disappearing
(2001) estimates that 50 percent of languages may be in various
degrees of endangerment while Michael Krauss (1992) believes
that up to 90 percent may be threatened. More research is needed
in order to understand the role of various factors, such as size (i.e.,
number of speakers), status, function, and so on, in supporting
or not supporting languages. Most languages are unwritten, not
302
Family Resemblance
Table 1. Percentages of languages according to continent of origin
having fewer than indicated number of speakers
Continent/
Region
< 150
< 1000
< 10,000
< 100,000
<1
Million
Africa
1.7
7.5
32.6
72.5
94.2
Asia
5.5
21.4
52.8
81.0
93.8
Europe
1.9
9.9
30.2
46.9
71.6
North
America
22.6
41.6
77.8
96.3
100
Central
America
6.1
12.1
36.4
89.4
100
South
America
27.8
51.8
76.5
89.1
Australia/
Pacific
22.9
60.4
93.8
99.5
World
11.5
30.1
59.4
83.8
94.1
100
95.2
maintenance, and transmission of human knowledge, the prospect of extinction raises critical issues about the survival of this
knowledge. Loss of linguistic diversity also threatens the scientific study of language by diminishing the range of structures
for constructing hypotheses about universals.
Suzanne Romaine
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Abley, Mark. 2003. Spoken Here: Travels among Threatened Languages.
Toronto: Random House of Canada.
Crystal, David. 2000. Language Death. Cambridge: Cambridge University
Press.
Gordon, Raymond G., Jr., ed. 2005. Ethnologue: Languages of the World.
15th ed. Dallas: SIL International. Available online at: http://www.
ethnologue.com/.
Krauss, Michael. 1992. The worlds languages in crisis. Language
68: 410.
Nettle, Daniel, and Suzanne Romaine. 2000. Vanishing Voices: The
Extinction of the Worlds Languages. Oxford: Oxford University Press.
UNESCO. 2001. World Atlas of the Worlds Languages in Danger of
Disappearing. Paris: UNESCO.
Wolfram, Walt, and Natalie Schilling-Estes. 1998. Endangered dialects: A
neglected situation in the endangerment canon. Southwest Journal of
Linguistics 14: 11731.
F
FAMILY RESEMBLANCE
Ludwig Wittgensteins Philosophical Investigations ([1953] 1958,
hereafter PI) is one of the most influential and well-known
texts of twentieth-century philosophy. A number of terms
have been introduced into the philosophical lexicon by this
work. Alongside language-game, the related term family
303
Feature Analysis
The suggestion of family resemblance as a way of understanding the similarities between different language-games or different
uses of words helps bring to consciousness our unacknowledged
enthrallment to a picture of language-as-necessarily-havingan-essence, that is, a picture of something being essential to
all instances of language use and to all uses of a word (such as
game). Family resemblance, as with all the terms introduced
by Wittgenstein in PI, serves, and should stay subservient to,
the therapeutic task (pace authors such as R. Bambrough and E.
Rosch and C. B. Mervis). Methodological readings and employments of the term, as in Bambrough (1960) and in Rosch and
Mervis (1975) (see prototype), therefore, fundamentally misunderstand the language-game in which family resemblance has
its home in Wittgensteins work.
Phil Hutchinson
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Baker, Gordon. 2004. Wittgensteins Method: Neglected Aspects.
Oxford: Blackwell. Hugely important text on Wittgensteins vision of
philosophy.
Bambrough, R. 1960. Universals and family resemblances. Proceedings
of the Aristotelian Society 61: 20722.
Rosch, E., and C. B. Mervis. 1975. Family resemblances: Studies in the
internal studies of categories. Cognitive Psychology 7: 573605.
Wittgenstein, Ludwig. 1922. Tractatus Logico-Philosophicus. London:
Routledge.
. [1953] 1958. Philosophical Investigations. Oxford: Blackwell.
. 1958. The Blue and Brown Books: Preliminary Studies for the
Philosophical Investigations. Oxford: Blackwell
FEATURE ANALYSIS
Feature analysis extends to the cognitive domain the interest of
early Greek philosophers such as Democritus and Plato, in identifying the fundamental building blocks (atoms) of the physical
world (Greenberg 1967). Cognitive feature analysis starts from
the observation that animals appear to organize their perceptual
worlds in terms of finite sets of discrete abstract elements rather
than simply storing and manipulating unanalyzed streams of
continuous gradient stimuli. Evidence for the existence of discrete representational features governed by abstract combinatorial principles (as opposed to concrete [e.g., phonetic]
categories emerging from lower-level gradient processes) comes
primarily from areas related to phonology but also from other
domains of human cognition, such as vision (Pylyshyn, Blaser,
and Holcombe 2000); computation of object similarity (Tversky
1977), induction (Sloman 1993), typicality, asymmetry, diversity (Heit 1997), speech perception (Stevens 2002; Poeppel,
Idsardi, and van Wassenhove 2007), sign language phonologies, morphology, semantics, alphabet processing, and
vision and object perception (see Morgan 2003 and Vaux 2008
for references).
First explored formally by A. M. Bell (1867) for alphabets,
Alfred Kroeber (1909) for kinship systems, and A. G. Bell (1911)
for speech, feature theory was most famously developed for
phonology by Roman Jakobson, Serge Karcevsky, and Nikolaj
Trubetzkoy (1928 and [1939] 1958), with significant extensions
by Jakobson, Gunnar Fant, and Morris Halle ([1952] 1963),
304
Feature Analysis
Similar feature-based perceptual distinctions have been
found to take place in the auditory cortex of adults (Phillips,
Pellathy, and Marantz 2000) and in lexical access (MarslenWilson and Warren 1994) and masked phoneme identification
(Miller and Nicely 1955). The fact that humans analyze speech
signals in terms of distinctive features may be connected to the
quantal nature of auditory responses to sound, such as responses
to acoustic discontinuities and closely spaced spectral prominences (Chistovich and Lublinskaya 1979; Delgutte and Kiang
1984; Stevens 1972, 1989, 2002; Clements and Ridouane 2006).
The same features have been implicated in speech production as well, notably in studies of speech errors (Fromkin 1973;
Goldrick 2004).
In addition to the natural class patterns discussed, phonological evidence for distinctive feature theory comes from considerations of economy: Languages appear to organize their feature
systems so as to minimize the number of features employed to
distinguish among both consonants and vowels (Archangeli
and Pulleyblank 1994; Clements 2008; Poeppel, Idsardi, and van
Wassenhove 2007).
305
Feature Analysis
by Chomsky and Halle (1968). The Toronto School employs
essentially the same system, but with privative rather than
binary features; as a result, the more features a segment contains, the greater its degree of markedness. (See Steriade 1995,
Reiss 2003b, and Clements 2008 for further discussion.)
Bert Vaux
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Anderson, John, and Colin Ewen. 1987. Principles of Dependency
Phonology. Cambridge: Cambridge University Press.
Anderson, Stephen. 1999. The nature of phonetic representations. Talk
given at Keio University. Available online at: http://bloch.ling.yale.
edu/Public/Phonetic_Reps.pdf.
Archangeli, Diana, and Douglas Pulleyblank. 1994. Grounded Phonology.
Cambridge, MA: MIT Press.
Avery, Peter, and Keren Rice. 1989. Segment structure and coronal
underspecification. Phonology 6: 179200.
Baltaxe, Christiane. 1978. Foundations of Distinctive Feature Theory.
Baltimore: University Park Press.
Bell, Alexander Graham. 1911. The Mechanics of Speech. New York and
London: Funk and Wagnalls.
Bell, Alexander Melville. 1867. Visible Speech: The Science of Universal
Alphabetics. London: Simkin, Marshall.
Browman, Catherine, and Louis Goldstein. 1989. Articulatory gestures
as phonological units. Phonology 6: 20151.
Chistovich, L., and V. Lublinskaya. 1979. The center of gravity effect in vowel spectra and critical distance between the formants: Psychoacoustical study of the perception of vowel-like stimuli.
Hearing Research 1: 18595.
Chomsky, Noam, and Morris Halle. 1968. The Sound Pattern of English.
New York: Harper and Row. See especially Chapter 7.
Clements, G. N. 1985. The geometry of phonological features. Phonology
Yearbook 2: 22352.
. 1991. Vowel height assimilation in Bantu languages. Working
Papers of the Cornell Phonetics Laboratory 5: 77123. Ithaca, NY: Cornell
University.
. 2008. The role of features in phonological inventories.
In Contemporary Views on Architecture and Representations in
Phonological Theory, ed. Eric Raimy and Charles Cairns. Cambridge,
MA: MIT Press.
Clements, G. N., and Rachid Ridouane. 2006. Quantal phonetics and distinctive features: A review. Proceedings of ISCA Tutorial and Research
Workshop on Experimental Linguistics, 2830 August 2006, ed. Antonis
Botinis, 1724. Athens: University of Athens Press.
Cohn, Abigail. 1990. Phonetic and Phonological Rules of Nasalization.
Ph.D. diss., University of Califonia at Los Angeles.
Delgutte, B., and N. Kiang. 1984. Speech coding in the auditory nerve: IV.
Sounds with consonant-like dynamic characteristics. Journal of the
Acoustical Society of America 75: 897907.
Dresher, B. E., G. L. Piggott, and K. Rice. 1994. Contrast in phonology: Overview. Toronto Working Papers in Linguistics 13.1: iii-xvii.
Fennell, Christopher, and Janet Werker. 2003. Early word learners ability to access phonetic detail in well-known words. Language and
Speech 46.2/3: 24564.
Fitzpatrick, Justin, Andrew Nevins, and Bert Vaux. 2004. Exchange rules
and feature-value variables. Presented at The 3rd North American
Phonology Conference, Concordia University, Montral, Qubec.
Flemming, Edward. 1995. Auditory representations in phonology.
Ph.D. diss., University of California at Los Angeles.
. 2005. Deriving natural classes in phonology. Lingua
115: 287309.
306
Feature Analysis
Lisker, Leigh. 1985. The pursuit of invariance in speech signals. Journal
of the Acoustical Society of America 77: 11991202.
Marslen-Wilson, William, and Paul Warren. 1994. Levels of perceptual
representation and process in lexical access: Words, phonemes, and
features. Psychological Review 101: 65375.
Meillet, Antoine. 1903. Introduction ltude comparative des langues
indo-europennes. Paris: Hachette.
Miller, George, and Patricia Nicely. 1955. An analysis of perceptual confusions among some English consonants. Journal of the Acoustical
Society of America 27.2: 33852.
Morgan, Michael. 2003. Feature analysis. In The Handbook of Brain Theory
and Neural Networks, ed. M. Arbib, 4447. Cambridge, MA: MIT Press.
Odden, David. 1991. Review of generative and non-linear phonology, by
Jacques Durand. Language 67.2: 3637.
. 2005. Introducing Phonology. Cambridge: Cambridge University
Press. Chapter 6, Feature theory, is particularly relevant.
Ohala, John. 1996. Speech perception is hearing sounds, not tongues.
Journal of the Acoustic Society of America 99: 171825.
Padgett, Jaye. 2002. Feature classes in phonology. Language
78.1: 81110.
Phillips, Colin, Tom Pellathy, and Alec Marantz. 2000. Phonological
feature representations in auditory cortex. Manuscript, University
of Delaware. Available online at http://www.ling.udel.edu/colin/
research/papers/feature_mmf.pdf.
Pierrehumbert, Janet. 2003. Phonetic diversity, statistical learning, and
acquisition of phonology. Language and Speech 46: 11554.
Poeppel, David, William Idsardi, and V. van Wassenhove. 2007. Speech
perception at the interface of neurobiology and linguistics. In
Proceedings of the Royal Society of London B 363: 107186.
Pulleyblank, Douglas. 2001. Defining features and constraints in
terms of complex systems: Is UG too complex? Paper presented at
the Workshop on Early Phonological Acquisition, Carry-le-Rouet,
Marseilles, France.
Pycha, Anne, Pawel Nowak, Eurie Shin, and Ryan Shosted. 2003.
Phonological rule-learning and its implications for a theory of vowel
harmony. In Proceedings of WCCFL 22, ed. M. Tsujimura and G.
Garding, 10114. Somerville, MA: Cascadilla Press.
Pylyshyn, Zenon, E. Blaser, and A. Holcombe. 2000. Tracking an object
through feature space. Nature 408: 1969.
Reiss, Charles. 2003a. Quantification in structural descriptions: Attested
and unattested patterns. Linguistic Review 20: 30538.
. 2003b. Accepting markedlessness: How non-phonological
symbolic computation shapes trends in attested phonological systems. In Proceedings of the 29th Annual Meeting of the Berkeley
Linguistics Society, Parasession on Phonetics Sources of Phonological
Patterns: Synchronic and Diachronic Explanations, 56981.
Ribot, Thodule. 1890. Psychologie de lattention. Paris: Alcan.
Saffran, Jenny, and Erik Thiessen. 2003. Pattern induction by infant language learners. Developmental Psychology 39: 48494.
Sagey, Elizabeth. 1986. The representation of features and relations
in non-linear phonology. Ph.D. diss., Massachusetts Institute of
Technology.
Sapir, Edward. 1929. The status of linguistics as a science. Language
5.4: 20714.
Shattuck-Hufnagel, S., and Dennis Klatt. 1979. The limited use of distinctive features and markedness in speech production: Evidence from
speech error data. Journal of Verbal Learning and Verbal Behavior
18: 4155.
Sloman, S. 1993. Feature-based induction. Cognitive Psychology
25: 23180.
Soli, S., and P. Arabie. 1979. Auditory versus phonetic accounts of
observed confusions between consonant phonemes. Journal of the
Acoustical Society of America 66.1: 4659.
Felicity Conditions
Steriade, Donca. 1995. Underspecification and markedness. In
Handbook of Phonological Theory, ed. John Goldsmith, 11474.
Oxford: Blackwell.
Stevens, Kenneth. 1972. The quantal nature of speech: Evidence from
articulatory-acoustic data. In Human Communication: A Unified
View, ed. P. Denes and E. David, 5166. New York: McGraw-Hill.
. 1989. On the quantal nature of speech. Journal of Phonetics
17: 346.
. 2002. Toward a model for lexical access based on acoustic landmarks and distinctive features. Journal of the Acoustical Society of
America 111.4: 187291.
. 2003. Acoustic and perceptual evidence for universal phonological features. In Proceedings of the XVth International Congress of
Phonetic Sciences, Barcelona, 338.
Swingley, D., and R. Aslin. 2002. Lexical neighborhoods and the wordform representations of 14-month-olds. Psychological Science
13: 4804.
Tatham, Mark. 1999. Distinctive feature theory. Available online at
http://www.essex.ac.uk/speech/teaching-01/documents/df-theory.
html.
Taylor, James. 1962. The Behavioral Theory of Perception. New Haven,
CT: Yale University Press.
Tettamanti, M., G. Buccino, M. Saccuman, V. Gallese, M. Danna, P.
Scifo, F. Fazio, G. Rizzolatti, S. Cappa, and D. Perani. 2005. Listening
to action-related sentences activates fronto-parietal motor circuits.
Journal of Cognitive Neuroscience 17.2: 27381.
Trubetzkoy, Nikolay. [1939] 1958. Grundzge der Phonologie.
Gttingen: Vandenhoeck and Ruprecht.
Tversky, Amos. 1977. Features of similarity. Psychological Review
84: 32752.
Vaux, Bert. 2008. The role of features in a symbolic theory of phonology. In Contemporary Views on Architecture and Representations in
Phonological Theory, ed. Charles Cairns and Eric Raimy. Cambridge,
MA: MIT Press.
Vaux, Bert, and Bridget Samuels. 2005. Laryngeal markedness and aspiration. Phonology 22: 395436.
Warren, Richard, and C. Obusek. 1971. Speech perception and phonemic restorations. Perception and Psychophysics 9: 35862.
Williamson, Kay. 1977. Multivalued features for consonants. Language
53.4: 84371.
Zonneveld, Wim. 1976. A phonological exchange rule in Flemish
Brussels. Linguistic Analysis 2: 10914.
FELICITY CONDITIONS
The distinction between felicity conditions and truth conditions was introduced by J. L. Austin as the basis for distinguishing between two types of utterances, performative
and constative, respectively. As such, it is central to the view
within pragmatics that utterances perform actions, an idea
that gave rise to the theory of speech-acts. Constatives such
as Snow is white have truth conditions, that is, they assign truth
or falsity to the proposition expressed (Snow is white is true if
and only if snow is white). Performatives, on the other hand,
have felicity conditions (i.e., pragmatic conditions that must
be met if the utterance is to achieve its intended goal) and will
only function felicitously or happily if these conditions of use
are met.
So, for an utterance such as I hereby pronounce you husband
and wife to be felicitous, the following conditions must apply
(Austin 1962, 15):
307
Felicity Conditions
(1) a) There must exist an accepted conventional procedure
having a certain conventional effect (in our example, changing the wording to I reckon youre man and wife would make
it infelicitous); and b) the particular persons and circumstances in a given case must be appropriate for the invocation
of the particular procedure invoked.
(2) The procedure should be executed by all participants correctly and completely. Stopping halfway through a marriage
ceremony, for example, would result in its being null and void.
(3) When the participants, as is often the case, are required
to have specific thoughts or feelings or when any subsequent
conduct is specified, the participants involved should actually (intend to) have these requisite thoughts and feelings and
conduct themselves appropriately.
Austin later abandoned the distinction between constatives and
performatives in favor of a model in which all utterances have
felicity conditions.
Acts that fail to meet criteria 1 and 2 are labeled misfires, and
those that fail to meet criterion 3 abuses. This is an important
distinction: Misfires simply result in the intended act not being
realized because of some deficiency in the procedure (e.g., the
speaker not having the requisite authority), whereas abuses
involve a conscious decision on the part of either the speaker (S)
or the hearer (H) to ignore or manipulate the action (e.g., when S
makes a promise that he/she does not intend to keep).
It should be noted that for Austin, the so-called uptake of the
action is an integral part of its being felicitous. So, for instance,
when S offers H something, as in Have a cookie, not only must S
have the requisite intention of wanting to do something for H (as
stipulated in condition 3a), but H must actually act accordingly,
that is, perform the appropriate uptake and respond to the offer
by either accepting or refusing. This notion is closely related to
the perlocution associated with the action, which is thus an
integral part of performing a speech-act happily, an aspect often
neglected by later speech-act theorists.
J. R. Searle (1969) used various classes of felicity conditions
as the main criterion for distinguishing between different types
of illocutionary acts. First of all, there are restrictions on the
propositional content expressed through the speech-act, which,
depending on the type of illocutionary force involved,
must be of a certain type, such as, for instance, a future act by S
(promises) or a past act by H (complaints). Secondly, there are
preparatory conditions associated with speech-act types, that
is, specific social preconditions; for a command to be felicitous,
for instance, it is a precondition that S is in a position of authority over H. Otherwise, the command will be ineffective. Thirdly,
Searles sincerity condition refers to the fact that S is supposed
to be sincere when performing a speech-act in order for it to be
felicitous, which crucially depends on the speaker having the
appropriate beliefs or feelings (as Austin had stated already). So,
when uttering a promise, S must sincerely intend to carry out
the future act. The fourth and most crucial criterion Searle calls
the essential condition, which summarizes the appropriate illocutionary force of the utterance. So, for a promise, the speaker
assumes an obligation to commit to the carrying out of the action
expressed in the promise. It is the essential condition that determines all the others: If something is a request, that is, an attempt
308
Field (Bourdieu)
to get H to do something, it follows that this must be a future act
(propositional content), that S really wants H to carry out this
future act (sincerity condition), and that H would not have done
so spontaneously (preparatory condition). In addition to these
conditions, there are general felicity conditions applying to any
speech-act: Both S and H must be able to speak and understand
the language, they must have no pathological conditions, and
the act should not occur in a parasitical context, such as a joke
or a play.
On the basis of these criteria, Searle (1975) distinguishes five
classes of acts: assertives or representatives (which describe
a state of affairs), directives (which attempt to get H to do
something), commissives (which commit S to some future
course of action), expressives (which express Ss psychological
state), and declarations (which effect institutional changes). As
S. C. Levinson points out, however, these classes are not really
built in any systematic way on felicity conditions (1983, 240).
The question remains how hearers determine the speakers
intended illocutionary force of the utterance in the absence of
any linguistic devices that mark it (such as performative verbs),
as is the case in indirect speech-acts.
Ronald Geluykens
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Austin, J. L. 1962. How to Do Things with Words. Oxford: Oxford University
Press.
Levinson, S. C. 1983. Pragmatics. Cambridge: Cambridge University
Press.
Searle, J. R. 1969. Speech Acts: An Essay in the Philosophy of Language.
Cambridge: Cambridge University Press.
. 1975. A taxonomy of illocutionary acts. In Language, Mind,
and Knowledge, ed. K. Gunderson, 34469. Minneapolis: University of
Minnesota Press.
FIELD (BOURDIEU)
Field is one of Pierre Bourdieus two fundamental concepts
the other being habitus and is addressed in several contexts
in his work, with precise definitions given. For example:
A field may be defined as a network, or a configuration, of objective relations between positions. These positions are objectively
defined, in their existence and in the determinations they impose
upon their occupants, agents or institutions, by their present and
potential situation (situs) in the structure of the distribution of
species of power (or capital) whose possession commands access
to the specific profits that are at stake in the field, as well as by
their objective relation to other positions (domination, subordination, homology, etc.). (Bourdieu 1992, 97)
Field (Bourdieu)
1960s. Increasingly, field became Bourdieus main theoretical
tool in analyzing a wide range of social phenomena.
Bourdieus early work was developed in opposition to two
salient intellectual traditions, both of which were highly influential during his formative years (the 1950s): existentialism and
structuralism. Existentialism is best represented by the
work of the French philosopher Jean-Paul Sartre, with its philosophy of personal liberation through the subjective choices
we make in defining our lives. Structuralism is best represented
by the work of the anthropologist Claude Lvi-Strauss and his
study of the objective rules that can be found across cultures
and which govern human behavior taboos, myths, and so on.
There is a philosophy of language at the base of both of these
traditions, albeit from an objective and subjective point of view.
Bourdieu referred to the divide between objectivism and subjectivism in the social sciences as the most fundamental, and
the most ruinous ([1980] 1990, 25). His entire theory of practice can be seen as an attempt to bridge this divide. He defined
his approach as a science of the dialectical relations between
objective structures and the subjective dispositions within
which these structures are actualised and which tend to reproduce them [1972] 1977, 3), and the relationship between field
and habitus as one of ontological complicity (1982a, 47). The
same complicity can be seen in the relation between langue and
parole.
The notion of structure affords just such a reconciliation. Both
habitus and field are structured. In other words, social spaces
need to be understood as differentiated and thus structural
in essence. Similarly, individual cognition arises from, generates, and is generated by mental structures that are also essentially structured because of their systems of differentiation. In
a seminal paper in 1966 Intellectual Field and Creative Project
([1966] 1971) Bourdieu builds on the discovery of the historian
E. Panofsky that there was a link between Gothic art, for example in the design of cathedral architecture, and the mental habits of those involved. In other words, each was symptomatic of
the other. Bourdieu used this principle to argue that there was
a structural homology between subjective thought and objective surroundings, the latter most noticeable in forms of social
organization. Such homologies exist because they are both generated by and generate the logic of practice of the field, itself
defined in terms of its substantive raison dtre. For Bourdieu,
therefore, social and mental structures were both structured
and structuring concrete and dynamic. Similarly, language is
both structured and structuring.
Much of Bourdieus own work subsequently brought this
understanding and methodology to a series of field studies: for example, the academic field ([1984] 1988); the artistic
field ([1987] 1993); the religious field (1982b); the judicial field
([1986] 1987); the bureaucratic field ([1989] 1996); the scientific
field (1975); the cultural field (1993); the literary field ([1992]
1996); the economic field ([2002] 2005); and the political field
(1981). Even the academic discipline of applied linguistics can
be regarded as a field (see Grenfell 2004a). Many of these are
shown to be defined in terms of overarching power structures
in society: class, the state, economic interests. However, there
are also fields within fields microcosms which exhibit the
same features as macro fields but in local contexts. For example,
although all fields are in some ways interlocking, they all possess
a degree of independence or autonomy. They are also bounded
with strict rules of entry. Moreover, they are ruled by specific
values, logics, and behaviors (mostly implicit but also explicit),
which determine the legitimate ways of thinking and doing
things within the particular field. Such legitimation is a necessary part of the functioning of the field according to its own
logic and purpose, as well as representing the interests of those
who hold position within it. Such social features also define the
orthodoxy of the field the doxa against which other nonorthodox forms oppose themselves heterodoxa. Similarly, fields
have a specific orthodox language, which can be expressed at
any level: phonetic and syntactic, as well as gesture and
expression.
Fields are dynamic and in a constant state of flux; whole
sections of society can be understood historically as the move
from one field structure and logic to another, for example, in
the way the ruling class in France moved from a situation of
basing their power on inherited money and industrial wealth
to one where intellectual and academic qualifications became
the main form of social legitimation. For Bourdieu, the currency of fields was capital social, economic, and cultural.
Different configurations are held by individuals and groups
in the field and can be used to buy social positioning. Fields
are therefore also the sites of struggle and conflict a process
by which they evolve according to the salient sociostructural
forces of society at large.
Fields can be studied internally or externally. In both cases,
position is all important. Bourdieu lists this method of field
analysis (1992: 1047) as
(1) Analyze the position of the field vis--vis the field of
power.
(2) Map out the objective structural relations between the
positions occupied by those in the field.
(3) Analyze the habitus of those involved.
In this way, both the logic of practice of the field and its procedures can be rendered visible, making what is normally misrecognized open to public scrutiny. Studying language in this
way allows for a focus on linguistic features, while at the same
time relating them to field position and the position of the fields
within the overall network of fields (see Encrev 1983; Fehlen
2004; Grenfell 2004b; and Vann 2004). Field positions and positioning can depend on economic power but also on symbolic
power acquired from taste, style, and language capital that
buys social prestige. Ultimately, field language articulates and
expresses the power order found within it. In social fields, this is
the language of the state and dominant social classes.
Michael Grenfell
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bourdieu, Pierre. [1958] 1961. Sociologie de lAlgrie. New rev. ed.
Paris: Que Sais-je.
. 1962. The Algerians. Trans. A. C. M. Ross. Boston: Beacon.
, with Jean-Claude Passeron. [1964] 1979. The Inheritors: French
Students and their Relation to Culture. Trans. R. Nice. Chicago: University
of Chicago Press.
309
Field (Bourdieu)
, with Luc Boltanski, Robert Castel, Jean Claude Chamboredon,
and Dominique Schnapper. [1965] 1990. Photography: A Middle-Brow
Art. Trans. S. Whiteside. Oxford: Polity.
. 1966. Champ intellectual et project crateur. Les Temps
Modernes 246: 865906.
. [1966] 1971. Intellectual field and creative project. In
Knowledge and Control: New Directions for the Sociology of Education,
ed. M. F. D. Young, 16188. London: Collier Macmillan.
, with Alain Darbel and Dominique Schnapper. [1966] 1990. The
Love of Art: European Art Museums and Their Public. Trans. C. Beattie
and N. Merriman. Oxford: Polity.
. with Jean-Claude Passeron. [1970] 1977. Reproduction in
Education, Society and Culture. Trans. R. Nice. London: Sage.
. [1972] 1977. Outline of a Theory of Practice. Trans. R. Nice.
Cambridge: Cambridge University Press.
. 1975. The specificity of the scientific field and the social conditions
of the progress of reason. Social Science Information 14.6: 1947.
. [1980] 1990. The Logic of Practice. Trans. R. Nice. Oxford: Polity.
. 1981. La representation politique: Elments pour une thorie du champ politique. Actes de la receherche en sciences sociaels
37: 324.
. 1982a. Leon sur une leon. Paris: Les Editions de Minuit.
, with Monique de Saint Martin. 1982b. La sainte famille.
Lpiscopat franais dans le champ de pouvoir. Actes de la recherch
en sciences sociales 44/45: 253.
. [1984] 1988. Homo Academicus. Trans. P. Collier. Oxford: Polity.
. [1986] 1987. The force of law: Toward a sociology of the judicial
field. Hastings Journal of Law 38: 20948.
. [1987] 1993. Manet and the institutionalisation of anomie. In The
Field of Cultural Production, ed. and introd. Randall Johnson, 23853.
Oxford: Polity.
. [1989] 1996. The State Nobility: Elite Schools in the Field of Power.
Trans. L. C. Clough. Oxford: Polity.
, with Loc Wacquant. 1992. An Invitation to Reflexive Sociology.
Trans. L. Wacquant. Oxford: Polity.
. [1992] 1996. The Rules of Art. Trans. s. Emanuel Oxford: Polity.
. 1993. The Field of Cultural Production. Oxford: Polity.
. [2000] 2005. The Social Structures of the Economy. Oxford: Polity.
Dreyfus, Hubert, and Paul Rabinow. 1993. Can there be a science
of existential structure and social meaning? In Bourdieu: Critical
Perspectives, ed. C. Calhoun, E. LiPuma, and M. Postone, 3544.
Oxford: Polity.
Encrev, Pierre. 1983. Le sens en pratique. Construction de la rfrence
et structure social de lintraction dans le couple question/rponse.
Actes de la Recherche en Sciences Sociales 46: 330.
Fehlen, Fernand. 2004 Pre-eminent role of linguistic capital in
the reproduction of the social space in Luxembourg. In Pierre
Bourdieu: Language, Culture and Education, ed. M. Grenfell and M.
Kelly, 6172. Bern: Peter Lang.
Grenfell, Michael. 1993. The linguistic market of Orlans. In
France: Nation and Regions, ed. M. Kelly and R. Bock, 7299.
Southampton: ASM & CF.
. 2004a. Language: Construction of an object of research. In
Pierre Bourdieu: Language, Culture and Education, ed. M. Grenfell and
M. Kelly, 2740. Bern: Peter Lang.
. 2004b. Bourdieu in the classroom. In Culture and Learning: Access
and Opportunity in the Curriculum, ed. M. Olssen, 4972. Wesport,
CT: Greenwood.
. 2006. Bourdieu in the field: From the Barn to Algeria a timely
response. French Cultural Studies 17: 22339.
Grenfell, Michael, and Cheryl Hardy. 2003. Field manoeuvres: Bourdieu
and the young British artists. Space and Culture 6.1: 1934.
. 2007. Art Rules. London: Berg.
310
Filters
indiscriminability, but if the cat is called Felix, this is a minimal unit, since its parts, such as lix, do not denote part of the
cat. Films are composed of pictures, but pictures do not have a
merely conventional relation to the world. From acquaintance
with at most a small number of pictures, one can go on to interpret correctly any other picture in the same style, a feature that
Flint Schier (1986) called natural generativity, grounded on the
fact that we use our object-recognitional capacities to recognize
pictures of those objects. In contrast, acquaintance with a small
number of words in a language does not allow one correctly to
interpret other words in that language: Language lacks natural
generativity.
Nor should we hold that the shot is analogous to the sentence.
Indeed, Metzs version of the claim is, taken strictly, incoherent
since he denies that there is anything corresponding to words
in cinema; yet a sentence is by definition composed of words.
Moreover, there is nothing in a photograph corresponding to
subject-predicate structure. A photograph of a black cat lacks
parts that pick out separately the cat and blackness, unlike the
sentence the cat is black.
The claim that films have a syntax is also untenable: if films have
nothing corresponding to words, they cannot have syntax since
syntax is what couples words together. Film structure does involve
conventions, such as the alternate syntagma. But not all communication conventions are structured linguistically: Pointing is a
convention, but pointing, though it is a referential device, is not
linguistic. And, in general, communication need not be linguistic,
as is shown by animal communication. Nor does the alternate
syntagma denote simultaneity in the way that a linguistic phrase
such as at the same time does. The alternate syntagma joins
shots together; the linguistic phrase joins sentences or clauses
together. However, as just noted, shots are not like sentences (or
clauses), and Metzs view that shots are like sentences that are not
composed of words is incoherent. So, though it is a convention,
the alternate syntagma is not a linguistic convention.
Philosophers of film have given good reasons to reject the film
as language hypothesis. Film is a form of communication that, in
its visual dimension, communicates as pictures do: In terms of
Peirces semiotics, films are composed of icons, not symbols (see
icon, index, and symbol).
Berys Gaut
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Currie, Gregory. 1995. Image and Mind: Film, Philosophy, and Cognitive
Science. Cambridge: Cambridge University Press.
Eisenstein, Sergei. 1949. The cinematographic principle and the ideogram. In Film Form, trans. Jay Leyda, 3844. San Diego, CA: Harcourt
Brace Jovanovich.
Harman, Gilbert. 1977. Semiotics and the cinema: Metz and Wollen.
Quarterly Review of Film Studies 2: 1524.
Metz, Christian. 1974. Film Language: A Semiotics of the Cinema. Trans.
Michael Taylor. New York: Oxford University Press.
Prince, Stephen. 1993. The discourse of pictures: Iconicity and film studies. Film Quarterly 47.1: 1628.
Pudovkin, V. I. 1958. Film Acting and Film Technique. Trans. Ivor
Montagu. London: Vision.
Schier, Flint. 1986. Deeper into Pictures: An Essay on Pictorial
Representation. Cambridge: Cambridge University Press.
FILTERS
Filters are output constraints of the form *xy that rule out illicit
sequences generated by a computational system. Filters can be
added to any computational system, allowing hardware to
be kept maximally general. Filters play different roles in different theories, ranging from statements ensuring satisfaction of a
universal property (the case filter [Chomsky 1981; Rouveret and
Vergnaud 1980), which requires an (overtly realized) noun phrase
(NP) argument to be associated with a case configuration), to the
ranked violable output constraints of optimality theory.
Many named filters, covering diverse phenomena, date from
the 1970s and 1980s. Although they appear to be reflections of
the underlying computational system and architecture, theoretical understanding must await future research.
The doubly-filled C filter (Chomsky and Lasnik 1977) excludes
the co-occurrence of a pronounced wh-phrase and complementizer in Spec (specifier/subject position), CP (complementizer
phrase) and C (complementizer) (I wonder who (*that/*whether)
she saw, *the [man [who that] you saw ). This filter relates to the
more general question of how the distribution of overt and covert
material over hierarchical structures is determined.
The that-t filter (Perlmutter 1971) prohibits an extraction site
next to the C (that, for): who do you think (*that) t left). That-t
reflects a very general, and theoretically not understood, problem with subject extraction.
The person case constraint (Perlmutter 1971) expresses the
interaction of case forms and person marking: It bans first and
second person accusative clitics or agreement markers in the
presence of a (third person) dative clitic or agreement marker.
(French: le(ACC)-lui (DAT) but *me(ACC)-lui(DAT) *me.ACC- te.DAT ).
It is widely assumed to follow from universal person and case
hierarchies, either structurally hardwired or encoded as soft
constraints.
Some filters deal with restrictions on recursion. The doubl-ing
filter (Ross 1977) restricts (complement) recursion of -ing forms
in English. A verb like continue can combine with an -ing complement (it continued raining), unless continue is the complement
of an -ing selecting head (*its continuing raining). The head-final
filter (Williams 1982) restricts recursion on the nonrecursive side
of the head. (a man [afraid of dogs], *an [afraid of dogs] man.)
There is no generally accepted theoretical understanding of how
restrictions on recursion emerge from the derivation (but see
Koopman 2002).
Hilda Koopman
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bresnan, Joan, Shipra Dingare, and Christopher D. Manning. 2001. Soft
constraints mirror hard constraints: Voice and person in English and
Lummi. Proceedings of the LFG 01 Conference. Stanford, CA: CSLI
Publications.
Chomsky, Noam. 1981. Lectures on Government and Binding. Dordrecht,
the Netherlands: Foris.
. 1995. The Minimalist program. Cambridge, MA: MIT Press.
Chomsky, Noam, and Howard Lasnik. 1977. Filters and control.
Linguistic Inquiry 8: 425504.
Koopman, Hilda. 2002. Derivations and complexity filters. In Dimensions of
Movement, ed. A. Alexiadiou et al., 15189. Amsterdam: John Benjamins.
311
Focus
Perlmutter, David. 1971. Deep and Surface Constraints in Syntax. New
York: Rinehart and Winston.
Ross, John. 1972. Doubl-ing. Linguistic Inquiry 3: 6186.
Rouveret, Alain., and Jean-Roger Vergnaud. 1980. Specifying reference
to the subject: French causatives and conditions on representations,
Linguistic Inquiry 11: 97202.
Williams, Edwin. 1982. Another argument that passive is transformational. Linguistic Inquiry 13:1603.
FOCUS
Focus refers to a constituent within a sentence that is highlighted
or emphasized by grammatical means. English sentences with
marked accent patterns provide prototypical examples, like (1)
(focus on the direct object):
(1) They ordered COffeeF at the bar.
312
Foregrounding
use cleftlike sentences or, more generally, marked constituent
order in corrections or answers (often in addition to intonation);
also common are special morphemes, as well as prosodic phrase
boundaries, to mark the edge of a focus.
Daniel Bring
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Beaver, David, and Brady Clark. 2008. Sense and Sensitivity: How Focus
Determines Meaning. Malden, MA: Blackwell.
Bring, Daniel. 2007. Intonation, semantics and information structure.
In The Oxford Handbook of Linguistic Interfaces, ed. Gillian Ramchand
and Charles Reiss, 44573. Oxford: Oxford University Press.
Kadmon, Nirit. 2001. Formal Pragmatics: Semantics, Pragmatics,
Presupposition, and Focus. Malden, MA: Blackwell.
Rooth, Mats. 1996. Focus. In The Handbook of Contemporary Semantic
Theory, ed. Shalom Lappin, 27197. Blackwell.
Schwarzschild, Roger. 1999. GIVENness, avoidF and other constraints
on the placement of accent. Natural Language Semantics 7: 14177.
FOREGROUNDING
Foregrounding is the patterned deviation from anticipated language, which is an important characteristic of literariness. In
general usage, the term sometimes describes the many conventional means that language provides for certain information to
be made prominent in discourse, such as by variation of intensity in speech or by given/new organization in syntax. In literary analysis, however, foregrounding refers specifically to the
effects achieved by textual devices that interrupt the automatic
processes of linguistic understanding. Foregrounding shifts a
readers or listeners attention away from linguistic meaning to
linguistic form so as to impart the sensation of things as they are
perceived and not as they are known, in the Russian formalist
Viktor Shklovskys words ([1917] 1988, 20).
At all levels of linguistic organization, perceptual processes
entail anticipation. Foregrounding devices defeat the processing benefits of anticipation either by creating unconventional
patterns or by deviating from established patterns. Soniclevel devices, such as alliteration, rhyme, assonance, (see
rhyme and assonance), and meter, are typically adduced as
examples of foregrounding because the repetition of individual
phones or regular metrical structures falls outside of languages
conventionalized means of sense making. Once these unconventional patterns are detected, some processing resources
are diverted to construing their significance. Devices operating
at more complex levels of structure must work differently, as
ellipsis, metaphor, and irony, for example, regularly occur
in normal speech. Foregrounding with these devices is said to
be achieved by an unusual degree of difficulty, which requires
the reader consciously to attend to construal. Deviation from
expected patterns can also be achieved at any level of linguistic
organization. Creative punctuation, neologism, oxymoron, conversational infelicity, and unusual narrative structuring are just a
few of the many devices that have been recognized as deautomatizing linguistic perception.
If foregrounding involves the creation of new patterns or
deviation from expected patterns, one must ask what gives rise
to the expectations against which foregrounding is recognized.
Forensic Linguistics
In most cases, expectations arise from experience with everyday language, such as those evidenced by garden-path phenomena in syntactic processing (see psycholinguistics),
but expectations can also arise from genre-specific knowledge.
Emily Dickinsons verse, for example, often metrically evokes the
religious hymns that would have been familiar to her community. The masculine end-rhyme patterns of these hymns establish an expectation against which the slant rhymes of her verse
stand out.
Foregrounding devices are not unique to literary texts but can
be found in traditionally nonliterary genres, such as advertising
copy or political speech. In literary texts, however, they usually
participate in larger patterns of meaning or coherence, such as
those established by thematization or iconicity (see literariness for a discussion of the latter two terms). Empirical research
has established some support for the contribution of foregrounding devices to the literariness of a text. In studies, segments of
short stories with a higher degree of foregrounding were read at
a slower rate and were rated as more striking and more emotionally evocative than segments with low foregrounding (Miall and
Kuiken 1994).
Claiborne Rice
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Miall, David S., and Don Kuiken. 1994. Foregrounding, Defamiliarization,
and Affect: Response to Literary Stories. Poetics Today 22: 389407.
Shklovsky, Viktor. [1917] 1988. Art as technique. Trans. Lee T. Lemon
and Marion J. Reis. In Modern Criticism and Theory, ed. David Lodge,
1630. London: Longman.
Van Peer, Willie. 1986. Stylistics and Psychology. London: Croom Helm.
FORENSIC LINGUISTICS
Forensic linguistics refers to the use of linguistic expert evidence
in legal proceedings, and more broadly, to linguistic research in
legal contexts. Most forensic linguistic work published in English
pertains to the common law adversarial legal system of the United
Kingdom, the United States, Canada, and Australia. Any area of
linguistics can have a forensic application. In order to do forensic linguistics, a person must qualify as a linguist, specializing in
a particular area, such as phonetics or sociolinguistics.
There are a few graduate programs in forensic linguistics, but
most practitioners are linguists with a doctorate in their specialization who apply this expertise to legal questions and contexts.
Forensic phonetics is used mostly in disputes over transcription of incriminating recorded speech, and over the identification of individuals who have committed a language crime such
as a threat, bribe, or hoax by means of a recorded voice message. acoustic phonetics is often combined with articulatory phonetics. Analysis of anonymous voice recordings
can be compared with recordings of suspects. Similarly, linguistic discourse analysis of a written text may help to identify
an author, for example, of a so-called suicide letter in a case
where police have grounds for suspecting murder, rather than
suicide. For example, if discourse analysis finds striking similarities between a suicide letter and previous (uncontested) correspondence of the spouse of the deceased, together with a striking
313
Formal Semantics
confusing, and compromises their ability to tell their story. Second
language speakers rely on interpreters, whose role is often misunderstood by legal professionals and witnesses. Microanalysis
demonstrates linguistic challenges involved in courtroom interpreting, finding that it is much easier for interpreters to accurately interpret propositional meaning than pragmatic meaning.
Second dialect speakers are disadvantaged by widespread ignorance about subtle dialectal differences. Where there are also
important cultural differences in communicative style, as with
Australian Aboriginal English speakers, the possibilities for miscommunication are disturbing. In Australia, forensic linguistics
has contributed to positive developments in delivery of justice
generally for Aboriginal people, and specifically in legal cases of
some individuals.
Forensic linguistics is also concerned with the language of
lawyer-client interviews and alternative legal processes, such as
mediation. For example, conversation analysis investigates
the extent to which mediators talk is neutral, as well as the ways
in which gender affects talk between lawyers and their clients.
Another recent focus is on the transformation of oral narratives
into written legal documents (see intertextuality). While
most forensic linguistics is concerned with criminal or civil law,
linguists concerns have recently been extended to immigration
law, notably immigration officials use of untrained analysts who
examine the speech patterns of asylum seekers to determine
their national origin.
Diana Eades
SUGGESTIONS FOR FURTHER READING
Eades, Diana. 2010. Sociolinguistics and the Legal Process. Bristol:
Multilingual Matters. Comprehensive examination of sociolinguistic
research, examining how language works and does not work in the
legal process.
Gibbons, John. 2003. Forensic Linguistics: An Introduction to Language
in the Justice System. Oxford: Blackwell. Accessible introduction to the
field.
Tiersma, Peter. 1999. Legal Language. Chicago: University of Chicago
Press. Authoritative textbook on written and spoken language in the
legal process.
FORMAL SEMANTICS
Formal semantics is an approach to semantics, the study of
meaning, with roots in logic, the philosophy of language, and
linguistics, and since the 1980s a core area of linguistic theory.
Characteristics of formal semantics treated in this entry include
the following: Formal semanticists treat meaning as mind independent (though abstract), contrasting with the view of meanings
as concepts in the head (see i-language and e-language
and meaning externalism and internalism ); formal
semanticists distinguish semantics from knowledge of semantics
(Lewis 1975, Cresswell 1978), which has consequences for the
notion of semantic competence. A central part of the meaning
of a sentence on this approach is its truth conditions, and
most, though not all, formal semantics is model theoretic, relating linguistic expressions to model-theoretically constructed
semantic values cast in terms of truth, reference, and possible
314
Formal Semantics
An Example
We illustrate the methods of formal semantics without formal
details by considering one aspect of the analysis of restrictive relative clauses like who fed Fido in (1a) and (1b).
(1)
315
Formal Semantics
{x: x is a boy and x fed Fido}. Different theories of the semantics
of determiners give different technical implementations of the
rest of the solution, but that first step settles both the syntactic
question and the core of the semantics. Sentence (1a) asserts
that the set of boys who fed Fido and the set of boys that I saw
overlap; (1b) says that the set of boys I saw is a subset of the set of
boys who fed Fido. See Partee (1995) for a fuller argument, and
Barwise and Cooper (1981), Heim (1982), and Kamp (1981) for
treatments of the semantics of determiners.
A Second Example
Also informally, we illustrate the beginnings of the more dynamic
semantics of Kamp (1981) and Heim (1982). Consider the contrasting minidiscourses in (4) and (5):
(4) A baby was crying. It was hungry.
(5) Every baby was crying. #It was hungry. (# means
anomalous.)
316
Formal Semantics
Forms of Life
FORMS OF LIFE
Forms of life, an expression associated with Ludwig
Wittgenstein, is not one he made that much use of, employing it
only five times in his Philosophical Investigations. It is not clear
from his usage exactly what he had in mind. Here are statements
in which the phrase occurs:
It is what human beings say that is true or false; and they
agree in the language they use. That is not agreement in opinions, but in forms of life (1958, 38), and
What has to be accepted, the given, is so one could say
forms of life (1958, 226), as well as
317
Forms of Life
I would like to regard this certainty, not as something akin
to hastiness or superficiality, but as a form of life (That is
very badly expressed and probably badly thought as well).
(1969, 46)
Its use may nonetheless be associated with Wittgensteins
attempt to humanize matters especially logic that philosophers
had been prone to idealize to the extent of making them seem to
surpass all possibility of creation by merely human beings, meaning that their objectivity must surely reside elsewhere than in the
contingent facts and incidental variety of human life. He seeks to
reassert the connection between seemingly immutable necessities (of logic and mathematics especially) and the contingencies
of human life. To counteract the idealization of logic into something almost superhuman, Wittgenstein is prepared to say that I
want to conceive of it [logic] as something that lies beyond being
justified or unjustified; as it were, something animal (1969, 47).
318
understood as a challenge: Try to imagine understanding a language without knowing something about how those who use the
language live their lives, some of the different sorts of things they
do. The language as a totality is an entirely contingent assemblage, accumulated ways of speaking that come from different
parts of life (as add, subtract, divide come from calculation, or
swing, putt, birdie from golf).
The direction in which natural human responses can be cultivated varies greatly, according to the multifarious ways in which
circumstances impact upon them, but Wittgenstein emphasized
both that and the importance of the natural, animal responses
of human beings. For example, consider the way in which the
human animal naturally responds to a pointing gesture (following the direction of the finger) when other animals (cats, cows)
do not, and the role that gesture can play when embedded in a
parentchild relation in associating a color name with a color
sample. This is an effective way of teaching words because children naturally respond to the rulelike connection made between
word and sample, and after a very few examples, can make the
same association for themselves. Such natural, and common,
reactions are what enable practices, as well as the development
of their associated language forms, to find purchase as standard
practices: Agreement in the way color samples are extended to
cases is the basis for the possession of standard color names.
On that same basis, of course, varied color vocabularies have
arisen.
Necessities
There can be a strong temptation to think that our institutions, practices, and ways of acting can (sometimes at least) be
explained in terms of necessities. Logic was especially important
to Wittgensteins reflections throughout his life. It was widely
assumed that the power of logical forms derived from the fact
that they reflected necessities intrinsic to the way things are
logic could be no other way than it is because it corresponds to
unalterable features of reality. Wittgensteins was no attempt to
do away with logical necessities altogether, only to dereify them,
and to persuade, more generally, that our institutions and practices do not follow from external necessities. Rather, our idea
of necessity flows from our institutions, practices, and so on.
Getting things the right way round is, as Peter Winch summarized it, a matter of understanding that logical relations among
propositions themselves depend on social relations between
Forms of Life
men (1958, 126). Wittgenstein suggested that when it seemed
that things could not be conceived of as being otherwise, this
was, if not an illusion, at most a function of the part that they
played in our ways of acting, not a sign of their essential relation
to any necessitating external reality. The temptation to appeal to
such externally imposed necessities could be alleviated by drawing attention to the degree to which seemingly transcendental
necessity is connected with some familiar, unobtrusive feature of
the human organism or its life, by imagining the consequences
of some gross change in the character of the human organism,
or by imagining a cultural practice significantly different from
our own. Thus, one need only imagine children not naturally
responsive to the rulelike link between instruction and sample to
quickly realize that color vocabularies and arithmetic (for example) depend on the trainable susceptibilities of children for their
existence as stable practices. Alternatively, imagine how another
people might have different practices from ours in counting or
measuring, and then appreciate that the specific forms that serve
these tasks in our society do not comprise the unique, eternal, or
universal essence of counting or measuring, and work as well as
they do only because they have connections with the practical
needs of our distinctive way of life.
Certainties
Certainty, too, is to be connected to natural human reactions,
to the way in which we commonly act without doubts or hesitations we go about many of our activities in a way in which doubt
is simply absent; it does not occur to us. The philosophers stock
idea of certainty is something that we are entitled to only after
we have arrived at some final, unquestionable justification, have
given grounds for our practices that allow no logical possibility
of error. Rather than accepting the skeptics query as to whether
it is logically possible to doubt, Wittgenstein prefers, instead,
to reflect on where we do doubt. We do not possess certainties
because we have arrived at them through critical reflections, but
acquire them as an integral part of coming to mastery in various
activities. Is this not, however, simply a complete concession to
the skeptic? All our practices lack justification; we cannot identify any ultimate, unquestionable self-evident truth that underpins them. No, it is a counterskeptical move: The certainties that
are incorporated into our learning and thus into our practices
are ones that serve as conditions of intelligibility, that shape our
capacity to understand what a doubt could be in an actual case.
These certainties are not held in place by an utterly uncritical attitude toward them but by their connection with our ways of doing
practical things, in which they act as the setting within which the
notions of doubt and justification can have content.
The doubts that the skeptic tries to induce are pretend ones
only, ones that by their very design are empty of empirical
content, and are, as genuine doubts, unintelligible. Skepticism
is right only insofar as its continuing pressure for further justification until a rock-bottom certainty is reached leaves us (quite
soon) in a position where we have no further justifications. It is
misguided to think that this shows that we have been forced to
admit our inability to respond to legitimate demands for justification. Rather, we run out of justifications at the same point at
which the skeptics demands for justification themselves pass
the limits of what can make sense as a doubt. We can settle the
Frame Semantics
doubts about the length of a table by recommending the use of a
tape measure as a dependable way of precisely determining this,
but what sorts of serious doubts are there about a tape measures
dependability in such a case? Our notion of reliable measurement is very much tied up with the ways we use tape measures,
but, at the same time, we should not think that our practices of
measurement are the only ones that could possibly make sense,
and that other people might not have ways of measuring for
which our tape measures would be useless.
Wes Sharrock
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Conway, Gertrude.1989. Wittgenstein on Foundations. New
Jersey: Humanities Press International.
Emmett, Kathleen. 1990. Forms of life. Philosophical Investigations 13
(July): 21331.
Garver, Newton. 1990. Form of life in Wittgensteins later works.
Dialectica 44.1/2: 175201.
Hanfling, Oswald. 2002. Wittgenstein and the Human Form of Life.
London: Routledge.
Hunter, John. 1968. Forms of life in Wittgensteins Philosophical
Investigations. American Philosophical Quarterly 5: 23343.
Winch, Peter. 1958. The Idea of a Social Science. London: Routledge.
Wittgenstein, Ludwig. 1958. Philosophical Investigations 2d ed.
Oxford: Blackwell.
. 1969. On Certainty. Oxford: Blackwell.
FRAME SEMANTICS
This term refers to a wide variety of approaches to the systematic description of natural language meanings. The one common feature of all these approaches which, however, does not
sufficiently distinguish frame semantics from other frameworks
of semantic description is the following slogan from Charles
Fillmore (1977a):
Meanings are relativized to scenes.
The background scene for the first sentence is a sea voyage, while
the second sentence refers to an interruption of air travel. This
319
Frame Semantics
illustrates Fillmores use of the term frame as an idealization of a
coherent individuatable perception, memory, experience, action,
or object (1977c).
In order to understand frame semantics, it is helpful to begin
with a brief history. From here we turn to an overview of the most
important theoretical concepts. Next, the relationship of frame
semantics to one specific version of construction grammar
is introduced and some examples analyzed. The entry ends with a
short summary of applications of frame semantics and a note on
formalization. Usually, frame semantics is taken to be a very informal approach to meaning, but nevertheless, some approaches
relating frame semantics to formal semantics exist.
History
There are at least two historical roots of frame semantics; the first
is linguistic syntax and semantics, especially Fillmores case
grammar; the second is artificial intelligence (AI) and the notion
of frame introduced by M. Minsky (1975) in this field of study.
A case frame in case grammar was taken to characterize a
small abstract scene that identifies (at least) the participants of
the scene and thus the arguments of predicates and sentences
describing the scene. In order to understand a sentence, the language user is supposed to have mental access to such schematized scenes.
The other historical root of frame semantics is more difficult to
describe. It relates to the notion of frame-based systems of knowledge representations in AI. This is a highly structured approach
to knowledge representation that collects together information
about particular objects and events and arranges them into a taxonomic hierarchy familiar from biological taxonomies. However,
the specific formalism suggested in the aforementioned paper by
Minsky was not considered successful in AI .
320
Table 1.
BUYER
buy
Subject
GOODS
(SELLER)
(PRICE)
object
From
for
from Pete
Angela
bought
the owl
Eddy
bought
them
Penny
bought
a bicycle
for $ 10
for $ 1
from Stephen
Table 2.
VERB
BUYER
GOODS
SELLER
MONEY
PLACE
Buy
subject
object
From
For
at
Sell
to
Cost
indirect
object
subject
Object
at
Spend
subject
on
Object
at
prototypical descriptions of scenes. A prototype has the advantage that it does not have to cover all possible aspects of the
meaning of a phrase; in other words, a prototype does not have to
provide necessary and sufficient conditions for the correct use of
a phrase. Fillmore (1977b) illustrates the use of prototypes within
frame semantics by an analysis of the concept widow. The word
widow is specified with respect to a background scene in which
people marry as adults, they marry one person, and their lives
are affected by their partners death and perhaps other properties. The advantage of a theory of meaning based on the prototype
concept, compared to a theory that insists on stating necessary
and sufficient conditions for the meaning of a phrase, is that it
does not have to care about certain boundary conditions; that is,
it does not have to provide answers for questions like Would you
call a woman a widow who has lost two of her three husbands but
who had one living one left? Fillmore (1977b). In a case like this,
whether the noun widow applies or not is unclear since certain
properties of the background frame for this concept are missing.
The concept prototype is not unproblematic either, however.
Note that Fillmore does not use this concept with respect to words
but with respect to frames or scenes. Some words like bird certainly have prototypes, but others may not have a corresponding
prototype. What is a prototypical vegetable, for instance, or a prototype corresponding to the adjective small? Moreover, applications of prototype theory often involve two different measures for
category membership. A penguin, for example, is certainly not a
prototypical bird, but nobody hesitates to judge it as a bird. The
other measure of category membership is typically used in the
analysis of vague predicates, for instance, color adjectives. It may
sometimes be hard or even impossible to assign a given object to
the category of pink or red entities .
Another central notion within frame semantics is the concept
profiling. R. Langacker (1987) uses the example of hypotenuse for
explaining this concept. One can easily draw a mental picture
of the concept hypotenuse. The interesting question concerning
this mental picture is this: Can you imagine what a hypotenuse is
without imagining the whole right triangle? The answer is clearly
no. The triangle and the plane in which it is included is a frame,
and the terms hypotenuse and right triangle are interpreted
Frame Semantics
with respect to this frame, but they profile different parts of the
frame.
The following example taken from Goldberg (1995) illustrates
lexical profiling of participants. Consider the following differences between the closely related verbs rob and steal.
(3) a. Jesse robbed the rich (of all their money).
b. *Jesse robbed a million dollars (from the rich).
(4) a. Jesse stole money (from the rich).
b. *Jesse stole the rich (of money).
These distributional facts can be explained by a semantic difference in profiling. In the case of rob, the victim and the agent
(the thief) are profiled; in the case of steal, the agent and the valuables are profiled. Representing profiled participants in boldface, A. Goldberg proposes the following argument structure for
rob versus steal:
rob <thief target goods>
steal <thief target goods>
The peculiarity of example (5) is due to the fact that the verb
bake, which normally has two arguments, is used with three
arguments here. Peculiar as this sentence is, we nevertheless can
make sense of it. Margaret baked some cookies with the intention to give them to Peter. Note that this interpretation helps us
to make sense of the recipient role, which is not provided by the
verb bake; that is, we think of this sentence as an instance of the
ditransitive construction of which a more standard example is
(7) John gave Mary a present.
Applications
Frame semantics has a wide range of applications reaching
from subfields of linguistic theorizing such as morphology,
to typolology, linguistic discourse analysis, and
language acquisition. However, the central and most successful application seems to be (computational) lexicography. In a
frame-based lexicon, the frame accounts for related senses of a
single word and its semantic relations to other words. A framebased lexicon, therefore, offers more comprehensive information than the traditional lexicon. An example is Petruck (1986), a
study of the vocabulary of the body frame in modern Hebrew. An
example of computational lexicography is the FrameNet-System
(see Boas 2002).
Formalization
Although frame semantics does not lend itself easily to formalization there is an early approach by J. M. Gawron (1983) in which
basic insights of frame semantics were formalized by notations
like those of the LISP programming language, in combination
with situation semantics. A more recent approach is presented in
van Lambalgen and Hamm (2005) in which scenarios a concept
closely related to the frame concept are formalized as certain
kinds of logic programs. An explicit formalization of the combination of frame semantics and construction grammar based on
this work can be found in Andrade-Lotero (2006).
Fritz Hamm
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Aitchison, J. 1994. Words in the Mind. Malden, MA: Blackwell.
Andrade-Lotero, E. 2006. Meaning and Form in the Event Calculus.
M.A. thesis, University of Amsterdam, MOL-200601.
Atkins, B. T. S. 1995. The role of the example in a frame semantics dictionary. In Essays in Semantics and Pragmatics, ed. M. Shibatani and
S. Thompson, 2542. Amsterdam: John Benjamins.
Boas, H. 2002. Bilingual FrameNet dictionaries for machine translation. In Proceedings of the Third International Conference on Language
Resources and Evaluation. Vol. 4. Ed. M. G. Rodriguez and C.P.S.
Araujo, 136471. Las Palmas de Gran Canaria, Spain: University of Las
Palmas de Gran Canaria.
Fillmore, C. 1977a. The case for case reopened. In Syntax and Semantics 8:
Grammatical Relations, ed. P. Cole, 5981. New York: Academic Press.
321
Framing Effects
. 1977b. Scenes-and-frames semantics. In Linguistic Structure
Processing, ed. A. Zambolli, 5582. Amsterdam: North Holland.
. 1977c. Topics in lexical semantics. In Current Issues in
LinguisticTheory, ed. R. W. Cole, 76138. Bloomington: Indiana
University Press.
Fillmore, C., and C. Baker 2000. Frame Net. Available online at: http://
www.icsi.berkeley.edu/framenet.
Gawron, J. M. 1983. Lexical representation and the semantics of complementation. Ph.D. diss., University of California, Berkeley.
Goldberg, A. 1995. Constructions: A Construction Grammar Approach to
Argument Structure. Chicago: University of Chicago Press.
Hamm, F., H. Kamp, and M. van Lambalgen. 2006. There is no opposition between formal and cognitive semantics. Theoretical Linguistics
32: 140.
Langacker, R. 1987. Foundations of Cognitive Grammar. Vol. 1. Theoretical
Prerequisites. Stanford, CA; Stanford University Press.
Minsky, M. 1975. A framework for representing knowledge. In The
Psychology of Computer Vision, ed. P. H. Winston, 21177. New York:
McGraw-Hill.
Petruck, M. 1986. Body part terminology in Hebrew. Ph.D. diss.,
University of California, Berkeley.
van Lambalgen, M., and F. Hamm. 2005. The Proper Treatment of Events.
Malden, MA: Blackwell.
For more information, the interested reader is advised to consult the
Frame Semantics Bibliography drawn up by Jean Mark Gawron.
FRAMING EFFECTS
A framing effect is usually said to occur when equivalent descriptions of a decision problem lead to systematically different
decisions. Framing has been a major topic of research in the psychology of judgment and decision making and is widely viewed
as carrying significant implications for the rationality debate
(e.g., Shafir and LeBoeuf 2002). Framing effects are commonly
taken as evidence for incoherence in human decision making
and for the empirical inapplicability of the rational actor models
used by economists and other social scientists. The first part of
this entry presents a brief review of the empirical phenomena;
the second part describes the standard normative interpretation
of these empirical effects. Although the literature has not typically focused on the structure of human conversational environments, framing effects involve utterances selected by a speaker
for a listener. A final section considers the possible implications
of communicative factors for a normative and descriptive understanding of framing effects.
Empirical Review
In this section, we follow I. P. Levin, S. L. Schneider, and G. J. Gaeths
(1998) taxonomy of framing effects into three categories: attribute
framing, risky choice framing, and goal framing.
In attribute framing, a single attribute of a single object is
described in terms of either a positively valenced proportion
or an equivalent negatively valenced proportion. The subject
is then required to provide some evaluation of the object thus
described. The typical finding is a valence-consistent shift (Levin,
Schneider, and Gaeth 1998): Objects described in terms of a
positively valenced proportion are generally evaluated more
favorably than objects described in terms of the corresponding negatively valenced proportion. For example, in one study,
322
beef described as 75% lean was given higher ratings than beef
described as 25% fat (Levin and Gaeth 1988); similarly, research
and development (R&D) teams are allocated more funds when
their performance rates are framed in terms of successes rather
than failures (Duchon, Dunegan, and Barton 1989). The valenceconsistent shift in attribute framing is a robust effect, observed in
a large range of experimental environments, with obvious implications for marketing and persuasion.
In risky choice framing, subjects are presented with two
options in a forced-choice task. The two options are typically
gambles that can be described in terms of proportions and probabilities of gains or losses. Usually, one of these options is a sure
thing (in which an intermediate outcome is specified as certain),
while the other is a risky gamble (in which extreme good and bad
values are both assigned non-zero probabilities). The gamble and
sure thing are both described either in terms of gain outcomes
and probabilities or in terms of equivalent loss outcomes and
probabilities. The two options are usually equated in expected
value (i.e., the mean outcome expected over many repeated trials), enabling the framing researcher to interpret observed patterns of preference in terms of subjects risk attitudes. Within this
rubric, preferences for the sure thing indicate risk aversion, and
preferences for the gamble indicate risk seeking. The best-known
risky choice framing problem is the so-called Asian Disease
Problem (Tversky and Kahneman 1981). In it, subjects first read
the following background blurb:
Imagine that the U.S. is preparing for the outbreak of an unusual
Asian disease, which is expected to kill 600 people. Two possible
programs to combat the disease have been proposed. Assume
that the exact scientific estimates of the consequences of these
programs are as follows:
Framing Effects
corresponding disadvantages of not participating. The most
common result is that subjects are more likely to engage in the
activity when the disadvantages of not engaging, rather than the
advantages of engaging, are emphasized (Levin, Schneider, and
Gaeth 1998).
Normative Analysis
Risky-choice framing effects have been put forward as positive
evidence for prospect theory (Kahneman and Tversky 1979),
a theory of choice that aims to be both formally tractable and
cognitively realistic. However, the focus in the framing literature has been largely on the negative evidence that framing
effects allegedly raise against classical expected utility theory
and other so-called rational actor models. The literature on
attribute framing, in particular, is concerned almost exclusively
with the normative and practical implications of the empirical effects. Framing effects, D. Kahneman has noted, are
less significant for their contribution to psychology than for
their importance in the real world and for the challenge they
raise to the foundations of a rational model of decision making (2000, xv). This raises the important questions: Are framing effects always counternormative? And if so, what norm or
norms do they violate?
In an important paper, A. Tversky and Kahneman argued that
framing effects violate a bedrock normative condition of description invariance [a]n essential condition for a theory of choice
that claims normative status so basic that it is tacitly assumed
in the characterization of options rather than explicitly stated as
a testable axiom (1986, S253). Any theory of rational choice, they
argued, must stipulate that the same problem will be evaluated
in the same way, regardless of how the problem is described
thus, equivalent descriptions should lead to identical decisions.
Expected utility theory, for example, satisfies this principle: it
evaluates choice options strictly as a function of probability and
outcome, with no specification of probability-outcome framing. This reducibility of decision problems to a canonical form
is clearly a theoretical convenience; the principle of description
invariance states that it is also a normative requirement. Because
the framing phenomena observed both in the laboratory and in
real-world situations violate the description invariance principle, these effects are taken to imply that no theory of choice
can be both normatively adequate and descriptively accurate
(ibid., S251).
described in the previous section, applies to listener effects without any consideration of associated speaker phenomena (i.e.,
regularities in how speakers choose frames in typical linguistic
environments). Researchers have tended to interpret the experimental effects as if the experimenter had somehow surgically
implanted a framing of the decision problem into the subjects
brain. However, because linguistic utterances are employed,
regularities in speaker behavior may be relevant to the normative and descriptive understanding of listener behavior: If speakers tend to choose different frames as a function of background
conditions, then listeners may reasonably draw inferences from
the speakers choice of frame. If knowledge of these background
conditions is relevant to the listeners choice, then the frames,
while logically equivalent, would not be information equivalent.
S. Sher and C. R. M. McKenzie (2006; cf. McKenzie and Nelson
2003) argued that the frames studied in the attribute framing
literature are commonly information nonequivalent, because
speakers tend to frame options in terms of attributes that are
relatively salient. For example, a generally impressive R&D team
is more likely to be described in terms of its success rate than a
generally incompetent team with the same success/failure rate.
A positive frame thus highlights the salience of the positive attribute in the speakers conception of the option information relevant to its evaluation.
Experiments convey information to subjects in framed statements, and researchers have generally assumed that the only
information content is logical information content. The framing
of the logical content is assumed not to convey information but
simply to influence the listeners construal of the logical content.
In this way, the usual normative analysis of framing experiments
leans on an implicit assumption of the information equivalence
of logically equivalent frames. However, while the logical equivalence of a pair of frames can usually be determined on inspection
(though see Jou, Shanteau, and Harris 1996), a determination of
information equivalence requires empirical study of the human
communicative environments in which speakers typically frame
objects and options. At least in the domain of attribute framing,
logical equivalence does not imply information equivalence.
Whether the study of communicative environments will have
similar implications for traditional normative conclusions drawn
in risky choice and goal framing is an open question. There also
remain important questions about how, and how flexibly, listeners use subtle information that is, in principle, available in particular framing experiments. However, an analysis of speaker
regularities in human communicative environments is likely to
be of some significance in any research area in which information presented to experimental subjects is evaluated against a
normative standard of information equivalence (cf. Hilton 1995;
McKenzie 2004; Sher and McKenzie 2008; Schwarz 1996).
Shlomi Sher and Craig R. M. McKenzie
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Duchon, D., K. J. Dunegan, and S. L. Barton. 1989. Framing the problem
and making decisions: The facts are not enough. IEEE Transactions on
Engineering Management (February): 257.
Hilton, D. J. 1995. The social context of reasoning: Conversational inference and rational judgment. Psychological Bulletin 118: 24871.
323
Frontal Lobe
Jou, J., J. Shanteau, and R. J. Harris. 1996. An information processing
view of framing effects: The role of causal schemas in decision making. Memory and Cognition 24: 115.
Kahneman, D. 2000. Preface. In Choices, Values, and Frames, ed.
D. Kahneman and A. Tversky, ixxvii. Cambridge: Cambridge
University Press.
Kahneman, D., and A. Tversky. 1979. Prospect theory: An analysis of
decision under risk. Econometrica 47: 26391.
Levin, I. P., and G. J. Gaeth. 1988. How consumers are affected by the
framing of attribute information before and after consuming the product. Journal of Consumer Research 15: 3748.
Levin, I. P., S. L. Schneider, and G. J. Gaeth. 1998. All frames are not
created equal: A typology and critical analysis of framing effects.
Organizational Behavior and Human Decision Processes 76: 14988. A
thorough taxonomy, review, and analysis of attribute, risky choice, and
goal framing effects.
McKenzie, C. R. M. 2004. Framing effects in inference tasks and
why they are normatively defensible. Memory and Cognition 32:
87485.
McKenzie, C. R. M., and J. D. Nelson. 2003. What a speakers choice of
frame reveals: Reference points, frame selection, and framing effects.
Psychonomic Bulletin and Review 10: 596602.
Schwarz, N. 1996. Cognition and Communication: Judgmental
Biases, Research Methods, and the Logic of Conversation. Mahwah,
NJ: Erlbaum. A summary of empirical evidence indicating that many
supposed shortcomings of human judgment are due to experimental subjects going beyond the literal meaning of the information
provided by the researcher and drawing on the pragmatic meaning.
Shafir, E., and R. A. LeBoeuf. 2002. Rationality. Annual Review of
Psychology 53: 491517.
Sher, S., and C. R. M. McKenzie. 2006. Information leakage from logically equivalent frames. Cognition 101: 46794.
. 2008. Framing effects and rationality. In The Probabilistic
Mind: Prospects for Bayesian Cognitive Science, ed. N. Chater and M.
Oaksford, 7996. Oxford: Oxford University Press. A summary of problems of information nonequivalence in the framing literature and of
the application of information equivalence to other areas of psychological research.
Tversky, A., and D. Kahneman. 1981. The framing of decisions and the
psychology of choice. Science 211: 4538.
. 1986. Rational choice and the framing of decisions. Journal of
Business 59: S25178. An influential discussion of risky choice framing
effects and their implications for models of rational choice.
FRONTAL LOBE
The evolution of the frontal lobes in humans is strongly associated with the advent of language, the primary behavior distinguishing humans from other primate species. Although not
much is known about how the frontal lobes evolved, we do know
that there was a marked expansion 3.3 to 2.5 million years ago
in Australopithecus Africanus, a prehistoric form of human. This
frontal lobe development, coupled with growth of the temporal
lobes also associated with language function, coincides with the
creation and use of tools. Tool use may have given rise to the
development of hand preference and eventual lateralization of
language functions. Verbal communication likely evolved from
hand signals to vocalizations, which progressively came to signify common referents (i.e., words) used in spoken language.
Thus, the genesis of a linguistic symbolic communication system influenced, and in turn was influenced by, brain growth and
324
reorganization. The primary role of the frontal lobes has developed from behavior regulation through language, giving rise to
abstract thinking and the ability to plan and execute complicated
actions. It may very well be that the development of language
and the frontal lobes propelled humans from a hunting and
predatory species to a social species, capable of modifying and
manipulating their environments.
Frontal Lobe
325
Frontal Lobe
Brocas area consists of the foot of the third frontal convolution (the inferior frontal gyrus), just in front of the motor strip.
Traditionally thought to be responsible for the production of language, it has also been implicated in comprehension of syntax.
As with any of the perisylvian regions involved in language, it
appears to contribute to lexical access as well.
The second major portion of the frontal lobe, in addition to
the motor areas just discussed, is the prefrontal cortex, located
in the foremost part of the frontal lobe, in front of the motor
area. The ratio of the prefrontal cortex to the rest of the cortex is
larger in humans than in any other species. The prefrontal cortex
is divided into three major regions: dorsolateral, orbitofrontal,
and medial prefrontal cortex. Although there is some controversy, currently many researchers associate different behavioral
changes and cognitive disturbances with each of these three
prefrontal regions. Dorsolateral prefrontal cortex is suggested to
have a role in working memory, attention, planning, and problem solving. Lesions in this area cause an inability to plan correctly and difficulty with multistep tasks. The orbitofrontal cortex
is associated with emotion and the initiation of behavior. Lesions
of this area lead to disinhibition, socially inappropriate behavior, and change of affect. Finally, the medial/cingulate prefrontal cortex is related to motivation and drive. Lesions in this area
result in apathy, loss of interest in ones life, and reduced spontaneous speech and movement.
The prefrontal cortex is extensively connected to the motor,
perceptual, and limbic systems of the brain. It sends and receives
great amounts of information and controls cognitive processes
so that appropriate behaviors are carried out at the correct time
and place. It takes part in higher aspects of motor control and
various cognitive functions, such as inhibition, problem solving, abstract thinking, emotional control, and appropriate social
behavior. Neuropsychologists consider the prefrontal cortex
particularly responsible for executive function, by which they
mean the ability to plan, monitor, and make inferences, as well
as related self-governing behaviors. The old notion of a computer with a central processor that determines how the rest of
the systems computational abilities are to be used is sometimes
employed when the role that executive function plays vis--vis
other cognitive abilities is discussed. When distinction is made
between automatic and controlled processes, it is the controlled
processes that the prefrontal cortex controls. Cognitive flexibility
is also considered a role played by the frontal lobes. working
memory the ability to keep a limited amount of information
available for analysis is either assumed to be a part of executive
function or to be closely linked to it.
Language behaviors that involve executive function are the less
automatic ones, such as appreciating humor, making inferences
from nonexplicit phrasing, appreciating the pragmatic aspects of
communication, monitoring ones speech for speech errors and
correcting them, monitoring ones interlocutors to rephrase or
shift register if they appear not to be comprehending, selecting
the right language or mix of languages to speak in bilingual and
polyglot situations, avoiding culturally taboo words or topics that
would be inappropriate in a given context, and reanalyzing input
when garden-pathed or on other occasions when it appears not
to make sense. One language task that appears to put substantial
burden on working memory is simultaneous interpretation.
326
Functional Linguistics
survived the serious accident, he was no longer himself. Before
the accident, Gage had been known as a hard-working, capable,
and sociable man, but his personality changed radically after the
accident. He uttered profanities, becoming impatient, obstinate,
fitful, and antisocial. Indeed, in Gages case and similar ones,
difficulty in socializing follows frontal lobe brain damage. Thus,
while no overt language problems regarding the form of language, such as grammar, or regarding language content, such as
semantic meaning, were reported in these patients, their use of
pragmatic language skills was impaired.
In sum, the frontal lobes play a crucial role in various speech
and language processing tasks, including initiation of articulatory movement, phonetic/phonological selection, morphosyntactic production, syntactic manipulation, and integration of
pragmatic function. Depending on the location and depth of the
frontal lobe lesion, a mixture of cognitive and behavioral symptoms can appear with the disorders of these speech and language
functions as well.
JungMoon Hyun, Elizabeth Ijalba, Teresa M. Signorelli,
Peggy S. Conner, and Loraine K. Obler
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Anderson, Steven W., Hanna Damasio, Daniel Tranel, and Antonio
R. Damasio. 2000. Long-term sequelae of prefrontal cortex damage acquired in early childhood. Developmental Neuropsychology
18: 28196.
Baddeley, Alan. 2003. Working memory and language: An overview.
Journal of Communication Disorders 36: 189208.
Damasio, Hanna, Daniel Tranel, Thomas Grabowski, Ralph Adolphs, and
Antonio R. Damasio. 2004. Neural systems behind word and concept
retrieval. Cognition 92: 179229.
Miller, Bruce L., and Jeffrey L. Cummings. 2006. The Human Frontal
Lobes: Functions and Disorders. New York: Guilford.
Penfield, W., and T. Rasmussen. 1950. The Cerebral Cortex of Man A
Clinical Study of Localization of Function. New York: Macmillan.
FUNCTIONAL LINGUISTICS
Functional linguistics, broadly defined, includes a wide range of
diverse approaches that highlight the interdependence of language structure and language function. In this view, structural
features of languages have evolved and continue to develop as
a result of competing cognitive, communicative, ecological, and
social pressures. (See also grammaticalization.) Functional
research treats language neither as a purely formal object nor as a
closed autonomous system; rather, it considers language form to
be tightly integrated with both the uses and users of language (see
also usage-based theory).
Within the context of current theories of linguistics, many
functional approaches differ sharply from so-called formal theories, such as government and binding or minimalism. While
lines of demarcation between formal and functional approaches
are, of course, not always clear-cut, formal theories tend to begin
with the premise that grammar is innate, to assume the autonomy of syntax, to hold to the modularity of language from
other cognitive faculties, and to impose a distinction between
competence and performance. Functionally oriented linguistics, on the other hand, tends to question the validity of these
327
Functional Linguistics
describe the conditions for the occurrence and nonoccurrence
of the complementizer that in a corpus of English conversation
(e.g., in sentences like You can tell that its going to rain versus
I think its going to rain). They find the distribution of that to
be probabilistically determined by functional factors, such as
topic management and strength of epistemic commitment. The
explanation of these findings rests in the observed blurring of
the distinction between main and subordinate clauses in these
contexts, whereby the so-called main clause in constructions
without that are actually best analyzed as epistemic adverbial
phrases (see also Thompson 2002). These findings suggest that
the traditional binary distinction between main and subordinate clauses is not categorical, as it emerges out of specific communicative situations in natural discourse. (See Cumming and
Ono 1997 for a thorough explication of the discourse-functional
approach, including further examples and references; see also
emergent structure.)
In addition to communicative and discourse factors, functional research of the last several decades has also focused on the
role of semantics in shaping grammar (Bolinger 1977 inter alia;
Halliday 1994). In these views, meaning is central in motivating
form, and, therefore, particular grammatical constructions have
the forms they do because of the particular meanings that they
convey. Two noteworthy paradigms emerging from this tradition include cognitive linguistics, which views language
form as motivated by meaning and conceptualization, and the
cross-linguistic approach known as typological-functional linguistics, which explores the ways in which similar meanings are
expressed in the grammars of languages around the world (see
cognitive grammar and typology.)
Other strands of functional research investigate the ways in
which neurological, cognitive, and physiological faculties of the
human user motivate and constrain grammar and phonology.
For example, Sidney M. Lamb (1999) outlines a connectionist,
relational network model of language based in what is known
about neurons and connections among neurons and the neurological processes of activation and inhibition. Functional theories
of phonology likewise take seriously the conceptual, neurological, and physiological aspects of language users, for example,
articulatory phonology (Browman and Goldstein 1986), which
views the abstract phonological system as constrained by the
physical system of articulatory phonetics. More recently,
Joan L. Bybee (2001) has proposed an exemplar-based theory
of phonology, which likewise treats phonology as grounded in
the concrete physiology of phonetics and considers speech as
a neural-motor activity subject to the same processes observed
in other complex motor activities such as routinization, simplification, and entrenchment based on frequency. (See also Bybee
2006 for an extension of this approach to grammar in general.)
Recent functional research has also begun to concentrate on
ways in which the inherent social and interactional nature of language users likewise motivates and constrains language form. One
socially oriented approach has become known as interactional
linguistics, incorporating insights from sociology especially the
subfields of ethnomethodology and conversation analysis.
Interactional linguistics views language as a form of highly systematic social action. Interactional research seeks to describe the
prosodic and syntactic forms that regularly accomplish specific
328
G
GAMES AND LANGUAGE
Rules and Language Games
One traditional view in philosophy and linguistics is that without rules of usage common to the speaker and the listener,
1st day
D
C
D
(1,1)
(0,4)
C
(4,0)
(3,3)
Pa Qb Qc
(1,1)
(1,1)
(1,1)
(1,1)
Thus, in the Prisoners dilemma game, (C,C) is a Nash equilibrium, but there is no Nash equilibrium in the Matching pennies
game. There are games that have two Nash equilibria, like the
following one where both (L,L) and (R,R) are Nash equilibria:
L
(1,1)
(1,1)
(1,1)
(1,1)
329
330
Extensive Games
It is customary to exhibit extensive games as a sequence
G = (N,H,P,(ui)iN)
We witness here a phenomenon of semantical dependence: Before an expression (She) gets an interpretation,
another expression that is its head (A woman) must get an
interpretation. The semantical games in this case are completely
analogous to the quantifier game (in fact, the game involves
quantifiers). The rules of the game will contain not only choices
prompted by quantified expressions but also choices prompted
by the anaphoric pronoun: She prompts a move by the existential player, who must now choose the same individual chosen
earlier as a possible value of the indefinite A woman.
Gabriel Sandu
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Dekker, Paul, and Robert van Rooy. 2000. Bi-directional optimality theory: An application of game-theory. Journal of Semantics 17: 21742.
Grice, Paul. 1975. Logic and conversation, In Syntax and Semantics 3:
Speech Acts, ed. P. Cole and J. L. Morgan. New York: Academic Press.
Hintikka, Jaakko, and Gabriel Sandu. 1991. On the Methodology of
Linguistics. Oxford: Basil Blackwell.
Lewis, David. 1969. Conventions. Cambridge: Harvard University Press.
Stenius, Erik. 1967. Mood and language-game. Synthese 17: 25474.
331
332
avoided by women and men. Some of the reasons for this differentiation have to do with the influence of gender norms and stereotypes. Girls may be instructed by parents and teachers that the
use of features that index modesty or deference is appropriate for
them, whereas boys may be ridiculed for using those same features; speakers in each gender group may develop an investment
in using the features to the extent that they also have an investment in being judged as gender appropriate. The approach, however, allows for the possibility that not everyone does develop
such an investment: There have been various studies of groups
whose behavior appears to be shaped by a conscious refusal of
gender appropriateness (e.g., Abe 2004; Bucholtz 1999; Okamoto
1995.) There are also speakers (e.g., some transgendered or
transsexual individuals) whose investment in using features that
will index their adopted gender identity is such that they produce
extreme gender stereotypes (Kulick 1999).
Other reasons for using or avoiding features that indirectly
index gender, though, have more to do with the demands of the
activities in which speakers are engaged. Bonnie McElhinny
(1995) reports that women police officers in Pittsburgh adopt a
relatively affectless style of interaction regarded by some observers as defeminizing, but that they are quite clear that they are not
trying to talk like men; they are trying to talk like police officers.
They also believe that the style of talk required is not simply a
contingent norm reflecting the historical domination of policing
by men but is intrinsically demanded by the nature of the work.
Although some styles do have gendered connotations (and histories) that can pose problems for individuals whose gender is
stereotypically incongruent with them, it is clear that the way
men and women behave in different contexts has as much to do
with the nature of those contexts as with gender per se.
It is also clear that gender itself is not always and everywhere
indexed in similar ways because there is cross-cultural and historical variation in the social roles allotted to men and women
and the qualities ascribed to them (linguistic markers of which
become secondary indices of gender). The Japanese association
of femininity with delicacy may seem natural to Westerners,
too, but the association is not made by, for instance, the villagers of Gapun in Papua, New Guinea, who characterize womens
language, like women themselves, as blunt, direct, and aggressive (Kulick 1993). Nor can it be automatically assumed that such
associations, ideologically powerful though they may be, determine the actual behavior of most speakers. For instance, recent
research analyzing the speech of working-class and rural women
in different parts of Japan points to the practical irrelevance for
many Japanese women of the idealized normative construct
womens language (Okamoto and Shibamoto Smith 2004).
Whereas early language and gender researchers often took
issue with prefeminist generalizations about mens and womens
language use, more recent researchers influenced by the shifts
just outlined have revisited many of the generalizations made
during the 1970s and 1980s. The classic claims of variationist
sociolinguistics about gender (that women are generally closer
than men to prestige norms and lead in change from above
because of their greater status consciousness, but are otherwise
conservative) have been substantially revised (e.g., Labov 2001).
While the new variationist orthodoxy is still a gender generalization (that women tend to lead in both change from above and
as men is one way to deal with this marginal social positioning. If so, it is evident that inequality, rather than just difference,
shapes the relationship of language use to gender.
A recent external development to which researchers are now
beginning to respond is the rise of scientific paradigms such
as evolutionary psychology, which, in addition to being
generally critical of feminist social constructionism, have made
specific claims (some of them empirically ill-founded see,
e.g., Hyde 2005) about malefemale linguistic differences as
the hardwired products of millennia of natural selection. The
recent resurgence of biological essentialism in both academic
and popular culture constitutes a challenge, both intellectual
and political, that language and gender researchers, in my view,
should not ignore. Yet while in the future there may well be more
discussion of the relationship between sex and gender, I think
it is unlikely that researchers will abandon the commitment to
(some variant of) the social constructionism that has proved so
productive in recent years.
Deborah Cameron
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Abe, Hideko. 2004. Lesbian bar talk in Shinjuku, Japan. In Okamoto
and Shibamoto Smith 2004, 20521.
Baxter, Judith, ed. 2006. Speaking Out: The Female Voice in Public
Contexts. Basingstoke, UK: Palgrave.
Brown, Penelope. 1980. How and why are women more polite? In
Women and Language in Literature and Society, ed. Sally McConnellGinet, Ruth Borker, and Nelly Furman, 11149. New York: Praeger.
Bucholtz, Mary. 1999. Why be normal? Language and identity practices
in a community of nerd girls. Language and Society 28: 20323.
Butler, Judith. 1990. Gender Trouble: Feminism and the Subversion of
Identity. New York: Routlege.
Cameron, Deborah. 2006. On Language and Sexual Politics.
London: Routledge
Eckert, Penelope. 1990. The whole woman: Sex and gender differences
in variation. Language Variation and Change 1: 24568.
. 2000. Gender and sociolinguistic variation. In Language and
Gender: A Reader, ed. Jennifer Coates, 6475. Oxford: Blackwell.
Eckert, Penelope, and Sally McConnell-Ginet. 1992. Think practically
and look locally: Language and gender as community-based practice.
Annual Review of Anthropology 12: 46190.
. 2003. Language and Gender. Cambridge: Cambridge University
Press.
Hellinger, Marlis, and Hadumod Bussman, eds. 20013. Gender Across
Languages: The Linguistic Representation of Women and Men. 3 vols.
Amsterdam: John Benjamins.
Holmes, Janet. 1995. Women, Men and Politeness. London: Longman.
. 2006. Gendered Talk at Work. Malden, MA: Blackwell.
Holmes, Janet, and Miriam Meyerhoff, eds. 2003. The Handbook of
Language and Gender. Malden, MA: Blackwell.
Hyde, Janet Shibley. 2005. The gender similarities hypothesis. American
Psychologist 60: 58192.
Johnson, Sally, and Ulrike H. Meinhof, eds. 1997. Language and
Masculinity. Oxford: Blackwell.
Kulick, Don. 1993. Speaking as a woman: Structure and gender in domestic arguments in a Papua New Guinea village. Cultural Anthropology
8: 51041.
. 1999. Transgender and language. GLQ 5: 60522.
Labov, William. 2001. Principles of Linguistic Change. Vol. 2. Social
Factors. Oxford: Blackwell
333
Gender Marking
Lakoff, Robin. 1975. Language and Womans Place. New York: Harper
and Row.
Maltz, Daniel, and Ruth Borker. 1982. A cultural approach to malefemale misunderstanding. In Language and Social Identity, ed. John J.
Gumperz, 196216. Cambridge: Cambridge University Press.
McElhinny, Bonnie. 1995. Challenging hegemonic masculinities: Female and male police officers handling domestic violence.
In Gender Articulated, ed. Kira Hall andMary Bucholtz, 21743.
London: Routledge.
Mills, Sara. 2003. Gender and Politeness. Cambridge: Cambridge
University Press.
Ochs, Elinor. 1992. Indexing gender. In Rethinking Context: Language
as an Interactive Phenomenon, ed. Alessandro Duranti and Charles
Goodwin, 33558. Cambridge: Cambridge University Press.
Okamoto, Shigeko. 1995. Tasteless Japanese: Less feminine speech
among young Japanese women. In Gender Articulated, ed. Kira Hall
and Mary Bucholtz, 297325. London: Routledge.
Okamoto, Shigeko, and Janet Shibamoto Smith, eds. 2004. Japanese
Language, Gender and Ideology. New York: Oxford University
Press.
Walsh, Clare. 2001. Gender and Discourse: Language and Power in
Politics, the Church and Organizations. London: Longman.
GENDER MARKING
Almost all languages have some grammatical means of dividing up their noun lexicon into distinct classes, with devices
or markers occurring in surface structures under specifiable
conditions and providing information about the semantic
characteristics of the referent of the nominal head of the noun
phrase. Gender marking is one such device, typically found
in languages with a fusional or agglutinating profile; other
devices are frequently grouped under the term classifiers and
are typically found in isolating languages. The term gender is
used both for the particular classes of nouns (a language may
have two or more genders, or noun classes) and for the whole
grammatical feature (a language may or may not have the feature of gender).
There is always some semantic basis to gender classification,
though it may be supplemented with additional formal (phonological and morphological) criteria. The semantic
criteria include humanness, animacy, sex, shape, form, consistency, and functional properties. A minimal gender system
consists of two genders (e.g., French), and this is the most common system; it was found in 50 languages of a sample of 256
(Corbett 2005). Three-gender systems (e.g., Russian) appear to
be roughly half as common, and larger systems are increasingly
less common. The largest system found so far is Nigerian Fula
with around twenty genders (the exact count depending on the
dialect). However, 144 of the 256 languages had no gender
system.
Semantic distinctions between classes of nouns, even lexical
derivations (e.g., the English poet versus poetess), do not in themselves make genders. This is because it is taken as the definitional
characteristic of gender that some constituent outside the noun
itself must agree in gender with the noun. Thus, gender refers
to classes of nouns within a language that are reflected in the
behavior of the associated words (Hockett 1958, 231), and a language has a gender system only if we find different agreements
334
Generative Grammar
dependent on nouns of different classes, regardless of whether
or not the nouns themselves bear gender markers.
Agreement in gender with the head noun can be found in
other words in the noun phrase (adjective, determiner, demonstrative, numeral, etc., even focus particle), in the predicate of the
clause, an adverb, and arguably in an anaphoric pronoun
outside the clause boundary. For example, in Polish the feminine noun skarpeta sock (the controller) requires that many
other elements (targets) in the clause agree with it in gender: ta
jedna stara porwana skarpeta, ktra leaa na pododze this.f
one.f old.f torn.f sock(f) which.f lay.f on floor. Markers of
gender often do not mark gender alone but may be portmanteau
markers that combine information about gender with number,
person, case, or other features.
If antecedentanaphor relations are accepted as agreement,
languages in which free pronouns present the only evidence
for gender (gender distinctions being absent from noun phrase
modifiers and from predicates) can be counted as having a
(pronominal) gender system. Such languages are rare the best
known example is English, which is typologically unusual (see
typology) in this respect (Corbett 2005); another is Defaka
(Niger-Congo).
Anna Kibort
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Corbett, Greville G. 1991. Gender. Cambridge: Cambridge University
Press.
. 2005. Number of genders. In The World Atlas of Language
Structures, ed. Martin Haspelmath, Matthew S. Dryer, David Gil, and
Bernard Comrie, 1269. Oxford: Oxford University Press.
Hockett, Charles F. 1958. A Course in Modern Linguistics. New York:
Macmillan.
GENERATIVE GRAMMAR
The approach to linguistics known as generative grammar
(GG) was initially introduced by Noam Chomsky in the 1950s,
and though it has developed continuously ever since, the core
assumptions have remained remarkably constant. For instance,
at the highest level, GG has always maintained that a grammar
effectively constitutes a set of formal rules that recursively enumerate all, and only all, the well-formed (i.e.,grammatical)
sentences of a language (see recusion, iteration, and
metarepresentation ). As part of the process of constructing such a model of linguistic knowledge, GG research has
consistently attempted to describe how a speaker-hearer can
generate and comprehend an infinite number of grammatical
sentences despite encountering only a finite amount of primary
linguistic data (PLD) while learning any given language. In
order to account for this apparent conundrum, GG standardly
assumes that the language capacity is a genetic endowment (see
innateness and innatism ) that is distinctive to homo sapiens and which specifies those aspects of linguistic knowledge
that are genetically determined, as opposed to those that must
be acquired via contact with PLD. As a result of this emphasis
on language acquisition, GG is often closely associated with
biolinguistics.
Generative Grammar
The following sections summarize the main stages in the
development of GG from the 1950s to the present.
335
Generative Grammar
Figure 1.
LEXICON
LEXICON
D-STRUCTURE
NARROW SYNTAX
Move
PHONETIC FORM
S-STRUCTURE
LOGICAL FORM
Figure 3.
PHONETIC FORM
LOGICAL FORM
Figure 2.
336
Generative Grammar
Generative Poetics
GENERATIVE POETICS
Generative poetics comprises all theories that seek to explain the
production and reception of literary works by reference to a set
of rules or algorithmic procedures. It is closely related to cognitive poetics. However, generative poetics has tended to draw
inspiration from Chomskyan generative grammar. In contrast, many writers in cognitive poetics have aligned their analyses with cognitive linguistics.
Early work in generative poetics tended to track developments
in generative grammar. In Noam Chomskys usage, a generative
grammar produces all and only the grammatical sentences of
a language. For the early theorists of generative poetics, then,
a narrative grammar should produce all and only well-formed
narratives. Moreover, in both cases, it was commonly thought
that this goal is best accomplished through a transformational grammar.
Much of this early work was illuminating and important.
However, there are several problems with modeling, for example, narratology on linguistic theory. First, linguistic theories
may change rapidly. If one bases ones narrative theory on any
specific syntactic theory, ones narrative theory will probably be
outmoded by the time it is published. Second, there is no reason
to believe, a priori, that rules for the generation of stories should
directly parallel rules of syntax anyway. Finally, it is not clear that
there is any narrative counterpart to grammaticality. In other
words, the ambition of generating all and only well-formed
stories may be misguided. (The situation is, of course, different
for areas of poetics that are directly governed by linguistic rules.
Indeed, much of the most important work in generative poetics has been done in such areas; see meter, verse line, and
poetic form, universals of.)
The point about well-formedness is worth considering in further detail. There are, of course, speech actions that are clearly
not stories and speech actions that are. The difficulty is that
there is a gradient of more or less marginal cases, rather than a
strict division between stories and nonstories. In short, story is
a prototype concept. Given the prototype nature of story, it
is not clear just how the relevant data are organized, thus, just
what needs explaining. In other words, it is not clear precisely
what structures the generative rules should generate.
More exactly, the outputs of a generative system clearly
need to match the relevant data. At least two aspects of the
Chomskyan approach are generalizable here. First, the system
should not overgenerate. For example, a generative grammar
might produce all grammatical sentences. But it is still invalid if it
also produces The the of at as a grammatical sentence. Second,
the grammar cannot produce only sentences that have already
occurred. Such a grammar would be falsified as soon as someone
uttered a new sentence. In short, the data comprise all possible
speech actions of the relevant type (sentences, narratives) and
no impossible ones. This returns us to the issue of intermediate cases. However we determine the data, a generative system
337
Generative Poetics
should produce structures in keeping with the categorial organization of the data. If the data have a sharp well-formed/not wellformed division with few intermediate cases, the system should
generate structures divided in this way. If the data involve a more
gradual gradient from excluded to included cases as we find
with stories then the system should generate structures organized in this way.
In addition to this difference in the data, there is a normative
component to our concept of stories that is largely absent from
our concept of sentences. We routinely refer to some stories as
better than others. However, we do not commonly think of some
sentences as better than others.
Although it is not usually referred to as generative poetics,
some recent work in cognitive science and literature does fall
into that category as defined here. This work avoids the problems of earlier approaches by developing theoretical principles
independently of particular grammatical theories and by recognizing the prototype nature of our ordinary language concept of
stories.
In order to understand this recent work, we need to consider
what constitutes a generative rule system for a speech action.
Most obviously, a generative system needs a productive component and a receptive component. Parts of these components
will be directly parallel. In other words, many rules of production and reception have to be systematically coordinated so that
when I say Could you ask John to pass the salt? my addressee
will understand the question in such a way that I will most often
end up getting the salt. On the other hand, some parts of these
components will necessarily be different. For example, not everything involved in producing an alliterative, rhyming, metered
verse line is also involved in reading that line.
Whether speaking of productive or receptive speech actions,
there are several ways in which we could organize the rules that
constitute a generative system. No matter how we do this, we are
likely to have processes (e.g., activation), structures in which processes operate (e.g., episodic memory), and elements on which
processes operate (e.g., memories of particular experiences that
we activate from episodic memory). Note that these divisions
need not be absolute. Processes could themselves be the elements on which other processes (meta-rules) operate. Moreover,
some elements may incorporate processes operating on other
elements. We see this in the case of scripts, such as the script
for eating at a restaurant. When these scripts are activated, they
guide our speech and behavior by integrating various preexisting
processes and other elements. For example, when we decide to
eat at a restaurant and follow our script for doing so, such processes as activating our prototype for a menu are involved so
that we can recognize menus and respond appropriately when
the server holds one out to us. Finally, different processes may
only operate at different derivation levels. A common case of this
sort would be the distinction between basic construction rules
and rules that adjust the outputs of the basic construction rules.
For example, certain rules of politeness might not affect our
initial production of a sentence, but may enter as adjustments
performed while we are speaking.
As the preceding analysis suggests, there are different levels
of commonality for the components and operations of any rule
system. The main levels would be as follows: 1) universal, 2)
338
Generative Poetics
Generative Semantics
GENERATIVE SEMANTICS
339
340
341
Genetic studies of typical forms of specific language impairment have identified other genomic regions that are likely to be
relevant to language, and researchers are focusing considerable attention on those chromosomal sites in the hope of pinning down particular genes (e.g., The SLI Consortium 2002). For
developmental dyslexia, a disorder primarily characterized by
reading disability, but underpinned by subtle persistent deficits
in language processing, it has been possible to home in on several candidate genes (DYX1C1, KIAA0319, DCDC2, and ROBO1).
These genes differ from FOXP2 in that there have been no specific causal mutations identified instead, it is thought that the
increased risk of dyslexia stems from as-yet unknown variants in
the regulatory parts of those genes that govern their expression
(Fisher 2006). Nevertheless, there are some striking parallels
with FOXP2; each of the dyslexia candidate genes shows widespread expression patterns in multiple circuits in the brain, and
each is active in additional tissues, not only the brain. None of
the genes is unique to humans; for example, highly similar versions of each are found both in other primates and in rodents. At
this stage, there is little understanding of why alterations in these
genes should have relatively specific effects on reading abilities,
although their basic neurobiological functions are beginning to
be defined; three of the genes (DYX1C1, KIAA0319, and DCDC2)
have been linked to neuronal migration, and the fourth (ROBO1)
codes for a receptor protein involved in signal transduction,
which helps regulate axon/dendrite guidance.
At the time of writing, there are indications that alterations
in gene dosage may emerge as a general theme underlying overt
speech/language deficits. For example, it has been recently
shown that duplications of a specific region on chromosome
7 (far away from the site of the FOXP2 gene) can cause speech
deficits (Somerville, Mervis, et al. 2005). What is especially interesting about this finding is the fact that the relevant part of chromosome 7, which contains several different genes, corresponds
to the region that is most commonly deleted in cases of Williams
syndrome, a well-studied disorder in which language skills can be
relatively well preserved as compared to other abilities. In other
words, while deletion of that part of chromosome 7 (i.e., reduced
gene dosage) tends to spare language, duplication of this same
set of genes (increased gene dosage) leads to speech disruptions.
Similarly, there is evidence to suggest that the number of functional copies of a chromosome 22 gene called SHANK3, recently
implicated in autism spectrum disorders, may be critical for
speech development (Durand, Betancur, et al. 2007). Language,
like many aspects of biology, is likely to depend on a precise balance among many different molecules.
342
Genes, like species, are the product of the process that Darwin
called descent with modification. Each gene has an evolutionary history, with its current function a modification of earlier
functions. To the extent that the language system is the product
of descent with modification, most genes that are associated with
language can be expected to have counterparts in nonlinguistic
species. As such, comparisons of gene sequences and expression
patterns in different species can help cast light on language evolution, identifying which of the relevant neurogenetic pathways
are shared with other species and which have been modified
data that will soon emerge, both in terms of sheer data analysis
and in terms of relating these data to linguistic functions.
Another exciting prospect is the use of genetic manipulation in order to find out more about the functions of genes that
are involved language, for example, by examining the function
of nonhuman counterparts to those genes (White, Fisher et al.
2006). In this way, individual genes may provide the first molecular entry points into neural pathways involved in human communication, and a direct way to understand how the twin processes
of descent and modification led to the remarkable and uniquely
human faculty for complex language.
Gary F. Marcus and Simon E. Fisher
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bishop, D. V. 2001. Genetic and environmental risks for specific language impairment in children. Philos Trans R Soc Lond B Biol Sci
356.1407: 36980.
Durand, C. M., C. Betancur, et al. 2007. Mutations in the gene encoding
the synaptic scaffolding protein SHANK3 are associated with autism
spectrum disorders. Nat Genet 39.1: 257.
Enard, W., M. Przeworski, et al. 2002. Molecular evolution of FOXP2, a
gene involved in speech and language. Nature 418.6900: 86972.
Fisher, S. E. 2006. Tangled webs: Tracing the connections between genes
and cognition. Cognition 101.2: 27097.
Fisher, S. E., C. S. Lai, et al. 2003. Deciphering the genetic basis of speech
and language disorders. Annu Rev Neurosci 26: 5780.
Fisher, S. E., and G. F. Marcus 2006. The eloquent ape: Genes, brains and
the evolution of language. Nature Reviews Genetics 7: 920.
Gopnik, M., and M. B. Crago 1991. Familial aggregation of a developmental language disorder. Cognition 39.1: 150.
Hurst, J. A., M. Baraitser, et al. 1990. An extended family with a dominantly inherited speech disorder. Dev Med Child Neurol 32.4: 3525.
Lai, C. S., S. E. Fisher, et al. 2001. A forkhead-domain gene is mutated in
a severe speech and language disorder. Nature 413.6855: 51923.
Lai, C. S., D. Gerrelli, et al. 2003. FOXP2 expression during brain development coincides with adult sites of pathology in a severe speech and
language disorder. Brain 126 (Part 11): 245562.
Lehmann, O. J., J. C. Sowden, et al. 2003. Foxs in development and disease. Trends Genet 19.6: 33944.
MacDermot, K. D., E. Bonora, et al. 2005. Identification of FOXP2 truncation as a novel cause of developmental speech and language deficits. Am J Hum Genet 76.6: 107480.
Marcus, G. F. 2004. The Birth of the Mind: How a Tiny Number of Genes
Creates the Complexities of Human Thought. New York, Basic Books.
Newbury, D. F., E. Bonora, et al. 2002. FOXP2 is not a major susceptibility gene for autism or specific language impairment. Am J Hum Genet
70.5: 131827.
Shu, W., H. Yang, et al. 2001. Characterization of a new subfamily of
winged-helix/forkhead (Fox) genes that are expressed in the lung and
act as transcriptional repressors. J Biol Chem 276.29: 2748897.
SLI Consortium, The. 2002. A genomewide scan identifies two novel
loci involved in specific language impairment. Am J Hum Genet
70.2: 38498.
Somerville, M. J., C. B. Mervis, et al. 2005. Severe expressive-language
delay related to duplication of the Williams-Beuren locus. N Engl J
Med 353.16: 1694701.
Terrace, H. S., L. A. Petitto, et al. 1980. On the grammatical capacity
of apes. In Childrens Language 2, ed. K. E. Nelson, 371495. New
York: Gardner.
Vargha-Khadem, F., D. G. Gadian, et al. 2005. FOXP2 and the neuroanatomy of speech and language. Nat Rev Neurosci 6.2: 1318.
343
Gesture
Vargha-Khadem, F., K. E. Watkins, et al. 1998. Neural basis of an
inherited speech and language disorder. Proc Natl Acad Sci U S A
95.21: 12695700.
Vernes, S. C., J. Nicod, et al. 2006. Functional genetic analysis of mutations implicated in a human speech and language disorder. Hum Mol
Genet 15.21: 315467.
Watkins, K. E., N. F. Dronkers, et al. 2002. Behavioural analysis of an
inherited speech and language disorder: Comparison with acquired
aphasia. Brain 125 (Part 3): 45264.
White, S. A., S. E. Fisher, et al. 2006. Singing mice, songbirds, and
more: Models for FOXP2 function and dysfunction in human speech
and language. J Neurosci 26.41: 103769.
GESTURE
no ideas, just irritable mental gestures.
(remark attr. to Lionel Trilling,
New York Times, June 21, 2006, A1)
What Is Gesture?
Lionel Trilling, in this non-motto, invokes an all-too-common
view of gesture. The very phrase hand waving suggests triviality.
But let us imagine Trillings own gesture. It would have been (we
can predict) what Cornelia Mller has called the palm up open
hand (PUOH), the hand seeming to hold a discursive object,
holding, in fact, Trillings view. These kinds of gestures have been
linked to the conduit metaphor the metaphor whereby language
or cognition is a container holding some content.
The PUOH is also one of a species of gesture termed by Kendon
gesticulation, one of several kinds of gesture he distinguished
and that I arranged on Kendons Continuum:
Phenomenon
Stuttering
Blindness
Fluency
Information exchange
344
Gesture
2 2769
1 2740
bends it way ba
ck]
Figure 2. Phases of a gesture timed with and he bends it way back. The insert is a frame counter (1 frame = 1/30
sec.). The total elapsed time is about 1.5 seconds. Panel 1: Preparation. Panel 2: A prestroke hold while saying
he. Panel 3: Middle of stroke bends it way ba(ck). Panel 4: End of stroke and beginning of the poststroke
hold in the middle of back.
Properties of Gestures
THE UNBREAKABLE SPEECH-GESTURE BOND. Synchronized
speech and gesture comprise virtually unbreakable psycholinguistic units, unbreakable as long as speech and gesture are
coexpressive. A diverse range of phenomena show the inseparability of the two modes; Table 1 summarizes some of them. In
each case, some disruption to speech-gesture combination is
resisted; it holds despite the disruption. To break this bond, one
has to drain the combination of meaning for example, through
rote repetition.
PHASES AND THEIR SIGNIFICANCE. Gesture phases are organized
around the stroke: everything is designed to present it in proper
synchrony with its coexpressive speech segment(s). Figure 2
shows all gesture phases except retraction. The full span, from
the beginning of preparation to the end of retraction, brackets
what can be thought of as the lifetime of a specific idea unit in
language-geared imagery. We see the image in a state of activation that did not exist before and does not exist after this span.
The dawn of the idea unit is seen in the beginning of the preparation, and the idea unit itself is the unit formed of the synchronized coexpressive speech and stroke (called a growth point).
345
Gesture
WHEN DO GESTURES OCCUR? Somewhat surprisingly, the timing
of gestures in relation to speech has been the subject of controversy. The question is whether gestures tend to anticipate their
linked linguistic material or coincide with it. The anticipation
view is often accompanied by a further idea that gestures take
place during speech pauses. The synchrony view, clearly, implies
that gestures and speech are co-occurring. When the question
is examined with careful attention to the distinction between
preparation and stroke, the facts are clear: The preparation for
the gesture precedes the coexpressive linguistic segment (with
a pause or not); the stroke coincides with this segment about
90 percent of the time. Holds ensure that this synchrony is
preserved.
346
From his earliest work, Chomskys answer to (1) posited a computational system that provided statements of the basic phrase
structure patterns of languages (phrase structure rules) and
operations for manipulating these basic phrase structures (transformations). GB strongly focused on question (2) by positing
heavier and heavier restrictions on the computational system,
thus limiting the choices available to the learner.
An innovation was the development of trace theory, which
proposed that when movement transformations operate, they
leave behind traces, silent placeholders marking the position
from which movement took place, as schematized in (3).
(3) Linguistics, I like t
Under trace theory, the earlier importance of deep structure (the initial representation in the syntactic derivation in the
standard theory) for semantic interpretation is ultimately
eliminated. It was already known that some aspects of meaning
(including scope of quantifiers, anaphora, focus) depend on surface structure. Once surface structure is enriched with traces, even
grammatical relations (subject of, object of, etc.) can be determined at that level of representation. Using the term LF (logical
form) for the syntactic representation that relates most directly
D-Structure
Transformations
S-Structure
PF
LF
Modularity
The GB theory displayed a high degree of modularity. Complex
phenomena were seen as the result of interactions of simple
modules. The phrase structure module was virtually reduced to
x-bar theory (originally developed in Chomsky [1970] 1972),
with specific instantiations following from properties of particular lexical items. Further, the X-bar schema itself was extended
from just lexical categories (noun, verb, adjective, etc.) to functional categories. For example, a sentence came to be analyzed
as the projection of an inflectional head, Infl, containing tense
and agreement information.
The transformational module is also dramatically simplified
in comparison with its predecessors. The GB framework replaced
the earlier numerous specific transformations with very general
operations, Move (displace any item anywhere), or even Affect
(do anything to anything). There is thus very little transformational syntax that the child has to learn. A grammar this simple
and general would seem to massively overgenerate, producing
countless numbers of unacceptable sentences. To deal with this
overgeneration problem, GB theorists, further developing a line
of research begun in the 1960s, posited general constraints on
the operation of transformations (locality constraints, especially
subjacency, part of bounding theory) and also conditions
on the output of the transformational component (including
filters).
Parameters
The postulated universal (wired-in) parts of the computational system are called principles. The (limited) ways in which
languages can differ syntactically are called parameters. The system is fundamentally based on principles and parameters.
The child learning a language is preequipped with the principles
and needs only to set the values of the parameters. The standard
assumption is that there are few parameters, they are very simple, and their values can be determined by the child on the basis
of readily available primary linguistic data.
Case Theory
S-structures result from the transformational component operating on D-structures. Given the generality of Move , derivations often seem to yield ungrammatical sentences. One module
reining in this overgeneration by regulating S-structure is case
theory. There are characteristic structural positions that license
particular cases. In many languages (such as Latin, Russian,
German), these case distinctions are overtly manifested. In
English, only pronouns show an overt distinction between nominative and accusative, for instance, but Case Theory posits that all
noun phrases (NPs) have abstract case (henceforth, Case), even
when it is not phonologically visible. The requirement that all
NPs occur in appropriate Case positions is the Case Filter, a wellformedness condition on the S-structure level of representation.
Government
The GB approach always sought regularities and generalizations.
The notion government is itself a generalization of the X-bar
theoretic head-complement relation. The basic definition is as
follows:
(5) A head H governs Y if and only if every XP (highest projection of X) dominating H also dominates Y and conversely.
[Domination is ancestry in a phrase structure tree
diagram.]
By (5), a head governs its complement and also its specifier. Case
licensing then is under government, with the governor licensing the governee. A transitive verb governs its direct object NP; a
preposition governs its complement NP; Infl governs its specifier
(the subject of the clause). Thus, a Case-licensing head licenses
Case on a nominal expression that it governs. For example, a
transitive verb, such as prove licenses (accusative) Case on its
complement direct object.
Types of Movement
The transformational module of the theory recognizes three
major subtypes of movement. A-movement is movement to an
argument-type position (i.e., an A-position, especially subject
position. (non-A)-movement is movement of an XP (highest
projection of X, for variable X) to a non-A position. The movement of an interrogative expression, such as Who in (6) (WHmovement) is a central exemplar:
(6) Who will they hire t
347
(13) *She thinks Mary will solve the problem [with She intended to
refer to Mary]
CP
|
C'
C
NP
|
Susan
I
V
|
is
NP
|
a linguist
Figure 1.
a. Susan is a linguist
b. Is Susan a linguist
Binding
The binding part of government and binding theory has as
its core anaphoric relations, circumstances under which one
expression can or cannot take another as its antecedent, that is,
pick up its reference from the other. In (9), him can take John as
its antecedent, while in (10), it cannot.
(9) John said Mary criticized him
(10) John criticized him
That is, (10) has no reading corresponding to that of (11), with the
pronoun him replaced by the anaphor himself (see anaphora).
(11) John criticized himself
348
Grammaticality
Grammaticality Judgments
GRAMMATICALITY
A sentence that is well formed according to a given grammar
is grammatical. By analogy, the property can also be assigned
to constituents of a sentence, for example, noun phrases. In
generative linguistics, the term grammar has been used with
a systematic ambiguity: We must be careful to distinguish the
grammar, regarded as a structure postulated in the mind, from
the linguists grammar, which is an explicit articulated theory
that attempts to express precisely the rules and principles of the
grammar in the mind of the ideal speaker-hearer (Chomsky
1980, 220).
Each sense of grammar corresponds to a sense of grammaticality. If grammar is taken in the sense of competence,
a component of the speakers mind, grammaticality is a gradual property that is realized in judgments by the speaker (see
GRAMMATICALITY JUDGMENTS
Grammaticality judgments involve explicitly asking speakers
whether a particular string of words is a well-formed utterance of
their language, with an intended interpretation stated or implied.
Among the many kinds of data available to linguistics, grammaticality judgments are particularly useful in distinguishing
349
Grammaticalization
possible from impossible utterances (the latter conventionally
marked *) among those that are not spontaneously produced.
Intermediate degrees of well-formedness may also be of interest (e.g., ? questionable, ?? highly questionable). These
judgments can also bring to light knowledge of language in special populations whose production and even comprehension
may not evince it. Contra widespread belief, however, like all
performance data they bear on linguistic competence only
indirectly and call for the same attention to methodology as
other data sources.
Carson T. Schtze
GRAMMATICALIZATION
Grammaticalization is a historical process whereby fixed grammatical forms, such as prepositions, conjunctions, suffixes, and
auxiliaries, and the constructions of which they are part arise out
of what were previously independent categorial forms in freer
arrangements. The negative construction with ne pas in French
provides a good example: pas step, pace was originally a noun
that functioned to reinforce the ne that supplied the negative
meaning, as in il ne va pas he doesnt go (originally doesnt
go a step). Nowadays, pas has become grammaticalized as a
general marker of negation, as in il (ne) parle pas he doesnt
speak the ne being increasingly dropped altogether. As a field,
grammaticalization is the study of those linguistic changes that
result in specifically grammatical forms and construction. The
standard history of grammaticalization is Lehmann 1995.
Grammaticalization involves two aspects: structural change
and semantic change. These two kinds of change go hand
in hand, and it is impossible to assign priority to either of them.
Structural changes include reanalysis and phonological reduction. In reanalysis, adjacent forms are rebracketed: [I am going]
[to sell my pig] Im on my way to sell my pig becomes [I am going
to][sell my pig] and eventually [going to] assumes the meaning
of future tense. Characteristically, too, a major category, such as
verb or noun that was formerly the head of a phrase, is demoted to
a minor category, such as auxiliary or preposition, and becomes
a satellite to the new head; thus, when a cup full of flour becomes
a cupful of flour, the erstwhile head noun cup is reanalyzed as a
component of a quantifying expression, now reduced in status to
a determiner of the semantic head noun, and its place as semantic head noun is usurped by flour. Similarly, full, previously the
head of the adjective phrase full of flour, is reduced to the status
of a suffix on cup. This decategorialization of grammaticalized
forms (Hopper 1991, 303; Hopper and Traugott 2003, 10614)
means that the forms lose the typical attributes of the older category, such as ability to take modifiers and determiners, availability as an argument of the verb, and independent referential
status. Parallel restrictions are placed on verbs that become
auxiliaries.
Phonological reduction, or erosion (Heine and Reh 1984,
215), is a frequent but not inevitable accompaniment of grammaticalization. English lets is clearly derived from let us, but
now serves to introduce an adhortative predicate, as in lets
leave. When [going][to sell] was reanalyzed as [going to][sell],
going to became reduced to gonna. The French future tense in
350
Background
Over the past century, a great deal of work has been done on
questions about the anatomical and neural bases of language
production (i.e., speech and phonetics; see brain and lan-
351
352
percent among them. A more detailed analysis of the social content itself suggested that the vast majority was concerned with
factual personal experiences (~30% of all social conversation
time), personal social/emotional experiences (~30%), or with
third-party social/emotional experiences (~30%). The balance
was devoted to critical comments about third parties and seeking/giving advice.
More recently, an experimental study of language transmission in which groups of four participants were asked to relay
information passed on to them by a previous member of a chain
found that both gossip (in the racy sense) and social information (about the actions of people in a story) were transmitted
much more reliably than was factual information about a social
event (with no motivational content) or descriptive information
about a nonsocial event (for example, tourist information about
a location). This suggests that information with a strong social
(and perhaps emotional) content is more memorable in some
way than other kinds of information. The transmission rates for
social and gossip content did not differ significantly, suggesting
that the issue is not the raciness of the content but the fact that it
concerns individuals social and emotional lives.
Alternative Hypotheses
Two alternative hypotheses as to how language might function
in the social domain have been suggested: the symbolic contract hypothesis (proposed by Terrance Deacon in his book The
Symbolic Species) and the Scheherazade effect (proposed by
Geoffrey Miller). Both are concerned with aspects of reproductive behavior.
Deacon observed that humans have a unique mating system
based on pairbonds that are embedded within a large multimale/
multifemale social group. What makes this a particular problem
is that the division of labor means that mates are often separated
for long periods of time in the presence of potential rivals. Since
the presence of rivals creates risks for the sexual fidelity of a pairbond, Deacon argued that language must have been needed to
establish formal social contracts in which exclusive mating rights
are identified and publicly agreed upon (This is my mate; you
are not allowed to interfere with him/her while I am away).
Millers Scheherazade effect is also concerned with the business of mating, but in this case, the focus lies with the intrinsic
(as opposed to the extrinsic) dynamics of the pairbond: how to
woo and keep your mate interested, rather than merely guarding
him/her against rivals. Miller argued that the capacity to be witty
and entertaining (in the Arabian Nights sense) would have been
strongly selected for, especially in a context where more attractive rivals were readily available.
While both suggestions are undoubtedly plausible, for neither has any substantive evidence yet been adduced. However,
an alternative view might be to see both of these mechanisms
as being derivative of an initial situation for which language
had been selected as a more general mechanism for bonding
large social groups. One reason for this suggestion is simply that
Deacons paradox (the risk to pairbonds posed by the presence
of large numbers of rivals) can only be a problem when social
group size is large. But without some kind of bonding mechanism over and above social grooming, it is difficult to see how
our distant ancestors would have been able to create and hold
Habitus, Linguistic
together large social groups. Thus, we might see language evolving initially as a social bonding device and then subsequently see
the skills underpinning language having been exaggerated by
either or both of these two mechanisms.
R. I. M. Dunbar
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Deacon, T. 1997. The Symbolic Species: The Coevolution of Language and
the Human Brain. Harmondsworth, UK: Allen Lane.
Dunbar, R. I. M. 1993. Coevolution of neocortex size, group size and language in humans. Behavioral Brain Science 16:681735.
Dunbar, R. I. M., N. Duncan, and A. Marriot. 1997. Human conversational behaviour. Human Nature 8: 23146.
Mesoudi, A., A. Whiten, and R. I. M. Dunbar. 2006. A bias for social information in human cultural transmission. British Journal of Psychology
97: 40523.
Miller, G. 1999. Sexual selection for cultural displays. In The Evolution
of Culture, ed. R. I. M. Dunbar, C. Knight, and C. Power, 7191.
Edinburgh: Edinburgh University Press.
H
HABITUS, LINGUISTIC
Habitus is one of Pierre Bourdieus two fundamental conceptual tools of analysis; the other being field. If field relates to the
objective conditions of social space, then habitus is an expression
of subjectivity and is defined by Bourdieu as
systems of durable, transposable dispositions, structured structures predisposed to function as structuring structures, that is,
as principles which generate and organize practices and representations that can only be objectively adapted to their outcomes without presupposing a conscious aiming at ends or an
express mastery of the operations necessary in order to attain
them. Objectively regulated and regular without being in any
way the product of obedience to rules, they can be collectively
orchestrated without being the product of the organizing action
of a conductor. ([1980] 1990, 53)
353
Habitus, Linguistic
354
criticized
verb
HEA D
VFORM
fin
A UX
LOC
SYNSEM
LOC
NP
COMS NP
SUBJ
CA T
normi
normj
Figure 1.
...
...
(b)
PHON criticized
verb
VFORM fin
A UX
HEA D
SYNSEM LOC CA T
LOC
SUBJ
NPnormi
COMPS
CONT w(j)(i)
verb
HEA D VFORM fin
A UX
LOC CA T
NPj
criticized
Leslie
Figure 2.
NPnormi
PHON criticized
SYNSEM
VP
LOC
SUBJ NP
normi
COMPS NPnormj
Leslie
CONT w(j)(i)
355
(3) (a)
NP
PHON
LOC 1
LOC 1
1
SLA SH
SYNSEM
NONLOC|SLA SH 1
Terry
VP SLA SH
NP
I
V
NP
SLA SH
know
N
PP
SLA SH
stories
P
1
NP
LOC 1
about
SLA SH
Figure 3.
empty subj list. Thus, Robin criticized Leslie will be straightforwardly licensed.
Note further that *Robin criticized and *Robin criticized Leslie
certain books will both be ruled out. Neither of them will satisfy the
head-complement schema under the constraint imposed by the
valence principle. In the first case, the lack of a complement daughter will yield a result with a nonempty complements list, violating
the schema requirement. In the second case, the valence requirements on criticize will not equal the sum of those on the mother plus
the set of daughters that appear in the structure, as the full form of
the valence principle requires. Hence, both fail to be licensed.
The valence feature lists do not in fact specify information
about the entire sister sign selected by the head. Lexical selection
is universally blind to phonological form and also to the descending constituent structure of any selected phrasal constituent;
therefore, it makes sense to restrict valence specifications to synsem values. But the latter contain enough information to implement virtually all local dependencies as simple expressions of
selection. Thus, it is possible to select not only for coarse-grained
information, such as the number of valents and their respective
lexical category types, but also for information such as their case,
or the inflectional properties of their heads (on the assumption
that the latter are head features and will therefore be shared
between a phrasal valent and its lexical head daughter). The morphosyntactic dependency in English between the identity of auxiliary elements, on the one hand, and the inflectional form of the
verb that immediately follows them, on the other hand, follows
simply and directly by specifying that, for example, each auxiliary
have selects as its complement a VP, preserving, in its own feature
specifications, the inflectional class of its head daughter (specifically, the value psp, encoding past participial status).
the percolation of wh properties (as in so-called pied-piping phenomena), throughout arbitrarily large structures exist in natural
language and must be accounted for.
HPSG follows the central innovation introduced by Gerald
Gazdar a quarter of a century ago in treating filler/gap linkages
as a by-product of the diffusion of a specific feature slash, whose
value is correlated with that of the filler and which is driven
from mother to (at least one) daughter by the nonlocal feature
principle:
Nonlocal Dependencies
Remaining Issues
356
Nonlocal Feature Principle (NFP, simplified): For any nonlocal feature f the value of f on a phrase at any point below the
place in the structure where f was introduced is the union of
the values of f on the set of daughters.
The NFL states, in essence, that below the point where the nonlocal feature is introduced, at least one daughter in each twogeneration tree must bear the value of f specified. In the case of
filler/gap constructions, the feature slash is identified at the top
of the dependency with the relevant part (called the loc value,
illustrated in Figure 2) of the filler. The descendents of the highest category with a nonempty slash feature preserve this value
in their own specifications, until at the bottom of the dependency, the slash value is cashed out as an empty category. An
example is illustrated in (3)a, with the relevant lexical entry for a
slash terminal category listed in (3)b in Figure 3. This approach
to filler/gap dependencies comports well with familiar phenomena in languages that mark such dependencies by local flagging
of extraction pathways; it also has considerable advantages over
movement-based approaches in the analysis of multiple gap
linkages to a single filler, such as parasitic and across-the-board
extractions in coordinate structures.
Hippocampus
The syntax/semantics interface issue is not fully resolved in
HPSG, in the sense that there are a number of competing
approaches to a variety of issues, ranging from the nature of
the objects in HPSG representations that map to semantic
representations to the mapping rules themselves. Different
answers to these questions typically correspond to major differences in syntactic analyses.
Probably the most fundamental point of contention is the
disagreement within HPSG about the degree to which patterns in natural languages reflect, on the one hand, lexical properties and systematic mappings between lexical
descriptions, or restrictions on specific constructions, as
encoded by constraints imposed on types belonging to very
elaborate ontologies, on the other hand. At one extreme,
HPSG includes a set of lexical heads with very abstract
properties and possibly no phonological realization, as in
earlier treatments of relative clauses. At the other extreme
are intricate multi-inheritance hierarchies in which properties of constructions are essentially posited as underived
primitives, or are derived by combining a number of such
underived primitives.
These and a number of other foundational issues are currently
under intense discussion and debate within the HPSG research
community, and there is no reason to expect a consensus on any
of these questions in the near future.
On the basis of the relatively brief history of the framework
so far, it seems very possible that divergences in approaches
to these fundamental matters will in time yield major schisms
within the theory, leading in the end to two or more major versions of the theory, in much the way that categorial grammar has
split into combinatory categorial grammar, on the one hand,
and the type-logical version based on the Lambek calculus, on
the other. It remains to be seen whether the theory will develop
along these lines.
Robert D. Levine
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Borsley, Robert. 1996. Modern Phrase Structure Grammar.
London: Blackwells.
Levine, Robert, and Thomas Hukari. 2006. The Unity of Unbounded
Dependency Constructions. Stanford, CA: CSLI
Pollard, Carl, and Ivan Sag. 1994. Head-Driven Phrase Structure Grammar.
Chicago: University of Chicago Press.
HIPPOCAMPUS
The hippocampus is a bilaterally symmetric subcortical structure
adjacent to the lateral ventricle in the medial temporal lobe
(mtl). Researchers in 1968 reported dramatic memory phenomena associated with hippocampal-MTL damage, and data
reported from 1998 to 2006 have indicated parallel phenomena
for many other aspects of cognition, including language. I first
discuss patient H. M., the initial source of data for the hippocampal-memory and hippocampal-language links. I then discuss
related patient groups and the theoretical significance of the
hippocampal-language link.
357
Hippocampus
and failures to follow experimenter requests for clarification.
Research also indicated comprehension failure involving an
initial meaning for sentences, ambiguous or not.
To summarize, in a wide range of tasks involving many fundamental aspects of sentence comprehension, H. M. exhibited
deficits not caused by his memory problems (for corroborating
evidence on H. M.s comprehension deficits, see Corkin 1984;
Lackner 1974; and Schmolck, Stefanacci, and Squire 2000).
However, H. M.s comprehension deficits were selective rather
than across the board: Experiment six in MacKay et al. (2007)
demonstrated that H. M. comprehended familiar words and
phrases in isolation without deficit despite large deficits in comprehending these same stimuli when embedded within sentences. Besides demonstrating selectivity, these results indicated
that H. M.s deficits were not attributable to low motivation, to
failure as a child to learn the meaning of the critical words and
phrases, or to failure to understand and follow instructions for
the task.
H. M. also exhibited significant production deficits when
describing the meanings of familiar words that he comprehended without deficit in MacKay et al. (2007): Judges blind to
speaker identity rated H. M.s meaning descriptions as reliably
more redundant, less coherent, less grammatical, and less comprehensible than those of controls. These findings replicated
earlier results indicating deficits in H. M.s production of novel
or non-clich sentences (see MacKay, Stewart, and Burke 1998).
Again, however, H. M. exhibited selective production deficits that
mirrored his memory deficits, for example, spontaneously producing clich phrases such as in a way (familiar from before his
surgery) without errors (ibid.).
H. M. also exhibited similar deficits and sparing in the seemingly simple task of reading sentences aloud (MacKay and James
2001): He produced abnormal pauses at major syntactic boundaries unmarked by commas in the sentences, but normal pauses
at syntactic boundaries marked with commas, a prosodic marker
that H. M. had learned prior to his operation. H. M. also produced
abnormal pauses within unfamiliar phrases in the sentences, but
normal pauses within frequently used phrases. These and other
selective deficits indicated that he has difficulty with the process
of reconstructing novel aspects of sentence structure when reading aloud.
H. M. also exhibited similar deficits and sparing in visual
cognition: When detecting target figures hidden in concealing
arrays, he performed reliably worse than controls for unfamiliar
targets but not for familiar targets (MacKay and James 2000). In
short, H. M. exhibits similar selective deficits in visual cognition,
episodic memory, sentence-level comprehension, and sentence
production when speaking and reading aloud; impaired processing of never previously encountered events, visual figures,
phrases, and propositions; but spared processing of information familiar to him before his lesion and used frequently since
then.
Why are these parallels important? One reason is that H. M.
is not unique: Other patients with hippocampal-MTL damage
exhibit identical parallels, reinforcing the links among hippocampus-MTL, language, and memory. For example, other amnesiacs
exhibit deficits in detecting the two meanings in ambiguous sentences (Zaidel et al. 1995) and make errors resembling H. M.s
358
Hippocampus
Historical Linguistics
HISTORICAL LINGUISTICS
Historical linguistics is the study of how languages change over
time. We can approach the study of change in various ways. One
is by studying the histories of individual languages. An example
would be analyzing the changes that have taken place in English
over the last thousand years. A second approach involves comparing various related languages in order to draw inferences
about the types of changes that have occurred since the time
they split from their common ancestor. We can further study the
frequency and naturalness of the changes that the languages are
hypothesized to have undergone and the effects of change in
one area on other parts of the language. Finally, historical linguistics is also concerned with finding explanations for change,
including why languages change and how a particular change
was actuated in a particular circumstance. Historical linguistics
also intersects with other fields. For example, a linguist trying
to reconstruct the geographic extent of a proto-language may
also make use of data from both archaeology and historical
anthropology.
The following discussion begins with an abbreviated history
of the field, from classical times to the early twentieth century.
From there, we turn to contemporary research and conclude
with a look at possible future directions for the field.
359
Historical Linguistics
anas duck is related to the verb nre to swim. Furthermore,
although numerous similarities between Latin and Greek were
noted, it was assumed that all such words were direct borrowings into Latin from Greek. Shared common ancestry from a language no longer spoken was never considered (see extinction
of languages). Such an assumption was in keeping with the
strong cultural debt that the Roman world owed to the Greek
(see Law 2003).
The classical etymological method continued to be employed
throughout the Middle Ages, where it was joined by theories of
language change and diversity built on the biblical story of the
destruction of the Tower of Babel (Gen. 11). A summary of the
theory can be found in Dantes de Vulgari Eloquentia (for one
English translation, see Shapiro 1990). Such a theory of diversity
specifies both a cause of language change and a partial model
of the origin of modern linguistic diversity. Work within a model
of change laid the foundation for much later linguistic scholarship, for it led to questions about the language that was spoken
by those erecting the Tower of Babel (and, therefore, what the
first human language was) and exactly how modern attested languages related to one another.
The de Vulgari Eloquentia is also a founding discussion of
relationships among the vernacular languages of Europe. The
rise of the study of Romance vernaculars led to an examination
of systematic differences among those languages, as well as comparison with Latin (for example, why Latin de is a preposition
meaning from, but in French and Italian it marks possession).
As R. H. Robins (1968, 100 ff) notes, it was this examination that
allowed the development of an adequate framework for diachronic linguistics because of a chance to study change where
the parent language was already well understood.
In the early Middle Ages, there was also a highly sophisticated
Arabic comparative linguistic tradition. Ibn Hazm (9941064)
noted regular correspondences among Hebrew, Arabic, and
Syraic, and in the Ihkam Ibn Hazm further identified changing
pronunciation and language contact as driving forces in the creation of linguistic diversity (see contact, language).
In the sixteenth to eighteenth centuries, we begin to see the
study of language change linked to other branches of linguistics,
such as typology. Scholars such as Konrad Gessner ([1555]
1974), Joseph Justice Scaliger ([1599] 1610), and later Andreas
Jger (1686) and Peter Simon Pallas (1786) collected and compared vocabularies of the languages available to them and made
hypotheses about linguistic relationships on this basis. However,
the comparisons are unsystematic and based mostly on very few
features. For example, Scaliger divides the languages of Europe
into four major classes, depending on whether their word for
god is based on deus, theos, gott, or bog (roughly corresponding to Latin/Romance, Greek, Germanic, and Slavic, respectively). Gessner ([1555] 1974, 110) deduces that Armenian is
closely related to Hebrew because of the similarity of words such
as lezu tongue (Hebrew laschon in Gessner) and hhatz cross
(Hebrew etz or hetz). Moreover, until G. W. Leibniz, there is no
conception that languages could have been descended from a
language that is no longer attested.
The beginning of modern historical linguistics and the comparative method is often said to date from a speech given by Sir
360
Historical Linguistics
A further methodological coup is due to Ferdinand de
Saussure ([1915] 1972). Saussure hypothesized, on the basis of
both internal evidence of root structure and varied correspondences among vowels in Indo-European languages such as Greek
and Sanskrit, that there had once been a further set of laryngeal
consonants (possibly /h/, // and /w/) that had disappeared in
all environments in attested languages. Saussures contribution
is very important to historical linguistics for two reasons: First, it
is the foundation of internal reconstruction, that is, the
hypothesis of reconstructions based on synchronic patterns in
one language, rather than a direct comparison of forms between
languages. Second, it demonstrates very clearly the power of the
comparative method and the importance of regularity in correspondence sets. The subsequent decipherment of inscriptions in
Hittite and Luwian confirmed Saussures hypothesis since two of
the laryngeals are preserved in these languages precisely where
we should expect to find them.
Historical linguists in the twentieth century made progress in
the reconstruction of families outside Europe, in the growing use
of quantitative methods in modeling and describing language
change, and in historical syntax (see syntactic change). Most
of the methods used today were developed through the reconstruction of proto-Indo-European; however, the methods have
also been successfully applied to other families. Much early historical work was done on Finno-Ugric languages (e.g., Sajnovics
[1770] 1968; Gyarmathi 1799), and more recently, there has been
considerable progress in the reconstruction of Austronesian
(Pawley and Ross 1993), Niger-Congo (cf. Hombert and Hyman
1999), and numerous families in North America (see, for example, Campbell 1997).
While the family tree model has been very influential in historical linguistics, other models of language change are also
used. Perhaps the most commonly cited is the wave theory of
J. Schmidt and Jules Guilliron (see Guilliron 1921), who proposed that sound changes diffuse through the lexicon, gradually
affecting more and more instances of phonemes in a given environment. Regularity in correspondences is thus epiphenomenal
and only appears once a change is complete. Another common
appeal to wave theory is in subgrouping, where it is argued that
the family tree is not an accurate representation of language splitting. Rather, linguistic differentiation occurs through the gradual
building up of isoglosses and changes affecting individual lexical
items.
work on historical syntax concerns word order change in the synchronic analysis of ancient languages or the causes of syntactic
change, rather than reconstruction per se. Influential here has
been the work of David Lightfoot (e.g., 1999), who has developed
a theory of change that assigns the primary cause of change to
a childs acquisition of the language. Since children are exposed
to linguistic data that is slightly different from what their parents
were exposed to, they therefore draw slightly different conclusions about the syntactic structure of their language. We see this
reflected in the historical record as a syntactic change. Others are
less comfortable in ascribing change solely to grammar change at
acquisition and argue that syntactic changes also occur in adult
speakers as a result of exposure to new languages and dialects,
changing prestige, and other factors. There is ongoing debate
over the extent to which a person may spread change as an adult,
since there are also clearly generational differences in linguistic
production (see age groups).
The study of reconstructions within historical phonology,
morphology, and syntax may also be used to classify languages
into genetic families. Language classification must also take ideas
of language contact into account. Extensive contact between two
unrelated languages may over time lead to enough similarities
that it is difficult to tell whether they are related or not. Several
languages have been misclassified on this basis. For example,
Armenian was originally classified as an Indo-Iranian language
rather than as its own branch of Indo-European because of the
number of loans it exhibits.
There are other less widely accepted methods of investigating linguistic prehistory. One is lexicostatistics, which involves
estimating genetic relatedness by comparing the percentage of
vocabulary common to pairs of languages. Underlying the method
is the assumption that languages that share more common material are likely to be more closely related. Glottochronology uses
the estimations from lexicostatistics to estimate the time depth
of a particular family. Mass comparison (e.g., Greenberg 1987)
involves using large-scale word lists to reconstruct further back
than the strict application of the comparative method allows, by
granting more exceptions to the regularity of sound change and
greater latitude in semantics. In each case, the methods are not
widely accepted. For example, gross similarity between two
languages may be caused by several factors apart from common
genetic inheritance, including chance and borrowings, and only
detailed reconstruction of correspondences by the comparative
method allows us to choose between them.
Future Prospects
Currently, our knowledge of the history of different language
families in the world is very uneven. Some families including Indo-European, Finno-Ugric, and Algonquian have been
reconstructed in detail. In other cases, we are not even sure
which languages belong to the family, let alone what the protolanguage looked like.
There are still many pressing concerns and active areas of
research. The first is in the reconstruction of various language
families. Basic original reconstruction research is needed for
much of the world. Secondly, there is an ever-increasing concern
with questions about how and why languages change. We have
long moved away from arguments of language change involving
361
Historical Linguistics
Historical Reconstruction
362
HISTORICAL RECONSTRUCTION
Like all complex things in nature, languages change over
time. Historical reconstruction is a process of inference by
which changes are undone so as to recover certain aspects
of historically nonattested linguistic structure and content in
hypothetical form. Although it can be applied to languages
for which written documentation is available (e.g., the reconstruction of proto-Romance next to the extant records of Latin)
or to single languages by means of internal reconstruction, historical reconstruction typically is directed to prehistoric languages and requires the simultaneous comparison of multiple
witnesses.
The set of procedures used in historical reconstruction is
called the comparative method. After the elimination of
chance, borrowing, and universals as plausible causes of crosslinguistic similarity, it can be shown that the resemblances
between two or more languages must result from a common
origin (usually called genetic relationship), coupled with divergent descent. In this respect, historical reconstruction in linguistics shows striking parallels to the study of the evolutionary
history of natural species, and much of its terminology is conceptually modeled on that of biological taxonomy. Once it has
Historical Reconstruction
been determined that languages are genetically related, a more
exact picture of their historical connection can be achieved by
the reconstruction of a proto-language or hypothetical common
ancestor.
Although some progress has been made with other aspects
of historical reconstruction, it is widely agreed that the
comparative method has been successfully applied only in
phonology . For reconstruction to begin, it is necessary to
identify a corpus of cognate morphemes , that is, morphemes
that have a common historical origin, identified inductively
as forms of similar meaning that exhibit recurrent sound
correspondences. Distinct sound correspondences that are
found in the same or closely similar environments normally
must be attributed to different proto-phonemes , and the
inventory of proto-phonemes so inferred forms a hypothesis
about the sound system of the proto-language of a language
family . Since proto-phonemes can only be reconstructed in
lexical forms, phonological and lexical reconstructions
are inextricably bound together. Reconstructed phonemes
and reconstructed words (proto-forms) are preceded by an
asterisk to indicate their hypothetical status. Some linguists
take the position that the phonetic substance of such symbols is beyond recovery and that proto-phonemes are, therefore, little more than abstract formulas used to summarize
sound correspondences (the formulaic position). The majority view (the realist position) is more sanguine; although the
phonetic nature of some proto-phonemes clearly is controversial, many others permit little latitude in interpretation.
A related, though distinct, issue concerns the structure of
reconstructed phonological systems, since some of these
have violated implicational universals in typology .
Where this occurs, most historical linguists today would
question the validity of the reconstruction, using the typological generalizations of the present as a guide to inferences
about the past.
Once lexical reconstructions are available, it becomes possible to determine sound changes in a large number of daughter languages. No topic in linguistics has a longer history than
the study of sound change, which commenced during the first
quarter of the nineteenth century with the pioneering studies of
Rasmus Christian Rask and Jacob Grimm. A major point of controversy in the study of sound change is the issue of regularity.
It is now generally agreed that the strict Neogrammarian position, which ruled out the possibility of unconditioned phonemic splits, is overly restrictive. A second issue, which is yet to be
resolved, is whether all sound change is phonetically (or phonologically) motivated.
An examination of sound changes leads not only to theoretical models of how (and why) this process occurs but also to
evidence for subrelationship within a language family. Sound
changes that are exclusively shared (exclusively shared innovations) form the basis for linguistic subgroups. Subgrouping
allows linguists to move beyond the mere recognition of a
language family as an internally undifferentiated collection
of related languages to the reconstruction of a family tree that
defines the historical order of splits of major and minor groups
of languages within the family. The structure of a family tree,
in turn, supports inferences about the most likely center of
363
364
Theoretical Approaches
At the onset, it is important to understand the possible
claims about grammatical knowledge during the holophrastic period. The most conservative claim would be, basically,
what you see is what you get, that is, that children only produce single words because that is all they know at that point.
This approach has come to be known as a lean interpretation
of childrens knowledge. At the other extreme, one can claim
that childrens grammatical knowledge is greater than what is
reflected in single-word productions. This rich interpretation
is based on the fact that language acquisition is rapid and that
children must have innate language learning mechanisms
that enable them to determine the grammatical characteristics of their language at a very early age (see innateness
and innatism ).
These two positions can be demonstrated by again looking at
a childs production of the words doggie and eat. A lean interpretation would be that the child has some initial semantic categorization of word meaning, such that doggie represents
an emerging category of animate objects, and eat represents
an emerging category of actions. A rich interpretation would
propose that innate principles enable the child to establish
explicit grammatical knowledge from these semantic categories.
One principle would be that languages universally categorize
things as nouns and actions as verbs. Another principle would
be that categories like noun and verb are heads of larger units
or phrases, that is, noun phrases (NPs) and verb phrases (VPs).
Another principle would be that NPs and VPs are semantically
connected through semantic relations such as agent and action.
Seeing the doggie eating would lead to the establishment of the
sentence category. This process of building grammatical knowledge from semantics has been called semantic bootstrapping
(Pinker 1984).
How, then, does one decide between these two positions? The
answer depends on the importance placed on supporting childrens grammatical advances with observable changes in their
linguistic behavior. The rich interpretation deals with grammatical development as a logical problem. Since the complexities of
language are acquired so rapidly, it seems reasonable to assume
that linguistic principles must be at work at the very onset of
language acquisition. Researchers who study childrens actual
comprehension and production of language, however, set an
additional requirement. Claims about childrens abilities must
be supported in the way children are comprehending and producing language. The remainder of the present entry discusses
the evidence for grammatical knowledge during the holophrastic stage, based on studies of childrens comprehension and
production.
Language Comprehension
Studying the language comprehension of one-year-olds can be
difficult since they are not capable of responding to the tasks typically used for older children and adults. Researchers, however,
have come up with several clever ways to get at least a general
idea of childrens understanding at this age. If childrens knowledge of language is limited to single words, then one would predict that they should do as well or better responding to utterances
understand the nature of multiword utterances. They can process relations between at least two words in some sentences and
are aware that sentences contain both stressed lexical words
and unstressed words. They may not be aware of the nature of
the latter words, but they know that sentences require them to
be well formed.
Language Production
The claim that preliminary knowledge of grammar takes place
during the holophrastic stage would be strengthened if evidence
could also be found in childrens spoken language. At first glance,
this would seem impossible since holophrastic children are only
producing a single word at a time. There are, however, aspects
of childrens spoken language at this stage that, taken together,
indicate emerging grammatical knowledge as well.
It is well known that holophrastic children are not very intelligible, the result of the fact that they are limited in their phonetic skills and are, in most cases, mixing their single-word
productions with babbling . A. Peters (1983) also pointed out
that some children are not exclusively single-word producers.
He identified children who do not just attempt single words
but attempt to imitate and repeat longer sentences. These longer utterances are often hard to interpret and may be identified in many cases as some form of jargon, that is, attempts
to produce sentence-length productions without meaning.
Peters found, however, that some of these jargon productions
may be meaningful, though the meaning may be missed by
their parents. These productions do not represent evidence
that holophrastic children know grammar, but they support
the previously stated results on well-formed commands, that
is, that the children know that sentences consist of more than
single words.
Peters (1983) and others also drew attention to the fact that
even single-word productions are not exclusively single words.
With the advent of advanced tape recorders, researchers found
in their phonetic transcriptions that words often were preceded by brief phonetic material that was often difficult to hear
or interpret. For example, a child who was saying book, was
actually saying something like uh book or uhm book. These
brief phonetic instances have been called several terms, such as
filler syllables, phonetically consistent forms, and presyntactic devices. The last term reflects the opinion of many researchers that these filler syllables are not yet syntactic units, such
as articles or auxiliaries, but they are evidence that children
are taking notice that they exist and noting their distributional
characteristics.
A further characteristic of holophrastic speech is that single words are produced in sequences. Bloom (1973) examined the successive single-word utterances of her daughter to
see if these sequences reflected later word combinations. For
example, does a child who says eat cookie when she begins
word combinations show earlier single-word sequences, such
as eat, cookie, during the holophrastic stage? Such cases
would provide evidence of grammatical knowledge during
the holophrastic stage. Bloom found that her daughters early
sequences did not show these relations, in that each word had
its own context, that is, distinctly associated action. Later, however, toward the end of the holophrastic period, sequences on
365
Summary
Children go through a period of language acquisition of six
months or so during the second year of life when they produce
single words at a time. This time of acquisition has been referred
to as the holophrastic stage. The reason is that the single words
used often communicate the idea of a sentence, that is, the
meaning of the word expressed and the communicative intent
of the utterance. The term is suggestive that preliminary grammatical acquisition is taking place during this period. Research
on childrens comprehension and production during this stage
suggests that this may be the case.
David Ingram
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bloom, L. 1973. One Word at a Time. The Hague: Mouton.
Ingram, D. 1989. First Language Acquisition. Cambridge: Cambridge
University Press.
Katz, N., E. Baker, and J. Macnamara. 1974. Whats in a name? A study of
how children learn common and proper names. Child Development
45: 46973.
Leopold, W. 193949. Speech Development of a Bilingual Child: A
Linguists Record. 4 vols. Evanston, IL: Northwestern University
Press.
Miller, J., R. Chapman, M. Bronston, and J. Reichle. 1980. Language
comprehension in sensorimotor stages V and VI. Journal of Speech
and Hearing Research 23: 284311.
Peters, A. 1983. The Units of Language Acquisition. Cambridge: Cambridge
University Press.
Pinker, S. 1984. Language Learnability and Language Acquisition.
Cambridge: Harvard University Press.
Sachs, J., and L. Truswell. 1978. Comprehension of two-word instructions by children in the one-word stage. Journal of Child Language
5: 1724.
Shipley, E., C. Smith, and L. Gleitman. 1969. A study in the acquisition of
language: Free responses to commands. Language 45: 32242.
366
Lvi-Strauss, Claude. 1969. The Raw and the Cooked. Trans. John and
Doreen Weightman. New York: Harper. This is the first volume of
Introduction to the Science of Mythology.
I
ICON, INDEX, AND SYMBOL
The nineteenth-century American philosopher C. S. Peirce developed extensive sign theories in order to explain reference,
meaning, communication, and cognition. One of the central
and most innovative features of his theories was the icon, index,
symbol classification of signs.
A crucial aspect of understanding Peirces icon, index, symbol division is his account of sign structure. According to Peirce,
any instance of signification consists of three interrelated parts: a
sign, an object, and an interpretant. For the sake of simplicity,
we can think of the sign as the signifier, for example, a written
word or an animals footprint. The object, on the other hand,
is whatever is signified, for example, the object denoted by the
written word or the animal that left the print. The interpretant is
the understanding or interpretation that the sign/object relation
generates, for example, that the word or utterance is meant to
refer to its object or that the animal track signifies the presence
of the animal that made it. The importance of the interpretant
for Peirce is that signification is not a simple dyadic relationship
between sign and object: A sign signifies an object only if it can
be interpreted as such.
With this structure in mind, Peirce was interested in classifying the various ways in which the sign/object relation might
generate an interpretant. In particular, he thought that a sign
might come to signify its object, and so generate an interpretant,
in three possible ways. First, a sign may be understood as signifying in virtue of similarities or shared qualities between it and its
object. As Peirce says, I call a sign which stands for something
merely because it resembles it an icon (1935b, 362). His own
preferred examples of icons are portraits or mathematical diagrams indeed, he thought icons were especially important to
mathematical thought. However, we can also include examples
such as color swatches, sculptures, and so on. What is central to
iconic signification is that the qualities of the sign are also qualities of the signified object and that this sharing of qualities is crucial in enabling the sign to signify.
The second way in which a sign might be understood as signifying is in virtue of some physical or causal connection between
it and its object. Such a sign is an index. Peirces own description
of an index is as a sign which refers to the object that it denotes
by virtue of being really effected by that object (1935a, 248).
Again, there are numerous and wide-ranging examples, including demonstratives and indexical expressions, weather vanes,
barometers, fever as a sign of an underlying illness, or smoke
as a sign of fire. What is crucial to indices is that the object has
a causal effect upon the sign (as in the case of fire causing the
smoke that indicates it) or has some spatio-temporal proximity
to its sign, which can be used to aid an interpreter of the sign
to grasp that object (as in the case of pointing to some nearby
object).
The third way in which a sign might be understood as signifying is in virtue of some convention or law that connects it to its
object. Peirces own description of a symbol is as a sign which
refers to the object that it denotes by virtue of a law, usually an
association of general ideas, which operates to cause the symbol to be interpreted as referring to its object (1935a, 249). There
are numerous examples of symbols, from the various words and
utterances in human languages to such things as road signs.
What is crucial in the case of symbols is that there exists some
underlying convention, agreement, habit, or law that means
that invoking some symbol invokes its associated object. For
instance, a red traffic lights being symbolic of a lack of priority
at a road junction works because we have all agreed (by habit, by
convention, and by imposing traffic regulations) to use red traffic
lights this way.
Throughout his life, Peirce made numerous alterations to his
account of signs (see, for instance, Short 2004), but the broad
division among icons, indices, and symbols tends to find a place
throughout. There are, of course, some subtleties to Peirces
account. For instance, it is not clear that there are very many
examples of signs that are purely iconic, indexical, or symbolic
that is, which do not overlap with one or both of the other elements of the trichotomy. As an example, take a painted portrait
as a sign of the person it depicts. This sign is an icon in that it signifies its object in virtue of the qualities it shares with that object
the skin and hair color of the depicted person are replicated in
the painting. But, of course, many of the things that make a portrait a successful depiction of its sitter are due to particular conventions governing paintings and how particular blocks of color
in two dimensions can stand for some subject. This seems to
make the painting look as though it has symbolic elements, too.
Similar considerations hold for indices such as barometers
although such signs indicate their objects in virtue of a causal
and physical connection with their object; conventions about
how we should interpret this physical connection also seem to
play a part in signification. Whats more, there are clear instances
of symbols that have some iconic element. Obvious examples
might include forms of writing, such as Chinese, that involve
pictograms, at least partially. Even onomatopoeic words such as
cuckoo present clear cases of symbols with a strong iconic element the phonic qualities of the object are aped by the phonic
qualities of the word.
Peirce was aware of the various overlaps among icons, indices, and symbols, and at some point proposed to call icons and
indices with symbolic elements hypo-icons and subindices as
a way of acknowledging this. However, in any case where more
than one of the three elements is present, one will be most
prominent. Consequently, we can think of Peirces trichotomy
as dividing signs according to whether they are predominantly
iconic, indexical, or symbolic.
The main influence of Peirces division is in semiotics,
where his work is considered foundational. However, the icon,
index, symbol distinction has had some influence in philosophy, particularly through the work of Arthur Burks (1949), and
has even been used in such diverse areas as literary theory (see,
for example, Sheriff 1989), film theory (see, for example, Wollen
1969; see also film and language), and musicology (see Turino
1999; see also music, language and). The use and relevance of
this distinction to linguistics are similarly diverse, but it features
most prominently in analyses of the relation between animal
communication and human language and in some explanations of the evolution of language.
367
368
369
370
statistically based examination of social networks to more interpretative examination of communities of practice, defined
as an aggregate of people who come together around mutual
engagement in an endeavor (Eckert and McConnell-Ginet
1992, 464). In the course of this endeavor, shared beliefs, norms,
and ideologies emerge, including, though not limited to, linguistic and communicative behavior. This line of research is thus
continuous with another one that has focused more directly
on the normative beliefs or ideologies by which national and
other group identities are maintained (see Verschueren 1999;
Blommaert 1999; Kroskrity 2000).
Other features of recent work on language and identity
include the view that identity is something constructed, rather
than essential, and performed, rather than possessed features
that the term identity itself tends to mask, suggesting as it does
something singular, objective, and reified. Each of us performs
a repertoire of identities that are constantly shifting and that we
negotiate and renegotiate according to the circumstances.
371
372
Idioms
of criticism is about the sorts of texts that analysts choose to
subject to ideological analysis. To be blunt, if we know a text to
be ideologically problematic at the outset, then any subsequent
linguistic analysis will only confirm what we already know, and
any linguistic feature uncovered through the analysis can by
imputation be passed off as ideologically insidious. This deterministic approach connects with the third major area of concern, which is simply that studies of ideology and language tend
to be elitist. If the main purpose of the analysis is to uncover and
challenge the repressive discourse practices of powerful, interested groups, then what needs to be considered before anything
else are the effects of these practices on ordinary (nonacademic)
people. Reactions of ordinary communities to what the analysts
deem ideologically insidious discourse are rarely considered;
instead, the academic analyst comfortably assumes the perspective of those for whom the text was intended, moving seamlessly
in and out of the multiple interpretative positions of specialist
and nonspecialist alike. It is still early to say how these serious
and far-reaching criticisms will affect the ways in which scholars
investigate the widespread interconnections between ideology
and language.
Paul Simpson
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Althusser, L. 1971. Lenin and Philosophy and Other Essays.
London: NLB.
Blommaert, J. 2005. Discourse: A Critical Introduction. Cambridge:
Cambridge University Press.
Charteris-Black, J. 2004. Corpus Approaches to Critical Metaphor Analysis.
Basingstoke and New York: Palgrave MacMillan.
Chilton, P. 1985. Language and the Nuclear Arms Debate: Nukespeak
Today. London: Pinter.
Fairclough, N. 1992. Discourse and Social Change. Cambridge: Polity
Press.
. 2001 Language and Power. 2d ed. London: Longman.
Foucault, M. 1984. Truth and power. In The Foucault Reader, ed.
P. Rabinow, 5175. Harmondsworth, UK: Penguin Books.
Fowler, R. 1991. Language in the News: Discourse and Ideology in the
Press. London: Routledge.
Fowler, R., R. Hodge, G. Kress, and T. Trew. 1979. Language and Control.
London: Routledge.
Marx, K. [with Frederick Engels]. [1933] 1965. The German Ideology. Ed.
and trans. S. Ryazanskaya. London: Lawrence and Wishart.
Simpson, P. 1993. Language, Ideology and Point of View.
London: Routledge.
. 2004. Stylistics. London: Routledge.
Stubbs, M. 1997. Whorfs children: Critical comments on critical discourse analysis (CDA). In Evolving Models of Language, ed. A. Ryan
and A. Wray, 10016. Swansea: BAAL.
Widdowson, H. 1995 Discourse analysis: A critical view. Language and
Literature 4.3: 15772.
. 1996. Reply to Fairclough: Discourse and interpretation: Conjectures and refutations. Language and Literature
5.1: 5770.
IDIOMS
Idioms hold an important place in the class making up fixed, nonliteral expressions. These have a perplexing characteristic: They
373
Idioms
374
and
the
Lexicon.
375
376
Note that the partitions in (1c) and (2c) do not contain a component of illocutionary force. They are also the meanings of embedded interrogatives like (1b) and (2b), and embedded clauses are
standardly assumed not to have illocutionary force.
Thus, the meaning of a declarative is a proposition and the
meaning of an interrogative is a partition. Let us consider how
illocutionary force of two kinds may be added to these meanings: i) illocutionary force as a statement: the speaker expresses
that (s)he believes something to be true, and ii) illocutionary force
as a question: the speaker wants to learn from the addressee the
truth in an as-yet open issue. Let us begin by assuming that the
choice between statement and question is pragmatically inferred
rather than grammatically triggered.
THE ILLOCUTIONARY FORCE OF DECLARATIVES. The meaning of a
declarative (It is raining) as a proposition (that it is raining) makes
it likely that the speaker intends a statement interpretation of his/
her utterance (the speaker conveys that [s]he believes that it is raining), perhaps in connection with the maxim of quality in Grice
(1975), which requires such truthfulness (see also cooperative
principle). If we assume that this statement interpretation is not
hardwired into declaratives, we correctly allow declarative questions, as in It is raining? Here, a declarative sentence with question intonation is used as a question. Unlike in statements, the
speaker does not convey that he/she believes that it is raining.
THE ILLOCUTIONARY FORCE OF INTERROGATIVES. No single state
of affairs (ordinary proposition) is offered in the interrogative
meaning (the true one among , see [1c] and [2c]), and so
there is no specific content for an interpretation as a statement.
The interrogative meaning makes a question interpretation
particularly likely, in which the speaker wants the addressee
to pick the true one among the possibilities. If we assume that
this illocutionary force is not hardwired into root interrogatives,
we correctly allow the flexibility for untypical uses. In monological questions, such as in a lecture, the speaker later provides the answer him-/herself: Who was behind these reforms?
Everything points towards Bismarck (Brandt et al. 1992). In rhetorical questions like Am I your servant? the true one among
the given possibilities is to be inferred and need not be stated.
377
378
Image Schema
can be approximated in terms of a commitment by a salient individual that is expressed or assumed. Sentence mood (such as
imperative vs. indicative) plays a further crucial role in conditioning the illocutionary interpretation of a clause.
Hubert Truckenbrodt
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Austin, John L. 1962. How To Do Things with Words. Oxford: Clarendon.
Brandt, Margareta, Marga Reis, Inger Rosengren, and Ilse Zimmermann.
1992. Satztyp, Satzmodus, und Illokution. In Satz und Illokution 1,
ed. Inger Rosengren, 190. Tbingen: Niemeyer.
Grice, H. P. 1975. Logic and conversation. In Syntax and Semantics
3: Speech Acts, ed. P. Cole and J. Morgan, 4158. New York: Academic
Press.
Groenendijk, Jeroen, and Martin Stokhof. 1982. Semantic analysis of
WH-complements. Linguistics and Philosophy 5: 175233.
Gunlogson, Christine. 2001. True to form: Rising and falling declaratives
as questions in English. Ph.d. diss., University of California at Santa
Cruz.
Heycock, Caroline 2006. Embedded root phenomena. In The Blackwell
Companion to Syntax. Vol. 2. Ed. M. Everaert and H. van Riemsdijk,
174209. Oxford: Blackwell.
Karttunen, Lauri 1977. Syntax and semantics of questions. Linguistics
and Philosophy 1: 344.
Knig, Ekkehard, and Peter Siemund. 2007. Speech act distinctions in
grammar. In Language Typology and Syntactic Description. Vol. 1. Ed.
T. Shopen, 276324. Cambridge: Cambridge University Press.
Portner, Paul. 1997. The semantics of mood, complementation, and
conversational force. Natural Language Semantics 5: 167212.
Schwager, Johanna Magdalena. 2005. Interpreting imperatives. Ph.d.
diss., University of Frankfurt am Main.
Searle, John R. 1969. Speech Acts. An Essay in the Philosophy of Language.
Cambridge: Cambridge University Press.
. 1975. A taxonomy of illocutionary acts. In Language, Mind,
and Knowledge, ed. K. Gunderson, 34469. Minneapolis: University
of Minnesota Press. Repr. John R. Searle, Expression and Meaning
(Cambridge: Cambridge University Press, 1979), 129.
Truckenbrodt, Hubert. 2004. Zur Strukturbedeutung von
Interrogativstzen. Linguistische Berichte 199: 31350.
. 2006a. On the semantic motivation of syntactic verb movement
to C in German. Theoretical Linguistics 32: 257306.
. 2006b. Replies to the comments by Grtner, Plunze and
Zimmermann, Portner, Potts, Reis, and Zaefferer. Theoretical
Linguistics 32: 387410.
IMAGE SCHEMA
A foundational concept in cognitive linguistics, image
schemas are associated most closely with the work of Mark
Johnson (1987) and his collaborator George Lakoff (1987). Image
schemas are thought to play a key role in the acquisition, structure, and use of language; a central tenet of cognitive linguistics
is that the perceptual interactions and motor engagements with
the physical world shape the way we think. An image schema
can be defined as a condensed representation of bodily motor
experience for purposes of mapping spatial structure onto conceptual structure. According to Johnson, these patterns emerge
as meaningful structures based on our bodily movements
through space, our manipulations of objects, and our perceptual interactions (1987, 29). Jean Mandler aptly likens image
Image Schema
schemas to the representations one is left with when one has
forgotten most of the details (2004, 7981).
Lack of specificity and content makes image schemas highly
flexible preconceptual and primitive patterns used for reasoning
in an array of contexts (Johnson 1987, 30). A partial inventory of
image schemas is as follows: container; balance; blockage;
counterforce; restraint removal; source-path-goal; link;
center-periphery; verticality; scale; part-whole; support.
Since the publication of the influential works of Johnson and
Lakoff, the notion of image schemas has been theorized, investigated, and applied in multiple domains of inquiry. Within
cognitive linguistics, the notion of an image schema has taken
center stage in research in semantic and grammatical analysis, psycholinguistics, cognitive development, and neurocomputational modeling, to name the most prominent areas of
activity in the language sciences. Controversy has ensued over
the definition of and criteria for positing something as an image
schema, over its status as conscious or unconscious representations, and even over its status vis--vis individual and social cultural cognition.
Implicational Universals
motion; and caused motion emerge as preverbal, perceptually
based conceptual primitives underlying cognitive development
and language acquisition. Chris Sinha and K. Jensen de Lopez
take a slightly different tack in emphasizing the social-cultural
and artifact dimensions of the acquisition of locative prepositions in and under in Danish-acquiring, English-acquiring, and
Zapotec-acquiring children. They conclude that the sociocultural environment plays a greater role in cognitive development
than previously thought. Image schemas may in fact be distributed throughout the local environment, with considerable differences appearing among cultures with very different material
makeups.
Finally, image schema theory is a central pillar of the neural
theory of language project initiated by Jerome Feldman (2006),
Lakoff (1987), Srini Naryanan, and others. A crucial component
of this neural computational model is the execution or x-schema
protocol for representing human actions, which include models
for enacting drop-schema that simulate the neural computational activity involved. For instance, a computational representation for distinguishing between verbs such as push and shove
invokes the slide x-schema but differs in the microdetails of
body-part movement and acceleration.
Todd Oakley
WORKS CITED AND SUGGESTION FOR FURTHER READING
Cienki, Alan. 1998. Straight: An image schema and its transformations.
Cognitive Linguistics 9.2: 10749.
Feldman, Jerome. 2006. From Molecule to Metaphor. Cambridge,
MA: MIT Press.
Gibbs, Raymond W. 2005. The psychological status of image schemas.
In Hampe 2005, 11335.
Hampe, Beate, ed. 2005. From Meaning to Perception: Image Schemas
in Cognitive Linguistics. Berlin: Mouton De Gruyter. An edited collection of essays presenting a wide range of views on the nature of image
schemas.
Johnson, Mark. 1987. The Body in the Mind. Chicago: University of
Chicago Press. This book is considered the locus classicus of image
schema theory.
Lakoff, George. 1987. Women, Fire, and Dangerous Things.
Chicago: University of Chicago Press.
Mandler, Jean. 2004. The Foundations of Mind. New York: Oxford
University Press.
Matsumoto, Yo. 1995. From attribution/purpose to cause: Image
schema and grammaticalization of some cause markers in Japanese.
In Lexical and Syntactic Constructions and the Construction of
Meaning, ed. Marolijn Verspoor, K. Dong, and E. Sweetser, 287307.
Amsterdam: John Benjamins.
Sinha, Chris, and K. Jensen de Lopez. 2000. Language, culture, and the
embodiment of spatial cognition. Cognitive Linguistics 11.1/2: 1741.
Turner, Mark. 1991. Reading Minds. Princeton, NJ: Princeton University
Press.
Verspoor, Marolijn. 1995. Predicate adjuncts and subjectification.
In Lexical and Syntactic Constructions and the Construction of
Meaning, ed. Marolijn Verspoor, K. Dong, and E. Sweetser, 43349.
Amsterdam: John Benjamins.
IMPLICATIONAL UNIVERSALS
typological universals seek to capture the limits of grammatical variation, that is, the extent to which particular traits
379
Implicational Universals
or features of languages may covary, both within languages
and across languages. Typological universals may be classifed
along several dimensions. The most important distinction is
between substantive and implicational universals. The former deal with a single property and reflect what occurs in all
languages (e.g., all languages have vowels). Since such unrestricted universals tell us little about variation, they are considered to be of moderate interest. For the study of variation,
implicational universals are of considerably greater interest
since they reflect not only the existence of variation but, crucially, the constraints imposed on it.
Implicational universals relate two (or more) logically independent features or properties and apply to some subset of
languages for which the given features or properties obtain.
Implicational universals may be absolute or statistical.
Absolute implicational universals hold for all languages; that is,
they specify that if a language has feature A, then it will have
feature B. By contrast, statistical implicational universals specify relationships that hold only at a certain level of probability.
Statistical implicational universals take the form, If a language
has feature A, then with greater than chance frequency it will
have feature B. Examples of the two types of implicational universals are given in (1a) and (2a), respectively, and the distribution of features that each universal reflects is presented in (1b)
and (2b) (in the form of a tetrachoric table). (The data depicted
in (2b), where the numerals reflect the number of languages
with the specified features, are taken from Greenberg (1963,
Appendix II.)
(1)
a. If the demonstrative follows the head noun, then the relative clause also follows the head noun
b.
DemN
NDem
RelN
+
NRel
+
+
(2)
Note that the five prepositional SOV languages are counterexamples to the universal. Although many absolute implicational
universals have been posited in the typological literature, most,
particularly of the simple kind (see the following), have turned
out to be in fact statistical.. Needless to say, these may be assessed
from the point of view of their relative strength (the number of
overall languages considered, the number of languages displaying the features in question, the number of exceptions to the
universals) and their validity in relation to criteria such as those
presented in Stassen (1985, 201).
Implicational universals may be monodirectional or bidirectional. The absolute implicational universal in (1) is monodirectional since NRel order is compatible with both DemN and
NDem. Consequently, while NDem order entails NRel, NRel
order does not entail NDem. The statistical implicational universal in (2), on the other hand, is bidirectional; not only does SOV
order favor postpositions over prepositions, but postpositions
380
also favor SOV order over non-SOV order. However, given the
data in (2b), the strength of the implicational universal with SOV
as the antecedent and postpositions as the consequent is greater
(95%) than the converse universal with postpositions as the antecedent and SOV order as the consequent (91%).
The implicational universals in (1) and (2) are simple ones,
as they specify a dependency between only two traits. Relations
between several traits are captured by means of complex universals that may involve conjunctions or disjunctions of traits within
the antecedent and/or the consequent, as depicted in (3).
(3)
a. X (Y & Z)
b. (X or Y) Z
Indeterminacy of Translation
Cristofaro, Sonia. 2003. Subordination. Oxford Studies in Typology and
Linguistic Theory. Oxford: Oxford University Press.
Croft, William. 2003. Typology and Universals. 2d ed.
Cambridge: Cambridge University Press.
Greenberg, Joseph. 1963. Some universals of grammar with particular reference to the order of meaningful elements. In Universals of
Human Language, ed. Joseph Greenberg, 73113. Cambridge, MA:
MIT Press.
Hopper, Joan Bybee. 1976. An Introduction to Natural-Generative
Phonology. New York: Academic Press.
Keenan, Edward, and Bernard Comrie. 1977. Noun phrase accessibility
and universal grammar. Linguistic Inquiry 8: 6399.
Maslova, Elena. 2000. A dynamic approach to the verification of distributional universals. Linguistic Typology 3/4: 30733.
. 2003. A case for implicational universals. Linguistic Typology
7.1: 1018.
Stassen, Leon 1985. Comparison and Universal Grammar. Oxford: Basil
Blackwell.
INDETERMINACY OF TRANSLATION
In Word and Object (1960), one of the most influential books in
the history of American philosophy, W. V. O. Quine presents an
argument along the following lines. One may always formulate
mutually exclusive hypotheses regarding the meaning of a linguistic item such that there is no fact as to which hypothesis is
correct. Put differently, there are always divergent ways of defining a given word such that no one of these ways is correct. This
highly influential, and highly controversial, indeterminacy of
translation thesis extends Quines argument against analyticity (the view that some statements are true simply due to the
meaning of the terms see Quine 1961; see also meaning and
belief) and points toward his conception of ontological relativity (that reference is not absolute, but relative to a coordinate
system [Quine 1969, 48], that Factuality is internal to our
theory of nature [Quine 1981, 23]).
To explore Quines idea, we might begin with a simple distinction between decidability criteria and demarcation criteria. A decidability criterion allows us to chose among different
hypotheses regarding a words meaning (see word meaning).
A demarcation criterion distinguishes in principle between the
validity of two such hypotheses. A demarcation criterion presumably isolates a factual difference; a decidability criterion, first of
all, isolates an evidential difference (or, more broadly, a distinction by the lights of scientific method; thus, it may incorporate
simplicity or other adjudicative criteria). Quine argues that there
is no strict decidability criterion for meaning. We may refer to
this as uncertainty. But he also maintains that for at least some
set of mutually exclusive hypotheses, there is no demarcation
criterion either. The latter is indeterminacy proper.
Quines arguments are bound up with his holism, the
view that hypotheses of any sort, including hypotheses regarding meaning, are part of larger complexes of belief and are not
understandable outside those complexes. Crucially, this means
that apparent counterevidence regarding one part of the whole
need not have simple, local consequences. Such counterevidence may be accommodated by alterations elsewhere in the
whole. Suppose I am a field linguist investigating a new language.
I encounter the word vagavai, which I take to mean gold. I then
381
Indexicals
psychological fact about the degree of vagueness or ambiguity of my usage at particular times and places.
These arguments, then, point toward global, though still adjudicable uncertainty along with local, subencoding indeterminacy
in the limited sense of vagueness or ambiguity. Both are significant and both are highlighted by Quines discussions. However,
neither is radical indeterminacy.
On the other hand, to make these arguments is to go against a
range of Quines other views for example, his behaviorism and
his commitment to ontological relativity. Like all theories, Quines
own theories operate holistically. Even if he were to accept some
version of the preceding arguments, he could accommodate
these arguments by alterations elsewhere in the system. Thus,
despite such arguments, radical indeterminacy remains at least
a continuing and important philosophical challenge.
deictic use (see deixis), but indexicals have many other uses.
I is used demonstratively when (1) is written near a picture of
Truman with an arrow pointing at the picture. It is used anaphorically in (2) as opposed to (3):
(2)
(3)
I am Timothy,
Je means I in French.
(6)
INDEXICALS
An indexical is a linguistic expression whose reference (extension) may vary from one context of use to another even when
used with the same meaning (sense) and with respect to the
same possible world or story. An example is the pronoun I,
which I use to refer to myself and you to yourself. Sentence (1)
itself is indexical: true when used by Harry Truman in 1951 but
not when used by Churchill.
(1)
I am president.
382
In a semantic use like (5), it is crucial that I be used with the first
person meaning we have been focusing on; if it means iodine,
(5) is false. Sentence (6) illustrates the quantificational use, with
pronouns bound by quantifiers like variables in quantification theory.
Indexicals present a problem for formal theorists because
assigning extensions and intensions (functions from possible
worlds to extensions) to indexicals cannot adequately represent
their role in determining the truth conditions of sentences (see
truth conditional semantics). David Kaplans ([1977]
1989) seminal solution was to represent the meanings of indexicals by assigning them characters (functions from contexts to
intensions). For I, Kaplan had in mind the function satisfying the following condition: i(c) is the intension whose value in
any world w is the speaker uttering I in c. Thus, the value of
i(c) in any context in which Theodore Roosevelt used I is the
constant function tr(w) whose value is Theodore Roosevelt for
every world. The value of i(c) in any context in which Franklin
Roosevelt used I is the constant function fdr(w) whose value
is Franklin Roosevelt for every world.
The character theory of meaning has many of the difficulties
of referential theories generally. For example, I and the person identical to me have the same character, but not the same
meaning (Freges problem). The character function for he would
seem to be undefined when it is used in After Santa came down
the chimney, he left the presents; Santa does not exist (Russells
problem). Quantificational and semantic uses are nonreferential
for different reasons. Another difficulty is the Enterprise problem.
Suppose Mary points at the bow of a big ship and says That is
the Enterprise and then points at the stern and says That is not
the Enterprise. The character function for that would assign it
the same intension in both these contexts the one whose value
at every world is the Enterprise. Yet Mary is not contradicting
herself the way she would if the same pointing gesture accompanied both her utterances. The essential indexical problem is that
sentences like (2) and (7) may differ in truth value, as in cases of
amnesia or delusion.
(7)
varieties of language, varieties that did not offer the same social,
cultural, and political rewards as others.
The central insight here is indeed the parallelism between
linguistic variation and social differentiation, between linguistic differences and social hierarchies and forms of stratification
often organized around the dynamics of prestige and stigma.
This insight has been developed by a number of scholars in sociolinguistics and related fields of study (Irvine 1989). I first give a
brief survey of the development of this topic in sociolinguistics
and then turn to a discussion of two authors: Pierre Bourdieu
and Dell Hymes.
Inequality in Sociolinguistics
Bernsteins thesis about elaborated and restricted codes became
a deeply controversial one, though often for the wrong reasons. It
coincided with the emergence of modern sociolinguistics, a discipline that saw itself initially as devoted to the study of linguistic
variation and distribution as a horizontal exercise: an exercise
in mapping linguistic varieties over fragments of a population,
based on an assumption of the fundamental equivalence of every
linguistic variety. Thus, Bernsteins vertical image of stratified
variation variation that comes with differential value attribution was countered by William Labov (1972) and others, who
argued that a linguistic variety such as black American English
should not be seen as bad English (which Labov chose to equate
with Bernsteins restricted code) but as a complex, sophisticated
code used by virtuoso speakers. Labov, like Bernstein, started
from an awareness of real inequalities in language: The black
American English variety was stigmatized in U.S. education and
its speakers were often negatively categorized. But Labovs efforts
were aimed at demonstrating the intrinsic linguistic equivalence of these different varieties, whereas Bernstein addressed
the extrinsic attributional, ideological nonequivalence of
the varieties. Seen from a historical distance, both efforts were
connected by a joint concern for how particular language varieties disenfranchized speakers, not because of their intrinsic
inferiority but because of social and political perceptions of the
varieties and their speakers. In other words, both were addressing
a language-ideological effect in which linguistically equivalent
varieties indexed socially unequal features and categories.
Similar ideas were to be found in the work of many of the early
sociolinguists. John Gumperz (1982) stressed the different ways
in which intercultural misunderstandings were not a result of
participants intentions but were an effect of small differences
in linguistic and communicative structures in a particular context of different contextualization cues. In his work as well, we
encounter an awareness of relativity: Different language varieties
can fulfill the same functions, but the perception of these functions by people not sharing the contextual conventions of a particular variety can differ. Here again, we encounter the tension
between intrinsic equivalence and extrinsic nonequivalence.
From a different vantage point, sociolinguists concerned
with language policy stressed the fact that the political
institutionalization of language in multilingual societies (see
bilingualism and multilingualism) usually involved a
stratification based not on intrinsic superiority of one language
but on prestige hierarchies and ideological images of society.
J. Fishman (1974) presented studies in which former colonial
383
State of Affairs
In all of the work discussed so far, there is an awareness of
inequality, notwithstanding that the emphasis would be on difference and variation, rather than on the power and inequality
dimensions of difference and variation. Other scholars developed
full-blown theories of linguistic and communicative inequality,
and two stand out: Bourdieu and Hymes.
Bourdieu (1991) emphasized the ways in which language is
part of the superstructural apparatus of society; as a nonmaterial resource, it can nevertheless be imbued with an economic
value: social or cultural symbolic capital. As in hard economic
sectors, the value of different resources is different and can fluctuate, and not every resource can be traded for another one; the
terms of exchange are unpredictable. Bourdieus work on language fitted into his larger program of analyzing the economies
of taste, culture, and ideas from within a generalized materialism
and with the aim of redefining (and empirically substantiating)
the notion of social class. In his view, the traditional distinction between material and immaterial resources in society was
unjustified since the same forces of differential value allocation
appeared to operate on both: Class is both a material and an
immaterial thing, and language differences do play a role in this
value allocation. Consequently, the stability of class as a material
complex is also there in the immaterial aspects: A lack of real capital is often paralleled by a lack of symbolic capital, and social hierarchies play as much into the symbolic diacritics as they do into
the material ones. (See also field and habitus, linguistic.)
It is not difficult to see the similarities between the project
developed by Bourdieu and the (more limited) one presented
by Bernstein. Both treat language not just as an opportunity a
positive feature of humans but also as a constraint, as something inherently limited and limiting. And rather than enabling
people to perform particular acts, language also restricts them
and prevents them from performing other acts. Underlying this
idea of language as constraint is a model of society as nonegalitarian and stratified, with relatively stable strata, some of which
are tightly controlled as to access while others are more democratically accessible. Particular language resources are required
to move into the more controlled strata, and membership of
these strata for example, elite, professional milieux requires
384
Evaluation
Despite its implicit focus on inequality, sociolinguistics privileges difference and variation as its objects and avoids explicit
analyses of how such variation converts into social inequality.
This is an effect of the descriptive bias in sociolinguistics, as well
as of the hesitant and ambivalent relationship between sociolinguistics and social theory (Williams 1992). Many studies of
variation, consequently, consider differences merely in terms of
spread and distribution and see the power and inequality relationships among language varieties in simplified terms. The fact
that speaking with a stigmatized accent is not merely a matter
of distribution but a matter of social opportunities, determined
by the way in which others rank such accents in an ideologically
informed hierarchy, could be more central to sociolinguistic
reflections.
This is a critical insight and eminently applicable to many
sociolinguistic and discursive phenomena in the contemporary
385
386
In Bs response, the noun phrase (NP) Mary in the first sentence (or, more accurately, the piece of referential information represented by the NP Mary) is given information, as it
was mentioned in the immediately preceding turn; John, on
the other hand, is new information, as it cannot be assumed to
be in the hearers consciousness. In Bs second sentence, both
Mary and John are now given information, which explains why
they can be pronominalized as she and him, respectively. At
the same time, it could be argued that both sentences are more
about Mary than about John and that Mary is, therefore, the
theme of both sentences, the remaining information being rhematic (i.e., providing additional information about the theme).
Although both levels are functionally independent, they do
appear to correlate strongly, in that themes tend to consist of
given information.
Example (1) also illustrates some tendencies with regard to
word order and prosody. First of all, new information tends to
occur toward the right of the sentence, whereas given information tends to be initial (in English, therefore, being an subjectverb-object (SVO) language, there is a strong correlation between
givenness and grammatical subjecthood, as shown by Mary in
Bs contribution). Secondly, thematic information tends to be
sentence-initial. In fact, some linguists argue that the theme
is initial by definition (e.g., Halliday 1994). Thirdly, the main
387
388
Information Theory
INFORMATION THEORY
This is a general purpose and abstract theory for the study of
communication. Standard information theory was founded
by Claude Shannon (1948) and is based on a communication
framework in which a) the sender must transform a message
into a code and send it through a channel to the receiver, and
b) the receiver must obtain a message from the received code.
Communication is successful if the message of the sender is the
same as the message obtained by the receiver. For instance, a
speaker utters a word for a certain meaning and then the hearer
must infer the meaning that the speaker had in mind. Noise can
alter the code when traveling from the sender to the receiver
389
390
S1
S0
S2
Si
Sn
Sn
Sn
Figure 1.
D1
D2
Di
Dn
Dn+1
Dn+2 ........
a reasonable length, the number of well-formed sentences in
English is astronomical. It has been estimated that the number
of grammatical English sentences of 20 words and less is 1020
(Levelt 1967). (Note that this very normal sentence is exactly 21
words and costs nine seconds to pronounce). At an average of six
seconds per sentence, it will cost 19 trillion years to say (or hear)
them all. In the case of nonstop listening, the percentage of these
a child could have heard in six years time is 0.000000000031,
clearly still a gross overestimation as compared to what the child
can actually be expected to hear. So on the basis of at most such
an extremely small percentage of the potential input, the child
gets to know how language works.
There are many further practical complications we ignored,
such as lack of homogeneity and the presence of errors in the
data. If we were to take these into account, the task would only
become more formidable. This sets the stage for the logical
problem of language acquisition (for instance, Chomsky 1965,
1986), which can be formulated as the projection problem:
Consider a given finite set taken from some (infinite) superset.
Determine (the characteristic function of) this superset on the basis
of this subset.
Like anyone can see, this task is in its generality impossible. For
any finite given set, the projection problem has infinitely many
solutions.
For a concrete illustration, consider the following task involving the completion of a series of numbers:
(1)
1,2,3,4,5,6,7, .
One might say that finding the next number is easy. It should
obviously be 8. But of course, this is not at all guaranteed. It is easy
to think of a perfectly well-behaved function that enumerates the
first seven natural numbers, followed by their doubles, triples,
quadruples, and so on. This illustrates a very simple point: There
is no general procedure to establish the correct completion of
some initial part of a sequence, whether a sequence of numbers
as in (1) or a data sequence (D1, D2, Di, Dn) as in Figure 1.
This fact reflects a poverty of the stimulus in a fundamental
sense, as a trivial logical truth.
The completion task may become possible, however, if it is
redefined as the task to find a solution within a restricted space of
possible solutions. In that case, certain instances of the projection
problem become even trivial. For instance, (1) can be trivially
completed if it is given that there is a constant difference between
each member of the series and its successor.
As E. Gold (1967) showed, even highly restricted hypothesis
spaces may not ensure a solution of the projection problem as
defined. The task may become easier if the input does contain
systematic evidence as to what is not in the target language.
Since, as is generally acknowledged, the input to the child does
391
392
Applied to learning:
Learning a recursive step by presentation only, without restrictions
on the hypothesis space, is as impossible as creating the perpetuum
mobile.
Integrational Linguistics
depth of misunderstanding involved: It is an empirical problem,
but analyzing the logical problem is essential for solving it.
All this does not demonstrate that the restrictions necessary
for language acquisition are specific to language. It does show
that such restrictions are there and have to be studied if we are to
understand language acquisition. No insight can be gained unless
precise and substantive hypotheses are formulated and tested in
a way that reflects what we already know about language.
Eric Reuland
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Chomsky, Noam. 1957. Syntactic Structures. The Hague: Mouton.
. 1965. Aspects of the Theory of Syntax. Cambridge, MA: MIT Press.
. 1980. On cognitive structures and their development. In
Language and Learning, The Debate between Jean Piaget and Noam
Chomsky, ed. M. Piattelli-Palmarini, 3552. Cambridge: Harvard
University Press.
. 1986. Knowledge of Language: Its Nature, Origin and Use. New
York: Praeger.
. 1995. The Minimalist Program. Cambridge, MA: MIT Press.
Deacon, Terrence. 1997. The Symbolic Species. New York: Norton.
Fodor, Jerry. 1983. The Modularity of Mind. Cambridge, MA: MIT Press.
Gold, E. 1967. Language identification in the limit. Information and
Control 16: 44774.
Hirsch, C., and K. Wexler. 2007. The late development of raising: What
children seem to think about seem. In New Horizons in the Analysis
of Control and Raising, ed. W. Davies and S. Dubinsky, 3770.
Heidelberg: Springer.
Kandel, Eric, James Schwartz, and Thomas Jessell. 2000. Principles of
Neural Science. 4th ed. New York: McGraw-Hill.
Lenneberg, Eric H. 1967. Biological Foundations of Language. New
York: Wiley.
Levelt, Willem. 1967. Over het Waarnemen van Zinnen. Groningen,
Germany: Wolters.
Pullum, Geoffrey K., and Barbara C. Scholz. 2002. Empirical assessment
of stimulus poverty arguments. Linguistic Review: 19.1/2: 950.
Quine, Willard Van Orman. 1960. Word and Object. Cambridge, MA: MIT
Press.
Reuland, Eric. 2005. Binding conditions: How are they derived? In
Proceedings of the HPSG05 Conference Department of Informatics,
University of Lisbon, ed. Stefan Mller. Stanford, CA: CSLI Publications.
Available online at: http://csli-publications.stanford.edu/
Tomasello, Michael. 2003. Constructing a Language. Cambridge: Harvard
University Press.
Wexler, Kenneth, and Peter Culicover. 1980. Formal Principles of
Language Acquisition. Cambridge, MA: MIT Press.
INTEGRATIONAL LINGUISTICS
This is the application of integrational semiology to the study
of language. Integrational linguistics is based on the assumption that human communication, whether verbal or nonverbal,
involves the creation of signs in particular contexts whereby two
or more individuals engage in joint integration of their respective activities. Thus, for example, speech communication would
be impossible without integration of the biomechanically separate activities of vocalization and hearing. Written communication requires the production of marks on a surface that can be
integrated with programs of optical scanning. The appropriate
integration of these and many other activities is one of the major
393
Integrational Linguistics
functions of the cerebral cortex, and failure to achieve the integrational proficiency required for social fluency in communication is commonly perceived as a defect or handicap of some
kind (e.g. deafness, dyslexia, etc.) when due to physiological
factors. Integrationists differentiate forms of communication
according to the range of activities typically integrated by the
participants, and the kinds of integrational proficiency typically
required for participation. A blind person, for obvious reasons,
lacks the integrational proficiency presupposed in various forms
of visual communication (see blindness and language).
From this perspective, the term language does not correspond
to any one mode of communication but straddles or conflates
several. In this respect, integrational linguistics differs radically
from mainstream schools of thought in linguistics and neighbouring disciplines, which tend to assume that language is a single human faculty, common to all humanity, and that languages
(English, French, Latin, etc.) are different social codes enabling
individual users to exercise this faculty. For integrationists, on
the other hand, individuals are not language users but language
makers. They make language by their creative integration of verbal signs into a myriad diverse activities, in both expected and
unexpected ways, with due regard for the circumstances, just as
they make human relationships by the ways in which they interact with others in particular cases.
394
Integrational Linguistics
a case in which A insulted B, who then struck A, from a case in
which A struck B, who then insulted A. As individuals, we are
time-bound agents in all our activities. Our linguistic acts do
not have some special time track of their own. There is no such
thing as a contextless linguistic sign.
This does not mean that the context simply is a given
sequence of events. Integrationists take a very different view of
context from that usually found in orthodox linguistics. Context
is not to be equated with situation (which may be irrelevant)
and even less with preceding or following speech-acts. It is
not some kind of local backdrop against which communication
takes place. Context, for the integrationist, is always the product
of contextualization, and each of us contextualizes in our own
way. The individual participants in any communication situation
will each contextualize what happens differently, as a function
of the integrational proficiency each exercises in that situation.
This does not mean that we can never reach communicational
agreement, but it explains why we often do not. It is not enough
to say that every act of communication is unique. Each such act
is in principle subject to multiple contextualizations and recontextualizations. That is what makes it essential in linguistic analysis for the analyst to specify what forms of contextualization are
presupposed (a requirement that the great majority of orthodox
linguists ignore). For integrationists, a sign is not a sign until it
has been contextualized: The act of contextualization and the
establishment of the sign are one and the same.
Meaning
It follows, then, that integrationists take a quite different view of
meaning from that which informs most work in orthodox linguistics, where the meaning of a linguistic form is usually construed
as some kind of concept (as in the definitions of conventional
dictionaries) or, even more vaguely, mental representation.
This is traditionally imagined to yield a more or less permanent
value attached to the form and known to all competent speakers and writers of the language in question (even if the language
is no longer living). Thus, the meaning of classical Latin aqua is
treated as timeless: It still means for Latin specialists what it
meant in the days of Julius Caesar and will mean in a thousand
years from now.
For integrationists, this assumption conflicts with the principle of cotemporality. There are no atemporal invariants in language. Meaning is made by participants as part of the process of
communication. It is thus subject to the principle of cotemporality in just the same way as all other aspects of communication.
There are no fixed meanings. There is nothing in language to provide us with a miraculous guarantee of the stability of meaning(s)
over time or even from one moment to the next. To demand
such a guarantee for any mode of communication is as futile as
demanding that a currency remain stable in value from one day
to the next. (The indeterminacy holds regardless of whether people shopping in High Street are aware of it.)
We have no option but to interpret particular episodes of
communication by integrating them into the unique temporal
sequence of events that constitutes our previous experience.
Thus, where there are two or more participants, what is communicated must be open to two or more interpretations. These
cannot be guaranteed to coincide. People are no more obliged
Parameters of Communication
For integrationists, there are certain capacities required of an
individual in order to participate in communication. These
capacities are of three kinds: biomechanical, macrosocial, and
circumstantial. The first relates to the human organism and the
ability to integrate activities requiring a wide variety of physiological and mental processing (e.g., the very different biomechanical requirements involved in speech production and
hearing). The second relates to the human ability to integrate
particular activities into sets of assumptions provided by social
conventions of various kinds. The third relates to the ability to
integrate ones activities into whatever else is going on at the
time.
A simple illustration of all three is provided by what happens
when a motorist encounters a pedestrian about to cross the road.
At the macrosocial level, there are assumptions about the road,
the conventions of the highway code, and so on. At the biomechanical level, there are questions about what the motorist sees
within a certain range of visibility and the alertness necessary for
initiating the appropriate physiological actions for driving the
vehicle. At the circumstantial level, even when all of these aspects
have been taken into account, the individual motorist must constantly be prepared to modify any action taken, depending on the
moment-to-moment behavior of the pedestrian and other road
users.
This does not mean that the levels of integration are independent. Circumstantial factors intervene across the board. An
accident may be caused because the pedestrian is shortsighted
or the road badly lit. The drivers exercise of biomechanical skills
may be affected by knowing or not knowing that the brakes of
this particular vehicle are not very reliable.
Much recent work in integrational linguistics has focused on
analyzing how, at the macrosocial level, societies succeed in constructing supercategories, which integrate what would otherwise
be quite separate activities in such a way as to set up a common
framework for the intellectual and practical pursuits of society as
a whole. These supercategories include science, art, history, and
religion, each with a dedicated discourse of its own.
Roy Harris
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Harris, R. 1998. Introduction to Integrational Linguistics. Oxford:
Pergamon. A concise survey of the whole field.
Harris, R., and C. Hutton. 2007. Definition in Theory and Practice.
London: Continuum. An integrationist approach to problems of definition, with particular reference to lexicography and the law.
Love, N., ed. 2006. Language and History: Integrationist Perspectives.
London: Routledge. Collection of papers bearing on the construction
of the macrosocial supercategory history.
Toolan, M. 1996. Total Speech: An Integrational Linguistic Approach to
Language. Durham, NC: Duke University Press. Includes interesting
discussions of literal meaning, metaphor, and related issues.
395
396
For if (3) and (4) have the same meaning, then they are either
both true or both false. But (3) may be true while (4) is false for
example, if Lola does not realize that Mark Twain and Samuel
Clemens refer to the same man. Since (3) is just like (4) except for
an occurrence of Samuel Clemens instead of Mark Twain, Mark
Twain and Samuel Clemens differ in meaning.
Such are two arguments that Mark Twain and Samuel Clemens
differ in meaning. But if the extension of a proper noun is its
referent, then Mark Twain and Samuel Clemens have the same
extension. Thus, some (but not all) conclude, there is a distinction between two aspects of meaning: intension and extension.
397
Intentionality
Frege, Gottlob. [1892] 1997. On sense and reference. In The Frege
Reader, ed. M. Beaney, 15171. Oxford: Blackwell.
Gamut, L. T. F. 1991. Logic, Language and Meaning. Chicago: University
of Chicago Press.
Garcia-Carpintero, Manuel, and Josep Maci, eds. 2006. Two-Dimensional
Semantics. Oxford: Oxford University Press.
Heim, Irene, and Angelica Kratzer. 1998. Semantics in Generative
Grammar. Oxford: Blackwell.
Kaplan, David. 1989. Demonstratives. In Themes from Kaplan, ed. J.
Almog, J. Perry, and H. Wettstein, 481563. New York: Oxford University
Press.
Katz, J. J. 1990. The Metaphysics of Meaning. Cambridge, MA: MIT Press.
Kripke, Saul. Naming and Necessity. 1980. Cambridge: Harvard University
Press.
Larson, Rich, and Gabriel Segal. 1995. Knowledge of Meaning. Cambridge,
MA: MIT Press.
Mill, John Stuart. [1843] 1872. A System of Logic. 8th ed.
London: Longmans.
Montague, Richard. 1974. Formal Philosophy: Selected Papers of Richard
Montague, ed. Richmond Thomason. New Haven, CT: Yale University
Press.
Porphyry. 2003. Porphyrys Introduction. Trans. and commentary by
Jonathan Barnes. Oxford: Oxford University Press.
Putnam, Hilary. 1975. The meaning of meaning. In Philosophical
Papers. Vol. 2. Mind, Language and Reality. Cambridge: Cambridge
University Press.
Quine, W. V. 1953. Two dogmas of empiricism. Philosophical
Review 60: 2043. Reprinted in From a Logical Point of View, 2d ed.
(Cambridge: Harvard University Press, 1961).
. 1960. Word and Object. Cambridge, MA: MIT Press.
von Fintel, Kai, and Irene Heim. 2007. Intensional Semantics. Manuscript,
Massachusetts Institute of Technology.
INTENTIONALITY
Aboutness
The closest thing to a synonym for intentionality is aboutness;
something exhibits intentionality if and only if it is about something. The relevant sense of about is best elucidated by example: The sentence Saul Kripke is a philosopher is about Saul
Kripke; my belief that the weather in South Bend is dreary is about
the city of South Bend, Indiana; the black lines and curving blue
stripe on the map in my hand are about the streets of South Bend
and the St. Joseph River; the position of the needle on the gas
gauge in my car is about the amount of gasoline in its tank. While it
is difficult to find an uncontroversial and illuminating paraphrase
of the relevant sense of about, its hard to deny that there is some
reasonably clear sense of aboutness common to these examples.
This characterization of intentionality as aboutness is only
true to a first approximation because something can exhibit
intentionality without being about anything if it purports to be
about something. Zeus is not about, does not represent, anything; this name, unlike Saul Kripke, does not have a worldly
correlate. Nonetheless, Zeus counts as an example of intentionality by virtue of the fact that it (in a difficult to explain sense)
aims to be about something, even if it does not succeed.
398
Intentionality
distinctions between intentionality and intensionality, on the
one hand, and intentionality and intentions, on the other.
Intensionality is a property of sentence contexts. Given any
context in a sentence, we can then ask: Can we, by replacing
one expression or phrase in that context with another that has
the same reference, change the truth value of the sentences as a
whole? If so, then the context is said to be intensional.
So far, the connection between intentionality and intensionality may seem to be merely orthographic. But it has been claimed
that the latter is a criterion for the former: that descriptions of
intentional phenomena will always include an intensional context (Chisholm 1957). For descriptions of propositional attitudes like beliefs, this seems plausible. For example,
John believes that the worlds most famous sheep is famous.
is false, even if the worlds most famous sheep and Dolly the
sheep have the same reference. But the criterion seems to fare
less well in other cases. For example,
The thick blue line on my map of South Bend represents the
St. Joseph River.
appears to ascribe the right sort of aboutness to qualify as a sentence about intentionality, but the sentence does not contain
any intensional contexts. And many sentences that do contain
intensional contexts dont seem to be descriptions of intentional
phenomena. For example,
Mammals have a greater chance of heart failure than flatworms
because they are cordates.
and
Mammals have a greater chance of heart failure than flatworms
because they are renates.
differ in truth value, even though cordates and renates have the
same reference.
A second potential source of confusion is the similarity of
intention, and intentionality. Intention, like belief and
desire, is the name of a type of mental state. Like beliefs and
desires, intentions exhibit intentionality, but they are no more
essential to intentionality than other mental states.
399
Intentionality
using them (Grice 1957, 1969; Schiffer 1972, 1982). On this
view, what a speaker means by uttering an expression on an
occasion (speaker-meaning) is a function of the beliefs that
that speaker intends to bring about in his or her audience via
their recognition of that communicative intention, and, further, what an expression means in a community is a function
of what speakers mean or would mean by using the expression
on various occasions. By this two-part reduction (of expression-meaning to speaker-meaning, and speaker-meaning to
communicative intentions), the intentionality of language is
explained in terms of the intentionality of intentions. Critics
of this approach have focused on its inability to explain uses
of language in thought and apparently normal examples of
communication in which speakers lack the requisite communicative intentions (Chomsky 1975; Schiffer 1987). But despite
the problems faced by specific versions of this reductive program, there is widespread agreement that there is some way of
explaining the intentionality of language via the intentionality of the mental states of language users if not their intentions, then perhaps their beliefs (Lewis 1975). (Opposed views
of the source of the intentionality of language are defended in
Laurence 1996 and Brandom 1994.)
400
Intentionality
Byrne, A. 2001. Intentionalism defended. Philosophical Review
110: 199240.
Carnap, R. 1947. Meaning and Necessity: A Study in Semantics and Modal
Logic. Chicago: University of Chicago Press.
Chisholm, R. 1957. Perceiving: A Philosophical Study. Ithaca, NY: Cornell
University Press.
Chomsky, N. 1975. Reflections on Language. London: Temple-Smith.
Churchland, P. 1981. Eliminative materialism and the propositional
attitudes. Journal of Philosophy 78.2: 6790.
Crane, T. 1998. Intentionality as the mark of the mental. In Current Issues
in Philosophy of Mind, ed. A. OHear, 22951. Cambridge: Cambridge
University Press.
Cummins, R. 1989. Meaning and Mental Representation. Cambridge,
MA: MIT Press.
Dretske, F. 1981. Knowledge and the Flow of Information. Cambridge,
MA: MIT Press.
Fodor, J. 1975. The Language of Thought. Hassocks, UK: Harvester.
. 1987. Psychosemantics: The Problem of Meaning in the Philosophy
of Mind. Cambridge, MA: MIT Press.
. 1990. A theory of content, II: The theory. In A Theory of Content
and Other Essays, 89136. Cambridge, MA: MIT Press.
Frege, G. 1892. On sense and reference. In Translations from the
Philosophical Writings of Gottlob Frege, ed. P. Geach and M. Black,
5678. Oxford: Basil Blackwell.
. 1918. Thought. In The Frege Reader, ed. M. Beaney, 32545.
Oxford: Basil Blackwell.
Grice, P. 1957. Meaning. Philosophical Review 66.3: 17788.
. 1969. Utterers meaning and intentions. In Studies in the Way of
Words, 86116. Cambridge: Harvard University Press.
Harman, G. [1988] 1999. Wide functionalism. In Reasoning, Meaning,
and Mind, 23543. Oxford: Clarendon.
Husserl, E. [1901] 2002. Logical Investigations. 2 vols.
London: Routledge.
Kaplan, D. 1989. Demonstratives. In Themes from Kaplan, ed.
J. Almog, J. Perry, and H. Wettstein, 481563. Oxford: Oxford
University Press.
Kripke, S. 1972. Naming and Necessity. Cambridge: Harvard University
Press.
. 1982. Wittgenstein on Rules and Private Language: An Elementary
Exposition. Cambridge: Harvard University Press.
Laurence, S. 1996. A Chomskian alternative to convention-based semantics. Mind 105: 269301.
Lewis, D. 1975. Languages and language. In Language, Mind, and
Knowledge, ed. K. Gunderson, 335. Minneapolis: University of
Minnesota Press. Reprinted in Lewis 1983, 16388.
. 1983. Philosophical Papers. Vol. 1. Oxford: Oxford University
Press.
Millikan, R. 1989. Biosemantics. Journal of Philosophy 86.6: 28197.
Moran, D. 1996. Brentanos thesis. Proceedings of the Aristotelian
Society Supplementary Volume 70: 127.
Perler, D., ed. 2001. Ancient and Medieval Theories of Intentionality.
Boston: Brill.
Pietroski, P. 2003. The character of natural language semantics. In
Epistemology of Language, 21756. Oxford: Oxford University Press.
Quine, W. V. 1953. Two dogmas of empiricism. In From a Logical Point
of View, 2046. Cambridge: Harvard University Press.
. 1960. Word and Object. Cambridge, MA: MIT Press.
Rey, G. 1995. A not merely empirical argument for a language of
thought. Philosophical Perspectives 9: 20122.
Rorty, R. 1979. Philosophy and the Mirror of Nature. Princeton,
NJ: Princeton University Press.
Russell, B. 1905. On denoting. Mind 14: 47993.
Schiffer, S. 1972. Meaning. Oxford: Oxford University Press.
Internal Reconstruction
. 1982. Intention-based semantics. Notre Dame Journal of Formal
Logic 23: 11956.
. 1987. Remnants of Meaning. Cambridge, MA: MIT Press.
Searle, J. 1983. Intentionality. New York: Cambridge University Press.
Stalnaker, R. 1984. Inquiry. Cambridge, MA: MIT Press.
1990. Mental content and linguistic form. In Context and Content,
22540. New York: Oxford University Press.
Tye, M. 1995. Ten Problems of Consciousness: A Representational Theory
of the Phenomenal Mind. Cambridge, MA: MIT Press.
INTERNAL RECONSTRUCTION
Internal reconstruction (IR) is a method, or group of methods,
used to establish unattested earlier forms of languages. It differs
from the comparative method, which has similar aims, in
being based on the features of one language, without external
reference, and it is therefore appropriately applied to languages
without relatives, such as language isolates or reconstructed
protolanguages. Languages reconstructed by these means are
referred to as prelanguages, as opposed to the protolanguages
that result from the application of the comparative method.
IR arose in the later nineteenth century as an extension of
the comparative method. For example, Ferdinand de Saussures
reconstruction of sonant coefficients later termed laryngeals
for Indo-European was based largely on this method, though it
was not recognized as legitimate at the time, and his reconstructions were largely ignored. The method was identified in the early
twentieth century but was considered too speculative, methodologically unsound, and lacking controls on its results. It was not
until documentary evidence for laryngeals was found in Hittite
that its value was recognized, and then not universally.
IR is not a single method but a group of approaches having
in common reliance on internal features of the language. It is
possible to identify three such approaches, which are not necessarily completely distinct nor mutually exclusive: historical
morphophonemics, regularization of systems, and universal and
typological reconstruction.
Historical Morphophonemics
This approach, applicable only to phonology, is fairly well
defined methodologically. It is based on the structuralist
principle of the morphophoneme, which also corresponds to
the underlying phoneme of classical generative phonology.
Morphophonemes comprise a set of morphologically related
phonemes, that is, occurring in different forms of the same
morpheme. The principle is that the different phonemes result
from the application of different phonological changes in different contexts; by reducing these alternating phonemes to a single
form, we effectively establish the original phoneme from which
they are derived. For example, in Latin /s/ became /r/ between
vowels, giving alternations such as flos/flor-. Synchronically, this
provides us with a morphophoneme {S}; historically, we assume
that this corresponds to an original phoneme /s/. A weakness of
this method is that it assumes that there was no morphophonemic alternation in earlier stages of languages.
Regularization of Systems
While historical morphophonemics works by eliminating
alternations and apparent irregularities, the regularization of
401
Internal Reconstruction
systems method involves, in effect, the generalization of this
process to whole systems. Thus, where there are inconsistencies
in paradigms for example, with different classes of nouns or
verbs the method seeks to eliminate these differences in order
to produce a single paradigm, on the assumption that inconsistencies arise through change. For example, the different declensions or conjugations of Latin and Greek or the forms of Germanic
strong verbs can be reduced to single patterns. Again, a weakness
of this approach is that it assumes complete consistency in the
prelanguage.
402
more, bears on which of several possible interpretations is ultimately most satisfactory. Bush might be understood as having
advanced a moral principle that is implicitly qualified so that
there is no inconsistency with his military policies. Alternatively,
there might have been unnoticed inconsistency as all humans
exhibit some inconsistency. Finally, one might consider whether
Bush might have noticed an inconsistency that he conveniently
omits to mention.
The central point concerns how diverse information allows
one to decide among such alternative interpretations: It bears
on the explanation of what is said, done, and thought in the episode interpreted. The political concerns of Bush and his advisers might help to explain the assertion as political boilerplate.
Certainly the staking out of political positions, the responsiveness to constituencies, and the like can lead politicians to cast
about for simple (simplistic) formulations of sweeping principles
in which to wrap the desired policy. Is this the explanation, and
interpretation, of Bushs assertion? Understood and explained
as boilerplate, these questions about the exact content of the
asserted principle (whether or how it is implicitly qualified), the
way it then squares with various policies, and the obviousness of
inconsistency (if any) may become less pressing as they matter
less to the explanation of the episode. What is then central is the
sense of certain politicians for how taking the moral high road
will play in the relevant constituencies. But perhaps one senses
that Bush is a politician whose public face is here tied to his
own moral view of the world. Then, the degree of inconsistency
and how he could be insensitive to it become more central as a
matter of explanatory concern. In either case, one is informed
both by a general sense for what makes humans tick (cognitively and otherwise) and for an antecedently formed impression of this person in particular by the sedimented results of
past interpretations. This is to draw upon generic resources for
explaining human action and thought and on resources more
specific to this individual. Again, each is the fruit of past practice that is both interpretive and explanatory. One has a sense for
how humans think and act and for variations. One has a sense
for Bushs character as one variation, one that is the product
of his biography, and one that is evinced in past interpreted and
explained practice.
Just how one ends up interpreting the assertion in question
and the associated political or moral act depends on the explanation for this episode that one judges to be most likely, given all that
one knows about the agent or agents involved and about humans
as agents. Interpretation, it seems, is a matter of abduction or
inference to the best explanation of an ongoing, always revisable, sort. (Readers should find it easy to explore many further
alternatives and wrinkles in understanding the Bush case and
to assure themselves that their plausibility devolves onto the
question of greater or lesser explicability.)
The interdependence of interpretation and explanation, their
being two faces of the same coin, is widely appreciated. What is
contested is how to understand the explanation of thought and
deed. There are two broad schools of thought here, although representatives of each have been diverse.
Some think of explanation as subsumption as a matter of
deploying generalizations to show that the case or phenomenon
in question was dependent on certain antecedents. To conceive
403
404
Interpretive Community
. 1982. Tradition and modernity revisited. In Rationality and
Relativism, ed. Martin Hollis and Steven Lukes, 20160. Cambridge,
MA: MIT Press.
Leach, E. R. 1954. The Political Systems of Highland Burma. London: Bell.
Risjord, Mark. 2000. Woodcutters and Witchcraft. Albany: State University
of New York Press.
Salmon, Wesley. 1989. Four decades of scientific explanation. In
Scientific Explanation, Minnesota Studies in the Philosophy of Science.
Vol. 13. Ed. P. Kitcher and W. Salmon, 3219. Minneapolis: University
of Minnesota Press.
Stolberg, Sherly Gay. In Rare Threat, Bush Vows Veto of Stem Cell Bill.
New York Times, 21 May 2005. Stueber, Karsten. 2006. Rediscovering
Empathy. Cambridge, MA: MIT Press.
Turner, Stephen. 1980. Sociological Explanation as Translation.
Cambridge: Cambridge University Press.
van Fraassen, Bas. 1980. The Scientific Image. Oxford: Oxford University
Press.
Woodward, James. 2000. Explanations and invariance in the special
sciences. British Journal for the Philosophy of Science 51: 197254.
Intertextuality
same person writes two different texts for himself when reading
from different interpretive communities, for he understands the
Hebrew texts as prophesies of Jesus Christ only after his conversion at Damascus.
Fishs theory has been criticized for making words have no
meaning. He responds with just the opposite: Words always have
meaning, in fact many meanings, all of which are constructed by
situated readers in various communities. Fish adds that his theory is sociological, not normative, that is, it describes only what
people say (or think) a text means; it does not prescribe how
we ought to interpret texts. Finally, to the objection that some
authors use certain techniques to ensure that their texts convey
certain meanings, he responds that those meanings come to fruition only if the reader belongs to the same interpretive community as that author.
Jeffrey R. Wilson
WORK CITED
INTERPRETIVE COMMUNITY
In 1976, literary critic Stanley Fish used this term to describe
the unspoken (often unknown) alliances among readers who
share similar strategies for determining what a text means. This
theory of pragmatics, he says, is the explanation for both the
stability of interpretations among different readers (they belong
to the same community) and for the regularity with which a single reader will employ different interpretive strategies and thus
make different texts (he belongs to different communities) (Fish
1980, 171).
The notion of interpretive community insists upon the primacy of situated readers, and it can be thought of as a theory
of creative reading. Fish says that a set of general assumptions
on how one ought to interpret a text precedes every act of interpretation; thus, a reader always perceives a given text within an
already in-place hermeneutical framework. One does not read
the words on a page and then decide what those words mean
because no temporal separation exists between acts of perception and interpretation. Instead, ones community conditions
how its members read those words in the first place. As such,
readers actually write a text for themselves as they read, for they
have a tool kit of interpretive strategies always at work determining what certain words will mean should they arise in a given
context. Readers using the same tool kit belong to the same
community.
One can see interpretive communities at work in Christian
typology, a mode of biblical exegesis that aims to square Old
Testament texts with the events recounted in the New Testament.
For the typologist, the belief that Jesus was God combines with
other assumptions in order to form the exegetes set of interpretive strategies. Other readers who share these strategies make up
this exegetes community even if they do not know one another,
which explains how two Christians might independently interpret some events in the Old Testament as prophesies of Jesus
Christ. Of course, a Jew, Gnostic, or pagan produces a much different meaning of those same Hebrew texts because he or she
works from a community that reads/writes those texts differently. And finally, a look at Paul of Tarsus demonstrates how the
Fish, Stanley Eugene. 1980. Is There a Text in This Class? The Authority of
Interpretive Communities. Cambridge: Harvard University Press.
INTERTEXTUALITY
Building on Mikhail Bakhtins (1981) discussion of the dialogic
nature of language, Julia Kristeva (1986) coined the term intertextuality for the multiple ways in which texts refer to and draw
on other texts. This notion highlights the interconnectedness of
texts and challenges deep-rooted literary values, such as autonomy, uniqueness, and originality (Allen 2000, 56). An intertextual perspective views text production as a social practice in
which different texts, genres, and discourses are drawn upon and
text consumption as a process in which readers may bring additional texts not only those that have shaped production into
the interpretation process (Fairclough 1992, 845). The study
of intertextuality does not focus solely (or even primarily) on
the specific prior texts that are brought into play in a given text;
rather, it also examines the implicit texts underlying production
and interpretation (e.g., presuppositions, genre conventions)
(Culler 1976, 1388). Thus, a newspaper crime report has intertextual links not only to eyewitnesses accounts and previous
reports on the same and/or similar events but also to newswriting conventions, propositions that the journalist takes as given,
and even the journalists/readers understanding of crimes in
general.
Reported speech, a prime example of intertextuality, has been
extensively studied in sociolinguistics. Reporting speech is
always a reformulation of the original act. Even if prior speech is
reported verbatim, the reporting speaker may use prosodic features like stress and intonation to indicate his/her interpretation of the utterance, or he/she may frame the reported speech
in such a way as to manipulate the addressees perception of the
reported speaker. In some cases, material represented as reported
speech is not spoken by anyone at all. These observations have
led Deborah Tannen (1989) to conclude that reported speech is
primarily the creation of the reporting speaker and serves to create a sense of interpersonal involvement between the reporting
speaker and the addressee in the reporting context.
405
Intertextuality
Reported speech is an example of what Norman Fairclough
(1992) calls manifest intertextuality. Manifest intertextuality
refers to the way in which specific other texts are overtly drawn
upon within a text. In addition to reported speech, it covers such
phenomena as irony, negation, and presupposition. In
contrast, constitutive intertextuality also known as interdiscursivity refers to the way in which texts draw on abstract sets of
conventions like genres and styles. In their research on interdiscursivity, several critical discourse analysts have
noticed a widespread appropriation of conversational styles in
public discourse. Focusing on a consumer feature about lipstick
from a British teenage magazine, Mary Talbot (1995) examines
how the text producer exploits features of conversational speech
(e.g., the pronoun you, expressive punctuation) to establish an
informal, friendly relationship with the reader. This practice,
however, is far from benign. Under the guise of offering sisterly advice, the consumer feature serves as covert advertising
and trains teenage readers to become consumers of cosmetic
products.
Generic intertextuality, a notion developed by anthropologists
Richard Bauman and Charles Briggs, can be viewed as a particular kind of interdiscursivity. They define generic intertextuality as
the construction of the relationship between a text and a genre,
and they are interested in how and for what purposes this relationship is established in communicative practice (Briggs and
Bauman 1992). They see genre as a speech style that serves as an
orienting schema for the production and reception of discourse.
Genre interacts with such factors as the interactional context and
the speakers/writers communicative goals in shaping a given
text. In turn, these factors may lead to the selective adoption of
the constituent features of the generic framework and create an
intertextual gap between the text and its generic model. Briggs
and Bauman argue that research on strategies for manipulating
intertextual gaps between texts and their generic schemas can
shed light on issues of power, ideology, and political economy.
In cultures with traditional genres that are invested with great
power, speakers/writers often minimize the distance between
their texts and these genres. This serves as a powerful strategy
for creating textual authority. At times, however, speakers/writers may maximize the intertextual gaps between texts and their
generic models. They may do so to resist the hegemony of established genres or to claim authority in cases where creativity is
highly valued.
Appropriation, a specific case of intertextuality, refers to the
practice of adopting words, expressions, or ways of speaking
that are generally thought to belong to someone else. Many
white American teenagers use elements of African-American
vernacular English (AAVE) in their speech so as to align themselves with hip-hop culture and/or to project an urban youth
identity by exploiting certain connotations of AAVE (e.g., toughness). Appropriation, however, may also serve disaffiliating or
even denigrating purposes. In the United States, some monolingual Anglos use what anthropologist Jane Hill (1993) call Mock
Spanish that is, a subregister of colloquial English made
up of (pseudo-)Spanish expressions (e.g., maana) to project a congenial persona. Yet to make sense of Mock Spanish,
one also requires access to certain racist beliefs about Spanish
speakers. Maana works as a humorous substitute for later
406
Intonation
only because of the stereotype of Spanish speakers as lazy
and procrastinating.
Several issues continue to dominate sociolinguistic research
on intertextuality. One, to which I have alluded, focuses on
the relation of intertextuality to power. Others are concerned
with the conditions that make decontextualization and recontextualization possible, as well as the semantic and functional
changes that texts undergo as a result of recontextualization.
Intertextuality also raises interesting issues about authorship.
If all texts are created out of prior texts and conventions, what
is an author, and who is responsible for what is said/written?
These issues are likely to be worked out differently in different
cultures.
Andrew Wong
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Allen, Graham. 2000. Intertextuality. London: Routledge.
Bakhtin, Mikhail. 1981. The Dialogic Imagination. Ed. Michael Holquist,
trans. Caryl Emerson and Michael Holquist. Austin: University of Texas
Press.
Bauman, Richard. 2004. A World of Others Words: Cross-Cultural
Perspectives on Intertextuality. Malden, MA: Blackwell.
Briggs, Charles, and Richard Bauman. 1992. Genre, intertextuality, and
social power. Journal of Linguistic Anthropology 2: 13172.
Culler, Jonathan. 1976. Presupposition and intertextuality.
MLN: Modern Language Notes 91: 138096.
Fairclough, Norman. 1992. Discourse and Social Change. Cambridge,
UK: Polity.
Hill, Jane. 1993. Hasta la vista, baby: Anglo Spanish in the American
Southwest. Critique of Anthropology 13: 14576.
Kristeva, Julia. 1986. Word, dialogue, and the novel. In The Kristeva
Reader, ed. Toril Moi, 3461. New York: Columbia University Press.
Talbot, Mary. 1995. A synthetic sisterhood: False friends in a teenage
magazine. In Gender Articulated, ed. Kira Hall and Mary Bucholtz,
14365. Routledge: New York.
Tannen, Deborah. 1989. Talking Voices. New York: Oxford University
Press.
INTONATION
this term refers to the fundamental frequency (or its perceptual correlate, pitch) contour associated with phrases and other
large prosodic units. Language communities use intonation to
serve a wide range of functions, both grammatical and discourse
based. For example, intonation is used to signal prosodic boundaries: The ends of utterances are characteristically associated
with terminal pitch excursions, either a rise or a fall depending
on semantic factors. Another function of intonation is to cue
many types of semantic distinctions, such as the difference
between yes/no questions and neutral declarative statements
in many languages. Intonation is also used to convey emotional
(see emotion and language) and expressive states, as well as
pragmatic information.
Intonation is a universal property in that speakers of all languages manipulate pitch to communicate linguistic and paralinguistic functions (see paralanguage and phonology,
universals of). Even languages that use pitch to differentiate
individual lexical items, for example, tone languages such as
Mandarin Chinese and pitch accent languages such as Swedish,
Intonation
also have intonation systems that are evident when words are
grouped into larger prosodic constituents or uttered in isolation.
Intonation systems may vary substantially, however, from language to language and also potentially from speaker to speaker.
Thus, while a yes/no question is associated with a terminal rise in
pitch in many languages (e.g., German, Japanese, and Korean),
there are other languages (e.g., Finnish and Chickasaw) that
mark yes/no questions with a final pitch fall. One nearly universal property, however, is the lowered pitch characterizing the
end of semantically neutral declarative utterances.
The study of intonation has witnessed many important theoretical advances over the last 30 years. Whereas certain schools
of intonation (e.g., the British school) described pitch contours
in terms of their overall shape or gestalt, many current models
of intonation capture changes in pitch in terms of discrete tonal
sequences, thereby bringing the study of intonation into line with
the analysis of segmental and word-level phonological phenomena. This type of approach assumes that peaks and troughs in a
fundamental frequency contour are attributed to phonological
high and low tones aligned with various phonological elements.
Actual surface phonetic intonation patterns result from interpolation between these high and low tonal targets.
As one of the pioneering works in this school of intonation,
Janet Pierrehumbert (1980) developed an analysis of English
intonation in which a wide range of pitch contours conveying
several different semantic and pragmatic functions are captured
as a sequence of phonological high and low tones associated
with hierarchically arranged prosodic constituents. A fundamental insight of Pierrehumberts analysis is that tones may be
classified into two groups: those that are associated with prominent, that is, stressed, syllables and those that are associated with the periphery, especially the end, of prosodic domains.
Tones that are associated with stressed syllables are termed pitch
accents. Pitch accents differ according to whether they consist of
a single tone, high or low, or a sequence of tones, such as H + L
or L + H, which phonetically yield a tone fall or rise, respectively.
In addition to pitch accents on certain stressed syllables, tonal
excursions are often observed at the end (and potentially the
beginning) of phrases. These phrase-level tonal movements are
attributed to boundary tones, which may be associated with relatively small phrases, termed intermediate phrases, or with larger
phrases, termed intonation phrases or intonation units. Like
pitch accents, boundaries may be characterized by a single tone
or a sequence of tones. For example, intonation phrase boundaries in Korean consist of as many as five tonal targets, for example,
LHLHL, which conveys a sense of annoyance on the part of the
speaker (Jun 2005a).
One of the challenges facing linguists interested in the
typological investigation of intonation is the relative dearth
of reliable descriptions of intonation on a broad cross section of languages. Fortunately, recent years have witnessed
a dramatic expansion of cross-linguistic studies intonation.
Daniel Hirst and Albert Di Cristo (1998) studied intonation
in 20 languages, including several non-Indo-European languages. Jun (2005a) compiled investigations of 13 languages
all analyzed within a Pierrehumbert-type framework. A typologically and geographically diverse array of languages is discussed in this work, including languages with word-level stress
Irony
(Dutch, English, and German), tone languages (Cantonese and
Mandarin Chinese), pitch accent languages (Japanese, Swedish,
and Serbo-Croatian), languages lacking word-level stress, lexical tones, or pitch accents (Korean and French), and indigenous languages of North America and Australia (Chickasaw and
Bininj Gun-wok, respectively).
An issue common to the intonation systems of all languages
is the mapping between intonational tunes and their meanings.
Because intonation is used to convey many subtle differences in
meaning, often in gradient fashion, it is a challenge to determine
which differences in intonation merit different phonological analyses. Phonological distinctions between intonational tunes must
be captured as differences either in the sequence of tones comprising the tunes or in the alignment of those tones with words.
Yet another topic inextricably linked to intonation is prosodic constituency. Since pitch excursions are often observed
at prosodic boundaries, a comprehensive analysis of intonation hinges on the characterization of the types of constituents
constituting utterances. Research has shown that prosodic constituency and the mapping between constituency and intonation
vary from language to language. For example, some languages
divide utterances into groupings of words that are characterized
by a tonal template. Thus, phrases in Chickasaw (Gordon 2005)
are associated with a LHHL sequence, whereby the first and the
last low tonal targets associate with the beginning and the end of
phrases, respectively, and the two high tones associate with the
second syllable (or the first syllable if it contains a long vowel or
ends in a sonorant consonant) and the beginning of the final syllable, respectively.
A number of researchers have published books providing
overviews of these and other issues in the study of intonation,
including Ladd (1996), Cruttenden (1997), and Gussenhoven
(2004).
Matthew Gordon
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Cruttenden, Alan. 1997. Intonation. Cambridge: Cambridge University
Press.
Gordon, Matthew. 2005. Intonational phonology of Chickasaw. In Jun
2005b, 30130.
Gussenhoven, Carlos. 2004. The Phonology of Tone and Intonation.
Cambridge: Cambridge University Press.
Hirst, Daniel, and Albert Di Cristo, eds. 1998. Intonation Systems: A Survey
of Twenty Languages. Cambridge: Cambridge University Press.
Jun, Sun-Ah. 2005a. Korean intonational phonology and prosodic transcription. In Jun 2005b, 20129.
Jun, Sun-Ah, ed. 2005b. Prosodic Typology the Phonology of Intonation
and Phrasing. New York: Oxford University Press.
Ladd, D. R. 1996. Intonational Phonology. Cambridge: Cambridge
University Press.
Pierrehumbert, Janet. 1980. The phonology and phonetics of English
intonation. Ph.D diss., Massachusetts Institute of Technology.
Reproduced by the Indiana University Linguistics Club, Bloomington,
1987.
IRONY
There are several types of irony described in the literature, all of
which rely on an incongruity or discrepancy between appearance
407
Irony
and reality. In dramatic irony, for instance, the incongruity is created by having the audience aware of information about which a
character in a play is ignorant (such as in Sophocles Oedipus Rex).
In situational irony and irony of fate, the disconnect is between
ideal expectations of justice and actual (or idealized) outcomes,
such as would occur if Bill Gates, one of the worlds most wealthy
individuals, won a lottery or as exemplified by Beethovens inability, on going deaf, to hear his own musical masterpiece.
The form of irony most studied in the language sciences
is verbal irony, traditionally conceptualized as a trope in
which the meaning one intends to communicate is opposite
of that expressed literally by the words that are used. Thus,
in Shakespeares Julius Caesar, when in his famous soliloquy (Act 3, Scene 2) Anthony states: For Brutus is an honorable man / So are they all, honorable men, it is understood
that Anthony is emphasizing the opposite of being honorable,
namely, the dishonorable action of Brutus and the other conspirators. In principle, with irony one should be able to convey
negative attitude by expressing something positive (as in the
Shakespearean example above) or positive attitude by stating
something negative.
There is also a general recognition that because the expressed
utterances are literally plausible, the recovery of ironic intent is
consequently highly context dependent and facilitated by signals fashioned (or unintentionally employed) by the ironist.
Tone of voice is one such hint in spoken language, but because
irony can be detected even when ironic intonation is not
employed (such as when reading text), this cue is not necessary.
Other cues include hyperbole, understatement, and excessive
politeness, but it is generally agreed that there is no signal that
points exclusively to irony.
The context necessary for the recovery of ironic intent traditionally has been limited to the discourse in which the irony is
embedded, but in more recent years, the concept of context has
been widened, even for verbal irony, to encompass an ironic
environment that includes social-cultural factors, such as those
dependent on discursive communities that share knowledge,
beliefs, values, and communicative strategies (e.g., Hutcheon
1994).
Over the past three decades, the standard pragmatic
approach to nonliteral language processing (see metaphor )
has framed much of the discussion on irony. Based on the pragmatics of conversation and on speech-act theory, irony is
described accordingly as the outcome of a conversational
implicature initiated by a violation, or based on exploitation, of H. P. Grices (1975) Maxim of Quality: The recognition
that the literal expression does not make sense in the context
in which it is produced leads one to initiate a search for a context-appropriate interpretation in which the literal sense of the
expression is denied, suppressed, and replaced by the logical
opposite.
There have been challenges over the years regarding the traditional emphasis on the initial processing of the literal expression and then substituting it with its opposite meaning, on both
logical and empirical grounds. Consider, for instance, the analysis of a passage from Voltaires Candide given by Dan Sperber
and Deirdre Wilson: When all was over and the rival kings were
celebrating their victory with Te Deums in the respective camps
408
L
LANGUAGE, NATURAL AND SYMBOLIC
The paradigm of language is natural language, a naturally evolved
system of human communication using spoken or signed words
according to the various ways they can be combined. But in an
extended sense, many species of nonhuman animals also have
language, in this case, a means of communication through inarticulate sounds; and humans similarly have various nonverbal means of expression and communication through facial
expressions, gestures, and body language more generally, and
through music, dance, and the imitative arts (see animal com-
409
410
the child as language learner in the initial state, that is, before
language experience.
These components, which were proposed as necessary for
language learning, were formulated as accounting for the mapping from primary language data (data which is necessarily
finite and inherently variable) presented at the initial state to
complex, infinite, and systematic language knowledge in the end
state. They characterized the foundations of the specific innate
abilities that make this achievement possible for the language
learner (Chomsky 1965, 27).
At the same time, the components in Figure 1 were to indicate
what a linguistic theory would have to treat in order to support
an acquisition model and thus attain explanatory adequacy
(see descriptive, observational, and explanatory
adequacy). For example, such a theory would need to provide a representation of possible sentence in order to support
(i), a definition of structural description for (ii), generative
grammar for (iii), a method for determining structural descriptions for (iv), and an evaluation metric for (v).
The original Chomsky proposal is often interpreted as though
it described a realistic device. However, it is best viewed as providing an explication of the logical foundations that any comprehensive model for language acquisition would have to presuppose
and account for, with the precise nature of each of the components to be subsequently determined as an empirical matter
(Chomsky 1965, 37). Perhaps at least partially because of this
divergence in interpretation, subsequent language acquisition
research over the last decades has frequently pursued divergent
paths, one pursuing the logical problem of language acquisition consisting of linguistic analyses of potential data mapping
(e.g., Baker and McCarthy 1981), another pursuing a realistic
approach consisting of empirical studies of language development (initiated largely by Roger Brown and his students [1973]).
In fact, certain components in Figure 1 proved particularly
challenging to a realistic model. For example, if there were a realistic device, it would have to include a mechanism for (i) and (ii).
If (iii) were to be implemented in terms of an enumeration of
the class G1, G2 of possible generative grammars (Chomsky
1965, 31), then there is the de facto impossibility of predefining
given grammatical hypotheses across 6,000 to 7,000 actually
potentially innumerable human languages, as well as the risk
of begging the question of language acquisition, that is, how the
individual grammars arise. Similarly, the nature of an evaluation
411
412
Sound Change
The symmetry of the terms used to describe sound changes
assimilation/dissimilation, weakening/strengthening, deletion/
insertion makes it seem as though all directions of change were
equally probable. The facts now known from a wide array of languages show instead that assimilation, weakening, and deletion
are vastly more common than dissimilation, strengthening, and
insertion. If sound change is rigidly defined as language internal,
phonetically motivated, and lexically general, then the less common types of change show up even more rarely.
Common assimilation changes involve consonants taking
on the properties of adjacent vowels, for instance, palatalization before front vowels, labialization or rounding before round
vowels. Or vowels can take on properties of consonants, as
when vowels become nasalized in the same syllable with nasal
consonants. For some of these changes, more detailed hierarchies of contexts for change can be established. Palatalization
of coronal consonants (such as [t], [d], [s]) occurs first and most
commonly before a high front glide ([j]), occurs next before a
high front vowel ([i]), and then progresses to the lower front
vowels.
Paths of reductive change can also be established. For
instance, a common consonant reduction involves the loss of [p]
via the pathway p > f > h > . Parts of this path are documented
in different languages: Japanese has undergone a change that
reduced all prevocalic instances of /p/ to a fricative that assimilates to the place of articulation of the following vowel. Spanish
and other Romance languages have undergone a change that
reduced word-initial [f] to [h] and further to . The positions in
which such reductions occur also show regularity: Syllable-final
position favors the reduction of consonants, as does intervocalic
position. Word-initial position is least likely to condition reduction of consonants.
It is commonly assumed that such changes are caused by
phonetic tendencies. In the articulatory domain, some general
principles of change have been proposed that deal with the way
articulatory gestures change during production. The primary
directions in phonetic change are the reduction of the magnitude
of the gestures, leading to reduction or loss, and their increased
overlap, leading to assimilation (Browman and Goldstein 1992;
Mowrey and Pagliuca 1995).
As mentioned earlier, the importance of universals of language
change is that they can predict and thus explain synchronic patterns across languages. For instance, the presence of nasal vowels
in a language is almost always due to assimilation to a nasal consonant. Sometimes this consonant is lost, leading to phonemic nasal
vowels. The diachronic source explains why nasal vowels always
have a more restricted distribution and frequency compared to
oral vowels (Greenberg 1966). Similarly, the fact that some languages lack the phoneme [p] can be explained by its tendency to
undergo reduction. The restrictions against certain consonants
in syllable-final position can be attributed to their propensity for
loss in that position (Vennemann 1988, Blevins 2004).
413
Language Families
Maczak, Witold. 1980. Laws of analogy. In Historical Morphology ed.
J. Fisiak, 2838. Berlin: Mouton de Gruyter.
Mowrey, Richard, and William Pagliuca. 1995. The reductive character
of articulatory evolution. Rivista di Linguistica 7.1: 37124.
Vennemann, Theo. 1988. Preference Laws for Syllable Structure and the
Explanation of Sound Change. Berlin: Mouton de Gruyter.
LANGUAGE FAMILIES
A language family is a set of languages that developed from the
same ancestral language. The best-known example is the IndoEuropean family, which comprises more than a hundred languages that, even in premodern times, extended from the Indian
subcontinent to northwestern Europe. This family is well known
not only because it contains many of the worlds most widely
spoken languages, such as Bengali, English, French, German,
Hindi, Portuguese, and Russian, but also because it was the main
focus of research in the nineteenth century, when linguistics
was established as a modern science. However, many other language families are subjects of intense research today, such as the
following:
Afro-Asiatic, which includes Arabic, Hebrew, and several
languages of northern Africa, including Ancient Egyptian,
Hausa, and Somali
Algic, which includes several native languages of North
America, such as Wiyot, Yurok, Cheyenne, Ojibwa, and
Shawnee
Austronesian, which includes more than a thousand languages spoken from Madagascar to Polynesia, such as Bahasa
Indonesia, Fijian, Tagalog, and Tahitian
Dravidian, which includes most of the non-Indo-European
languages of India, such as Malayalam, Tamil, and Telugu
Niger-Congo, which includes most of the languages of subSaharan Africa, including Igbo, Swahili, and Zulu
Sino-Tibetan, which includes Burmese, Chinese, and Tibetan
Tupi, comprising several languages of South America, including Guarani
Uralic, which includes most of the non-Indo-European languages of Europe, such as Estonian, Finnish, Hungarian,
Nenets, and Sami (Lapp)
This list is only a sample of the hundreds of known language
families and of the languages included in each one. For a comprehensive listing, see Gordon (2005).
Familial metaphors are the standard terms of art. Languages in
the same family are said to be genetically related to one another.
A language from which other languages developed is called an
ancestor, or parent, of those languages. Words that descend from
the same form in an ancestral language are related, or cognate.
This homey terminology is undergoing some competition with
that of modern cladistics as used in biology, but certain linguistic
concepts do not translate well. In biology, it is understood that
all biological taxa are related to one another, and family is but
a midlevel taxon. Linguists assume that relationships between
languages must be proved, and a language family is a maximal
taxon. In principle, even isolates languages that do not group
with other languages can trivially be considered families by
reinterpreting their dialects as separate languages.
414
Traditional Methods
The traditional technique is the comparative method. The
linguist studies characteristics that rarely recur across languages,
such as grammatical paradigms and the associations between
sound and meaning in morphemes. Efforts are made to discard
loans and onomatopoeia, although the former is a difficult and
often intractable problem. Matching morphemes across languages by meaning, one looks for recurrent sound correspondences. For example, English f corresponds to Latin p in father
= pater, feels = palpat, few = pauca, and many other words. If a
large number of recurrent correspondences are found, the languages are related. The recurrences are also used to reconstruct
the protolanguage (see historical reconstruction).
After a language family is identified, the next step is subgrouping, identifying the branches or groups within the family. Subgrouping seeks to uncover the history of the divergence
(cladogenesis) of a language family. If the family contains three
or more languages, the linguist looks for evidence that some
proper subset of those languages may have descended from
an intermediate common ancestor. This is done by looking for
shared innovations (synapomorphies) sound changes or new
words or grammatical constructions that were not in the ancestor language but are found in two or more of the descendant
languages. For example, the fact that English, German, Swedish,
Language Families
and several other languages have f where Proto-Indo-European
had p is a shared innovation that indicates that those languages
may have a shared intermediate ancestor that underwent this
change; otherwise, we would have to assume that each of those
languages separately innovated the change of p to f or borrowed
the innovation from another language. In fact, the preponderance of evidence supports such an intermediate language and
a branch (clade) of languages descending from it: the Germanic
languages. Other branches of Indo-European include the BaltoSlavic (including Bosnian, Lithuanian, Polish, and Russian),
Celtic (including Breton, Irish, and Welsh), Italic (including Latin
and the Romance languages), Indo-Iranian (including Bengali,
Farsi, Pashto, and Urdu), and the extinct Anatolian branch,
which included Hittite, Luvian, and Lycian. Several other languages, including Greek, Albanian, Armenian, and half a dozen
extinct languages (see extinction of languages), do not
share an agreed-upon intermediate branch at all.
Associating language history with external facts entails pinning
the protolanguage to a particular time and place its homeland
and demonstrating how it spread from there. The time depths
under consideration mean that written records are rarely, if ever,
available. The primary linguistic tool is to look for words found
in multiple branches of a language family and exhibiting all of
the regular sound correspondences; they are assumed to date
back to the protolanguage and, therefore, to name objects found
in its environment. For example, a pan-Indo-European word for
wheel suggests that the protolanguage split up no earlier than
the invention of the wheel, some six thousand years ago (Mallory
1989). Another technique is to look for areas of greatest linguistic diversity. The fact that the Austronesian languages are much
more diverse in Taiwan than anywhere else supports the theory
that they developed there longest, that is, that Taiwan was the
homeland for the Austronesian family (Blust 1999). A third technique is to seek archaeological evidence of population movements that may have disseminated a language family. In the
case of Austronesian, knowledge of how people spread through
the Pacific and Indian Oceans is consistent with the theory of
a Taiwan homeland. In the case of Indo-European, it has often
been noted that early adopters of horse-drawn wheeled chariots
would be in an ideal position to spread their languages throughout much of Europe and Asia. A well-received theory points to
the chariot users who lived in the PonticCaspian region about
six thousand years ago (Mallory 1989).
415
Language Families
sound correspondences. This has the advantages both of being
closer to the traditional comparative method and of generating
correspondences useful for subgrouping and reconstruction.
Disappointingly, however, none of these neo-Greenbergian
techniques found evidence for the deep relations that were
advertised for the original, impressionistic method.
Other new techniques have concentrated on subgrouping.
Lexicostatistics (Swadesh 1955) was an early attempt to facilitate
subgrouping and also assign dates to protolanguages. The idea
was that if languages replace a constant number of words per
century with new words, then by measuring the percentage of
a list of words that is cognate between languages, one could calculate when the languages diverged and even construct a family
tree. Although these assumptions were mostly wrong and were
therefore rejected by most linguists, many people still use lexicostatistical techniques as a rough indication of a languages history in the absence of more compelling data.
Arranging many shared innovations into a binary tree is an
extremely laborious undertaking, especially given the possibility that some identical innovations are independent (homoplastic). The recent development of computational cladistic
methods similar to those used in biology (e.g., Ringe, Warnow,
and Taylor 2002) is a tremendous advance in helping the linguist find optimal trees. In addition, several solutions to the
problem of borrowing have emerged in the form of programs
that construct networks instead of trees. Shared innovations
that cannot be cleanly attributed to a shared ancestor are taken
as evidence of contact, obviating somewhat the need to make a
priori judgments about whether borrowing was involved (e.g.,
Bryant, Filimon, and Gray 2005; Nakhleh, Ringe, and Warnow
2005).
The problems of finding homelands and tracing the spread
of languages still requires one to resort to data that are often
suggestive but not definitive. Colin Renfrew (1987) added a new
perspective when he theorized that languages may be spread
by the movement of culture, rather than by the movement of
people. He suggested that Indo-European languages were
spread from Anatolia along with the adoption of agriculture.
Most linguists have not accepted this theory, in part because it
is incompatible with such linguistic data as an Indo-European
word for the wheel, which postdates the spread of agriculture
by millennia. Recently, further data are afforded by genetic
analyses of populations (genes and language). The presence of a Pontic genetic component in Europe is compatible
with the idea that invaders from the PonticCaspian region
brought Indo-European languages into Europe (Cavalli-Sforza,
Menozzi, and Piazza 1994).
Prospects
Recent computer techniques add simplicity, reproducibility,
and quantitative rigor to methodologies for proving relationships between languages, but so far there has been no noticeable increase in power over what experts are able to do by hand.
Failure to corroborate the sort of deep relationships conceived
by Greenberg may mean that better techniques need to be developed, or that the languages are not in fact related, or that the
answer is unknowable. Because languages are always changing,
they constantly lose information that links them to their relatives;
416
Language-Game
Oswalt, Robert L., 1998. A probabilistic evaluation of North Eurasiatic
Nostratic. In Nostratic: Sifting the Evidence, ed. J. C. Salmons and B. D.
Joseph, 199216. Amsterdam: Benjamins.
Renfrew, Colin. 1987. Archaeology and Language: The Puzzle of IndoEuropean Origins. London: Pimlico.
Ringe, Don A., Jr. 1992. On Calculating the Factor of Chance in Language
Comparison. Philadelphia: American Philosophical Society.
Ringe, Don, Tandy Warnow, and A. Taylor. 2002. Indo-European and
computational cladistics. Transactions of the Philological Society
100: 59129.
Swadesh, Morris. 1955. Towards greater accuracy in lexicostatistic dating. International Journal of American Linguistics 21: 12137.
Thomason, Sarah Grey, and Terrence Kaufman. 1988. Language Contact,
Creolization, and Genetic Linguistics. Berkeley: University of California
Press.
LANGUAGE-GAME
At the beginning of the Philosophical Investigations, Ludwig
Wittgenstein questions the Augustinian or traditional picture of
the essence of human language, according to which the meaning
of a word is the object for which it stands, so that a word to which
no object corresponds has no meaning. Learning language consists in giving names to objects, and the association between
the word and the object is established by ostensive teaching of
words. Wittgenstein attacks this conception as both confused
and reductive. It makes naming seem like a queer connection of
a word with an object, a kind of baptism of the object, as if meaning existed separately from the word and was attached to it by
a mental process: a remarkable act of mind or occult process.
Moreover, making the correspondence between name and object
a condition of meaning has the absurd consequence that when
the object no longer exists, the word no longer has any meaning.
Here, the meaning of a name is confounded with the bearer of
the name, whereas in fact when Mr. N. N. dies, one says that the
bearer of the name dies, not that the meaning dies (Wittgenstein
[1953] 1997, 40). The Augustinian picture of language is also
reductive. Even if we were to correct its conception of ostensive
definition (e.g., by replacing the occult process with training), it
would still err in being an oversimplification: Ostensive definition is too narrow; it does not describe everything that we call
language. We think that language consists in giving names to
objects, whereas in fact we do the most various things with our
sentences: Think of exclamations alone, with their completely
different functions. Water! Away! Ow! Help! Fine! No!
([1953] 1997, 27).
The functions of words, suggests Wittgenstein, are as diverse
as the functions of tools in a toolbox ([1953] 1997, 11). And
these are countless. But his toolbox analogy does more than
underscore the diversity of the uses of language; it also suggests
that words work much like tools in that their meaning resides in
their use: We remain unconscious of the prodigious diversity
of all the everyday language-games because the clothing of our
language makes everything alike ([1953] 1997, 224) and so it
is use that is the distinguishing mark. In reaction to the mentalist conception of meaning, which sees it unilaterally as a mental
connection between words and objects, Wittgenstein affirms
that the meaning of a word or sentence resides in the use we
make of it. He introduces the term language-game to highlight
the interplay, in the determination of meaning, between language and the actions into which it is woven and to bring into
prominence the fact that the speaking of language is part of
an activity, or of a form of life ([1953] 1997, 23; see forms of
life ).
In fact, Wittgenstein nowhere provides a well-rounded definition of language-game because his employment of the term
evolves and because it is what he calls a family resemblance
concept. He employs the term
i. to circumscribe various more or less broad domains of language. Here, we can speak of single and specific language-games,
such as those we play in our use of particular words or concepts (e.g., fear, game, hand, knowing) or in specific
activities (e.g., lying, thanking, cursing, making a joke,
following a rule, giving orders and obeying them); but these
specific language-games are subsumed under more general uses
of the term either a generic use: he calls the whole, consisting of
language and the actions into which it is woven, the languagegame ([1953] 1997, 7) or an anthropological use, that is, what
he calls the human language-game (1977, 554) (as opposed, say,
to the language-game of alien tribes).
ii. to describe the degree of sophistication of language use.
Here, Wittgenstein speaks of primitive or complicated languagegames.
iii. to describe how language works.
It is in its encapsulation of how language works that the
expression language-game is most eloquent. The expression
is due to the analogy Wittgenstein makes between language and
games, which supersedes the calculus analogy of the Tractatus,
thereby signaling the switch from his own conception of language
as a fixed and timeless symbolism to a conception of language as
essentially embedded in human practice language as essentially in use. In the Blue Book, he begins to question the idea that
speaking a language is, in all cases, to apply a calculus according
to strict and exact rules; rather, using language is much like playing a game. The game analogy is more fitting than the calculus
analogy in several ways:
1. Like game, language is a family resemblance concept. Just
as there is a multiplicity of games with nothing common to all,
there is nothing common to all of our uses of language that
makes them into language or parts of language, but they are
related to one another in many different ways, and it is because
of this relationship that we call them all language ([1953]
1997, 65).
2. Language is an activity, and it is essentially connected with
practice or use: our language-game is a piece of behaviour
(1980, 151); a language-game incorporates both language and
the actions into which it is woven ([1953] 1997, 7). To use language meaningfully is analogous to making a move in a game;
to understand a word is to know how to use it. Just as we learn
how to play a game by learning what the permitted moves are
in the game, we learn the meaning of words by learning what
is accepted as a meaningful use of the word. And here, it is not
the application of explicit rules but training (1970, 186) and
repeated exposure that are needed to play a game properly or
to use a word meaningfully. The concept of language-games
highlights the idea that the mastery of language is an acquired
skill or know-how, not a systematic (innate or acquired)
417
Language-Game
application of rules: To understand a language means to be
master of a technique ([1953] 1997, 199).
3. Like games, languages are rule governed, but this does not
mean that there are strict and precise rules for each languagegame: Just as there are not rules to legislate for everything in
a game (e.g., there are no rules for how high or how hard one
throws the ball in tennis), language, too, is not everywhere circumscribed by rules. And as in games, the learning of explicit
rules is not always necessary someone can have learned a
game without ever learning or formulating rules ([1953] 1997,
31) and so, too, in language can the game be learned purely
practically, without learning any explicit rules (1977, 95).
The constitutive rules of language are those of grammar
(1974, 184). What Wittgenstein means by grammar is not, however, what grammarians mean by grammar: It is neither a taxonomy of the structural features of a language nor a science
that describes or prescribes the correct or standard usage of
words or arrangement of words. Wittgensteinian grammar is a
generic term for the publicly determined (though this determination is not due to a concerted, but to an unconcerted, consensus) conventions or conditions (1974, 138, 88) that govern
our meaningful use of words or expressions. Languages are rule
governed, but the rules that govern them are not metalinguistic
norms that exist in advance of use; learning the meaning of a
word is learning how the word is used. Moreover, if grammatical rules determine what it makes sense to say, they cannot
themselves belong to the language-game: a grammatical rule
is a preparation (1993, 72) for a language-game, except in
heuristic language-games (e.g., pedagogical language-games;
language teaching) where the formulation of some rules is the
object of the language-game (the distinction here is an instance
of the use and mention distinction). To learn grammatical
rules is to learn what the conventional conditions and constraints of the uses of a word are, which linguistic moves are
meaningful and which are not. Just as the rules of a game constitute the game and its allowable moves, grammar constitutes
language and its allowable moves:
4. Like games, language is embedded in our social, cultural and natural ways of living that is, in our form of life.
Languages cannot be abstracted from the context in which
they live: words have meaning only in the stream of life
(1982, 913). Language is a normative practice, but it is also a
social practice. Any language is founded on convention ([1953]
1997, 355); it is embedded in the shared activities of the language users in a given community: To obey a rule, to make a
report, to give an order, to play a game of chess, are customs
(uses, institutions) ([1953] 1997, 199). To understand a language-game, one must be either immersed in the community
in which it is embedded or knowledgeable about that communitys practices: Someone not accustomed to, or aware of, the
practice in some Semitic cultures of saying Hamsa! (Five!)
a verbal conjuring of the five fingers of the hand that protects
against the evil eye would not understand the purpose of
the utterance. For a language to emerge or be possible, there
has to be something shared. What is shared is a distinct form
of life: the particular biosocial conditions and activities that
make particular languages possible. The human form of life
could not have produced a feline language, nor a feline form
418
of life a human language. Language and form of life are internally related: To imagine a language means to imagine a form
of life ([1953] 1997, 19), and to imagine a human language is
necessarily to imagine a human form of life, a human way of
being and acting that essentially involves both our biological
makeup and our social behavior. For our language-games are
impacted by the facts of our natural history, such as our biological human constraints; for example, the language-game
with colors is characterized by what we can and what we cannot do (1970, 345). Therefore, if certain very general facts of
nature were different from what they are, so would our concepts and language-games be. But to say that our languagegames are conditioned by certain facts (1977, 617), is not to
say that our language-games are justified by, or answerable to,
these facts.
5. Like games, language and language-games are autonomous. By this, Wittgenstein means that although our languagegames are rooted in our form of life, they are not accountable
to, or rationally grounded in, any reality: The language-game is
not based on grounds. It is not reasonable (or unreasonable). It
is there like our life (1977, 559). Rather than speak of symbols,
words, or sentences as the primary or elementary units of meaning (as logicians, including Wittgenstein himself in the Tractatus,
had done), the later Wittgenstein views the language-game as
the basic unit in linguistic activity; he urges us to look on the
language-game as the primary thing ([1953] 1997, 656), that
which does not have to be explained by any fact. Here, he can
be viewed as broadening Gottlob Freges context principle: The
context necessary for meaning is not the proposition but the
language-game (e.g., a sound is an expression only in a languagegame [(1953)1997, 261]).
In his last work, On Certainty, Wittgenstein dwells on the
importance of unmoving foundations for the possibility of
language-games: It is essential for our language-games that
no doubt appears at certain points (cf. 1977, 524). He argues
that some basic certainties such as The world exists or
This is a hand or Cats dont grow on trees are necessary,
unmoving foundations of our language-games (cf. 1977, 403,
411), that the whole language-game rests on this kind of certainty (cf. 1977, 446). This kind of certainty, however, is nonepistemic; he views it as a kind of trust: I really want to say
that a language-game is only possible if one trusts something
(1977, 509).
Danile Moyal-Sharrock
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Baker, G. P., and P. M. S. Hacker. 1980. Wittgenstein: Understanding
& Meaning: An Analytical Commentary on the Philosophical
Investigations. Vol. 1. Oxford: Basil Blackwell.
Canfield, John V. 1993. The Living Language: Wittgenstein and
the Empirical Study of Communication Language Sciences 15.3:
16593.
Wittgenstein, Ludwig. [1953] 1997. Philosophical Investigations. Trans.
G. E. M. Anscombe. 2d ed. Oxford: Blackwell.
. 1970. Zettel. Ed. G. E. M. Anscombe and G. H. von Wright, trans
G. E. M. Anscombe. Berkeley: University of California Press.
. 1974. Philosophical Grammar. Ed. R. Rhees, trans. A. Kenny.
Oxford: Blackwell.
Language-Learning Environment
. 1977. On Certainty. Ed. G. E. M. Anscombe and G. H. von
Wright, trans. D. Paul and G. E. M. Anscombe. Amended 1st ed.
Oxford: Blackwell.
. 1980. Remarks on The Philosophy Of Psychology. Vol. 1. Ed.
G. E. M. Anscombe and G. H. von Wright, trans. G. E. M. Anscombe.
Oxford: Blackwell.
. 1982. Last Writings on the Philosophy of Psychology. Vol. 1. Ed.
G. H. von Wright and Heikki Nyman, trans. C. G. Luckhardt and
M. A. E. Aue. Oxford: Blackwell.
. 1993. Moores Wittgenstein lectures in 19301933. In
Philosophical Occasions: 19121951, ed. J. C. Klagge and A. Nordman,
46114. Indianapolis: Hackett.
LANGUAGE-LEARNING ENVIRONMENT
This term refers to the linguistic and sociocultural environment
in which children learn to talk and, in particular, to the language
to which they are exposed. The first systematic studies of the language-learning environment were conducted in response to the
claim that, like the language used among adults, the language
heard by children was grossly defective full of false starts, grammatical errors, and misleading pauses and as such represented
a very poor sample of the language that the child must eventually
learn. These studies showed that speech addressed to children
was largely clear, well formed, and semantically and syntactically simpler than speech addressed to adults, leading
some researchers to argue that in simplifying their speech, parents were presenting their children with graded language lessons
that could bear at least some of the burden of explanation for the
childs remarkably swift progress in language learning (see Pine
1994 for a review).
In fact, there are a number of problems with this view. First,
it confuses the notions of facilitating interaction and facilitating
acquisition. Thus, some of the adjustments made by parents to
facilitate interaction (e.g., the extensive use of questions) probably have the effect of increasing the complexity of the language
to which children are exposed during the early stages. Second,
it is unclear to what extent the adjustments made by Western
middle-class parents generalize across cultures. Indeed, ethnographic researchers have identified cultures in which parents
react with mirth or horror to the idea of holding conversations
with young language-learning children (see Lieven 1994 for a
review). Third, the notion that the language-learning environment somehow facilitates acquisition is theoretically rather vacuous in the absence of a reasonably well-specified theory of how
the child is learning from this environment.
Cognizant of this last problem, more recent work has focused
on the kind of information that is present in the environment
and the way in which this might be exploited by the childs
language-learning mechanisms. Thus, work in computational
modeling has shown that there is a great deal of information in
the statistical structure of human languages that could, in principle, be used to solve particular language-learning problems.
For example, work on segmentation has shown that it is possible to use information about the stress pattern, the phonotactics, and the transitional probabilities between syllables in
a language to identify boundaries between words (Brent 1999),
and work on the acquisition of syntactic categories has shown
that it is possible to categorize words quite successfully on
419
Language of Thought
Jusczyk, Peter. 1999. How infants begin to extract words from speech.
Trends in Cognitive Sciences 3: 3238.
Lewis, John, and Jeffrey Elman. 2001. Learnability and the statistical
structure of language: Poverty of stimulus arguments revisited. In
Proceedings of the Twenty-Sixth Annual Boston University Conference
on Language Development, 35970. Somerville, MA: Cascadilla.
Lieven, Elena. 1994. Crosslinguistic and crosscultural aspects of
language addressed to children. In Input and Interaction in
Language Acquisition, ed. Clare Gallaway and Brian Richards, 5673.
Cambridge: Cambridge University Press.
Mintz, Toby. 2003. Frequent frames as a cue for grammatical categories
in child-directed speech. Cognition 90: 91117.
Pine, Julian. 1994. The language of primary caregivers. In Input and
Interaction in Language Acquisition, ed. Clare Gallaway and Brian
Richards, 1537. Cambridge: Cambridge University Press.
Redington, Martin, Nick Chater, and Steven Finch. 1998. Distributional
information: A powerful cue for acquiring syntactic categories.
Cognitive Science 22: 42569.
Richards, Brian. 1990. Language Development and Individual
Differences: A Study of Auxiliary Verb Learning. Cambridge: Cambridge
University Press.
LANGUAGE OF THOUGHT
This is a special language that has been postulated by a number
of writers G. Harman (1972), J. Fodor (1975, 1987) to explain
how humans and many animals represent and think about the
world. The language of thought (LOT) is claimed to be coded,
or entokened, in their brains, rather than in the way certain
formal languages are entokened in the circuitry of a computer.
What makes the LOT a language is that it possesses semantically
valuable, causally efficacious logico-syntactic structure: That is, it
consists, for example, of names, predicates, variables, quantifiers
(all, some) logical connectives (not, and, only if) and
operators (possibly, probably) that are combined to form
complex sentences that can be true or false.
The LOT need not be the natural language (e.g., English,
Chinese), if any, that a creature speaks, although some writers
have supposed that in adult humans the two may coincide (providing an interesting perspective in the Sapir-Whorf hypothesis
that thought is determined by language). Indeed, given that the
relevant sorts of intelligent behavior are displayed by many infralinguistic creatures infants, chimpanzees its postulation need
not be confined to natural language users. Nor need the LOT be
in the least conscious or introspectible: A persons thought processes might be explained by a LOT, while introspectively his or
her mental life might seem to consist wholly of images, feelings, and inarticulate impulses. Most importantly, processing a
LOT need not require any sort of intelligent creature to read and
understand the sentences being processed in the brain. Along
lines set out by Alan Turing, the processing of the LOT symbols
can be executed purely mechanically.
As in discussing any language, there are two issues raised by
such a postulation: i) There are syntactic issues regarding the
character of the actual symbols and the computations defined
over them. This is the kind of issue regularly addressed in rich
detail by programmers dealing with artifactual computers, and
with psychologists dealing with naturally occurring, living creatures, in their concern with data structures and algorithms for
dealing with them, for example, for vision or reasoning; and
420
ii) there are semantic issues regarding the meaning or interpretation of the symbols and data structures. In the case of artifactual
computers, this issue is usually conveniently settled by stipulation: The artifactor gets to say what the symbols represent (e.g.,
bank balances, chess moves). In the case of natural creatures,
there is, of course, no relevant artifactor, and so the meaning
of the symbols must be determined by some natural facts or
relations.
There are two (not necessarily exclusive) kinds of candidates
for a theory of meaning of a LOT:
i. Meaning is spelled out in terms of some of the symbols crucial causal/ conceptual roles in relation to other symbols, mirroring patterns of inference. This is a natural suggestion for logical
symbols, such as and and not: A symbol, @, might mean
and because states entokening sentences p and q separately
might cause and in turn be caused by a state entokening p@q
by itself (see Block 1986; Peacocke 1994).
ii. Meaning is spelled out in terms of causal relations that
the individual symbols bear to phenomena in the world. For
example, a symbol S might be entokened in a creatures brain in
a way that covaries with the presence of a certain shape before
the creatures eyes in various circumstances, such as under ideal
conditions, under evolutionarily selective ones, or under ones
that display a certain counterfactual structure (see Dretske 1988;
Stalnaker 1984; Fodor 1987).
The chief rivals to the LOT hypothesis are either one or another
form of interpretativism, according to which a creature has propositional attitudes only because ascription of such states permits
the most rational interpretation of their behavior (see Davidson
[1973] 1984; Dennett 1987), and/or the brain structures underlying that ascription are of a radical connectionist nonsyntactically structured sort (see Smolensky 1988). Its sometimes also
thought that some kind of system of imagistic representations
would not only be truer to introspection but also explain various
response-time results that suggest that people think in images
(see Kosslyn 1994).
The main reasons for believing in a LOT as opposed to these
rivals is that it would explain a number of interesting phenomena associated with the mind. Salient among these are the
following:
1. The propositional structure of attitudes: The standard
object of, for example, a thought, belief, desire, hope, or expectation is some kind of truth-valuable object, most perspicuously
expressed by a sentence; neither images nor connectionist
networks are able to systematically represent logically complex thoughts, involving, for example, negations and nested
quantifiers (what image could express Not everyone loves
someone?).
2. The causal efficacy of thought: A thought can cause bodily
states and movements because it is a structure entokened in the
brain.
3. The productivity of thought: People can potentially understand an infinitude of different thoughts formed by logical combinations of simpler ones, for example, Its possible for every
cat to chase some rat that eats some cheese that that lives in
the house that Jack built, since the LOT entokened in their brain
can (under reasonable idealization) produce a corresponding
infinity of sentences.
Language Policy
4. The systematicity of thought: If people can think some
thought p, then they can also think all logical permutations of
p; for example, people can think If John leaves, then someone
insulted Mary if and only if they can think If someone leaves, then
Mary insulted John, since they can mechanically recombine the
simple expressions of a sentence.
5. The intensionality of thought: People can think of things
in one way without thinking of those very things in another; for
example, they can think that the morning star is Venus without
thinking that the evening star is, and they can think about different nonexistent things, such as Zeus and Santa Claus. Indeed,
the LOT can explain the hyperintensionality of thought: People
can think about things that are even necessarily identical, as
when one thinks that Mark Twain but not Sam Clemens is
funny, because they employ correspondingly different LOT
symbols.
6. The multiple roles of attitudes: Different attitudes can be
directed upon the very same thought; for example, people can
believe, desire, suspect, doubt the same thought, that God exists.
Fodor (1975, 1987) and Georges Rey (1997) have argued that
these and other phenomena cannot be explained by the rival
views without substantial, additional ad hoc assumptions, for
example, that certain images or networks express logically complex thoughts and are causally related in the aforementioned
ways. For these reasons, the language of thought is to be preferred on empirical grounds (for more of an a priori argument
see, e.g., Davies 1991).
Georges Rey
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Block. N. 1986. Advertisement for a conceptual role semantics. In
Studies in the Philosophy of Mind. Ed. P. French, T. Uehling, and
H. Wettstein, 61578. Minneapolis: University of Minnesota Press.
Davidson, D. [1973] 1984. Radical interpretation. In Inquiries into Truth
and Interpretation. Oxford: Clarendon.
Davies, M. 1991. Concepts, connectionism, and the language of thought.
In Philosophy and Connectionist Theory, ed. W. Ramsey et al., 22956.
Hillsdale, NJ: Erlbaum.
Dennett, D. 1987. The Intentional Stance. Cambridge, MA: MIT Press
Dretske, F. 1988. Explaining Behavior: Reasons in a World of Causes.
Cambridge, MA: MIT Press
Fodor, J. 1975. The Language of Thought. New York: Crowell.
. 1987. Psychosemantics, Cambridge, MA: MIT Press.
Harman, G. 1972. Thought. Princeton, NJ: Princeton University Press.
Kosslyn, S. 1994. Image and Brain: The Resolution of the Imagery Debate.
Cambridge, MA: MIT Press.
Peacocke, C. 1994. A Study of Concepts. Cambridge, MA: MIT Press.
Rey, G. 1997. Contemporary Philosophy of Mind. Oxford: Blackwell.
Smolensky, P. 1988. On the proper treatment of conectionism.
Behavioral and Brain Sciences 11: 5974.
Stalnaker, R. 1984. Inquiry. Cambridge, MA: MIT Press.
LANGUAGE POLICY
Definition and History
Language policy studies the regular choices among varieties and
variants within a speech community the language practices
of members of the community, their beliefs about the values to
be assigned to the varieties, and efforts by individuals or groups
421
Language Policy
management, ranging from individual self-correction to the
organized management of all micro and macro levels. Detailed
descriptions of language policy such as Grenoble (2003), Kaplan
and Baldauf (2003) and Zhou (2004) have started to clarify the
complex dimensions that a theory must handle.
Territoriality
A second solution to problems associated with having multiple
major languages within a single nation-state is territoriality.
India tried to base its internal political divisions on language.
The partition into a Hindu-dominated India and Moslemdominated Pakistan was paralleled by a language-management
effort to divide what was previously considered a single language, Hindustani, into two Hindi written in Devanagari script
and Urdu with Perso-Arabic script (Annamalai 2001). The splitting up of India into states reflected major language differences,
although it could not capture the complexity of a nation with
422
Language Policy
the United Kingdom ( Laoire 1996; Coupland et al. 2006), and
Canada (Bourhis 2001). There are also efforts in New Zealand
(Spolsky 2005; May and Hill 2005), in South America (Hornberger
and King 2001), and among other indigenous peoples (McCarty
2003; Omoniyi 2003).
Speakers of major languages also fear language loss. This
can be seen in Spain, with the sensitivity of its Academy to language change; in France, with its growing number of regulations
and language agencies; in Russia, with its refusal to recognize
non-Cyrillic alphabets for minority languages and its claim of
defending the language rights of Russian-speakers in former
Soviet states; and even in the United States, where an Englishonly movement is struggling against what it sees as the threat of
Spanish and other immigrant languages to the survival of what
most people believe to be the strongest language in the world
(Baron 1990).
Language policy is a new and rapidly developing field, the
urgency and seriousness of which has resulted in activism by
scholars who feel responsible for correcting what they see as
injustices or blindness to the potential loss of linguistic diversity,
as well as in academic attempts to develop theories to explain
data from increasingly detailed descriptions of situations and
policies.
Bernard Spolsky
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Ager, Dennis E. 1999. Identity, Insecurity and Image: France and Language.
Clevedon and Philadelphia: Multilingual Matters.
Annamalai, E. 2001. Managing Multilingualism in India: Political and
Linguistic Manifestations. New Delhi: Sage.
Aunger, Edmund A. 1993. Regional, national and official languages
in Belgium. International Journal of the Sociology of Language
104: 3148.
Bamgbose, Ayo. 2000. Language and Exclusion: The Consequences of
Language Policies in Africa, Beitrage zur Afrikanistik. Mnster and
Hamburg: LIT Verlag.
Baron, Dennis E. 1990. The English Only Question. New Haven, CT: Yale
University Press.
Bourhis, Richard Y. 2001. Reversing language shift in Quebec. In Can
Threatened Languages be Saved? ed. J. A. Fishman 10141. Clevedon
and Avon: Multilingual Matters.
Cooper, Robert L. 1989. Language Planning and Social Change.
Cambridge: Cambridge University Press.
Cooper, Robert L., ed. 1982. Language Spread: Studies in Diffusion and
Social Change. Bloomington: Indiana University Press.
Coupland, Nicholas, Hywel Bishop, Betsy Evans, and Peter Garrett. 2006.
Imagining Wales and the Welsh language: Ethnolinguistic subjectivities and demographic flow. Journal of Language and Social Psychology
25.4: 35176.
Fishman, Joshua A. 1973. Language and Nationalism: Two Integrative
Essays. Rowley, MA: Newbury House.
. 1977. Comparative study of language planning: Introducing a survey. In Language Planning Processes, ed. J. Rubin, B. H. Jernudd, J. Das
Gupta, J. A. Fishman and C. A. Ferguson, 3140. The Hague: Mouton.
. 1990. What is reversing language shift (RLS) and how can it succeed? Journal of Multilingual and Multicultural Development 11.1/2:
536.
. 1991. Reversing Language Shift: Theoretical and Empirical
Foundations
of
Assistance
to
Threatened
Languages.
Clevedon: Multilingual Matters.
Fishman, Joshua A., Charles A. Ferguson, and Jyotirinda Das Gupta. 1968.
Language Problems of Developing Nations. New York: Wiley.
Grenoble, Lenore A. 2003. Soviet Language Policy. Dordrecht, the
Netherlands: Kluwer Academic Publishers.
Harlow, Ray. 2004. Switzerland. In Encyclopedia of Linguistics, ed.
P. Strazny. London: Taylor and Francis.
Haugen, Einar. 1966. Language Conflict and Language Planning: The
Case of Modern Norwegian. Cambridge: Harvard University Press.
Hoffmann, Charlotte. 1995. Monolingualism, bilingualism, cultural
pluralism and national identity: Twenty years of language planning
in contemporary Spain. Current Issues in Language and Society
2.1: 5990.
Hornberger, Nancy H., and Kendall A. King. 2001. Reversing language
shift in South America. In Can Threatened Languages be Saved? ed.
J. A. Fishman, 16694. Clevedon and Avon: Multilingual Matters.
Hrdegen, Stephan. 2001. The Fribourg linguistic case controversy
about the language of instruction in schools in the light of freedom
of language and equal educational opportunities in Switzerland.
European Journal for Educational Law and Policy 5: 7382.
Jernudd, Bjoern, and J. V. Neustupn. 1987. Language planning: For
whom? In Proceedings of the International Colloquium on Language
Planning, ed. L. LaForge, 7184. Quebec: Presses de lUniversite Laval.
Kaplan, Robert B., and Richard B. Baldauf. 2003. Language and
Language-in-Education Planning in the Pacific Basin. Dordrecht, the
Netherlands: Kluwer Academic Publishers.
Katsuragi, Takao. 2005. Japanese language policy from the point of view
of public philosophy. International Journal of the Sociology of language 175/176: 4154.
Kloss, Heinz. 1966. German-American language maintenance efforts.
In Language Loyalty in the United States, ed. J. Fishman, 20652. The
Hague: Mouton.
Krauss, Michael. 1991. The worlds languages in crisis. Language
68.1: 410.
Laitin, David D., and Robert Reich. 2003. A liberal democratic
approach to language justice. In Language Rights and Political
Theory, ed. W. Kymlicka and A. Patten, 80104. Oxford: Oxford
University Press.
Landau, Jacob, and Barbara Kellner-Heinkele. 2001. Politics of Language
in the Ex-Soviet Muslim States: Azerbaijan, Usbekistan, Kazakhstan,
Kyrgyzstan, Turkmenistan and Tajikistan. London and Ann Arbor:
C.Hurst & Co. and University of Michigan Press.
Lewis, E. Glyn. 1972. Multilingualism in the Soviet Union. The
Hague: Mouton.
May, Stephen. 2005. Language rights: Moving the debate forward.
Journal of Sociolinguistics 9.3: 31947.
May, Stephen, and Richard Hill. 2005. Mori-medium education: Current
issues and challenges. International Journal of Bilingual Education
and Bilingualism 8.5: 377403.
McCarty, Teresa L. 2003. Revitalising indigenous languages in homogenising times. Comparative Education 29.2: 14763.
Nekvapil, Ji. 2006. From language planning to language management.
Sociolinguistica 20: 92104.
. 2007. Language cultivation in developed contexts. In Handbook
of Educational Linguistics, ed. B. Spolsky and F. M. Hult, 25165.
Oxford: Blackwell.
Neustupn, J.V., and Ji Nekvapil. 2003. Language management
in the Czech republic. Current Issues in Language planning
4.3/4: 181366.
Laoire, Muiris. 1996. An historical perspective of the revival of Irish
outside the Gaeltacht, 18801930, with reference to the revitalization of Hebrew. In Language and State: Revitalization and Revival in
Israel and Eire, ed. S. Wright, 5175. Clevedon and Avon: Multilingual
Matters.
423
Laws of Language
Omoniyi, Tope. 2003. Local policies and global forces: Multiliteracy and
Africas indigenous languages. Language Policy 2.2: 13352.
Ozolins, Uldis. 2003. The impact of European accession upon language
policy in the Baltic States. Language Policy 2.3: 21738.
Phillipson, Robert. 1992. Linguistic Imperialism. Oxford: Oxford
University Press.
Pranjkovic, Ivo. 2001. The Croatian standard language and the Serbian
standard language. International Journal of the Sociology of Language
147: 3150.
Ricento, Thomas, ed. 2001. Ideology, Politics and Language Policies: Focus
on English. Amsterdam and Philadelphia: John Benjamins.
Skutnabb-Kangas, Tove, Robert Phillipson, and Mart Rannut. 1995.
Linguistic Human Rights: Overcoming Linguistic Discrimination. Berlin
and New York: Mouton de Gruyter.
Spolsky, Bernard. 2004. Language Policy, Key Topics in Sociolinguistics.
Cambridge: Cambridge University Press.
. 2005. Maori lost and regained. In Languages of New Zealand, ed.
A. Bell, R. Harlow, and D. Starks, 6785. Wellington: Victoria University
Press.
Zhou, Minglang, ed. 2004. Language Policy in the Peoples Republic
of China: Theory and Practice since 1949. Dordrecht, the
Netherlands: Kluwer Academic Publishers.
LAWS OF LANGUAGE
The Concept of Law
The philosophy of science defines the term scientific law as a
meaningful universal hypothesis that is systematically connected
to other hypotheses in the field and, at the same time, well corroborated on relevant empirical data (cf. Bunge 1967). A law is
called universal because it is valid at all times, everywhere, and
for all objects of its scope.
A system of laws is called a theory. The construction of a theory is the highest and most demanding goal of scientific research
and can be undertaken only if and when a number of interrelated laws have been found. There is much confusion about the
term theory, especially in linguistics, where all kinds of formalisms, thoughts, approaches, descriptive tools, definitions, and
concepts are called theories. The philosophy of science distinguishes two kinds of theories: 1) the axiomatic theories of logics
and mathematics and 2) the empirical theories in the factual sciences. While the first ones make statements only within a given
axiomatic system and can be used only to construct analytical
truths, the latter ones make statements about parts of the world.
The truth of an empirical theory and of its elements, the laws,
depends not only on internal correctness but also on the correspondence with the facts of reality although every empirical
theory must have an axiomatic kernel.
The value of theories and their components, the laws, lies not
only in their role as the containers of scientific knowledge but
also in the fact that there can be no explanation without at least
one law: A valid scientific explanation (the so-called deductivenomological explanation; cf. Hempel and Oppenheim 1948) is a
subsumption under laws taking into account boundary conditions. Laws must not be confused with rules, which are either
prescriptive or descriptive tools without any explanatory power;
hence, grammars and similar formalisms also cannot explain
anything. Another significant difference is that rules can be violated laws (in the scientific sense) cannot.
424
f (r) =
K
(b+ r)
where f (r) is the frequency, r the rank, b and empirical parameters, and K a normalizing constant that makes the probabilities
sum up to 1.0.
Since the seminal works of Zipf and Mandelbrot, numerous
laws have been found. Other examples of distributional laws
are (in morphology and lexicon) the distribution of length,
polysemy, synonymy, age, part of speech (see word classes),
and so on, (in syntax) the frequency distribution of syntactic constructions, the distribution of their complexity, depth of
embedding, information, and position in mother constituent; (in
semantics) the distribution of the lengths of paths in semantic
networks (see also semantic fields), semantic diversification, and so on. Any property and any linguistic unit studied so
far displays a characteristic probability distribution.
The second kind of law is called the functional type, because
these laws link two (or more) properties. An illustrative example
of this kind is Menzeraths Law (also called Menzerath-Altmann
Law), which relates the size of linguistic constituents to the size
of the corresponding construct. Thus, the (mean) length of the
syllables of a word depends on the number of syllables the
word consists of; the (mean) length of the clauses in a sentence
depends on the length of the sentence (measured in terms of the
number of clauses it consists of). The most general form of this
law is given by formula (2):
(2)
y = Ax be-cx
Laws of Language
3.2
3.0
2.8
2.6
2.4
2.2
2.0
(/vart/ > /vurde/) in the time period from 1445 to 1925. As the
graph shows, the replacement was very limited for the first 200
years, but even a much shorter time span can provide enough
information to predict the development over the next several
hundred years.
Another variant of this third kind of law is based on (discrete)
linguistic instead of (continuous) physical time. The simplest
way to operationalize linguistic time is the reference to text
position. In oral texts, there is a direct correspondence of the
sequence of linguistic units to physical time intervals.
Several linguistic characteristics can be investigated using
indices that relate their frequency to current text position, among
them the type-token-ratio (TTR). At each text position, the number of types occurred to that point is counted, which yields a
monotonously increasing curve, because the number of words
used before a given text position cannot decrease in the course
of the rest of the text. A straightforward theoretical derivation of
this law was given by Gustav Herdan (1966), represented by the
simple formula (4):
(4)
p=
c
1 + ae bt
y = ax b
Theory Construction
Currently, there are two approaches to the construction of a linguistic theory (in the sense of the philosophy of science): 1) synergetic linguistics and 2) Gejza Wimmer and Gabriel Altmanns
unified theory.
The basic idea behind synergetic linguistics (cf. Khler 1986,
2005) is the aim to integrate the separated laws and hypotheses found so far into a complex model that not only describes
the linguistic phenomena but also provides a means to explain
them. This is achieved by introducing the central axiom that
language is a self-regulating and self- organizing system .
An explanation of the existence, properties, and changes of
linguistic, more generally semiotic, systems is not possible
without the aspect of the (dynamic) interdependence of structure and function. The genesis and evolution of these systems
must be attributed to repercussions of communication upon
structure (cf. Bunge 1998 and Khler and Martinkov 1998).
425
Laws of Language
Types
100
50
0
0
300
600
900
1200
1500
Tokens
426
Learnability
LEARNABILITY
generative grammar shifted the foundations of theoretical
linguistics away from discovery procedures the automatic construction of an optimal grammar to the problem of language
learnability, the question of how a natural language could be
learned in principle. This statement of the problem is so vague
as to be useless; so let us break the question into subparts and
consider them in turn.
The study of language learning might proceed from the observation and investigation of actual human children engaged in the
process of learning their mother tongue or tongues. However fascinating and useful this approach is, it is fraught with a number
of difficulties. Real children are engaged in a number of different
tasks and are changing along a number of different dimensions
while in the process of learning their first language. The developmental psycholinguist must, then, be careful of all these different
factors.
The investigation of learnability seeks to circumvent these difficulties by considering language learning as a problem in computational logic. The researcher in language learnability seeks to
construct an explicit algorithm that will produce a grammar for
a target language after finite exposure to evidence from that language. Such a researcher takes the rarefied view of language as
a set of strings, corresponding to the grammatical sentences
of that language. He or she supposes that the learner is actually
an algorithm that takes as input a text, an infinite sequence of
sentences. The text is constructed by drawing strings from the
language and presenting them, one at a time, to the learner. In
this case, the learner is presented with positive only evidence;
he/she is given information about sentences that are in the language but no information about sentences that are outside the
language. An alternative learning setting would be to allow the
learner to be tutored by giving him/her strings that are marked
for grammaticality. Such information demonstrably simplifies the learning task, but in real learning, the child is unlikely
to receive systematic evidence about grammaticality. As a result,
learnability research has generally proceeded from the assumption of positive-only evidence.
After each example is presented to the learner, that learner
makes a guess about the grammar of the target language. A
learner is said to converge to a grammar for a language just in
case the learner hypothesizes that grammar after finite exposure
to the text and never alters the hypothesis after that. If the grammar is correct, then the learner is said to have learned the target
language; truly, learning a language means that the learner has
hit the correct grammar and never changes his/her mind after
that. We only require, at this point, that the grammar generate
the correct set of sentences; we have not said anything about
how the grammar does so, and so we place no constraints on the
structural descriptions assigned to sentences.
A language is learnable if there exists a learner who, upon
finite exposure to the language, learns the language in the aforementioned sense. A set of languages is learnable if the leaner can
learn every language in the set.
We can turn to an early learnable result from E. M. Gold
(1967; see also Osherson, Stob, and Weinstein 1986). Imagine a
set of languages defined as follows: We define a language, call it
427
Learnability
Costa Florencio (2003) has shown that the full range of categorial
grammars can be learned if the learner is presented with unlabeled structures.
Another response to Golds result is to impose different types
of constraints on the learners hypothesis space. For example,
the idea that universal grammar consists of a set of invariant
principles whose expression is regulated by a finite set of parameters has played a seminal role in linguistic theory over the past
quarter of a century (see principles and parameters theory and language acquisition). The learner is taken as
being faced with the finite task of discovering the correct value
of each parameter, where each parameter can take on one of a
finite set of parameters, given a text consisting of simple grammatical sentences.
The parametric approach usually assumes that the learner
is given positive-only input. After a sentence is presented to
the learner, it produces a hypothesis, possibly by changing the
value(s) of one or more parameters. It is obvious that even a relatively small set of parameters could produce an enormous space
of languages. The space of languages might, for example, contain local maxima. A local maximum might look correct to the
learner since it would always yield a structural analysis for any
input sentence, but it would systematically give the sentence an
incorrect structural analysis, for example. If this happened, then
the set of languages defined by that parameter space would not
be learnable relative to a simple learning device with positiveonly evidence.
In response to this problem, a number of different algorithms
were proposed. R. Clark (1992) proposed using a kind of artificial
evolution to converge to the target. A population of grammars
would be exposed to the input text, with the best performers
allowed to combine and produce offspring that had inherited
properties from the parent grammars. This approach uses the
parallelism implicit in a population to avoid the problem of local
maxima. This approach to learning falls into the class of probably
approximately correct (or PAC) learning; in this framework, the
learner is guaranteed to converge within a margin of error, where
the margin of error can be made arbitrarily small, but not zero.
Clark and I. Roberts (1993) extended this work to try to account
for language change. P. Niyogi (2006) has developed a more
sophisticated computational approach to this problem (see also
Yang 2002).
E. Gibson and Wexler (1994) tried to develop a learner that
used an algorithm that tests to see if resetting a parameter to
a new value actual improves the learners performance on the
input example. In order to avoid local maxima, they proposed
ordering the parameters according to a maturational sequence.
Readers should consult Frank and Kapur (1996) for an extensive
critique. S. Kapur (1991) developed a learning algorithm that
avoids local maxima by using a statistical model of indirect negative evidence. This algorithm is, once again, clearly within the
PAC learners.
Others have proposed using triggering evidence to set parameters (Dresher and Kaye 1990). On this model, each value of a
parameter would be associated with the abstract description of
a piece of triggering evidence that would cause the parameter
to be set to that value. The learner would scan the input text,
searching for examples that matched the description of a trigger;
428
A Real-Time Perspective
The most current approach to brain-language relations as we
near the end of the first decade of the twenty-first century is formulated in terms of processing metaphors such as activation
and maintenance that require a real-time analysis. It has been
suggested that within the LH, Brocas area is required for fastacting, relatively automatic and reflexive processing routines;
more frontal areas are critical for executive functions underlying language processing, including selecting among alternatives;
left posterior temporal areas seem important for activating and
maintaining argument structure, an aspect of lexical-semantics
or conceptual structure.
Take, for example, the activation of meaning-related word
forms during sentence comprehension. It has been argued that
unimpaired individuals initially activate multiple meanings of
ambiguous words, regardless of the context of the sentence.
There is evidence that this exhaustive access is controlled immediately by the LH (Burgess and Simpson 1988). Soon after, a
lexical choice is made on the basis of frequency and context.
Evidence from aphasia also supports the role of the left anterior
frontal cortex in lexical access. For example, individuals with
Brocas aphasia appear to show a slow rise time of the initial
activation of multiple meanings, while those with Wernickes
aphasia evince normal patterns (Prather et al. 1991). Other activation accounts suggest that individuals with Brocas aphasia
(with damage to LIFG) underactivate lexical forms, while those
with Wernickes aphasia (with damage to STG) overactivate
(Blumstein and Milberg 2000).
The role of LIFG in real-time processing also appears to extend
to the comprehension of sentences that contain displaced arguments or those with filler-gap dependencies, for example, in
object relative (OR) constructions (e.g., The audience liked the
wrestler that the priest condemned *____ for foul language) where
a direct object argument or filler (e.g., wrestler) has been displaced from its canonical, post-verb position or gap (noted by *).
Individuals with Brocas aphasia do not activate the filler at the
gap in real time, unlike what is observed for neurologically intact
individuals and those with Wernickes aphasia or RH lesions
(Swinney et al. 1996). Thus, a real-time processing deficit may
underlie the inability for Brocas individuals to ultimately comprehend these constructions when they are probed with simple
sentence-picture matching tasks or grammaticality judgments.
Variability
Much of the work detailing the role of the LH in language processing has been based on descriptions of neuroanatomy conducted in the later third of the nineteenth and early twentieth
centuries. K. Brodmann (1909) suggested that the most functionally relevant parcellation of the brain is by cytoarchitectonics (cellular composition), but the map of Brodmanns Areas
is based on manually drawn borders of a single brain. K. Amunts
and colleagues (1999) examined 10 postmortem brains, and the
borders for each brain were automatically drawn and superimposed on a template to produce a group cytoarchitectonic
map. Large intersubject variability was uncovered, perhaps partially explaining why so much variability exists in the mapping
between behavior and anatomy from both lesion and functional
imaging studies.
Another possible contributor to intersubject variability is the
assumption of dead tissue only in and around the lesion. For
many years, investigators have assumed that structural lesions
(and the ischemic penumbra surrounding the lesion) were the
primary loci contributing to language deficits. With the advent
of more refined neuroimaging technology, such as perfusion
weighted and diffusion tensor imaging, researchers have been
investigating areas of the brain that are found to be structurally
intact yet not receiving an optimal supply of blood flow. These
hypoperfused regions give way to functional lesions inside seemingly intact neural tissue (Hillis 2007; Love et al. 2002).
429
Translational Research
Finally, the investigation of language processing in the LH has
yielded a growing enterprise devoted to mapping recovery of
function in aphasia following brain damage. Converging evidence from clinical studies along with functional neuroimaging studies have demonstrated that, depending on individual
factors such as size and extent of lesion, premorbid handedness, and so on, recovery of language function may include the
undamaged regions of the LH language processing network or
homologous RH regions (Heiss et al. 1999; Kinsbourne 1971).
Treatment of language disorders for individuals with brain
damage, and the subsequent behavioral and neural changes
430
that are observed, may help illuminate brain-behavior mapping in both the left and the right hemispheres (e.g., Thompson
and Shapiro 2007).
Tracy Love and Lewis P. Shapiro
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Amunts, K., Axel Schleicher, U. Brgel, Hartmut Mohlberg, Harry Uylings,
and Karl Zilles. 1999. Brocas region revisited: Cytoarchitecture
and intersubject variability. Journal of Comparative Neurology 412
(August): 31941.
Beeman, M., and C. Chiarello, eds. 1997. Right Hemisphere Language
Comprehension: Perspectives from Cognitive Neuroscience. Hillsdale,
NJ: Lawrence Erlbaum.
Blumstein, S., and W. Milberg. 2000. Comprehension in Brocas and
Wernickes aphasia: Singular impairment. In Language and the
Brain, ed. Y. Grodzinsky, L. P. Shapiro and D. Swinney, 16783. San
Diego: Academic Press.
Brodmann, K. 1909. Vergleichende Lokalisationslehre der Grohirnrinde
in ihren Prinzipien dargestellt auf Grund des Zellenbaues. Leipzig:
Barth JA.
Burgess, C., and G. Simpson. 1988. Cerebral hemispheric mechanisms
in the retrieval of ambiguous word meanings. Brain and Language 33
(March): 86103.
Caplan, D., and G. Waters. 1999. Verbal working memory capacity and
language comprehension. Behavioral and Brain Sciences 22.1: 7794.
Friederici, A. 2002. Towards a neural basis of auditory sentence
processing. TRENDS in Cognitive Sciences 6.2: 7884.
Friedmann, N. 2006. Speech production in Brocas agrammatic
aphasia: Syntactic tree pruning. In Brocas Region, ed. Y. Grodzinsky
and K, Amunts, 6382. New York: Oxford University Press.
Geschwind, N. 1965. Disconnection syndromes in animals and man.
Brain 88: 585644.
Grodzinsky, Y. 2006. A blueprint for a brain map of syntax. In Brocas
Region, ed. Y. Grodzinsky and K. Amunts, 83107. New York: Oxford
University Press.
Heiss, W., J. Kessler, A. Thiel, et al. 1999. Differential capacity of left
and right hemispheric areas for compensation of poststroke aphasia.
Annals of Neurology 45.4: 4308.
Hickok, G., T. Love-Geffen, and E. Klima. 2002. Left temporal lobe
supports sign language comprehension. Brain and Language
82.2: 16778.
Hillis, A. 2007. Magnetic resonance perfusion imaging in the study of
language. Brain and Language 102.2: 16575.
Humphries, C. , T. Love, D. Swinney, and G. Hickok. 2005 . Response
of anterior temporal cortex to syntactic and prosodic manipulations during sentence processing. Human Brain Mapping
26.2 : 12838.
Kinsbourne, M. 1971. The minor cerebral hemisphere as a source of
aphasic speech. Archives of Neurology 25.4: 3026.
Love, T., D. Swinney, E. Wong, and R. Buxton. 2002. Perfusion imaging
and stroke: A more sensitive measure of the brain bases of cognitive
deficits. Aphasiology 16.9: 87383.
Prather, P., L. Shapiro, E. Zurif, and D. Swinney. 1991. Real-time examinations of lexical processing in aphasics. Journal of Psycholinguistic
Research 20.4: 27181.
Stowe, L., A. Paans, A. Wijers, F., et al. 1999. Sentence comprehension
and word repetition: A positron emission tomography investigation.
Psychophysiology 36.6: 786801.
Stromswold, K., D. Caplan, N. Alpert, and S. Rauch. 1996. Localization
of syntactic comprehension by positron emission tomography. Brain
and Language 52.3: 45273.
Legal Interpretation
Swinney, D., E. Zurif, P. Prather, and T. Love. 1996. Neurological distribution of processing operations underlying language comprehension. Journal of Cognitive Neuroscience 8.2: 17484.
Thompson, C., and L. P. Shapiro. 2007. Complexity in treatment of
sentence deficits in aphasia. American Journal of Speech-Language
Pathology 16: 3042.
Zurif, E., and A. Caramazza. 1976. Psycholinguistic structures in
aphasia: Studies in syntax and semantics. In Studies in Neurolinguistics,
ed. N. Avakian-Whitaker and H. Whitaker, 26092. New York: Academic
Press.
LEGAL INTERPRETATION
The judicial practice of legal interpretation provides the model
for other versions of legal interpretation, such as the interpretation of law by practicing lawyers. Interpretation of three types
of texts prior judicial decisions (which serve as precedents for
a present interpretive exercise), statutes, and constitutions
provides the model for interpretation of other texts, such as
regulations and treaties. It has a temporal and an institutional
dimension. Decision makers interpret texts written earlier,
sometimes substantially earlier, and produced by institutions
different from the ones engaged in the interpretive enterprise
(if only because of changes in personnel). These characteristics
generate some problems common to all three forms of legal
interpretation.
Statutory Interpretation
Questions of statutory interpretation as such arise only when
someone an enforcement official or a judge, for example has
some question about what a statutes terms mean. Where statutory terms are thought to be unambiguous, officials simply apply
the statutes, an operation that to them seems preinterpretive
(see also philology and hermeneutics). Application rather
than interpretation is likely to be more common soon after a
statutes adoption, because people will generally be familiar with
what the statutes enactors were trying to do unless, as happens
with some frequency, the adopters deliberately left specific provisions in the new statute unclear.
There are three prominent approaches to statutory interpretation in the United States, with parallels in other legal systems.
(For an overview of the contemporary discussion in the United
431
Legal Interpretation
States, see Vermeule 2006.) The textualist approach interprets a
statutes terms by asking what the words would mean to an ordinary reader (usually, a reader at the time the statute was enacted)
who is reasonably well informed about the meaning of the technical terms and about the entire statutory environment within
which the contested term is located.
Textualism is a rather bare-bones interpretive approach,
which to its critics requires the interpreter to ignore real and
accurate information about what a statute is designed to do.
Proponents of textualism often claim more clarity for the outcomes they reach than there actually is. Where ambiguity persists after considering the sources to which textualists limit
themselves, some other basis is needed for resolving the controversy. The most prominent candidate emerges from the assumption that ambiguous legislation should not be taken to disturb
the status quo. This assumption is sometimes expressed as a
canon of construction that ambiguous statutes should not be
construed to change the common law (that is, the background
rules that would apply if the legislature took no action). An alternative defense of textualism is comparative: that it reaches better
outcomes overall than alternative approaches that ask interpreters to assess information that they are not well equipped to handle, even though in particular cases the use of one or the other
approaches might produce a better result than textualism.
The intentionalist approach shifts the focus from the reasonable reader to the enacting legislature. It asks what the legislature intended to accomplish by enacting the statute. In its least
controversial version, the intentionalist approach directs the
interpreters attention to the problem the statute was designed
to solve, producing an interpretation that, in the judges view,
solves the problem as well as possible within the bounds set by
the statutes words as reasonably understood. Intentionalists in
the United States, more than in the United Kingdom, are willing to consult documents produced as the statute proceeded
through the enactment process (the statutes legislative history), such as reports by the committees that considered the
legislation and statements by the statutes supporters and opponents, to determine what the legislature meant by the terms it
used.
Textualists criticize these more expansive versions of intentionalism. Most narrowly, they note that materials drawn from
the legislative history are readily deployed strategically by advocates, who present only the materials that support the interpretation that will yield the result they favor, and selectively by judges,
who refer only to those parts of the legislative history that favor
that result the judges prefer for reasons independent of the interpretive enterprise. Critics also observe that referring to legislative history gives some degree of authority to committees and
individual members, whereas only the entire legislature has any
authority to enact law. Finally, critics question the coherence of
invoking intentionalist terms with respect to multimember legislative bodies. Some legislators might have favored the adoption
of a statutory provision because they thought that it solved an
important public policy problem, others because their constituents favored it, still others because important contributors to
their campaigns did so. How can these varying states of mind be
aggregated into an intent of the legislature? Continental legal
theorists elide this question by referring to the legislator in the
432
Legal Interpretation
Canons of interpretation can fit into each of the interpretive
approaches. For the textualist, the canons are part of the general
background the ordinary reader is assumed to know as he or she
reads a contested statutory text. The intentionalist can defend the
substantive canons on the ground that legislatures typically do
not intend to infringe on the policy goals embodied in substantive canons. The purposivist has an easy time with substantive
canons, which directly embody judgments about good policy,
and can treat the other canons as similarly reflecting good policy
judgments, rather than imputations of legislative intent.
Constitutional Interpretation
Questions of statutory interpretation typically involve specific
and detailed provisions of complex statutes. Constitutional
interpretation, in contrast, typically involves the application of
general and abstract constitutional terms, such as freedom of
speech, due process of law, and equal protection of the laws
(to take examples from the U.S. Constitution) to specific problems. Here, too, there are two primary families of approaches.
The first family includes varieties of originalism. One version holds that constitutional provisions should be interpreted
to conform to the intent of the constitutions adopters. As with
intentionalism in statutory interpretation, original-intent
approaches run into many difficulties, such as the problem of
aggregating individuals intentions. In addition, when, as with
the U.S. Constitution, major provisions were adopted two centuries earlier, the task of identifying what any particular individual
understood a provision to mean is extremely difficult.
Finally, the abstract terms that constitutions use pose an additional problem: Should the provisions be interpreted according
to the abstract or the concrete understandings of their adopters?
Ronald Dworkin uses the term concept to describe the abstract
understanding, conception the concrete one (Dworkin, 1977).
Consider, for example, a constitutional provision dealing with
equality. Does that provision enact into fundamental law the
particular understandings the adopters had about equality, such
as the understanding that laws could treat men and women differently while still providing equality, or does it enact equality
itself, that is, the best understanding an interpreter can devise at
the moment of interpretation?
Some proponents of originalism responded to these and
other problems with an original-intent approach by arguing that
constitutional provisions should be interpreted according to the
original public meaning of their terms. The interpreter should
identify what the terms meant to a reasonable and well-informed
member of the public when the provisions were adopted. This
does not completely eliminate the evidentiary problems associated with original-intent approaches, because one is still searching for what the terms meant to individuals, but it substantially
expands the range of relevant materials to include uses of the
terms in newspaper discourse and the like. Similarly, it shifts terminology about abstract versus concrete intentions to references
to abstract versus concrete public meanings.
The temporal problem is perhaps the most difficult one facing originalist approaches. The problem is that the adopters or
the general public years ago could have no intentions about, or
understanding of, what the terms they used meant, in connection with developments they did not and could not anticipate.
433
Lexical Acquisition
judges using originalism and perfectionism relatively unsystematically; the practice in Germany is more perfectionist, and that
in Australia is formalist.
Legal theorists have recurrently been attracted to the idea
that law, and legal interpretation, could become a science. The
direct invocation of sciences, such as linguistics, psychology, and
more recently neuroscience, to understand legal interpretation
has produced relatively little enlightenment, to the point where
it seems more likely than not that whatever science of law there
might eventually be, it will not be a science on the model of the
physical or biological sciences.
Mark Tushnet
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Dworkin, Ronald. 1996. Freedoms Law: The Moral Reading of the
American Constitution. Cambridge: Harvard University Press.
. 1977. Taking Rights Seriously. Cambridge: Harvard University
Press.
Ely, John Hart. 1980. Democracy and Distrust: A Theory of Judicial Review.
Cambridge: Harvard University Press.
Fish, Stanley E. 1994. Theres No Such Thing as Free Speech: and Its a
Cood Thing, Too. New York: Oxford University Press.
Goldsworthy, Jeff. 2006. Interpreting Constitutions: A Comparative Study.
New York: Oxford University Press.
Lasser, Mitchel de S.-O.-LE. 2004. Judicial Deliberations: A Comparative
Analysis of Judicial Transparency and Legitimacy. New York: Oxford
University Press.
Llewellyn, Karl. 1950. Remarks on the theory of appellate decision and
the rules or canons about how statutes are to be construed. Vanderbilt
Law Review 3: 395406.
Marshall, Geoffrey. 1997. What is binding in a precedent. In Interpreting
Precedents: A Comparative Study, ed. Neil MacCormick and Robert S.
Summers, 50317. Brookfield, VT.: Ashgate/Dartmouth.
Rumble, Wilfrid E. 1968. American Legal Realism: Skepticism, Reform,
and the Judicial Process. Ithaca, NY.: Cornell University Press.
Strauss, David A. 1996. Common law constitutional interpretation.
University of Chicago Law Review 63: 877935.
Vermeule, Adrian. 2006. Judging Under Uncertainty: An Institutional
Theory of Legal Interpretation. Cambridge: Harvard University Press.
LEXICAL ACQUISITION
Children start to produce their first recognizable words between
12 and 18 months of age, and typically understand more than
they produce. This asymmetry is a lifetime effect. The forms of
their earliest words often depart radically from the adult versions
(consider ga for squirrel) and may be hard to understand. But
as their pronunciation becomes more skilled (see phonology,
acquisition of), children add rapidly to the vocabulary at their
disposal. They add words for people, animals, everyday objects,
toys, food, and various activities and by age two generally produce between 200 and 800 distinct words.
Researchers have taken one of two main approaches to the
study of word acquisition in the last two decades: With the first
approach, they have postulated built-in constraints that would
limit the hypotheses children entertain about possible meanings, typically limited to noun meanings only (Markman 1989),
constraints that must later be overridden since they are incompatible with semantic relations, such as inclusion, overlap, and
434
Lexical-Functional Grammar
They take even longer to add a subject to the complement, as in I
want Anna to come. Constructions often appear to be built up on
single lexical items on a one-by-one basis. This takes time (Clark
and Kelly 2006; Tomasello 2003).
Finally, adult usage plays a crucial role in acquisition.
Children track the frequencies of constructions in parental
speech and acquire first those constructions that occur most
often. It is adult speakers who model word use, who offer children conventional terms for talking about types of objects,
activities, and relations. And it is adults who continually check
up with young children on just what meanings the children
intended to convey.
Eve V. Clark
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bloom, Paul. 2000. How Children Learn the Meanings of Words.
Cambridge, MA: MIT Press. This book reviews how meaning acquisition is linked to speaker intentions and theory of mind.
Clark, Eve V. 1993. The Lexicon in Acquisition. Cambridge: Cambridge
University Press. This book examines childrens word learning and
their ability to coin words to fill gaps in their current vocabulary.
. 2009. First Language Acquisition. 2nd ed. Cambridge: Cambridge
University Press. A review of first language acquisition, and how cognitive and social factors interact in a usage-based approach.
Clark, Eve V., and Barbara F. Kelly, eds. 2006. Constructions in Acquisition.
Stanford, CA: CSLI. This book reports on studies of how children
acquire constructions.
Markman, Ellen M. 1989. Categorization and Naming in Children.
Cambridge, MA: MIT. This book presents a constraints-based account
of lexical acquisition.
Tomasello, Michael. 2003. Constructing a Language. Cambridge: Harvard
University Press. This book presents a usage-based approach to early
syntactic acquisition (see syntax, acquisition of).
LEXICAL-FUNCTIONAL GRAMMAR
Lexical functional grammar (LFG) is what is known as a constraint-based parallel correspondence architecture for a theory of
language (Bresnan 1982, 2001; Dalrymple 2001). It was called lexical because certain relations between elements, like that between
an active and a passive verb, were dealt with in the lexicon, as a
relation between lexical items. This contrasts with the approach
in transformational theories. Functional ambiguously refers
to grammatical relations, which are prominent in the theory, and
mathematical functions, which are used in the LFG-formalism.
The LFG-formalism can be mathematically modeled and, hence,
analyses expressed within it are susceptible to computational testing (see Dalrymple et al. 1995). LFG is often referred to as a syntactic theory, but like many other syntactic theories, it is actually
a framework within which theories of language can be expressed.
As one of the founders of LFG puts it: [T]he formal model of LFG
is not a syntactic theory in the linguistic sense. Rather, it is an
architecture for syntactic theory. Within this architecture, there
is a wide range of possible syntactic theories and sub-theories,
some of which closely resemble syntactic theories within alternative architectures, and others of which differ radically from familiar approaches (Bresnan 2001, 43).
IP
DP
VP
435
Lexical-Functional Grammar
I takes an exocentric category S as its complement (for a discussion of the relevant data and potential constraints on the typological variation in constituent structure, see Nordlinger 1997).
F-structure is assumed to be reasonably invariant among
languages. It takes the shape of feature-value matrices, where
the features capture grammatical relations and functional features. The simplest features are those with atomic values, such as
[number plu]. Grammatical relations such as subj(ect), obj(ect)
or adj(unct) are represented as features that take f-structures as
their values. Each element that has lexical semantic content has
associated with it a feature pred, which has a semantic form as
its value. A verb such as tickle, for instance, has the feature-value
pair [pred tickle < (subj) (obj) >], that is, in the pred value of this
verb, tickle requires a subject and an object (for more details
on semantics within LFG, see especially Dalrymple 2001, 21754,
who develops an approach to semantic composition called glue
semantics). The pred feature also captures selectional properties,
which are based on functions and functional properties, rather
than on syntactic categories; a transitive verb selects for a subject
and an object, not for two noun phrases. An f-structure for the sentence Oscar tickled the cat can be found in (2).
(2)
PRED
SUBJ
OBJ
TENSE
)(
PRED Oscar
GEND masc
PRED cat
NUM sg
SPEC def
past
436
fill them from those that are restricted in this way; a subject, for
instance, can be associated with a large number of thematic
roles, whereas an oblique is restricted as to which thematic
role can fill it. LMT associates feature values to thematic roles
intrinsically for example, an agent is intrinsically associated
with [o] and by default for instance, the highest thematic
role according to a typologically motivated thematic hierarchy
is associated with [r]. Grammatical functions are then defined
in terms of these two features; a subject, for instance, is [-o] in
not being object-like and [-r] in not being restricted to any particular thematic role, whereas an oblique is [-o] but [+r]. Two
well-formedness constraints apply to the mapping; the function-argument bi-uniqueness condition which states that each
thematic role must be associated with exactly one function and
each function with exactly one thematic role and the subject
condition which states that every predicate must have a subject. For examples of how LMT works and how it can analyze
constructions such as locative inversion or complex predicates
in interesting ways, see, for instance, Dalrymple (2001, Chap. 8)
or Bresnan (2001, Chap. 14).
The mapping between c-structure and f-structure can be
structurally defined or identified through morphological elements. English is an example of a language wherein functions
are determined through their hierarchical position; the subject appears in the specifier of the IP. This is then captured
formally through the phrase structure rule in (3a), where the
up arrow should be read as the f-structure associated with
my mother node and the down arrow as the f-structure
associated with this node. The resulting tree can be found in
(3b), where indices have been inserted to identify f-structures;
the f-structure associated with the IP is referred to as f1, and
so on.
(3)
a.
IP f1
b.
DP
f2
(SUBJ)=
I' f 3
=
The arrows can now be replaced by the indices, and we get the
equations in (4).
(4)
f 1 SUBJ = f 2
f1 = f 3
These equations refer to three f-structures and define their relations; the f-structure f1 contains a feature subj which has as its
value the f-structure f2. The second equation states that f1 is identical to f3, which means that any featurevalue pair associated
with either of the two will also be associated with the other; the
two nodes IP and I will then be associated with one f-structure.
In fact, the categories that form the clausal backbone CP, IP
and VP will always share f-structure, so that featurevalue pairs
introduced to any of them will also be shared by the others. These
Lexical-Functional Grammar
categories are referred to as co-heads (Bresnan 2001, 102). The
equations in (4) give rise to the partial f-structure in (5).
(5)
SUBJ [ ] f
2
f
1 3
(7)
437
438
noun
/roz/
ROSE
During the past two decades, there has been remarkable progress in understanding the neural substrates of lexical processing,
mainly because of advances in two complementary approaches for
investigating the functions of specific brain structures: 1) the lesion
method, which, when used with ample numbers of patients who
are carefully studied both neuropsychologically and neuroanatomically, can yield indispensable insights about the neural systems that
are necessary for particular abilities; and 2) functional imaging techniques, such as fMRI, which allow researchers to identify with more
fine-grained spatial resolution the brain structures that are engaged
during the normal performance of certain tasks (see neuroimaging). Much more has been learned about the neural substrates of
lexical processing than can be summarized here, and so this review
concentrates on cortical regions that have been linked with the recognition and production of spoken and written word forms.
439
Figure 1. The LRM (i.e., Levelt, Roelofs, and Meyer) model of spoken word production. Left column: Word production tasks involving lead-in processes that enter the central word production architecture at different stages.
Middle column: Core processes of word production and their characteristic output. Right column: Example fragments
of outputs generated at each stage. Reprinted by permission from Elsevier, copyright 2004, from P. Indefrey and
W. Levelt, The spatial and temporal signatures of word production components, Cognition 92: 10144.
of the left middle temporal gyrus and typically occurs during a
time window of 150225 milliseconds (ms) post-stimulus in oral
picture-naming tasks (see Color Plate 4). They also found that
phonological form retrieval is linked with the posterior portions
of the left middle and superior temporal gyri and occurs during
a time window of either 200400 or 275400 ms, depending on
the studies that are considered. The three postlexical stages of
spoken word production are known to rely on a variety of motor-
440
the VWFA was reported by R. Gaillard et al. (2006; see also Martin
2006). In brief, prior to surgery for intractable epilepsy, the patient
exhibited normal single-word reading, including a lack of increase
in reading time for common words varying in length from three to
eight letters; moreover, fMRI revealed his VWFA to have normal
functional-anatomical characteristics, and local field potentials
recorded from implanted electrodes showed that this area was
sensitive to word frequency but not word length, again within
normal parameters. After excision of tissue just posterior to the
VWFA, the patients epileptic seizures were successfully eliminated, but his reading was markedly slow and inaccurate, with
reading times increasing linearly with word length (i.e., letterby-letter reading). In addition, the VWFA no longer responded
to printed words, even when they were contrasted with a simple
fixation point. This study, therefore, provides powerful new evidence that the VWFA is in fact necessary for access to the stored
orthographic forms of words during reading.
Writing also depends on a large network of widely distributed
brain regions (Hillis and Rapp 2004; Rapcsak and Beeson 2002).
Information about the neural basis of lexical access during written word production comes primarily from patients with lexical
agraphia, a disorder in which words with regular mappings
between phonology and orthography are spelled correctly, but
words with irregular mappings (e.g., choir) are misspelled. The
errors are usually phonologically plausible (e.g., circuit serkit)
and affect low-frequency words more than high-frequency
ones. Lexical agraphia is typically caused by damage to the left
temporo-parieto-occipital junction (Brodmann areas 37 and/or
39), although in some cases, there is involvement of the left ventral occipitotemporal region, close to, if not encompassing, the
VWFA. Several functional imaging studies with normal subjects
provide further support for a role of these cortical regions in written word production (e.g., Petrides Alvisatos, and Evans 1995;
Nakamura et al. 2000).
Conclusion
When people recognize and produce the spoken and written
forms of words, they usually concentrate on the meanings being
expressed and remain blithely unaware of the complex computations being executed by their brains in order to rapidly and
effectively process the lexical structures themselves. Cognitive
neuroscience is beginning to reveal the intricasies of these neural
systems, and dramatic advances are likely to happen in the coming years. Exciting new discoveries are appearing in the literature
almost daily, and this explosion of research will undoubtedly
provide fresh insights into the neurobiology of lexical processing, with significant implications for understanding and treating
disorders that result from brain injury.
David Kemmerer
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bohland, J. W., and F. H. Guenther. 2006. An fMRI investigation of syllable sequence production. NeuroImage 32: 82141.
Catani, M., D. K. Jones, and D. H. Ffytche. 2005. Perisylvian language
networks of the human brain. Annals of Neurology 57: 816.
Damasio, H., T. J. Grabowski, D. Tranel, R. D. Hichwa, and A. R. Damasio.
1996. A neural basis for lexical retrieval. Nature 380: 499505.
441
442
Lexical Relations
Price, C. J., and J. T. Devlin. 2003. The myth of the visual word form area.
NeuroImage 19: 47381.
Pulvermller, F., M. Huss, F. Kherif, F. M. del Prado Martin, O. Hauk,
and Y. Shtyrov. 2006. Motor cortex maps articulatory features of
speech sounds. Proceedings of the National Academy of Sciences
103: 786570.
Rapcsak, S. Z., and P. M. Beeson. 2002. Neuroanatomical correlates of spelling and writing. In The Handbook of Adult Language
Disorders: Integrating Cognitive Neuropsychology, Neurology, and
Rehabilitation, ed. A. E. Hillis, 71100. Philadelphia: Psychology
Press.
Rapp, B., and M. Goldrick. 2006. Speaking words: Contributions of
cognitive neuropsychological research. Cognitive Neuropsychology
23: 3973.
Shapiro, K., and A. Caramazza, A. 2004. The organization of lexical
knowledge in the brain: The grammatical dimension. In The Cognitive
Neurosciences. Vol. 3. Ed. M. Gazzaniga, 80314. Cambridge, MA: MIT
Press.
Skipper, J. I., V. van Wassenhove, H. C. Nusbaum, and S. L. Small. 2008.
Hearing lips and seeing voices: How cortical areas supporting speech
production mediate audiovisual speech perception. Cerebral Cortex
18: 243949.
Tomasello, M. 1999. The Cultural Origins of Human Cognition.
Cambridge: Harvard University Press.
Tranel, D., R. Adolphs, H. Damasio, and A. R. Damasio. 2001. A neural
basis for the retrieval of words for actions. Cognitive Neuropsychology
18: 65570.
Tranel, D., H. Damasio, G. R. Eichhorn, T. J. Grabowski, L. L. B. Ponto,
and R. D. Hichwa. 2003. Neural correlates of naming animals from
their characteristic sounds. Neuropsychologia 41: 84754.
Tranel, D., T. J. Grabowski, J. Lyon, and H. Damasio, H. 2005. Naming
the same entities from visual or from auditory stimulation engages
similar regions of left inferotemporal cortices. Journal of Cognitive
Neuroscience 17: 12931305.
Tranel, D., K. Manzel, E. Asp, and D. Kemmerer. 2008. Naming static and
dynamic actions: Neuropsychological evidence. Journal of Physiology,
Paris 102: 8094.
LEXICAL RELATIONS
Lexical relations are ways in which words, or lexemes, share
something in common. This broad definition includes relations
based on phonological relatedness, such as rhyming, and
morphological relatedness, like being the range of tensed
forms of a verb. However, in most contexts, the term is used to
refer specifically to semantic relations among words and, most
frequently, to paradigmatic semantic relations among words,
including synonymy, hyponymy, and antonymy. Such relations
are sometimes called sense relations (Lyons 1977), as it is usually
a single denotative sense of a word rather than every sense and
every aspect of the lexical item that is relevant to the relation.
Thus, we expect the postal, rubber, and stomp senses of
stamp to enter into lexical relations with different sets of words.
Lexical Relations
semantics build such relations into lexical entries, for example in the selectional restrictions of Jerrold J. Katz and Jerry A.
Fodor (1963) and the lexical functions of meaning-text theory
(Meluk 1996).
Paradigmatic relations are relations of substitutability; the
words in a semantic paradigm are different options for filling the
same phrasal slot. For example, red/yellow/blue are three options
for subject position in X is a primary color. Paradigmatic relations
are studied because of their role in logical relations among sentence meanings, such as entailment, and because of what they
might tell us about how the mental lexicon is organized.
synonyms, for example, when the same object is named independently by two subcommunities, either one word falls into disuse or one or both of the words become specialized to a slightly
different sense or context type. English, through its history of
contact with other languages, has come to have many such
near-synonyms, such as riseascend, smartintelligent, and dead
deceased. While these are substitutable for each other in many
contexts, they all differ in meaning and use. For example, while
a balloon could ascend or rise, a person rises, rather than ascends,
from a chair. And while people can be dead or deceased, plants
can only be said to be dead. Many other cases of not-quite-synonymy involve words that refer to similar, but not identical, things
for example, tapas and hors doeuvres or shovel and scoop.
The second problem with a truth-conditional definition of
synonymy is that it allows as synonyms items that have different non-truth-conditional content. For example, the nouns dog,
doggy, and pooch may all be truthfully applied to a particular
animal, but choosing pooch implies different things about the
animal and the speakers relation to it than dog does. Thus, some
apply a more restrictive substitution test that takes into account
connotational and social aspects of meaning, as well as the truth
conditional. To the extent that goodness of synonym relations
can be affected by non-truth-conditional issues like register
and morphological complexity, synonymy can be considered a
lexical relation, as well as a semantic relation.
Synonymy currently receives much attention in computational linguistics, as language generators and machine
translators require principled means for selecting the most
appropriate word for a context from a lexicon filled with nearsynonyms. Such studies can be particularly concerned with discerning ways in which synonyms or near-synonyms can differ
(see, e.g., DiMarco, Hirst, and Stede 1993).
HYPONYMY. Hyponymy, the type of relation, is the relation of
sense inclusion, although it is often defined in terms of referential, or categorial, inclusion. In sense terms, beer is a hyponym
of beverage because the sense of beer includes all the information that is in the sense of beverage, plus additional information
that identifies beers as special types of beverages. In categorial terms, everything that beer denotes is included in the set of
everything that beverage denotes. Hyponym relations are, thus,
linguistic reflexes of categorial inclusion relations, and, as such,
some theorists consider them to be semantic, but not lexical,
relations.
Unlike synonymy, hyponymy is an asymmetrical relation, in
that beer is a type of beverage but beverage is not a type of beer.
Hyperonymy is the term for the converse relation from beverage
to beer. Hyponymy is usually said to be a transitive relation. For
example, if beer is a hyponym of beverage and beverage is a hyponym of liquid, then beer is a hyponym of liquid. However, transitivity holds only in cases of proper sense inclusion not in all
cases that pass the is a type/kind of test (thus proving that the test
can be misleading). For example, speed-reading is a type of reading, and reading is a type of leisure activity, but speed-reading is
not usually considered to be a leisure activity. While reading can
function as a leisure activity, it is not defined in terms of being
leisurely, and thus reading is not a hyponym of leisure activity in
the logical sense of the term.
443
Lexical Relations
Inclusion relations, along with contrast relations, are central
to most approaches to lexical meaning, particularly in the treatment of noun meaning. For other word classes, hyponym relations are less frequent or less clearly paradigmatic. For example,
the adjectives happy and sad describe states that belong to the
category designated by emotion, but since emotion is a noun, it is
not substitutable for happy and sad.
ANTONYMY/CONTRAST. While perfect synonymy involves words
that always refer to the same thing, logical incompatibility
involves words that never overlap in reference. This logical relation contributes to two lexical-semantic relations, antonymy and
contrast.
Binary opposition seems to have special status in language
and conceptualization since even where more than one incompatible alternative is available, binary relations may be discerned. For example, although there are many emotions (happy,
sad, angry, afraid, disgusted), happy is generally understood as
having a single antonym, sad.
Logical approaches to antonymy distinguish various types.
Complementary (or contradictory) antonyms perfectly bisect
a semantic domain. For example, in relation to electrical
items, if something is not on, then it is necessarily off and vice
versa. Contrary antonyms designate the extremes of a scale.
For instance, something that is large is necessarily not small,
but something that is not small is not necessarily large. (Some
authors, including Lyons 1977 and Cruse 1986, restrict the term
antonym to just contraries.) Converse antonyms indicate different perspectives on the same relationship for example send/
receive, teacher/student, north/south. So, if X is north of Y, then Y
is south of X. Other types of antonymy have received some attention, but these tend not to show the same kinds of logical relations as those mentioned earlier. For example, kinship terms can
be opposed on the basis of gender (brother/sister) or generation
(mother/daughter).
In the broad sense of the term, antonymy is often defined as
a relationship of minimal difference; that is, antonymous words
share most of their semantic content, but for one difference that
makes the two terms incompatible. On such a definition, mother
and daughter are opposites because they only differ in the generation they refer to, while mother and son are not opposites
because they differ in both gender and generation.
As mentioned, the lexical relation antonymy is sometimes
contrasted to the semantic relation opposition, and antonym
pairs present the best evidence in favor of the position that relations between lexical items, not just senses, are represented in
the mental lexicon. Antonym pairings are particularly strong in
experiments such as word association tasks and lexical priming, and antonyms co-occur frequently in text leading some
(e.g., Jones 2002) to question whether antonymy is also a syntagmatic relation.
Larger sets of incompatible items exist, for example solid/
liquid/gas, but more linguistic-semantic attention is usually paid
to the not-necessarily-incompatible relation of co-hyponymy
or contrast, which, along with hyponymy, is basic to the taxonomic organization at the basis of semantic field and network
approaches. Such relations can also be defined in terms of minimal difference. For example, the basic color terms are similar
444
Lexical Semantics
in all being direct hyponyms of color but differ in the part of
the spectrum they designate. Still, they are not truly incompatible since they may overlap some shades of turquoise may be
considered to be both blue and green. The fact that most people
would insist that it must be one or the other, however, indicates
our preference for acting as if contrast sets are incompatible.
OTHER RELATIONS. The foregoing types of paradigmatic relation are generally held to be the most important to lexical and
semantic theories, but many more have been noted in the literature. The most discussed of these is meronymy, the part of
relation, though this can be thought of as a cover term for many
other types of relation, such as segmentwhole (slicecake),
materialwhole (woodtable), leaderorganization (captain
team), and so on.
Whether a precise taxonomy of relations is possible or necessary is an open question. While semantic field views of the lexicon rely on a small number of well-defined relations, theories
employing looser semantic networks or semantic domains might
allow for as many types of lexical relations as there are possible
relations among entities in the world. As the subtypes of antonymy and meronymy show, there is also the question of the level
of specificity that needs to be employed in representing these
relations in linguistic theory.
M. Lynne Murphy
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Clark, Eve V. 1992. Conventionality and contrast. In Frames, Fields, and
Contrasts, ed. Adrienne Lehrer and Eva Feder Kittay, 17188. Hillsdale,
NJ: Erlbaum.
Cruse, D. A. 1986. Lexical Semantics. Cambridge: Cambridge University
Press.
. 1994. Prototype theory and lexical relations. Rivista di Linguistica
6: 16788.
DiMarco, Crysanne, Graeme Hirst, and Manfred Stede. 1993. The semantic and stylistic differentiation of synonyms and near-synonyms.
Proceedings, AAAI Spring Symposium on Building Lexicons for Machine
Translation, Stanford, CA: 11421.
Gross, Derek, Ute Fischer, and George A. Miller. 1989. The organization of adjectival meanings. Journal of Memory and Language
28: 92106.
Jones, Steven. 2002. Antonymy. London: Routledge.
Katz, Jerrold J., and Jerry A. Fodor. 1963. The structure of a semantic
theory. Language 39: 170210.
Lyons, John. 1977. Semantics. 2 vols. Cambridge: Cambridge University
Press.
Meluk, I. A. 1996. Lexical functions: A tool for the description of
lexical relations in a lexicon. In Lexical Functions in Lexicography
and Natural Language Processing, ed. Leo Wanner, 37102.
Amsterdam: Benjamins.
Murphy, M. Lynne. 2003. Semantic Relations and the Lexicon.
Cambridge: Cambridge University Press.
LEXICAL SEMANTICS
Lexical semantics is often loosely described as the study of word
meaning, but both word and meaning require more precise
definition in this context. The term is usually used to describe the
study of lexical, or content, words or lexemes (including nouns,
Lexical Semantics
verbs, adjectives), rather than grammatical words (conjunctions,
determiners), which are more usually studied in the context of
sentential (or propositional) semantics. Lexical semantics can
also refer to the semantics of non-word lexical items, such as
idioms. The meaning aspect of lexical semantics most often
refers to denotative sense in particular that is, determining
what such words can and cannot refer to, as opposed to their
connotation or social import. Some of the main issues that concern lexical semanticists are the following:
How is the meaning of a word best represented in a model of
the mental lexicon?
Are different types of representation required for different
kinds of meaning?
How should multiple interpretations of a single word be
described and explained?
How are different words meanings related to one another?
445
Lexical Semantics
Lexicography
446
LEXICOGRAPHY
While lexicography is often thought of as a subfield of linguistics, it is a scholarly discipline in its own right with its own principles and practices. This discipline is divided into two subfields,
practical lexicography and theoretical lexicography. Practical
lexicography is concerned with compiling, writing, and editing
dictionaries, which serve a double function: as a record of the
vocabulary of the language and as a reference work to meet the
needs of users for information about words, their usage, and
their spelling. Theoretical lexicography is concerned with the
definition of general principles governing the compilation of
dictionaries.
Dictionaries differ in their selection of vocabulary and other
items that the editors believe merit inclusion, given the size
and purpose of the volume. While most dictionaries use alphabetized word lists, certain others such as Rogets Thesaurus of
English Words and Phrases are arranged by topic. It lists words
according to the ideas that they express, for example, abstract
relations (existence, relation, quantity, etc.), space (generally,
dimensions, etc.), and intellect (formation of ideas, communication of ideas, etc.), among others. Words and phrases are listed in
the main body of a thesaurus according to their word class, but
without a definition or any information about pronunciation or
etymology.
Dictionaries come in various formats. Besides traditional
print dictionaries, online dictionaries and dictionaries on solid
state media (such as CD or flash memory) have become increasingly popular during the last quarter of the twentieth century
because they facilitate rapid access to information, cross-referencing, and immediate updates with the latest vocabulary.
General-purpose monolingual dictionaries are organized
alphabetically and use the same language for both the object and
the means of description. While all dictionaries aim for comprehensiveness, the number and structure of lexical entries depends
upon the target audience, as well as constraints in funding and
time to complete the dictionary. General-purpose dictionaries
focus on the description of a standard language, aim to provide
an exhaustive coverage of the words in a language (abridged
dictionaries focus on somewhat shorter lists), and are typically
more linguistic than encyclopedic.
Dictionaries are usually divided into three parts: an introduction (including instructions on how to use the dictionary),
Lexicography
the body (the alphabetically ordered list of entries), and appendices (other information, such as weights and measures, punctuation, etc.). The arrangement of the entries of a dictionary
is referred to as its macrostructure. Dictionaries differ in the
placement of homonyms, derived words, compounds, and
phrases, which can be given independent entries or included
in an entry. The layout and organization of the individual entry
is referred to as the microstructure of the dictionary. Each dictionary differs in its conventions for structuring lexical entries.
Typically, the headword at the beginning of an entry is in bold
and indented by a few spaces. Bold or italic typeface may be
used to mark the part of speech at the beginning of a lexical
entry. Some dictionaries include the standard pronunciation of
headwords and spelling variants. When a word is polysemous,
its senses are often numbered, with the most frequently occurring sense first. Similarly, when a sense or a group of senses
belong to a different word class or subclass, the sense(s) are
labeled accordingly, in combination with definitions, examples, and usage notes. More technical, archaic, or obsolete
senses and idiomatic phrases usually appear toward the end of
the lexical entry.
The headword of a lexical entry consists of a lemma (the basic
word form) that conventionally represents all of the inflected
forms of the unit. Following the headword, lexical entries provide a definition of it in the form of a paraphrase (in the same
language), which is semantically equivalent to it. Lexicographers
aim to offer definitions that are simpler than the word itself.
Another goal is to avoid circularity of definitions, that is, to not
define two or more lexemes in terms of each other. The most
common type of definition is the analytic definition since it
aims at maximal inclusion and independence from the context.
Synonyms or antonyms are sometimes used as alternatives to
analytic definitions because they are short. However, they may
require the dictionary user to look up other definitions to understand their meanings. Depending on the scope and aim of the
dictionary, lexical entries also provide examples that illustrate
the headwords syntactic behavior or offer additional semantic
information.
Metalexicography has the goal to improve the structure
and content of dictionaries. One way of achieving this goal is
to be involved in research about lexicography. There are several professional organizations and academic journals devoted
to metalexicography. Another way is to suggest methods and
criteria for reviewing and evaluating dictionaries. Evaluating
and assessing dictionaries, also known as dictionary criticism,
is difficult because it is not always clear what types of criteria
should be applied. One solution to this problem has been to
take large sets of dictionary reviews and determine the range
of criteria applied by different reviewers in their evaluations.
These include reversibility, alphabetization, directionality,
coverage, reliability, currency, redundancy, retrievability, and
equivalents.
Another difficulty of dictionary criticism is the large number
of entries. Reviewers analyze only a small sample of entries or
focus on particular features of dictionaries. To begin with, they
typically focus on the preface of the dictionary that explains how
to use it, who the intended audience is, and what types of information are included in the main body. After flipping through the
Linguistic Relativism
dictionary to get an idea of the microstructure of lexical entries,
reviewers generally conduct an arbitrary sampling of dictionary entries. Depending on the time constraints, this procedure
may, for example, lead the reviewer to scrutinize every 10th main
entry on every 20th page for completeness, clearness, accuracy,
simplicity, and modernity. To ensure that the review is based on
a representative sample, it is necessary to check that the different parts of speech are adequately represented, that polysemy is
taken into account, and that there is a balance between words
from both the general and specific domains. In reviewing dictionaries and devising methods for improving the structure of
dictionaries, the reviewer notes the presentation of the text on
the page, as it plays a significant role in influencing the accessibility of the content. The range of vocabulary is also important
since users typically expect dictionaries to offer the latest words
from the domains of fashion, technology, and business, among
others, while at the same time including variants. Polysemy, the
structure of definitions, usage notes, examples, and etymological information are also important criteria for the evaluation of a
dictionarys content.
Hans C. Boas
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Atkins, Beryl T. S., and A. Zampolli. 1994. Computational Approaches to
the Lexicon. Oxford: Oxford University Press.
Bjoint, Henri. 2000. Modern Lexicography: An Introduction.
Oxford: Oxford University Press.
Hartmann, R. R. K., and G. James. 1998. Dictionary of Lexicography.
London and New York: Routledge.
Zgusta, Ladislav. 1971. Manual of Lexicography. The Hague: Mouton.
LINGUISTIC RELATIVISM
Linguistic relativism refers to the idea that language influences
thought and worldview (see also language of thought). In
essence, thinking and worldview are relative to the language one
learns to speak in childhood. Language, thought, worldview, and
influence are thus key concepts in linguistic relativism.
Various understandings of language have led to proposals for several levels of relativism, from semiotic relativity when
referring to the general faculty of language, to structural relativity when referring to the grammatical properties of languages,
and to functional relativity when referring to communicational
patterns of interaction within and across languages (Lucy 1997;
Hymes 1966). Most past and current research has concentrated
on structural relativity.
Several levels of thought have been posited as potentially
under the influence of language, including the neurological, cognitive, and propositional levels. To date, little work has
investigated neurological variation across speakers of different
languages (Gilbert et al. 2006). Most work addresses conceptualization by examining cognitive processes, such as memory,
categorization, inference, analogy, and emotions (Lucy
1992; Levinson 2003; Boroditsky, Schmidt, and Phillips 2003).
Very few studies, if any, have proposed that language may influence the actual content, or propositional level, of thought. This
idea, known as linguistic determinism, is commonly discredited
as scientifically untenable.
447
Literacy
Concerning the relationship between language and thought,
language has variably been suggested to influence, impact,
shape, mould, condition, limit, or channel thinking.
An important issue, then, concerns the scope of these effects.
Current research has been asking whether language effects are
on line in the process of producing and comprehending language (see Slobin 1996 on thinking for speaking) or whether
language effects pervade human cognition (i.e., effects exist
whether or not one is engaged in linguistic acts). Most research
has assumed the latter type of language effects on thought.
Finally, note that linguistic relativity was originally formulated
as a scientific principle by Benjamin Lee Whorf in 1940 (1956,
214, 221). The principle has since then been relabeled the SapirWhorf hypothesis, following an article by Harry Hoijer (1954)
referring to Edward Sapir and Whorf, who contributed to the
early development of linguistic relativity in the 1920s and1930s
(Sapir 1985; Whorf 1956).
Stphanie Pourcel
LITERACY
Prominent Research Frameworks
Written symbolic codes of languages may be alphabetic, syllabic, morphosyllabic, (Perfetti 2003; Ho et al. 2007), or alphasyllabic (Mishra and Stainthrope 2007). In each case, children
must first learn how the written code of their language embodies
spoken language units (Perfetti 2003, 17). Children must also
have language socialization experiences that promote thinking
and talking in more literate ways if they are to achieve academic
language proficiency (Wilkinson and Silliman 2000) or linguistic literacy (Ravid and Tolchinsky 2002). Contemporary research
frameworks and their related studies reflect different aspects
448
Literacy
of syntax (see syntax, acquisition of), to language variation and second language acquisition. Literacy learning
processes are also subjects of developmental language study
from differing viewpoints. Subareas include phonological
awareness, reading, composition, spelling, distinctions
between the spoken and written communication modes, and the
effectiveness of teaching reading and teaching writing
in educational programs that, internationally, span alphabetic
and nonalphabetic languages. Furthermore, literacy research
has expanded to incorporate neuroimaging in order to identify neurobiological correlates of dyslexia and effects on brain
function of scientifically based reading (SBR) interventions.
Behavioral studies have examined associations among oral language impairment, dyslexia, and text comprehension and related
disorders of reading and writing (for reviews, see Cain
and Oakhill 2007; Scarborough 2005).
Modern History
Research on literacy learning is relatively new, initiated in
the 1970s. Since its inception, one major principle has guided
this work: Children should have significant home and school
opportunities for the integration of oral and written language
experiences. These experiences support the development of literate stances in comprehension, speaking, reading, and writing.
Becoming literate is a social process, influenced largely by childrens search for meaning. Prior to school-based reading and
writing, childrens engagement in literacy-like actions in play,
such as scribbling on paper with crayons, and with adults, for
example, storybook reading, forms the foundation for later literacy learning.
Initial research on literacy was strongly shaped by studies of
classroom language. Nearly four decades ago, sociolinguists
launched a new direction for inquiry into language and literacy
learning, focusing on oral language use in classrooms. The first
research concentrated on language functions, the communicative demands of classrooms, individual differences, and the
social basis and social integration necessary for learning. Initial
reading studies addressed assessments of reading comprehension, whereas current studies emphasize effective instructional
models of decoding and reading comprehension. In the United
States with the 2001 passage of the No Child Left Behind Act
(NCLB), federal educational policies for the first time exerted a
profound influence on the way that beginning reading is taught.
The expectation was that explicit SBR instruction would guide
reading practices and curriculum development in phonemic
awareness, decoding, and fluency.
A broader array of language-related features is implicated in
more literate spoken and written language uses beyond phonemic awareness, however. These include the scope and density of
vocabulary knowledge, command of more advanced syntactic
constructions applied to diverse reading and writing purposes,
and familiarity with a variety of narrative and expository dialogue structures and their organization. Non-language factors
also contribute, such as working memory capacity for different kinds of information-processing demands, the motivation to
learn, inferencing and integration, and metacognitive strategies
for the self-regulation of literacy learning. The causal relationship of these language and non-language-related variables and
their interactions for proficient literacy learning remain unidentified (Cain and Oakhill 2007).
449
Literacy
Literariness
450
LITERARINESS
This term refers to the perceived distinctive quality of the language of literary, as opposed to nonliterary, texts. If the linguistic study of literature attempts to understand how linguistic
form is adapted to literary purposes, then identifying a text as
literary based on its language is one of the central problems.
There is currently among scholars of literature, both linguists
and literary critics alike, little agreement about the status of literature as language. Despite this lack of consensus, or maybe
because of it, the investigation of literature using linguistic
models has become a productive field within applied linguistics. The question of literariness is closely tied to the modern
concept of literature and its history, and so it seems best to proceed by examining first the history of the concept of literature
before turning to central issues that the question of literariness
has raised.
Although most cultures consider verbal art a separate,
recognizable class of speech, its rendering into print has subjected it to the transforming effects that mark the influence
of print on every aspect of modern culture (see print culture ). Literature (from Latin littera letter) in its restrictive, modern English sense of imaginative writing in the main
genres of poetry, prose, and drama which has claim to consideration on the ground of beauty of form or emotional effect
(Oxford English Dictionary), arose in the nineteenth century in
Literariness
conjunction with the increasing availability of authorship as a
profession. Before the advent of movable type printing, to write
or copy a book required a great investment of time and energy,
and so only highly valued items would have been widely circulated. Even after print technology began slowly to diffuse into
the wider culture, it was never doubted that only important
works would ever be printed, distributed, and saved. The rise of
industry and the middle class, with increased literacy rates and
an expanded market for writing, saw the normative concept of
literature emerge essentially as a means for differentiating traditionally sanctioned texts from those of supposedly ephemeral
quality. The reading and study of the superior texts were promoted in secondary and university curricula as profitable for
cultural improvement and development of national identity
(see nationalism and language ).
Arguments for the suitability of literature for this project
depended for their success on demonstrating that the privileged
texts possessed certain inalienable qualities. Matthew Arnold,
an English school inspector and poet, argued in 1880 that the
superior character of truth and seriousness in the matter and
substance of the best poetry, is inseparable from the superiority
of diction and movement marking its style and manner ([1880]
1988, 416). For Arnold, and the liberal humanism he has come to
represent, the reading of literature functions to stabilize society because it educates citizens in supposedly universal sociobehavioral norms.
Like Arnold, most influential theorists in the first half of the
twentieth century never questioned the status of literature as
an identifiable and beneficial form of linguistic behavior. They
expected that the literary text could be differentiated from nonliterary texts by some constellation of intrinsic linguistic characteristics, the discovery of which, Roman Jakobson argued in
1921 when he coined the term literariness, should be the goal of
literary linguistics. But when formalist investigators failed over
time to adduce a convincing set of characteristics necessary and
sufficient for identifying literature, attention turned to the role of
extrinsic factors, such as audience and medium, in establishing
literary distinctiveness. Many literary critics have resolved the
problem by historicizing the concept of literature itself, arguing
that the search for intrinsic features determinate of literariness
cannot be successful because wherever or whenever the category of literature arises, it does so from specific sociocultural
forces that situate readers differently with regard to the purposes
attributed to the literary texts within the cultural or theoretical
discourse that promotes the concept. Approaches grounded in
linguistics, however, prefer the term verbal art for their subject
because they also question the sufficiency of the traditional concept of literature to account for the variety of genres developed
around the world, by principally oral cultures, that function in literary ways for the culture in question. Most contemporary study
can thus be characterized as interactionist, refusing to privilege
intrinsic or extrinsic characteristics but understanding textual
genre categorization as a complex process involving interaction
among text, reader, performance situation, and various sociocultural practices.
Some intrinsic characteristics that have been recognized
as occurring across time and languages in texts that become
literary include iconicity and defamiliarization. Iconicity is an
important element in the organicist metaphors of literary structure developed principally by Robert Penn Warren and other
so-called New Critics in the United States, though by no means
unique to them. Organicist approaches see the form of a text as
highly responsive to its meaning, and the New Critics especially
valued the ability of a text to reconcile within itself the various
strands of meaning that its language evokes. In semiotic terms,
an icon is a sign in which the signifier somehow resembles its
referent. Iconicity operates to a limited extent in language, for
example, in onomatopoeia, where a word sounds like the thing
it represents, as is often the case for linguistic representations
of animal sounds English cows moo and sheep baa (but
see Haiman 1985 for arguments that iconic signs are widely
exploited in language). In poetry, sounds sometimes mimic the
thing being described, as in these lines from Tennysons The
Princess:
The moan of doves in immemorial elms
And murmuring of innumerable bees (7.2212)
The final line of the stanza does not rhyme, nor does it match the
iambic rhythm established in the first four lines. The disorder
spoken of in the line is acted out by the lines lack of sonic fit,
implying in the process a parallel between the form of the prayer
and its failure to penetrate. In the final line of the final stanza the
prayer at last comes around:
O cheer and tune my heartlesse breast,
Deferre no time;
That so thy favours granting my request,
They and my minde may chime,
And mend my rime.
When the speakers mind is aligned with Gods wishes, his prayer
also becomes formally complete. Iconicity is one way that patterns of linguistic elements can contribute to larger patterns
of meaning. Poetic iconicity is a local phenomenon, however,
which depends on the immediate environment of the signs
involved. The phrase and disorder is not without rhythm, nor is
it unrhymable, but in the context Herbert has created, the phrase
stands out for its ill fit. Organicist approaches value the extent to
which the local patterns of significance can be reconciled to one
another, making each poem a coherent emotional and semantic
whole. In the final stanza of Deniall, the local pattern of end
rhyme acts out an additive semantic logic, so that the notions
of chime and time together constitute the meaning of the word
rime, and at the same time the words function sonically within
451
Literariness
the rhyme scheme to act out the metonymy whereby my rime
refers to the poem as a whole.
The ideological content for the New Critical emphasis on
unity and coherence in literature is illuminated by comparison
to Viktor Shklovskys influential theory of ostranenie defamiliarization. Shklovsky was one of several scholars and writers who
met informally in Moscow and St. Petersburg. Often referred to
as the Russian Formalists, this group was the first to mix literary
analysis with an increasing awareness of the status of the literary text as a linguistic object and, therefore, subject to description by specifically linguistic tools (see Erlich 1955). Shklovsky
argued that the distinguishing characteristic of the literary text
was that it defamiliarized items or events the perception of which
had become automatized by the reader due to familiarity or
repeated exposure. Leo Tolstoy, as Shklovsky points out in several examples, makes the familiar seem strange by not naming
the familiar object, like the description in War and Peace of an
opera scene as pieces of painted cardboard, or when the narrator of Shame describes the sequence of actions in a flogging
without using the word ([1917] 1988, 21). Shklovsky called these
and similar techniques of creative distortion prim devices,
and Jakobson went so far as to claim in 1921 that if literary history wants to become a science, it must recognize the artistic
device as its only concern (quoted in Erlich 1955, 57; see also
foregrounding).
The young Jakobson, whose specifically linguistic theory of
poetics eventually became widely influential, believed that he
could justify his admiration for the futurist poetry of Velimir
Khlebnikov by explicating in linguistic terms the complex, suggestive, phonemic, and morphemic patterning of poems heavy in
neologisms, like Kuznechik The Grasshopper (Jakobson 1987,
252). Khlebnikovs poetry, and that of other so-called Russian
Futurists, so clearly eschewed traditional forms of poetic practice that there was no question of its being considered canonical. Like Arnold, the Futurists were interested in distinguishing
true literature from the products of the mass marketers, whom
they called traitors, and they also identified a strong nationalistic purpose in reading and writing literature. Unlike Arnold,
however, they saw literature as having a revolutionary, rather
than a stabilizing, purpose (influenced perhaps by the differing cultural conditions prevailing under the unstable czarist
regime in Russia and the imperative to govern as a global power
in England).
As Jakobsons poetics matured, he increasingly saw the role
of the device in terms of a Peircean semiotics. The devices do not
function alone to interrupt the direct awareness of the identity between sign and object (1987, 378), but the whole text is
revealed as a system of systems of equivalences, that, through
similarities and contrasts at all levels of linguistic organization,
up to and including the arrangement of the entire text, display
the text as primarily interested in the linguistic medium, the
materiality of its linguistic signs. This approach resembles the
organicist metaphors of the New Critics in that linguistic form
is motivated by poetic meaning, and devices are valued less for
their interruption of usual relationships of significance than for
the surplus of meaning that iconic relationships create.
In Jakobsons best-known literary-critical essays, however,
such as Baudelaires Les Chats, written with Claude Levi-
452
Literariness
among text, reader, and various sociocultural practices. Reuven
Tsurs sophisticated analysis of religious trance poetry, for example, identifies not only textual devices that contribute to a hypnotic
reading but also different types of cognitive styles of individual
readers, who are psychologically predilicted to react to certain
environmental variables in different ways and so read texts differently. Nigel Fabb, on the other hand, adopts a modularist
approach, identifying two types of structures in poetry, inherent
and communicated structure. Inherent structure is determinate
and is processed, rather than interpreted, by the appropriate
module (such as syntax or meter). Communicated structure is
implied by textual evidence in the context of the readers knowledge, so that being a sonnet holds of a text by inference, rather
than as an independent fact about the text. The important thing
for literariness is again that the text provides only some of the
cues necessary for categorizing it as literary. Communicated
structure is dependent on individual learning, so that responses
to any given textual characteristic might readily vary by individual. Fabb also argues that literary texts communicate more
about their form than verbal behavior generally does, and so in
a Jakobsonian sense draw more attention to form. Form in this
sense takes on the characteristics of meaning (because implied);
therefore, form is more likely to be ambiguous, indeterminate,
metaphorical, or ironic in literary texts.
Emotional responses to literature have always been important for the identification of literariness. Indian theories of
rasa, for example, categorized texts by the type of emotion they
prompted (see dhvani and rasa). Fresh empirical approaches
to the issue of literariness have focused on the role of emotion in
literary response. Miall identifies literariness as a combination
of formal elements in the literary text and the array of responses
these initiate when read (2006, 144). Readers who encounter
foregrounding devices experience them as defamiliarizing, a
cognitive response that has an associated affective dimension.
Subsequent changes in understanding seem to be guided largely
by the feeling evoked by defamiliarization, especially the feeling
of self-modification that, Miall argues, accompanies the recontextualization of the defamiliarized concept (see literature,
empirical study of).
In the last decade, literary critics interested in the ethical
dimensions of literary reading have begun to reconsider, to some
degree, the importance of the defamiliarizing effect of literature.
Derek Attridge (2004) has recuperated the concept of literariness
for literary studies by identifying it with the inventiveness that a
text manifests whenever it is enacted by a reader. For Attridge,
literariness inheres not in some fundamental unchanging core in
the work but in the inventiveness it shows over time, because it
remains open to change and porous to new contexts, continuing
to introduce what is unknown into the known. Readers experience that inventiveness anew when they perform the work, and
they desire to do justice to it by shifting the norms and habits
they use for dealing with the world.
As literature undergoes remediation into digital formats (and it should be noted that the Russian Formalists were
applying Shklovskys concept of ostranenie to film analysis
already before 1920), the notion of literariness is undergoing concomitant refinement. Motivated in part by new media
forms of verbal art, Jerome J. McGann and Lisa Samuels
453
454
Literary Universals
or narrative universals, which is modeled in part on the
study of language universals, focuses more on the construction
than the reception of texts. Such approaches tend to treat character types as narrative components that carry out action necessary for certain prototypical plot structures (see prototype). At
the same time, writers adopting this approach (e.g., Hogan 2003)
also consider the emotional effects elicited by different forms of
characterization in terms of empathy and identification. In addition, recent work by such writers as Toolan suggests that certain
contemporary grammars might provide better models of character types than the early generative grammar and transformational grammar upon which the theories of classical
narratology relied. For example, Toolan examines characters in
terms of the meaning-oriented grammar detailed by Michael
Halliday, which considers the different types of participants
in actions. These participants include a medium (the affected
participant), agent (a participant acting intentionally), force
(an inanimate agent), instrument (participant controlled by the
agent), beneficiary, and recipient (see also thematic roles).
Karen Renner
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bal, Mieke. [1985] 1997. Narratology: Introduction to the Theory of
Narrative. Toronto: University of Toronto Press.
Culler, Jonathan. 1975. Structuralist Poetics: Structuralism, Linguistics,
and the Study of Literature. Ithaca, NY: Cornell University Press.
Contains an excellent survey of structuralist approaches to narrative,
including a section on character, pp. 2308.
Gerrig, Richard J., and David W. Allbritton. 1990. The construction of literary character: A view from cognitive psychology. Style 24: 38092.
Greimas, A. J. [1966] 1983. Structural Semantics: An Attempt at a
Method. Trans. Daniele McDowell, Ronald Schleifer, and Alan Velie.
Lincoln: University of Nebraska Press.
Herman, David. 1997. Scripts, sequences, and stories: Elements of a
postclassical narratology. PMLA 112: 104659.
Hogan, Patrick. 2003. The Mind and Its Stories: Narrative Universals and
Human Emotion. Cambridge: Cambridge University Press.
Propp, Vladimir. [1928] 1968. Morphology of the Folktale. Trans. Laurence
Scott. 2d ed. Austin: University of Texas Press.
Schleifer, Ronald, and Alan Velie. 1987. Genre and structure: Toward an
actantial typology of narrative genres and modes. Modern Language
Notes 102: 112250.
Schneider, Ralf. 2001. Toward a cognitive theory of literary character: The dynamics of mental-model construction. Style 35: 60740.
Toolan, Michael. [1988] 2001. Narrative: A Critical Linguistic Introduction.
2d ed. New York: Routledge.
LITERARY UNIVERSALS
In parallel with language universals (see laws of language,
universal grammar, and universals, nongenetic,
among other entries), literary universals are, generally speaking,
patterns and structures exhibited widely by works of literature
across various familiar literary boundaries, whether national,
generic, or historical. Literary universals apply to various domains
besides world literature as a whole: individual regional and
national traditions; discrete literary forms and genres (poetry,
drama, narrative, etc.); separate literary histories and periods;
foundational literary concepts (plot, character [see literary
Literary Universals
character and character types], etc.); and common literary devices (e.g., metaphor), among others. Literary universals
may also be present as correlations across these various scalar
domains. Thus, Aristotles observation that tragedies have a
beginning, a middle, and an end applies to literary works more
broadly.
A specific call for literary universals appeared in Hogan (1994;
as a subtype of aesthetic universals), but the general concept
can be traced to Goethe, if not Aristotle. The Aristotelian binaries of universal/particular and substantial/nonsubstantial,
when intersected, generate a four-category ontology (see Lowe
2006). Literary universals include both nonsubstantial universals (abstract properties and relationships) and substantial universals: literary kinds (genres) and literary morphology, very
generally considered. Examples of the latter include metrical
analysis, biblical form criticism, thematics of creation stories,
Freytags pyramid, Proppian functions, Bakhtinian speech
genres, and so on. Substantial universals implies something
very different from the oxymoron concrete universals, sometimes invoked by New Critics and others, to suggest the possibility that literature can entirely transcend the universal/particular
dichotomy. Even when substantial (i.e., measurable, able to
be cataloged), literary universals work outside of any particular
instantiation or touchstone.
This entry briefly discusses five areas of literary universals: their
rationale and origins, some basic terminology, their relationship
to dominant strands of literary theory, various successful findings and limitations, and possible lines of future investigation.
The focus remains on universals as they apply directly to literary study, as opposed to universalist models conceived as more
purely linguistic or cognitive, and which merely invoke literary
terms (e.g., story grammar). In contrast, literary universals
seek to explore what unifies the incredible richness, beauty, and
diversity of global literatures, past and present.
Terminology
The basic vocabulary of literary universals stems from parallels in the theory of language universals. Absolute universals
are those that apply to all literatures, past and present. As with
absolute universals of language, these may be few in number
and difficult to substantiate completely since available information about both the languages and literatures of the world is far
from complete. Nevertheless, some absolute literary universals
do appear to exist. One simple absolute universal is that literature (including oral literature or orature) occurs in all known
cultures. Whether this is historically monogenetic, like language
in Homo sapiens sapiens may be, or polygenetic, like the invention of writing systems certainly was, remains an open question since oral literature long precedes the written record (see
oral composition). Another possible absolute is that all
literatures (eventually) develop fundamental generic differentiations, such as between poetry and prose (see poetic form,
universals of). Another content-oriented absolute is the universality of myth (stories of creation, flood, etc.) in the earliest
recorded traditions. Often, there are striking parallels between
very specific elements among even the most areally and genetically distinct myths, which may lend credence to the existence
455
Literary Universals
of a monogenetic mother literature. For instance, it is possible
to reconstruct a protoline of Indo-European epic poetry like he
killed the dragon (Watkins 1995), suggesting that heroic tales
are a common origin for global literatures.
Of course, common origin may imply prevalent rather than
across-the-board. If universals are not absolute, they are statistical, that is, occurring more often than chance alone would predict. The common distinction between poetry, drama, and fiction
is a statistical universal because these forms are widely but not
universally distinguished in the literary traditions of the world.
Universals that correlate (in ways also not influenced genetically or areally) are typological universals (see typology).
One typological universal may be that if a tradition has a category for nonfiction, it will also (n.b. the awkward non-) have
a category for fiction, or that drama presupposes poetry.
In other words, such literary categories may function like basic
color terms (Berlin and Kay 1991), with traditions that differentiate fewer kinds of literature, including the same few kinds.
Logical universals are typological universals that are logically
entailed by the nature of the given literary phenomena: Thus, a
narrative has only two options for recounting a plot sequence,
either temporally or out of that temporal order (e.g., a flashback).
This suggests, in turn, a less obvious but important statistical universal: very few plots are atemporal, far from the half expected by
random distribution.
Above all, literary universals are empirical they can be
(coldly) documented and are not the products of one cultural
point of view imposed upon another. Thus, any of the common
usages of universal(ity) in literary studies that imply normative, hegemonic, or totalizing judgments do not pertain to literary universals as discussed here but are instances of critical
contamination that parallel genetic and areal influence. (For
further discussion of this issue, along with other terminology of
literary universals, see Hogan 2006.)
456
concrete object of the work of art will elude our grasp (1977, 18).
In short, universal formulas have little purchase in any individual act of literary criticism. Yet, one page later, the fundamental
need for a universal theory reasserts itself: Like every human
being, each work of literature has its individual characteristics;
but it also shares traits with humanity. [Thus, the] characterization [of its individuality] can be accomplished only in universal
terms, on the basis of literary theory (1977, 19). The universal/
particular paradox of literary studies was already in place at the
fields birth around the turn of the twentieth century, when the
nomothetic/idiographic divisions of the German university were
dominant, and all disciplines were preconceived in the category
of Wissenschaften, sciences.
Linguistic approaches to literature, inaugurated by Ferdinand
de Saussure alongside the birth of modern structural linguistics, also imply universals. Though an odd literary digression
from his linguistic theory, the mysterious anagrams Saussure
culled from Latin texts might serve to show that literature
involves an equally systematic (and compensatorially non-arbitrary?) selection of signs (even if authorially unintended; see
Starobinski 1979). In any case, the literary situation is an instance
of language, and thus the linguistic universals of the moment of
communication will apply and will also be available for artistic exploitation. Roman Jakobson (1960) famously delineated
how each of six components of the communicative situation
(addresser, message, contact, etc.) could be exploited for different linguistic and literary purposes (emotive, poetic, referential,
etc.), with different purposes predictably dominant in different
literary genres. Among other brilliant readings, Jakobsons perspicacious exploration of the neuropsychology of the metaphoric
and metonymic poles of linguistic competence in relation to
the divergent styles of various Russian novelists (1956) also suggests that literary universals (pace Wellek and Warren) can illuminate seemingly idiographic cases in literary criticism.
The coming of various posts in late-twentieth-century literary theory (poststructuralism, postmodernism, postcolonialism) has added new complexity and urgency to the search for
universals in literature (see Carroll 1995). Even while prioritizing irreducible differences, polyvalent identities, and the local
particulars of each text and reader, current literary theory nonetheless proceeds by placing literature in the context of grand
sociopolitical, economic, or linguistic structures.
Literary Universals
begin to construct a universal metrics (1981, 266). Metrical
schemes are by nature universal formulas for generating and
judging individual poems: [W]e can only begin to state invariant facts about the iambic pentameter line if we state them in
terms of an abstract representation of the line, rather than by reference to any of the actual performances of the line (Fabb 2002,
6). Although the stanza has received some attention, the focus of
prosody is the verse line (or line group, e.g., the elegiac couplet), affording a vast corpus for analysis.
Other suggestive work on literary universals has been done
on the level of fundamental literary concepts. Evolutionary psychology has conflated Darwinian survival with literary conflict
(the primary problem or adversary to be overcome in a given
story), but this is only one of the bases of literature, being limited
primarily to narrative or dramatic works. In poetry, probably the
most ancient literary form, the dominant literary universals seem
to be imagery (regarding content) and meter (regarding form).
Some early formalists believed that imagistic language was the
root of all literature. However, literature is animated by both
imagistic tropes and also schemes (recalling the classical distinction within the rhetorical canon of style), that is, the deliberate
patterning of existing linguistic components (i.e., multitudes of
parallelism: rhyme, chiasm, ascending cola, etc.). Schemes
are, in Jakobsons vocabulary, either poetic or phatic (i.e.,
at play with the message itself or its instantiation in a linguistic
code), rather than referential, iconic, or invoking deliberate
visualization.
Tropes have found other universalizing uses since Peter
Ramus, Giambattista Vico, and other Enlightenment rhetoricians elevated four of these traditional figures of thought to the
status of master tropes: metaphor (see metaphor, universals
of), metonymy, synecdoche, and irony. Of these, metaphor
has been most ascendant (see, e.g., Kvecses 2005). The cognitivist theory of conceptual metaphor even suggests that the
fundamental metaphoric principle, to understand one thing in
terms of another, applies to all concepts, since the mind is universally embodied through the constant real-world experience
of comparison (Lakoff and Johnson 1999). Kenneth Burke took a
specific form, the drama, as a metaphor of all social action ([1945]
1969). Going a step beyond both Aristotles triad and Jakobsons
six factors of communication, Burkes pentad of dramatism
(scene, act, agent, agency, purpose; with a sixth term, attitude,
supplemented later) was conceived synecdochically so that its
terms could pair into any of ten ratios (e.g., the pathetic fallacy
is an instance of the scene-act ratio).
Even as previously literary things as metaphor and drama
show great power for comprehending human thought and society, however, there remains the problem of the way that such
findings feed back to illuminate literature per se. If we all use metaphor continually in language, what then sets apart the poetic
metaphor of a Shakespeare sonnet? Is all literature ultimately
dramatic? As the study of literary universals moves forward, it
must be remembered that ever since Saussurean structuralism
and Russian (and other Eastern European) formalisms, that the
fundamental theoretical agenda is literariness (literaturnost), rather than the particular literary work (which remains the
realm of criticism) the langue of literature, as it were, rather
than the parole.
Future Directions
As the aforementioned paths are furthered, two broad new avenues of investigation present themselves. The first has to do with
defining and integrating the fundamental disciplinary unit of
analysis of literature. In each discipline, the unit of analysis drives
the research paradigm, such as event and its time-context for
history, or the atom and its forces in physics. Literary studies are
blessed, or cursed, with an array of merely informative terms,
such as character, theme, genre, and reading, of which none by
itself provides the overarching anecdote for literary studies. In
practice, moreover, the field is divided among such competing
intradepartmental interests as criticism, theory, creative writing,
rhetoric and composition, education, film studies, and journalism. There may nonetheless be a grand theory of literariness
that unites these disparate subfields; if so, it would also likely
reveal the fundamental interconnections among all the sister
arts (literature, visual art, performing art, and new media). Will
there be a new grammatike, the ancient word for literary study,
now a more limited purview of linguistics?
The second little-explored territory for literary universals
(even since Carroll 1995) is the area of diachronic universals
(cf. language change, universals of). For instance, do all
literatures begin with poetry and then proceed to prose? Does
myth typically diverge into history (fact) and fiction (imagination)? Are there other universals of the rise and fall of various
genres over time? The word tradition, often used in this entry as
synonymous with a literature, implies an entire literary history,
and the historical mode is one of the literary disciplines oldest
and largest strands. Much of this standing evidence might be
mined for diachronic literary universals.
Christopher M. Kuipers
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Berlin, Brent, and Paul Kay. 1991. Basic Color Terms: Their Universality
and Evolution. 2d ed. Berkeley: University of California Press.
Brown, Donald E. 1991. Human Universals. New York: McGraw-Hill.
Burke, Kenneth. [1945] 1969. A Grammar of Motives. Berkeley: University
of California Press.
Caillois, Roger. 1961. Man, Play, and Games. Trans. Meyer Barash. New
York: Free Press.
Carroll, Joseph. 1995. Evolution and Literary Theory. New York: Cambridge
University Press.
Dissanayake, Ellen. 1992. Homo Aestheticus: Where Art Comes from and
Why. New York: Free Press.
Fabb, Nigel. 2002. Language and Literary Structure: The Linguistic
Analysis of Form in Verse and Narrative. Cambridge: Cambridge
University Press.
Hogan, Patrick Colm. 1994. The possibility of aesthetics. British Journal
of Aesthetics 34.4: 33750.
. 1997. Literary universals. Poetics Today 18.2: 22349.
. 2003. The Mind and Its Stories: Narrative Universals and Human
Emotion. New York: Cambridge University Press.
. 2006. What are literary universals? Literary Universals Project.
Available online at: <http://litup.unipa.it/docs/whatr.htm>.
Jakobson, Roman. 1956. Two aspects of language and two types of aphasia disturbances. In Fundamentals of Language, ed. Roman Jakobson
and Morris Halle, 5582. The Hague: Mouton.
. 1960. Closing statement: Linguistics and poetics. In Style in
Language, ed. Thomas Sebeok, 35077. Cambridge, MA: MIT Press.
457
Experimental Examples
In this section I discuss three themes that have been pursued
empirically and provide examples of the ways that readers
responses have been studied.
The term foregrounding refers to stylistically distinctive
aspects of literary texts. These may be apparent at the level of
458
Prospects
Although studies of the kind just described are not yet well
known among literary scholars, they point to a basis for rethinking the nature of literary studies and education. In contrast to
claims voiced by critics from I. A. Richards to E. D. Hirsch that
the responses of ordinary readers are ill-informed or whimsical, empirical studies demonstrate the presence of significant
regularities in readers responses, enabling significant conclusions to be drawn about the effects of literary language and form.
Empirical studies also shift the focus away from the interpretative
issues that have largely preoccupied literary scholars onto the
experiential aspects of literary reading. In this respect, they invite
a reconsideration of the role of literature in human culture.
David S. Miall
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bortolussi, Marisa, and Peter Dixon. 2003. Psychonarratology: Foundations
for the Empirical Study of Literary Response. Cambridge: Cambridge
University Press. An integrated approach to narrative theory and
the empirical study of reading, including exemplary studies by the
authors.
Kuiken, Don, David S. Miall, and Shelley Sikora. 2004. Forms of self-implication in literary reading. Poetics Today 25: 171203.
Larsen, Steen F., and Jnos Lszl. 1990. Cultural-historical knowledge and personal experience in appreciation of literature. European
Journal of Social Psychology 20: 42540.
Lszl, Jnos, and Steen F. Larsen. 1991. Cultural and text variables in
processing personal experiences while reading literature. Empirical
Studies of the Arts 9: 2334.
Martindale, Colin. 1990. The Clockwork Muse: The Predictability of Artistic
Change. New York: Basic Books.
Miall, David S. 2006. Literary Reading: Empirical and Theoretical Studies.
New York: Peter Lang. Chapter 3 in this book provides an introduction
to the methods of empirical study, while Chapter 7 surveys the principal research topics in empirical reception studies.
Miall, David S., and Don Kuiken. 1994. Foregrounding, defamiliarization, and affect: Response to literary stories. Poetics 22: 389407.
Oatley, Keith. 1999. Meetings of minds: Dialogue, sympathy, and identification, in reading fiction. Poetics 26: 43954.
459
460
context, the latter sentences contain at least one logical connective. In standard approaches, the truth conditions for molecular
sentences are given recursively, piggy-backing, as it were, on
the truth conditions for atomics. (An example follows.)
For present purposes, a language is a precise syntax (involving, among other things, a precisely defined set of sentences,
some of which are atomics, some molecular) coupled with truth
conditions, which, as noted, provide truth-in-a-case conditions
for all sentences. So, in addition to specifying a syntax, ones
specification of a language involves specifying a class of cases in
terms of which all sentences, provided by ones specified syntax,
enjoy truth-in-a-case conditions. (An example follows.)
Logical Form
With our language L so given, we can now see the sense in which
every language at least given suitable idealizations has a logic.
Applying (Val), we immediately see that, for any L-argument, it is
either truth preserving, in which case valid in L, or not (in which
case, invalid in L).
EXAMPLE. Consider the L-argument from premise (p&q) to
conclusion q. According to (Val), this argument is valid in L just
if theres no case in which (p&q) is true but q untrue. In L,
our cases are functions from the L-sentences into {1,0}. Is there a
case in which (p&q) is true but q not true? No. To see this, just
consider the truth conditions on L-sentences. According to those
conditions, a sentence of the form (A&B) is true in a case just if
both A and B are true in the given case. So, for any case c, if c(p&q)
= 1, then c(p) = 1 and c(q) = 1, in which case c(q) = 1. Hence, there
is no case c in which (p&q) is true but q isnt true.
What one also notices or would notice, on reflection about
L is familiar truth-functional behavior for negation. For example, given the truth conditions in L, and given (Val), it is easy to
see that p is a consequence of ~~p, and vice versa. In other
words, the double negation of a sentence is logically equivalent
to the given sentence.
One may proceed (as an exercise) to record the other logical
forms that are valid in our artificial language L. Once recorded,
one has a precise account of the logical behavior of & and ~
in L. In turn, one may evaluate whether such logical behavior
accurately captures the behavior of corresponding connectives
in ones natural language. In this respect, the artificial L serves
the role of idealized models in natural sciences: It gives a clear
account of how the target phenomena (in this case, logical connectives) behave. Rivalry among logical theories (more on which
follows) turns on the extent to which L is an accurate model.
Of course, L is but one example of a language (i.e., an idealized, artificial language), and a very simple one at that. Still, it
is not hard to see that, provided languages come equipped with
precisely defined cases and each sentence enjoys truth-in-acase conditions, (Val) quickly yields a logic for the language a
Any textbook on classical logic (of which there are many) will be
suitable further reading. From there, one might turn to textbooks
on nonclassical or so-called intensional logics (of which there
are many). As a first step, one might profitably peruse entries
under logic in the Stanford Encyclopedia of Philosophy, edited
by Edward N. Zalta, available online at: http://plato.stanford.
edu/.
LOGICAL FORM
The construction of systems in which valid inference can be characterized has been the central concern of logic since its inception with the ancients. Beginning with Gottlob Freges seminal
insights of the late nineteenth century (Frege 1879, 1893), it has
461
Logical Form
been understood that accomplishing this goal in a manner sufficiently rigorous that the inferred proposition can be taken to
be proven requires strict attention to the structural properties
of propositions, to their logical form. What was crystalized by
Frege was that this form can be revealed only in a language that
differs in two key respects from the superficial form of natural
languages: i) The structure of propositions is function-argument,
not subject-predicate, and ii) expressions of generality, unlike
proper names, do not occur as arguments, but rather bind variables that do. Together, these two differences afforded the first
adequate account of multiple generalization; by distinguishing xy(P(x,y)) from yx(P(x,y)) in terms of the scopes of
the universal and existential quantifiers, Frege was able to allay
one of the central problems that had plagued traditional logic.
(Cf. Kneale and Kneale 1962, 483ff.)
Freges insight, that grammatical form does not reliably reveal
logical form, was taken up by Bertrand Russell, most notably in
the theory of descriptions (Russell 1905). Russell proposed that
definite descriptions, as in The present King of France is bald,
is not to be understood in the manner of a proper name, that is,
standing as an argument, but rather as a complex term of generality. Thus, the proper logical form is not B(k), but rather
x(K(x) & y(K(y) x = y)) & B(x); that is, there is one and only
one present King of France and he is bald. By taking this to be
the proper logical form, Russell argued that a number of logical
issues could be directly addressed. For example, the ambiguity of
The present King of France is not bald could be accounted for
by taking the negation as having scope either inside or outside
the existential quantifier; negation having broader scope brings
the case into conformance with the law of the excluded middle,
as Russell observed.
Both Frege and Russell realized not only that the insights
about logical form being surveyed clarified the formal nature of
inference, but that these aspects of form also allowed for semantic elucidation, being directly related to an account of the conditions for the truth of propositions. For instance, for Russell,
a substantial virtue of the account of descriptions was that the
intuition that The present King of France is bald is false can be
directly accommodated. But that we can proceed beyond intuitive elucidation to a formally and materially adequate definition
of truth, based on the sort of conception of logical form pioneered
by Frege and Russell, is due to Alfred Tarski ([1936] 1956, 1944).
In the case at hand of quantification, Tarskis semantic clauses
run as follows: With respect to a universe of objects U, x(P(x))
is true just in case every sequence of objects of U satisfies P(x);
x(P(x)) is true iff this is so for some sequence. (A sequence S
satisfies an open formula P(x) iff there is an assignment of a
value a of S to the variable x such that a is P.) Because Tarskis
method iterates, it extends to multiple generalization, distinguishing the truth conditions of xy(P(x,y)) from those of
yx(P(x,y)), thus providing semantic foundation for the syntactical insights of Frege.
Central to the importance of Tarskis formalization of semantics is that it paved the way to metalogic, the study of the properties of logical systems, centrally their consistency, soundness,
and completeness. Tarski was clear to maintain that the definition of truth could be reliably applied only to those systems
whose propositions have the requisite logical form; like his
462
and
[someonej [everyonei [ti loves tj]]]
Logical Positivism
Mapping
LOGICAL POSITIVISM
Also known as logical empiricism, logical positivism was an
important philosophical movement in the first half of the twentieth century that reached its peak in the interbellum period and is
associated with the Vienna Circle (Wiener Kreis) and the Berlin
Circle (Berliner Kreis). The most prominent members of the
former were Moritz Schlick, Rudolf Carnap, Otto Neurath, Hans
Hahn, and Friedrich Waismann; of the latter, Hans Reichenbach,
Kurt Grelling, Carl Gustav Hempel, and Richard von Mises.
On the standard account (see, e.g., Alfred Ayer 1936), logical
positivism is committed to the following principles:
M
MAPPING
In what follows, mapping is used in the general mathematical
sense of a partial or total correspondence between elements,
relations, and/or structures in two sets.
463
Mapping
Much of the theoretical thinking in modern linguistics has
been strongly linked to the development since the 1950s of cognitive science, artificial intelligence, and neuroscience. The first
wave of cognitive science looked upon the brain as a sophisticated symbol-processing digital computer, and linguistic models
in the fifties and sixties took a largely algorithmic approach, with
a strong focus on syntax and logic.
In the 1970s and 1980s, there was a sharply different second
wave of thinking that launched a rigorous, empirically based
study of conceptual mappings: analogy, frames, metaphor, metonymy, grammatical constructions, and mental space projections. This original and ambitious research
program revisited from a modern point of view some fundamental issues that have been known since antiquity. It drew
on a powerful multidisciplinary mix of psychology, linguistics,
computational modeling, and philosophy. Names associated
with pioneering efforts in the new field of conceptual mappings include Douglas Hofstadter, Melanie Mitchell, Dedre
Gentner, and Keith Holyoak, for analogy; George Lakoff, Mark
Johnson, and Mark Turner, for metaphor and image schemas;
Erving Goffman and Charles Fillmore, for frames and frame
semantics; Ronald Langacker, Charles Fillmore, and Adele
Goldberg, for cognitive and construction grammars; Gilles
Fauconnier, Eve Sweetser, and John Dinsmore, for mental space
projections; and Geoffrey Nunberg for metonymic mappings
(pragmatic functions).
In the 1990s and up to the time of writing, there was substantial further evolution of our thinking on these issues. The creative
dimension of conceptual mappings was explored through the
study of conceptual blending and compression (Fauconnier
and Turner 2002; Coulson 2001) and through the modeling of
emergent structure in analogy (Hofstadter 1995; Hummel and
Holyoak 1997). The role of primary metaphors was discovered by
Joe Grady (1997); constraints on mappings were proposed within
metaphor theory and within blending theory.
Metaphor was once commonly viewed as literary, figurative,
poetic something exotic that we add to ordinary language to
make it more colorful, vivid, and emotional. But since the inception of conceptual metaphor theory, it is widely acknowledged that metaphor is, in fact, central to thought and language
and necessary for human language in its many forms. In order
to talk and think about some domains (target domains), we use
the structure of other domains (source domains) and the corresponding vocabulary (see source and target). Some of these
mappings are used by all members of a culture, for instance, in
English, TIME as SPACE. We use structure from our everyday conception of space and motion to organize our everyday conception
of time, as when we say Christmas is approaching, The weeks go
by, Summer is around the corner, The long day stretched out with
no end in sight. Rather remarkably, although the vocabulary often
makes the mapping transparent, we are typically not conscious of
the mapping during use unless it is pointed out to us. Though cognitively active, such mappings are opaque: The projection of one
domain onto another is automatic. Metaphoric mappings may
also be set up locally, in context, in which case they are typically
perceived not to belong to the language but rather to be creative
and part of the ongoing reasoning and discourse construction.
464
Mapping
metonymic chaining, flying home links to leaving the tournament, which links in turn to losing the match, itself caused by
the three lost points. Strikingly, very little of this is indicated
by the linguistic structure itself. It is constructed by means of the
cognitive models that we have for games, tennis, tournaments,
and travel and by applying to them the appropriate mappings.
The same sentence can take on completely different meanings if
we bring in different cognitive models.
Mental space projections link elements and relations in connected mental spaces. For instance, in saying Liz thinks her husband is tired, we build a mental space for Lizs reported beliefs,
with a counterpart for her husband and properties within that
space (tired) that may or may not be satisfied in connected
spaces: Liz thinks her husband is tired, but actually hes in great
shape. In saying Last year, Lizs husband was tired, we build a
mental space for last year, and in saying Liz thinks that last
year, her husband was tired, we build a space for last year embedded in a belief space, itself embedded in a base space. presuppositions (such as Lizs having a husband) can spread across
spaces: In the last example, we infer that Liz has a husband, that
she thinks she has a husband, and that last year, she also had
this husband. But any of these presuppositions can be prevented
from projecting by an explicit overriding entailment.
In mental space projection, the access principle allows a
description of an element to identify its counterpart in another
mental space. For example, if Liz got married to Bob yesterday,
we can say Last year, Lizs husband was tired, identifying Bob in
the mental space last year by means of his counterpart (Lizs
husband) in the mental space now.
Conceptual blending generalizes the notion of conceptual
mapping to arrays of multiple mental spaces with the creation
of new blended spaces and the emergence of novel structure.
Such arrays of connected spaces are called integration networks.
Partial mappings link the mental spaces in such networks, and
selective projection maps the spaces onto novel blended spaces.
The mappings are supported by a small number of vital relations, such as analogy, change, identity, rolevalue, causeeffect.
Compression is systematic in integration networks: A vital relation in one part of the network can be compressed into a different (or a scaled-down) vital relation in another part of the
network. Take, for example, My tax bill gets longer every year. The
inputs are the mental spaces corresponding to different years. In
each one, there is a tax bill. These input spaces are linked by the
vital relation of analogy: Each one is structured by the frame of
paying taxes in a particular year, and each tax-paying situation
is analogous to the others. The inputs are also linked by disanalogy: Each tax bill is different (longer than the previous one).
The analogous input spaces are integrated into a single blended
space, in which all the tax bills are fused into one: Analogy is
compressed into identity. Disanalogy is compressed into change.
In the blended mental space, there is a single tax bill that changes
over time.
Metaphors typically result from double-scope integration
networks, whereas metonymy turns out to be the compression of
one vital relation into another.
Conceptual mappings are not prompted only by spoken or
signed language. They are part of human thought, communication, and interaction quite generally; they are signaled through
multiple modalities (Alac 2006) and anchored by human cultural artifacts as part of socially distributed cognition
(Hutchins 1995).
Biologically, it is currently widely assumed that mappings are
effected by means of neural binding (Shastri 1996). Computational
models of such binding have been proposed within the neural
theory of language (Feldman 2006). Experimental techniques to
show the psychological reality of various mappings have been
devised by Lera Boroditsky (2000), Ray Gibbs (Gibbs et al. 1997),
and Seana Coulson (2001), among others
Gilles Fauconnier
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Alac, Morana. 2006. How brain images reveal cognition: An ethnographic study of meaning-making in brain mapping practice. Ph.D.
diss., University of California, San Diego.
Boroditsky, L. 2000. Metaphoric structuring: Understanding time
through spatial metaphors. Cognition 75.1: 128.
Coulson, Seana. 2001. Semantic Leaps. Cambridge: Cambridge University
Press.
Fauconnier, Gilles. [1985] 1994. Mental Spaces. Cambridge: Cambridge
University Press.
. 1997. Mappings in Thought and Language. Cambridge: Cambridge
University Press.
Fauconnier, Gilles, and Mark Turner. 2002. The Way We Think. New
York: Basic Books.
. 2008. Rethinking metaphor. In Cambridge Handbook of
Metaphor and Thought, ed. Ray Gibbs, 5366. Cambridge: Cambridge
University Press.
Feldman, Jerome. 2006. From Molecule to Metaphor. Cambridge,
MA: MIT Press.
Gentner, Dedre. 1983. Structure-mapping: A theoretical framework for
analogy. Cognitive Science 7: 15570.
Gentner, Dedre, Keith Holyoak, and Boicho Kokinov, eds. 2001. The
Analogical Mind: Perspectives from Cognitive Science. Cambridge,
MA: MIT Press.
Gibbs, R., J. Bogdonovich, J. Sykes, and D. Barr. 1997. Metaphor in idiom
comprehension. Journal of Memory and Language 37: 14154.
Goffman, E. 1974. Frame Analysis. New York: Harper and Row.
Grady, J. 1997. Foundations of meaning: Primary metaphor and primary
scenes. Ph.D. diss., University of California, Berkeley.
Hofstadter, Douglas. 1995. Fluid Concepts and Creative Analogies.
New York: Basic Books.
Hummel, J., and K. Holyoak. 1997. Distributed representations of structure: A theory of analogical access and mapping. Psychological Review
104: 42766.
Hutchins, Edwin. 1995. Cognition in the Wild. Cambridge, MA: MIT
Press.
Lakoff, George, and Mark Johnson. 1980. Metaphors We Live By.
Chicago: University of Chicago Press.
Lakoff, George, and Rafael Nez. 2000. Where Mathematics Comes
From: How the Embodied Mind Brings Mathematics into Being.
New York: Basic Books.
Liddell, Scott K. 2003. Grammar, Gesture, and Meaning in American Sign
Language. Cambridge: Cambridge University Press.
McDonald, Ian. 1992. King of Morning, Queen of Day. New York: Bantam
Books.
Mitchell, M. 1993. Analogy-Making as Perception. Cambridge, MA: MIT
Press.
Nunberg, G. 1978. The Pragmatics of Reference. Bloomington: Indiana
University Linguistics Club.
465
Markedness
Nez, Rafael. 2005. Creating mathematical infinities: Metaphor, blending, and the beauty of transfinite cardinals. Journal of Pragmatics
37: 171741.
Nez, Rafael, and Eve Sweetser. 2006. Looking ahead to the
past: Convergent evidence from Aymara language and gesture in the
crosslinguistic comparison of spatial construals of time. Cognitive
Science 30: 40150.
Shastri, Lokendra. 1996. Temporal synchrony, dynamic bindings, and
SHRUTI a representational but non-classical model of reflexive reasoning. Behavioral and Brain Sciences 19.2: 3317.
Sweetser, Eve. 1996. Reasoning, mappings, and meta-metaphorical
conditionals. In Essays in Semantics and Pragmatics, ed. Masayoshi
Shibatani and Sandra Thompson, 22134. Amsterdam: John
Benjamins.
Turner, Mark. 1991. Reading Minds. Princeton, NJ: Princeton University
Press.
Williams, Robert. 2005. Material anchors and conceptual blends in timetelling. Ph.D. diss., University of California, San Diego.
MARKEDNESS
The original insight of markedness was that many linguistic phenomena consist of polar opposed pairs for example, the phonological feature unvoicedvoiced and the grammatical relation
activepassive and that typically there is an asymmetry, such that
one term is more general and thus unmarked (given first in the
examples) and the other is more constrained and thus marked.
Markedness was first developed in phonology as an explanation
for asymmetries in phonological systems based on cross-linguistic
comparisons, with evidence from typology and universals for
example, more (unmarked) oral consonants than marked nasal
ones: Unmarked consonants occur in places of nonconditioned
neutralization (e.g., only unmarked voiceless consonants in wordfinal position in Russian). Later, markedness was used to study
grammatical semantics (where the unmarked term has a larger
semantic range than the marked term), to explain the order of phonological acquisition in child language (unmarked terms learned
before marked terms, e.g., stops before fricatives) and aphasia (marked terms lost before unmarked ones), and to identify
implicational universals, in which the presence of a marked
element implies the presence of the corresponding unmarked
element, but not vice versa (all of these in Jakobson 1990). Since
then, it has developed into an important (though controversial)
concept in other areas of linguistics, such as morphology, syn-
466
Market, Linguistic
Waugh, Linda, and Barbara Lafford. 2000. Markedness. In
Morphology: An International Handbook on Inflection and WordFormation, I: 27281. Berlin: Walter de Gruyter. An accessible treatment of the topic.
MARKET, LINGUISTIC
Pierre Bourdieu defines a linguistic market as a system of relations of force which determine the price of linguistic products and
thus helps fashion linguistic production (1989, 47). If linguistic
habitus is the subjective element of habitus connected with
language use, linguistic market represents the objective field
relations. As always with Bourdieu, the two are in a constant state
of dynamic interrelationship, as well as evolving dynamically as
a part of the transformation of social structures.
In positing a concept such as linguistic market, Bourdieu is
targeting traditional linguistics. His quarrel is with all linguistics going back to the work of Ferdinand de Saussure, which he
sees as treating language as an object of study rather than as a
practice. The concept thus constitutes language as logos rather
than praxis. Bourdieus critique extends to Noam Chomsky and
Chomskyan linguistics, with its discovery a of semibiological
language acquisition device, deep syntactical structure
(see underlying structure and surface structure),
and universal grammar. Bourdieu cannot accept the
Chomskyan precepts that linguistics should be concerned with
an ideal speaker-listener, a homogeneous speech community,
and perfect grammatical competence. Bourdieus alternative
can be summed up as follows:
In place of grammaticalness it puts the notion of acceptability,
or, to put it another way, in place of the language (langue), the
notion of legitimate language. In place of relations of communication (or symbolic interaction) it puts relations of symbolic power,
and it replaces the meaning of speech with the question of the
value and power of speech. Lastly, in place of specifically linguistic competence, it puts symbolic capital, which is inseparable
from the speakers position in the social structure. (1977, 646;
italics in original)
In other words, Bourdieu is seeking to socialize, or at least sociologicalize, all the major principles of traditional linguistics.
The linguistic market is, therefore, essentially an expression
of linguistic relations. However, like all markets, not everyone
to be found within it is equal, and linguistic knowledge is never
perfect. In reality, some are found to have greater practical mastery (connaissance). This knowledge is itself defined not simply
in terms of use but as an expression of legitimate language. In
most social contexts, there is a dominant language form. This
is most evident at a national level where there is received pronunciation and other standard language forms. However, it can
extend to sublevels and categories and field microcosms. In each
case, there is a right way of using language. This rightness is
defined by social common assent or common acknowledgment. The particularity of language is that while orthodox language forms are maintained by this consensus and recognized
as such reconnaissance not all can use them. There can be a
mismatch between any one individuals connaissance and reconnaissance, resulting from upbringing and proximity to legitimate
467
468
469
470
sum, meaning and belief are different (the distinctness component of the traditional view), and meaning allows the articulation
of belief (the expressive component).
In the last half-century or so, this picture of the relation
between meaning and belief has been challenged from a number of perspectives. One important challenge concerns the
interaction of meaning and belief, addressing such questions as
whether meaning is a relatively neutral vehicle for expressing
belief or something that may affect belief. This is the challenge of
linguistic relativism. Another concerns the validity of the
division between meaning and belief. This is the critique of analyticity. The two challenges point in somewhat contradictory
directions. (These are not by any means the only ways in which
meaning and belief have been discussed in recent years. For
example, Akeel Bilgrami [1992] addresses the issue of how to reconcile meaning externalism with certain subjective aspects
of belief. Unfortunately, a short entry can only point to a couple of
key issues that have arisen in connection with this broad topic.)
As to linguistic relativism, a number of writers (most famously
Edward Sapir and Benjamin Lee Whorf) have argued that meaning is not simply a means for articulating belief, but a means
of shaping belief (as well as emotion, action, even perception).
A popular version of this view is developed in George Orwells
novel 1984, where the government seeks to control peoples
ideas by changing their language. The idea of any conceptual
scheme relativism is difficult to sustain in global terms, as writers
such as Donald Davidson (1984; see Chapter 13) have pointed
out. However, it is clear that we do not have at least some beliefs
before we have some categories, and commonly those categories go along with words and meanings. For example, small children do not have beliefs about gravitation or terrorism because
they do not have the relevant concepts, and the concepts are
presumably something they acquire by learning the words and
their meanings. More importantly, it seems likely that peoples
beliefs about particular events are affected by the concepts (thus,
meanings) available to and salient for them. Thus, without the
concept of terrorism, perhaps Americans would have understood the events of September 11, 2001, as crimes. This would
have changed their beliefs about the nature of the event, proper
responses to the event (e.g., police investigation, extradition,
etc., rather than war), and so on.
This challenge to the expressive component of the commonsense view seems to preserve the distinctness component. In
other words, it seems to rely on a presumption that meaning and
belief are different. After all, if meaning and belief are not distinct, then it is difficult to tell exactly how meaning could affect
belief. This division between meaning and belief is precisely what
is challenged by the critique of analyticity.
There are two clear ways in which the meaning/belief division may be criticized. They relate to two obvious ways in which
the division itself may be formulated. One way concerns revisability. We might say that the belief component of our ideas
is revisable by reference to experience or facts. In contrast, the
meaning component is steady in the face of new experiences
or facts. However, W. V. Quine has argued famously that there
is no way of putting some truths into empirical quarantine and
judging the remainder free of infection. Thus meaning and
sensory evidence are inextricably intertwined (1976, 139; see
471
of items to which the noun refers (see intension and extension). One can only adjudicate the definition of a term by reference to an extension. For example, consider a definition of U.S.
state that involves the criterion of continuous land. One can
reject this definition by pointing to Hawaii, which is part of the
extension of U.S. state. But one can only adjudicate an extension
by reference to a definition. In other words, we rely on a definition of U.S. state to judge that Hawaii is a U. S. state. Thus, one
cannot adjudicate a definition and an extension simultaneously.
One of the two has to be established arbitrarily. By this argument,
there is no such thing as the real meaning of any term, including meaning. Meaning may be social, intentional, or whatever,
as we choose in particular contexts. Thus, whenever we engage
in an interpretive task, the type of meaning at issue should be
stipulated.
This argument disposes of one problem what meaning
really is. But it leads us to three other concerns.
The first is ontological just what sorts of meaning actually
exist. We may, for example, stipulate Platonic meaning as our
object of hermeneutic interest (see philology and hermeneutics). But we cannot actually interpret for Platonic meanings if they do not exist.
The second concern takes up the demarcation of our stipulative categories. These need to be adequately precise. For example, we might stipulate that we are concerned with intentional
meaning. But there are numerous sorts of intentional meaning
that should often be distinguished in the case of legal interpretation, the self-conscious intent of the author who drafted a
piece of legislation, the intents of the legislators who passed it,
the intents of the judges who gave opinions on its constitutionality, and so on.
The final concern bears on the particular purposes for which
we are interpreting. For example, for any given term in a law,
there may be variable social meanings. Ordinary people may use
a term with one meaning; scientists may use it with a slightly different meaning. In particular cases of interpretation, the meaning
associated with one or the other group may be more significant.
Note that in these cases, we are not trying to determine the real
meaning of the law. Rather, we are acknowledging that there are
many sorts of meaning and we are trying to determine which is
the most important in the case at hand.
Patrick Colm Hogan
472
473
Media of Communication
. 1996. Introduction. In The Twin Earth Chronicles: Twenty Years
of Reflection on Hilary Putnams The Meaning of Meaning, ed.
Andrew Pessin and Sanford Goldberg, xvxxii. London: M. E. Sharpe.
Tye, Michael. 1995. Ten Problems of Consciousness. Cambridge,
MA: Bradford Books, MIT Press.
MEDIA OF COMMUNICATION
This term refers to the means by which communication
takes place or the choice of substance for realizing a communicative act.
It has long been recognized that choosing between alternative means or media of communication can have consequences
for the linguistic form of the message. Thus, the choice of writing instead of speech as a medium of communication may entail
particular grammatical, syntactic, or lexical preferences over
others. Indeed some work suggests that certain kinds of linguistic
patterning may be distinctive to particular media (Biber 1988).
From Ferdinand de Saussure (1912) onward, much work
has been devoted to specifying the differences between speech
and writing. (See also oral composition; oral culture;
writing systems; writing, origin and history of).
One approach deals with speech and writing as alternative
expressions of the same underlying language system, realized
in differing ways depending on the medium adopted. Thus,
the written medium is associated with greater lexical density,
a wider range of grammatical structures, a greater degree of
embedding, and more varied forms of connectivity between sentences. Conversely, the spoken medium is associated with lexical repetition, low lexical density, vague or indefinite expressions
(thingymajig), high incidence of coordinated clauses linked by
common conjunctions (and, but), and selection of the active
rather than the passive voice. The character of these differences
has led one author to characterize speech as a process and writing as a product (Halliday 1985).
A more radical view suggests that sentence grammars generally have been implicitly biased toward the study of writing, and
grammars would be quite different if they were formulated from
the outset to take account of speech phenomena. As David Brazil
observes, if any part of the outcome [of a grammar of speech]
looks like a sentence, this comes about as an interesting by-product of the processes we are interested in, not as the planned outcome to which these processes owe their definition (1995, 39).
Certainly there is widespread agreement that the communicative potentialities of writing and speech are very different.
Speech typically takes place between interlocutors who are in
some way copresent to each other, and this enables them to
adjust their utterance in the light of the apparent reactions of
the other. The process of composing and planning speech goes
hand in hand with the act of speaking, and speaking, in turn,
goes hand in hand with the process of interpretation that must
keep pace with it. There is no time lag between production and
reception. Instead, speech is temporally bound, transient, and
dynamic, rooted in an unfolding context, with paralinguistic
behavior providing an important supplementary layer to communication (see paralanguage). Conversely, writing as a
semipermanent product enables a gap across time and space to
open up between participants. The process of composition may
474
Media of Communication
monologue, soap opera, various kinds of reality programing,
commercials, commentary on public events including sporting
occasions, argument, drama, talk shows, and phone-ins (several
of which have begun to attract systematic study; see Hutchby
2005; Tolson 2005). Although there may be generic antecedents
to these in the world of real-time, face-to-face communication,
certain properties seem to set broadcast genres apart from everyday nonmediated communication. For one thing, the idealized
speaker and hearer of the canonical speech situation, reciprocally exchanging roles and utterances, no longer easily applies
except in grossly simplified ways.
Instead, as Erving Goffman (1981) observes, broadcast communication takes shape from complex production formats and
participation frameworks in which the discourse is sometimes
scripted, sometimes relatively spontaneous, sometimes spoken,
sometimes written, sometimes written to be spoken, sometimes
single authored, sometimes multiply authored, sometimes dialogue, and sometimes monologue. Indeed, Goffman suggests
replacing the term speaker with notions of author, animator,
and principal. The author is the one who has selected the sentiments that are being expressed and the words in which they are
encoded (Goffman 1981, 144). The animator is the one who gives
voice to the words that have been selected, sometimes by someone else. The principal is whoever is potentially held to account
for the sentiments expressed. In many situations, the three roles
coalesce, but in broadcast communication in news programs,
for instance the presenter who reads the news from the autocue
may merely be animating a script authored elsewhere, by the
editorial team, and the ultimately accountable source for the discourse the principal may be the organization itself. Thus, in
the case of a BBC news bulletin, it may be the director general
or members of the board of trustees who resign their positions
should an item be called into question, not necessarily the news
editor, and certainly not the news presenter.
Just as various alignments are possible in terms of the production of broadcast communication, important distinctions apply
in its reception, where the potential participation framework
is equally complex. As Goffman again observes, an utterance
does not carve up the world beyond the speaker into precisely
two parts, recipients and non-recipients, but rather opens up an
array of structurally differentiated possibilities, establishing the
participation framework in which the speaker will be guiding
his delivery (1981, 137). Broadcast communication is quite frequently oriented to two kinds of recipient. In studio interviews,
for instance, in chat shows or news programs, there is the immediate recipient of the talk the interviewee or the interviewer
but beyond them is the overhearing audience numbering in size
from thousands to millions. In this way, in posing a question to
an interviewee, the discourse of the interviewer is bidirectional. It
is oriented in the first instance to the interviewee, but the design
of the question will also be shaped by the assumed concerns of
the broadcast audience beyond. Talk for an overhearing mass
audience in this way assumes characteristics distinctive to the
medium that are different from ordinary talk or conversation.
It should be noted also that in the broadcast media, the
foundational distinction in considering media of communication between speech and writing becomes confounded. Within
a continuous stretch of discourse, a language user may switch
475
476
477
478
elevator
exit-door
elevator
exit-door
woman
It follows that the man is on the left of the elevator and the
woman is on the right of the elevator. No alternative model of
the discourse is a counterexample to this conclusion, and so it
is logically valid; that is, it must be true, given the truth of the
premises.
Mental models govern our memory for discourse. Suppose
that this discourse continues:
The man standing in front of the office door was using a cell phone.
Later it states:
The man using the cell phone was wearing a suit.
Both assertions can be used to update our model. In an unexpected memory test, as Alan Garnham (1987) showed, we are not
likely to recall which of the following sentences occurred in the
discourse:
479
480
where not represents negation. The second assertion eliminates the second of these possibilities, and so it seems that there
is an ace in the hand. However, the connective or else means that
at the very least, one of the propositions that it connects may be
false. Given, say, the falsity of if theres a king in the hand then
theres an ace in the hand, we realize that its possible that theres
a king in the hand without an ace. So, even though the second
assertion tells us that there is a king in the hand, no guarantee
exists that theres an ace, too. The inference is fallacious. This
analysis relies only on two well-attested facts about our understanding: 1) or else allows that one of the clauses it connects is
false, and 2) the falsity of a conditional allows that its if-clause
can be true and its then-clause false. A computer program implementing the principle of truth led to the discovery of a variety of
illusions, and subsequent studies have corroborated their occurrence (Johnson-Laird 2006).
When you read the earlier description of the man and woman
standing in front of the doors, you might have formed a visual
image of the layout. Spatial relations are usually easy to visualize.
You might, therefore, assume that mental models are nothing
more than visual images. This assumption is wrong. Some relations, such as Pam is better than Viv, are impossible to visualize. You can imagine, say, Pam as further up a ladder than Viv is,
but nothing in your image or in any possible image can make
explicit the meaning of better than. Not all relations are rooted in
a sensory modality or have a spatial interpretation. Some relations, such as the cat is cleaner than the dog, are easy to visualize
but do not invoke a spatial representation. Reasoning with these
visual relations, which elicit images rather than spatial models,
takes longer than reasoning with other sorts of relation, and it
activates a region of the brain underlying vision.
The hypothesis that we represent discourse in mental models is
uncontroversial in psycholinguistics, though not all accounts
stress the iconicity of models (cf. Kintsch 1988; and Gernsbacher,
1990). We make a dynamic representation of entities, their properties, and the relations among them. The heart of the problem
in building a mental model is to recover the appropriate referent
for each expression. Speakers refer back to entities that they have
already introduced in the discourse, and they can use different
noun phrases, demonstratives, or pronouns to do so. The correct
interpretation of such anaphora depends on many factors. Given
a sentence like The man confessed to the priest because he wanted
absolution, we understand that he refers to the man rather than
the priest. We make this attribution because we know the purpose
of confession, because we have a preference for locating the antecedents of pronouns in the subjects of clauses, and because we also
have a preference for a parallel grammatical role of antecedent and
anaphor (Stevenson, Nelson, and Stenning 1995).
In computational linguistics, centering theory shows how the
focus on a local segment of discourse determines the antecedents of anaphora, especially pronouns (e.g., Grosz, Joshi, and
Weinstein 1995; Webber et al. 2003) Another factor is the semantic difference between antecedent and anaphor (Almor 1999).
The information load on interpretation increases when, unlike
Mental Space
Rinck, Mike, and G. Bower. 1995. Anaphor resolution and the focus of
attention in situation models. Memory & Language 34: 11031.
Schroyens, Walter, W. Schaeken, and G. dYdewalle. 2001. The processing of negations in conditional reasoning: A meta-analytic case study
in mental model and/or mental logic theory. Thinking & Reasoning
7: 12172.
Stevenson, Rosemary J., A. W. R. Nelson, and K. Stenning. 1995. The role
of parallelism in strategies of pronoun comprehension. Language and
Speech 38: 393418.
Webber, Bonnie, M. Stone, A. Joshi, and A Knott. 2003. Anaphora and
discourse structure. Computational Linguistics 29: 54587.
MENTAL SPACE
What Is a Mental Space?
Mental spaces are partial assemblies constructed as we think
and talk, for purposes of local understanding and action. They
contain elements and are structured by frames and cognitive
models. Mental spaces are connected to long-term schematic
knowledge, such as the frame for walking along a path, and to
long-term specific knowledge, such as a memory of the time you
climbed Mount Rainier in 2001. The mental space that includes
you, Mount Rainier, and your climbing the mountain can be activated in many different ways and for many different purposes.
You climbed Mount Rainier in 2001 sets up the mental space in
order to report a past event. If you had climbed Mount Rainier
sets up the same mental space in order to examine a counterfactual situation and its consequences. Max believes that you
climbed Mount Rainier sets it up again, but now for the purpose
of stating what Max believes.
Mental spaces are constructed and modified as thought and
discourse unfold and are connected to each other by mappings,
such as identity and analogy. It has been hypothesized that
at the neural level, mental spaces are sets of activated neuronal
assemblies and that the connections between elements correspond to coactivation bindings. On this view, mental spaces
operate in working memory but are built up partly by activating structures available from long-term memory. Connections
link elements across spaces without implying that they have the
same features or properties. When I was six, I weighed fifty pounds
prompts us to build an identity connector between the adult and
the six-year-old despite the manifest and pervasive differences.
Mental spaces are built up dynamically in working memory but
can become entrenched in long-term memory.
An expression that names or describes an element in one
mental space can be used to access a counterpart of that element
in another mental space (access principle).
481
Mental Space
Merge
[1985] 1994, 1997) was that it provided simple, elegant, and general solutions to problems such as referential opacity or presupposition projection that had baffled logicians and formal
linguists. Opacity results from the application of the access principle across mental spaces as discourse unfolds. What emerged
was a unified cognitively based approach to anaphora, presupposition, conditionals, and counterfactuals. Additionally, the
gestural modality of signed languages revealed other ways in
which mental spaces could be set up and operated on cognitively
and physically.
Shortly thereafter, J. Dinsmore (1991) developed a powerful
approach to tense and aspect phenomena, based on mental
space connections. The approach was pursued and extended
in fundamental ways by M. Cutrer (1994), who made it possible
to understand the role of grammatical markers as prompts to
deploy vast networks of connected mental spaces. Further generalizations were achieved in areas exemplified by the diverse
contributions to Spaces, Worlds, and Grammar (Fauconnier and
Sweetser 1996). Sophisticated research continues to be done in
all of the areas where mental space theory was first applied, in
particular on conditionals (see Dancygier 1998; Dancygier and
Sweetser 1996, 2005), scoping phenomena on locative and temporal domains (see Huumo 1996), grammar of sign languages
(see Liddell 2003), discourse (see Epstein 2001), and frame shifting (see Coulson 2001). But at the same time, there has been an
explosion of research triggered by the discovery of wide-ranging
phenomena whereby mental spaces are assembled, connected,
and constructed within networks of conceptual integration (see
conceptual blending). This area of research links linguistic and nonlinguistic phenomena in systematic ways that begin
to explain how and why there can be imaginative emergent
structure in human thought in its everyday manifestations,
as well as in its most original and singular sparks of creativity.
482
MERGE
Merge is the primitive combinatorial operation in the most recent
version of transformational grammar known as minimalism. In its most austere variety, merge is a generalized transformation that simply turns its input elements into a set with the input
elements as members (set-merge). Unlike the earlier government
and binding model of the principles and parameters theory, Noam Chomskys (1995) minimalist model does not assume
a deep structure (see underlying structure) representation
as a starting point of the derivation; instead, syntactic computation
starts out from individual words. Merge combines words as well as
syntactic objects it has already formed in a recursive manner (see
recursion, iteration, and metarepresentation), generating an infinite array of discrete expressions with a hierarchical
constituent structure. In principle, merge can freely apply
to elements available to it, but its application is constrained by
principles of computational efficiency and by output conditions
imposed by external systems of sound/gesture and meaning.
movement is construed as merge of a syntactic object with a
syntactic object contained in it, whence the term internal merge
(vs. external merge; see Chomsky 2004). Consider, for instance,
the derivation of the passive sentence (1a), where the underlying object is moved to the subject position. Here the expression
(1b), constructed by recursive applications of merge, undergoes
merge with its subset {a, house}, yielding (1c). (2) is a tree diagram representation of (1c).
(1) a. A house will be built
b. {will, {be, {built, {a, house}}}}
c. {{a, house}, {will, {be, {built, {a, house}}}}}
Merge
Metaphor
(2)
METALANGUAGE
house will
be
built
a
house
The syntactic object {a, house} has two copies or occurrences, but
it is realized phonetically only as a member of the (largest) set in
(1c). Note that the two occurrences resulting from movement are
not distinct syntactic objects. Rather, the same syntactic object
is a member of two sets, where one set is properly contained in
the other.
While there is currently no agreement regarding restrictions
on merge, or how to deduce them, the mainstream view holds
that merge is binary, always taking exactly two input elements
(entailing strict binary branching in syntactic trees), and it cannot alter set-membership relations that it has established before.
It is unresolved whether the output of merge should be enriched
to encode any asymmetry between its operands. On Chomskys
(1995) original definition, merge forms a set with the following
two members: the set of the input elements and the word functioning as the head of the constituent (also known as label),
thereby representing the asymmetry in the choice of the input
element that projects (see x-bar theory). Following this formulation, (1c) can be rewritten as (3). The structure in (3) is represented by the labeled tree diagram in (4).
(3) {will, {{a, {a, house}}, {will, {will, {be, {be, {built, {built, {a, {a,
house}}}}}}}}}}
(4)
will
will
a
a
house
will
be
be
built
built
a
a
house
Another asymmetry is that between an adjunct (e.g., an adverbial) and the host it is adjoined to (e.g., a verb phrase). Chomsky
(2004) suggests that when an adjunct and its host undergo
merge, the result is an ordered pair Adjunct, Host (also known
as pair-merge).
Balzs Surnyi
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Chametzky, Robert A. 2000. Phrase Structure: From GB to Minimalism.
Oxford: Blackwell.
Chomsky, Noam. 1995. The Minimalist Program. Cambridge, MA: MIT
Press.
Languages may be used to talk about languages. In the first sentence of this entry, English is used to mention (or talk about) languages. (See use and mention.) In general, a language under
discussion, on a given occasion, is called the object language,
and the language being used in the discussion (to talk about the
object language) is the metalanguage.
Alfred Tarski proved that no classical language (language with
classical logic) can express its own truth predicate. Accordingly,
the semantics of a classical language (with standard syntax)
can only be done in a richer metalanguage.
J. C. Beall
SUGGESTION FOR FURTHER READING
Tarski, Alfred. 1990. Logic, Semantics, Metamathematics: Papers from
19231938. 2d ed. Ed. J. Corcoran. Indianapolis, IN: Hackett.
METAPHOR
This term derives from the Greek metapherein, indicating a
transfer of meaning from one linguistic expression to a second,
semantically different expression. According to standard dictionary definitions, a metaphor is a figure of speech in which a
word or phrase that ordinarily designates one thing is used to
designate another to which it is not literally applicable. Thus,
in metaphor, there is an implicit comparison between unlike
things, as in the Shakespearean metaphor Juliet is the sun.
Employing the terminology introduced by I. A. Richards (1936),
the target concept (Juliet) is labeled the topic or tenor, the metaphoric source (sun) is the vehicle and the emergent meaning,
the ground (see also source and target). Despite this seemingly simple definition, the task of actually identifying metaphor
has proven to be a difficult enterprise because metaphors can be
implied (there is a vehicle but no specified topic), dead (a usage
so conventionalized that the metaphoric transference is no longer actively recognized), mixed (the conflation of two distinct
metaphors), submerged (in which the vehicle is implied but not
stated), or extended (suggested throughout a text). This difficulty
has been especially apparent in recent attempts with computer
applications, such those based on the identification of metaphor
in text corpora or in computer systems capable of generating
representations of metaphorical meaning.
483
Metaphor
with Eco, it is clear that interpretations of Aristotelean thoughts
have influenced much of the thinking on the topic for the past
several millennia, most notably a presumed distinction between
literal and nonliteral language and in framing the basic theoretical issues that have guided much of the subsequent scholarship: understanding the cognitive process that permits the
stretching of meaning from the literal to the metaphoric and in
determining the pragmatic reasons why metaphor would be
employed when literal counterparts could have been used.
In his Poetics and in Rhetoric, Aristotle situated metaphor
in the realm of language, a position that has been the basis for
subsequent theories but has been contested since 1980 by theorists working within cognitive linguistics (described later).
Aristotles basic premise is that with metaphor, one word (or
expression) is substituted for another. He described several categories of substitution, though the forms most studied are what
we would call today nominal and predicative metaphor, the
former in which one noun is substituted for another and the latter in which the substitution is of verbs. Aristotle provided some
explanation of the process involved in metaphor comprehension, namely, an innate tendency to see likeness in objects and
events that are, on the face of it, dissimilar (or in which the similarity is not transparent). Moreover, he provided some reasons
why metaphor might be employed, primarily, to serve a stylistic
and aesthetic function wherein the listener, forced to decode the
message, experiences a pleasurable reaction; and secondarily,
to serve the creative cognitive function of providing a name to
things that do not have proper names of their own. There has
been considerably elaboration of the seminal ideas of Aristotle in
the twentieth century, despite what some would consider a fatal
flaw in the logic of substitution as the basis for metaphor: If, for
example, with the nominal Shakespearean metaphor Juliet is
the sun the vehicle, sun is a substitution for another word that
falls within the same genus= as the topic, Juliet, what could
that word be?
TWENTIETH-CENTURY ELABORATIONS. Two interpretations consistent with Aristotle have been most influential. According to
the substitution position, the transfer from vehicle to topic is an
ornamental means of presenting some intended literal meaning,
so that when one states George is a wolf, it is merely an aesthetic
way of saying that George is fierce. The comparison approach is
less ornamental and closer to the second function described by
Aristotle in that, here, the listener must construct a way in which
properties of the vehicle are applicable to the topic. There have
been several variants of comparison theories proposed in the
literature over the past 50 years, but all include the notion that
the comprehension process involves the identification of a relevant set of preexisting features shared by topic and vehicle. These
theories all have shortcomings, including the failure to encompass the creation (and not mere identification) of similarity, the
problem in identifying the mechanisms that would select the
features assumed to be important for interpretation, and the failure in such theories to explain the asymmetry in meaning that
occurs when the topic and vehicle are reversed, as occurs when
one contrasts my lawyer is a shark with my shark is a lawyer.
Max Black attempted to address at least some of these shortcomings by postulating an interactive theoretical perspective in
484
which the novel meaning of a metaphor is not based on identifying a shared set of (possibly marginal) meanings of the words
being compared. Meaning, he argues, is generated by the interaction between a principal subject (the more literal usage of the
word, similar to the topic) and the complex of associations connected with a subsidiary subject (analogous to the vehicle). The
process is interactive inasmuch as reciprocal action between the
principal and subsidiary subject selects, emphasizes, suppresses
and organizes features of the principal subject by implying statements about it that normally apply to the subsidiary subject
(Black 1962, 46). The outcome is the creation of novel meaning
formed by a parallel implicational complex in which the topic
can be viewed in a radically different light and in which novel
emergent meanings can be created between words. Despite the
popularity of this general approach, it, too, has been subject to
various criticisms, notably regarding the ambiguity in defining
the theoretical terms employed and in determining which of the
terms is the principal and which is the subsidiary subject.
Subsequent psycholinguistic theories have attempted
to describe cognitive mechanisms that are consistent with
the interactive approach. Salience imbalance theory is a variant of traditional comparison models aimed at describing why
some statements are seen as literal and others as metaphoric by
assuming that the features shared by topic and vehicle differ in
relative level of salience: Literal statements are those in which
the shared features are salient to both terms, whereas with metaphor, they are salient to the vehicle but not the topic. Domain
interaction theory is an extension of a computational model of
analogy and assumes that metaphor involves the finding of
similarity both within and between the conceptual domains
evoked by words. Thus, a metaphor such as George Bush is a
hawk would be comprehended by finding a spot in semantic
space that would be consistent with the analogy George Bush
is to world leaders as hawks are to birds. In this model, ease
of comprehension is a function of the ease of finding a shared
similarity (i.e., in determining ways in which Bush is similar to
a hawk), whereas a sense of metaphor aptness increases as
the distance between conceptual domains, such as leaders and
birds, becomes greater. Finally, structure-mapping theory, also
emerging from computational work in analogy, is based on identifying a system of shared relations between the target and source
domains and not by merely identifying a feature shared by topic
and vehicle. Although there are psycholinguistic studies that
support each theory, each is based ultimately on finding similarity between the words presented as topic and vehicle and, consequently, heir to all of the criticisms of such models (reviewed, for
instance, in Glucksberg 2001).
Sam Glucksberg proposes a novel solution that rejects similarity as the basis for metaphor by arguing that in metaphor, one
does not look for a similarity between topic and vehicle (i.e., by
treating the comparison as an implicit simile). He avers, rather,
that metaphor should be understood as a class inclusion statement analogous in how we treat such statements as my dog is a
collie. He argues, and has presented convincing evidence, that
with metaphor, the vehicle has dual reference (both as the literal object and as indicative of higher-order categories) and that,
in comprehension, one classifies the topic to the category suggested by the vehicle. That is, in a metaphor such as my lawyer is
Metaphor
a shark, the vehicle shark stands for or exemplifies a category
to which lawyers could be assigned (such as aggressive, predatory, tenacious entities). In more recent expansion of the theory,
he and his colleagues have indicated how the topic plays a role
in identifying the appropriate category for which categorization
is appropriate.
THE PROCESS OF METAPHOR COMPREHENSION. Much of the theory and research described here is based on offline methodology.
Beginning in the late 1970s, researchers started to examine the
processing of metaphor online, measuring processes that were
happening during the act of comprehension. Most of the early
studies were based on the indexing of reading time or some other
measure of response latency; lately, studies have also employed
neurocognitive imaging techniques such as EEG and fMRI (see
neuroimaging). Much of the initial theorizing has been based
on speech-act theory, especially as espoused in the work of
John Searle. Following from the distinction between literal and
nonliteral language, the assumption has concerned the processing priority of literal meaning. According to what is now called the
standard pragmatic model, the model would be that the default
processing of language would be to its literal meaning, and that
those inferential processes that seek an alternative, nonliteral
interpretations are only triggered if one fails to find a contextappropriate literal interpretation. These assumptions have been
tested in psycholinguistic research that has concretized the standard model: It assumes that priority to a default literal meaning
would be demonstrated by more rapid reading (or other indices
of processing) of a metaphor in a discourse context that is consistent with its literal sense than in a context that is consistent with
the nonliteral sense; it also assumes that one should not find the
processing of metaphoric meaning in conditions in which the literal sense of an utterance is context appropriate.
More than 20 years of research have failed, in the main, to support the predictions arising from the standard pragmatic model,
instead showing that in appropriately elaborated contexts, one
can process the metaphoric sense as rapidly as the literal sense of
an utterance and that, using Stroop-like procedures and speedaccuracy analyses, the initiation of metaphoric interpretation
does not depend on a failure to find an appropriate literal interpretation. These findings, though sometimes complicated by the
level of conventionality of the metaphoric expression, have led
to a set of competing theories, all of which have some support,
including models based on the notion of resolving constraint
satisfaction and those that assume that the initial processing of
a word is at an underspecified schematic level. An increasingly
popular processing model by Rachel Giora attempts to maintain processing priority but places the emphasis not on literal
meaning (as Searle had it) but on the saliency of a word (as concretized by familiarity, conventionality, and frequency of use).
According to this theory, one is obligated to process the salient
sense of a word (or expression), regardless of context; contextual
constraints can boost the activation and meaning access of less
salient meanings but will not do so at a cost to the activation and
access of the more salient sense. The ultimate success and test of
these various theories are being contested, more often these days
with neuroimaging techniques that give a more fine-grained
analysis of online processing than available in the past.
485
Metaphor
Metaphor, Acquisition of
of a number of diverse disciplines, and although the on-line processing implications of the theory are still ongoing, the tests to
date have been encouraging, often employing brain neuroimaging techniques such as event-related potential (ERP) measurement (see Coulson 2001). Nonetheless, this theory has also been
subjected to criticisms that it, too, is incapable of being disproved
and that it is too indiscriminate, inasmuch as almost anything
that enters working memory can be considered a blend.
Evaluation
Treatments of metaphor as a linguistic and as a cognitive phenomenon coexist today, in much the same way that two species
of hominid have coexisted in our evolutionary history. It remains
to be seen whether the offspring of one approach will disappear. Despite the differences, there is a convergence between
approaches that should not be undervalued: This convergence
includes the undercutting of the distinction between literal and
nonliteral language; a recognition by both approaches of the
need to consider the richness of examples coming from literary or philosophical analysis, as well as the controlled rigor that
comes from experimental studies; an emphasis on the role of
cognition and pragmatics (see Carston 2002, for instance, for
an exposition from a relevance theory perspective); and
the growing sentiment (however conceptualized) that the construction of metaphoric meaning is flexible and involves more
of an active on-line interpretive process and less of a mere
arousal of entrenched meaning and that, ultimately, the battleground for theoretical supremacy (or for synthesis of the two
approaches) will depend on data generated and based in the
neurosciences.
Albert N. Katz
WORKS CITED AND SUGGESTED FURTHER READINGS
Black, M. 1962. Models and Metaphor. Ithaca, NY: Cornell University
Press.
Boroditsky, L. 2000. Metaphoric structuring: Understanding time
through spatial metaphors. Cognition 75: 128.
Carston, R. 2002. Thoughts and Utterances: The Pragmatics of Explicit
Communication. Oxford: Blackwell.
Coulson, S. 2001. Semantic Leaps. New York: Cambridge University
Press.
Eco, U. 1984. Semiotics and the Philosophy of Language.
Bloomington: Indiana University Press.
Fauconnier G. 1997. Mappings in Thought and Language.
Cambridge: Cambridge University Press.
Gibbs, R. 1994. The Poetics of Mind. New York: Cambridge University
Press.
Giora, R. 2003. On Our Mind: Salience, Context and Figurative Language.
New York: Oxford University Press.
Giora, R., ed. 2001. Models of figurative language. Metaphor and
Symbol 16.3/4 (Special Issue): 145333.
Glucksberg, S. 2001. Understanding Figurative Language. New
York: Oxford University Press.
Katz, A., C. Cacciari, R. Gibbs, and M. Turner. 1998. Figurative Language
and Thought. New York: Oxford University Press.
Lakoff, G., and M. Johnson. 1980. Metaphors We Live By. Chicago:
University of Chicago Press.
Ortony, A., ed. 1993. Metaphor and Thought. 2d ed. New York: Cambridge
University Press.
486
METAPHOR, ACQUISITION OF
Metaphor is a pervasive aspect of human language (Lakoff and
Johnson 1997). Metaphor also plays a central role in abstract
thought by structuring concepts (Gibbs 1994) and leading to
conceptual change (Gentner and Wolff 2000). As such, its rudimentary manifestation at the early ages and its continued growth
over developmental time has been the focus of scientific inquiry
for several decades. Research on the acquisition of metaphor
followed two main lines of inquiry, each bearing on a different
definition of the term. One approach defined metaphor as a similarity comparison between the perceptual features of objects or
actions, and explored how early children would understand and
produce these so-called perceptual metaphors (e.g., butterfly is
(like) rainbow). Another approach defined metaphor as a conceptual-linguistic mapping between the structural features of two
disparate knowledge domains a source domain, which serves
as the source of vocabulary and conceptual inferences, and a
target domain, to which vocabulary and inferences are extended
metaphorically, and it examined the age at which children begin
to develop an integrated understanding of such structural metaphors (e.g., time is motion along a path) as an amalgam of both
source and target domain meanings.
Following is a brief summary of the developmental changes
in childrens metaphorical ability, from the early onset of simple
perceptual metaphors to the later emergence of more complex
structural metaphors.
Metaphor, Acquisition of
go together, children were more likely to group a cherry lollipop
with a toy stop sign, which was similar in shape and color, rather
than matching it with a dissimilar object in the same category
(Mendelsohn et al. 1984). Similarly, when presented with event
triads, children were more likely to pair two events that were
alike (ballerina spinningtop spinning) than to match two events
that were of different types (ballerina leapingtop spinning; Dent
1984).
Moreover, five-year-old children could provide similaritybased explanations when asked about metaphorical expressions
that involve comparisons between objects such as a butterfly is a
flying rainbow, or a cloud is like a sponge (Gardner et al. 1975;
Gentner 1988; Billow 1975; Malgady 1977). For example, they
would explain the statement a cloud is like a sponge by saying that both clouds and sponges are round and fluffy (Gentner
1988), or they would complete the statement he looks as gigantic as by selecting from among multiple choice alternatives
an ending that draws on a feature-based comparison: he looks
as gigantic as a double-decker cone in a babys hand (Gardner
et al. 1975).
Thus, preschool children can both understand similarity
comparisons between two objects or events that are perceptually alike and spontaneously produce perceptual metaphors and
explanations based on such comparisons in their early communications. This ability constitutes an important milestone in
childrens language development. The ability to express similarities between objects and events based on shared perceptual
features is considered the earliest sign of metaphorical ability in
young children, and accordingly, children are believed to have a
rudimentary level of metaphorical ability as early as preschool
age (Billow 1981; Gardner et al. 1978; Vosniadou 1987; Winner
1979).
commonalities: Children, at all ages, have no difficulty understanding feature-based similarities between objects, but it is with
increasing age that they begin to understand cross-domain mappings based on relational structure and, accordingly, produce
explanations that reflect this understanding.
Others propose a developmental progression from an understanding of metaphor as involving only one domain to a conceptualization of metaphor as involving two domains (Asch
and Nerlove 1960; Cicone, Gardner, and Winner 1981; Schecter
and Broughton 1991; Winner, Rosenstiel, and Gardner 1976).
Thus, children initially focus only on the source domain of the
metaphorical mapping and gradually develop a more integrated understanding of metaphor as involving both a source
and a target domain. For example, in explaining the metaphorical statement the prison guard is a hard rock, children six
to eight years of age focused exclusively on the source domain
meaning of the mapping and provided literal interpretations for
metaphorical statements (e.g., The guard has hard muscles),
whereas older children and adults were able to consider both the
source and target domain meanings of the mapping, thus providing explanations that captured the metaphorical meaning
(e.g., The guard was mean and did not care about the feelings
of prisoners; Winner, Rosenstiel, and Gardner 1976). Similarly,
when asked to extend physical sensation terms onto psychological traits (e.g., Can a person be warm/ sweet/ soft?), children
three to seven years of age focused only on the source domain
of the metaphorical mapping and provided literal explanations
for metaphorical statements (e.g., Mommy is sweet because she
cooks sweet things), whereas older children focused on both
domains simultaneously and provided explanations that treated
metaphorical meaning as a different but related extension of
the literal meaning (e.g., Hard things and hard people are both
unmanageable; Asch and Nerlove 1960).
Yet another group of researchers argue that the ability to
understand structural metaphors is not determined solely by a
childs age but by a host of other factors, such as the nature of
the source or the target domain (Keil 1986), and the familiarity
of the metaphorical mapping or the source domain (zalkan
2007). For example, five-year-old children can correctly map
animate terms onto cars (e.g., the car is thirsty), but have difficulty understanding metaphors that involve mappings between
taste terms and people (e.g., she is a bitter person; Keil 1986).
Similarly, preschool children can both understand and explain
metaphors that are structured by motion (e.g., Time flies by,
Ideas cross my mind; zalkan 2005, 2007) a domain that
structures a wide range of abstract concepts across different languages of the world but have difficulty deciphering the meaning of metaphors that involve extensions of object properties
(e.g., The prison guard is a hard rock; Winner, Rosenstiel, and
Gardner 1976). From this perspective, the development of metaphorical ability shows different trajectories for different conceptual domains and metaphorical mappings, based on ones
knowledge of the source and/or the target domain and the familiarity of the metaphorical mapping.
In summary, research on childrens metaphor comprehension and production shows that children can both understand
and spontaneously produce perceptual metaphors that involve
similarity comparisons by preschool age. However, the ability
487
Metaphor, Acquisition of
488
zalkan, S. 2005. On learning to draw the distinction between physical and metaphorical motion: Is metaphor an early emerging cognitive
and linguistic capacity? Journal of Child Language 32.2: 128.
. 2007. Metaphors we move by: Childrens developing understanding of metaphorical motion in typologically distinct languages.
Metaphor and Symbol 22.2:14768.
zalkan, S., and S. Goldin-Meadow. 2006. X is like Y: The emergence of similarity mappings in childrens early speech and gesture.
In Cognitive Linguistics: Foundations and Fields of Application, ed. G.
Kristianssen, M. Achard, R. Dirven, and F. Ruiz de Mendoza, 22962.
Berlin: Mouton de Gruyter.
Schecter, B., and J. Broughton. 1991. Developmental relationships
between psychological metaphors and concepts of life and consciousness. Metaphor and Symbolic Activity 6.2: 11943.
Sinclair, M., and M. Stambak. 1993. Pretend play among three-year-olds.
Mahwah, NJ: Lawrence Erlbaum.
Vosniadou, S. 1987. Children and metaphors. Child Development
58: 87085.
Vosniadou, S., and A. Ortony. 1983. The emergence of the literal-metaphorical anomolous distinction in young children. Child Development
54: 15461.
Winner, E. 1979. New names for old things: The emergence of metaphoric language. Journal of Child Language 6: 46991.
. 1997. The Point of Words: Childrens Understanding of Metaphor
and Irony. Cambridge: Harvard University Press.
Winner, E., M. McCarthy, and H. Gardner. 1980. The ontogenesis of metaphor. In Cognition and Figurative Language, ed. R. P. Honeck and
R. Hoffman, 34161. Hillsdale, NJ: Lawrence Erlbaum.
Winner, E., A. K. Rosenstiel, and H. Gardner. 1976. The development of
metaphoric understanding. Developmental Psychology 12: 28997.
489
490
491
Metaphor, Universals of
. 2008. Hemispheric differences in processing the literal interpretation of idioms: Converging evidence from behavioral and fMRI studies. Cortex 44.7: 84860.
Oliveri, Massimiliano., Leonor Romero, and Costanza Papagno. 2004.
Left but not right temporal involvement in opaque idiom comprehension: A repetitive transcranial magnetic stimulation study. Journal of
Cognitive Neuroscience 16: 84855.
Ortony, Andrew, Diane L. Schallert, Ralph E. Reynolds, and Stephen J.
Antos. 1978. Interpreting metaphors and idioms: Some effects of
context on comprehension. Journal of Verbal Learning and Verbal
Behavior 17: 46577.
Rapp, M. Alexander, Dirk T. Leube, Michael Erb, Wolfgang Grodd, and
Tilo T. J. Kircher. 2007. Laterality in metaphor processing: Lack of evidence from functional magnetic resonance imaging for the right hemisphere theory. Brain and Language 100: 1429.
Searle, John. 1979. Expression and Meaning. Cambridge: Cambridge
University Press.
Schmidt, L. Gwen, Casey J. DeBuse, and Carol A. Seger. 2007. Right
hemisphere metaphor processing? Characterizing the lateralization
semantic processes. Brain and Language 100: 12741.
Stringaris, K. Argyris, Nicholas C. Medford, Vincent C. Giampietro,
Michael J. Brammer, and Anthony S. David. 2007. Deriving meaning: Distinct neural mechanisms for metaphoric, literal, and nonmeaningful sentences. Brain and Language 100: 15062.
Stringaris, K. Argyris, Nicholas C. Medford, Rachel Giora, Vincent C.
Giampietro, Michael J. Brammer, and Anthony S. David. 2006. How
metaphors influence semantic relatedness judgments: The role of the
right frontal cortex. NeuroImage 33: 78493.
Winner, Ellen, and Howard Gardner. 1977. The comprehension of metaphor in brain-damaged patients. Brain 100: 71729.
of evidence from a number of linguists who are native speakers of the respective languages, Zoltn Kvecses (2000) points
out that English, Japanese, Chinese, Hungarian, Wolof, Zulu,
Polish, and others possess the metaphor AN ANGRY PERSON
IS A PRESSURIZED CONTAINER, to various degrees. Ning Yus
(1995, 1998) work indicates that that the metaphor HAPPINESS
IS UP is also present not only in English but also in Chinese. The
system of metaphors called the event structure metaphor (Lakoff
1993) includes such submetaphors as CAUSES ARE FORCES,
STATES ARE CONTAINERS, PURPOSES ARE DESTINATIONS,
ACTION IS MOTION, DIFFICULTIES ARE IMPEDIMENTS
(TO MOTION), and so forth. Remarkably, this set of submetaphors occurs in such widely different languages and cultures as
Chinese (Yu 1998) and Hungarian (Kvecses 2005), in addition
to English. Eve Sweetser (1990) noticed that the KNOWING IS
SEEING and the more general the MIND IS THE BODY metaphors can be found in many European languages and are probably good candidates for (near-)universal metaphors. As a final
example, George Lakoff and Mark Johnson (1999) describe the
metaphors used for ones inner life in English. It turns out that
metaphors such as SELF-CONTROL IS OBJECT POSSESSION,
SUBJECT AND SELF ARE ADVERSARIES, and THE SELF IS A
CHILD are shared by English, Japanese, and Hungarian. Given
that ones inner life is a highly elusive phenomenon and, hence,
would seem to be heavily culture and language dependent, one
would expect a great deal of significant cultural variation in such
a metaphor. All in all, then, we have a number of cases that constitute near-universal or potentially universal conceptual metaphors, though not universal metaphors in the strong sense.
METAPHOR, UNIVERSALS OF
Universal Metaphors?
Native speakers of all languages use a large number of metaphors when they communicate about the world (Lakoff and
Johnson 1980). Such metaphorically used words and expressions may vary considerably across different languages. For
example, the idea that is expressed in English with the words
spending your time is expressed in Hungarian as filling your time.
The images that different languages and cultures employ to
code meanings can be extremely diverse. Given this diversity, it
is natural to ask: Are there any universal metaphors at all, if by
universal we mean those linguistic metaphors that occur in each
and every language? This question is difficult because it goes
against our everyday experiences and intuitions as regards metaphorical language in diverse cultures; it would also be extremely
difficult to study, given that there are 4,0006,000 languages spoken around the world today.
If we go beyond looking at metaphorically used linguistic
expressions in different languages, however, and look at conceptual metaphors instead of linguistic metaphors, we begin
to notice that many conceptual metaphors appear in a wide
range of languages. For example, Hoyt Alverson (1994) found
that the TIME IS SPACE conceptual metaphor can be found
in such diverse languages and cultures as English, Mandarin
Chinese, Hindi, and Sesotho. Many other researchers suggested
that the same conceptual metaphor is present in a large number of additional languages. Several other conceptual metaphors
appear in a large number of different languages. On the basis
492
Metaphor, Universals of
Ta xing congcong de.
he spirit rise-rise PRT
His spirits are rising and rising./Hes pleased and excited.
Zhe-xia tiqi le wo-de xingzhi.
this-moment raise ASP my mood
This time it lifted my mood/interest.
Hungarian:
happiness is up
Ez a film feldobott.
this the film up-threw-me
This film gave me a high.-This film made me happy.
Majd elszll a boldogsgtl.
almost away-flies-he/she the happiness-from
He/she is on cloud nine.
English, Mandarin Chinese, and Hungarian (a Finno-Ugric
language) belong to different language families, which
developed independently for much of their history. It is also
unlikely that the three languages had any significant impact on
one another in their recent history. This is not to say that such an
impact never shapes particular languages as regards their metaphors (e.g., the processes of globalization and the widespread
use of the Internet may popularize certain conceptual metaphors, such as TIME IS A COMMODITY), but only to suggest
that the particular HAPPINESS IS UP metaphor does not exist in
the three languages because, say, Hungarian borrowed it from
Chinese and English from Hungarian.
So how did the same conceptual metaphor emerge, then, in
these diverse languages? The best answer seems to be that there is
some universal bodily experience that led to its emergence. Lakoff
and Johnson argued early that English has the metaphor because
when we are happy, we tend to be physically up, move around, be
active, jump up and down, smile (i.e., turn up the corners of the
mouth), rather than down, inactive, and static, and so forth. These
are undoubtedly universal experiences associated with happiness
(or, more precisely, joy), and they are likely to produce potentially
universal (or near-universal) conceptual metaphors. The emergence of a potentially universal conceptual metaphor does not,
of course, mean that the linguistic expressions themselves will be
the same in different languages that possess a particular conceptual metaphor (Barcelona 2000; Maalej 2004).
Kvecses (1990, 2000) proposed, furthermore, that the universal bodily experiences can be captured in the conceptual
metonymies associated with particular concepts. Specifically,
in the case of emotion concepts such as happiness, anger, love,
pride, and so forth, the metonymies correspond to various kinds
of physiological, behavioral, and expressive reactions. These
reactions provide us with a profile of the bodily basis of emotion
concepts. Thus, the metonymies give us a sense of the embodied
nature of concepts, and the embodiment of concepts may be
overlapping, that is, (near-)universal, across different languages
and language families. Such universal embodiment may lead to
the emergence of shared conceptual metaphors.
Joseph Grady (1997a, 1997b) developed the Lakoff-Johnson
view further by proposing that we need to distinguish complex
metaphors from primary metaphors. His idea was that complex metaphors (e.g., THEORIES ARE BUILDINGS) are composed of primary metaphors (e.g., LOGICAL ORGANIZATION IS
PHYSICAL STRUCTURE). The primary metaphors consist of correlations of a subjective experience with a physical experience.
As a matter of fact, it turned out that many of the conceptual
metaphors discussed in the cognitive linguistic literature
are primary metaphors in this sense. For instance, HAPPY IS UP
is best viewed as a primary metaphor, wherein being happy is a
subjective experience and being physically up is a physical one
that is repeatedly associated with it. Other primary metaphors
include MORE IS UP, PURPOSES ARE DESTINATIONS, and
INTIMACY IS CLOSENESS. On this view, it is the primary metaphors that are potentially universal.
Primary metaphors function at a fairly local and specific level
of conceptualization, and, hence, in the brain. At the same time,
we can also assume the existence of much more global metaphors (see also generic- and specific-level metaphors).
For example, animals are commonly viewed as humans and
humans as animals; humans are commonly conceptualized as
objects and objects as humans, and so on. A famous example of
the objects as humans metaphor was described by Keith Basso
(1967), who showed that in the language of the Western Apache,
cars are metaphorically viewed in terms of the human body. In
addition, the work of Bernd Heine and his colleagues (Heine,
Claudi, and Hnnemeyer 1991; Heine 1995; Heine and Kuteva
2002) reveals other large-scale metaphorical processes that people seem to employ nearly universally; for example, spatial relations are commonly understood as parts of the human body (e.g.,
the head means up and the feet mean down). These conceptual
metaphors seem to be global design features of the brain/mind
of human beings.
It seems to be clear at this point that commonality in human
experience is a major force shaping the metaphors we have. It is
this force that gives us many of the metaphors that we can take
to be near-universal or potentially universal. But commonality
in human experience is not the only force that plays a role in
the process of establishing and using metaphors. There are also
countervailing forces that work against universality in metaphor
production.
493
Metaphor, Universals of
which the Euro-American conceptualization of anger (as well as
that of emotion in general) derived. The humoral view maintains
that the four fluids (phlegm, black bile, yellow bile, and blood)
and the temperatures associated with them regulate the vital
processes of the human body. The humors were also believed to
determine personality types (such as sanguine, melancholy, etc.)
and account for a number of medical problems. The humoral
view exerted a major impact on the emergence of the European
conception of anger as a hot fluid in a pressurized container. By
contrast, Brian King (1989) and Yu (1995 and 1998) suggest that
the Chinese concept of nu (corresponding to anger) is bound up
with the notion of qi, that is, the energy that flows through the
body. Qi in turn is embedded in not only the psychological (i.e.,
emotional) but also the philosophical and medical discourse of
Chinese culture and civilization. When qi rises in the body, there
is anger (nu). Without the concept of qi, it would be difficult to
imagine the view of anger in Chinese culture. Thus, emotion
concepts, such as anger in English, dh in Hungarian (the two
representing European culture), and nu in Chinese, are in part
explained in the respective cultures by the culture-specific concepts of the four humors and qi, respectively. It appears that the
culture-specific key concepts that operate in particular cultures
account for many of the specific-level differences among the various anger-related concepts and the PRESSURIZED CONTAINER
metaphor.
As an example of how differences in human concern can
create new metaphors, consider some well-known conceptual metaphors for sadness: SADNESS IS DOWN, SADNESS IS
A BURDEN, and SADNESS IS DARK. The counterpart of sadness is depression in a clinical context. Linda McMullen and
John Conway (2002) studied the metaphors that people with
episodes of depression use and, with one exception, found the
same conceptual metaphors for depression that nondepressed
people use for sadness. They identified the unique metaphor as
DEPRESSION IS A CAPTOR. Why dont merely sad people talk
about sadness as being a captor? Most people do not normally
talk about being trapped by, wanting to be free of, or wanting
to break out of sadness, although these are ways of talking and
thinking about depression in a clinical context. It makes sense to
suggest that people with depression use this language and way of
thinking about their situation because it faithfully captures what
they experience and feel. Their deep concern is with their unique
experiences and feelings that set them apart from people who
do not have them. It is this concern that gives them the CAPTOR
metaphor for depression (see also emotion and language).
People can employ a variety of different cognitive operations
in their effort to make sense of experience. For example, what I
call experiential focus can have an impact on the specific details
of the conceptual metaphors used, and what is conceptualized
metaphorically in one culture can predominantly be conceptualized by means of metonymy in another (Kvecses 2005). The
universal bodily basis on which universal metaphors could be
built may not be utilized in the same way or to the same extent
in different languages. What experiential focus means is that
different peoples may be attuned to different aspects of their
bodily functioning in relation to a metaphorical target domain
(see source and target) or that they can ignore or downplay
certain aspects of their bodily functioning with respect to the
494
Metaphor, Universals of
Barcelona, Antonio. 2000. On the plausibility of claiming a metonymic
motivation for conceptual metaphor. In Metaphor and Metonymy at
the Crossroads, ed. A. Barcelona, 3158. Berlin: Mouton de Gruyter.
Basso, Keith H. 1967. Semantic aspects of linguistic acculturation.
American Anthropologist, n.s., 69.5: 4717.
Charteris-Black, Jonathan. 2003. Speaking with forked tongue: A comparative study of metaphor and metonymy in English and Malay
phraseology. Metaphor and Symbol 18.4: 289310.
Geeraerts, Dirk, and Stephan Grondelaers. 1995. Looking back at
anger: Cultural traditions and metaphorical patterns. In Language
and the Cognitive Construal of the World, ed. J. Taylor and R. MacLaury,
15379. Berlin: Gruyter.
Grady, Joseph. 1997a. Foundations of meaning: Primary metaphors and
primary scenes. Ph.D. diss., University of California at Berkeley.
. 1997b. Theories are buildings revisited. Cognitive Linguistics
8: 26790.
Haspelmath, Martin. 1997. From Space to Time: Temporal Adverbials in
the Worlds Languages. Munich and Newcastle: Lincom Europa.
Heine, Bernd. 1995. Conceptual grammaticalization and prediction. In
Language and the Cognitive Construal of the World, ed. J. Taylor and R.
MacLaury, 11935. Berlin: Mouton de Gruyter.
Heine, Bernd, Ulrike Claudi, and Friederike Hnnemeyer. 1991.
Grammaticalization: A Conceptual Framework. Chicago: University of
Chicago Press.
Heine, Bernd, and Tania Kuteva. 2002. World Lexicon of
Grammaticalization. Cambridge: Cambridge University Press.
King, Brian. 1989. The conceptual structure of emotional experience in
Chinese. Ph.D. diss., Ohio State University.
Kvecses, Zoltn. 1990. Emotion Concepts. Berlin and New York: SpringerVerlag.
. 2000. Metaphor and Emotion. New York and Cambridge:
Cambridge University Press.
. 2005. Metaphor in Culture: Universality and Variation. Cambridge
and New York: Cambridge University Press.
Lakoff, George. 1993. The contemporary theory of metaphor. In
Metaphor and Thought, ed. A. Ortony, 20251. Cambridge: Cambridge
University Press.
Lakoff, George, and Mark Johnson. 1980. Metaphors We Live By.
Chicago: University of Chicago Press.
. 1999. Philosophy in the Flesh. New York: Basic Books.
Levenson, R. W., P. Ekman, K. Heider, and W. V. Friesen. 1992. Emotion
and autonomic nervous system activity in the Minangkabau of West
Sumatra. Journal of Personality and Social Psychology 62: 97288.
Maalej, Zouhair. 2004. Figurative language in anger expressions in
Tunisian Arabic: An extended view of embodiment. Metaphor and
Symbol 19.1: 5175.
McMullen, Linda, and John Conway. 2002. Conventional metaphors for
depression. In Verbal Communication of Emotion: Interdisciplinary
Perspectives, ed. S. Fussell, 16781. Mahwah, NJ: Lawrence
Erlbaum.
Mithen, Steven. 1996. The Prehistory of the Mind: A Search for the Origin
of Art, Science and Religion. London and New York: Thames and
Hudson.
. 1998. A creative explosion? Theory of mind, language, and the
disembodied mind of the Upper Paleolithic. In Creativity in Human
Evolution and Prehistory, ed. S. Mithen, 16591. London and New
York: Routledge.
Sweetser, Eve. 1990. From Etymology to Pragmatics. Cambridge and New
York: Cambridge University Press.
Yu, Ning. 1995. Metaphorical expressions of anger and happiness in
English and Chinese. Metaphor and Symbolic Activity 10: 5992.
. 1998. The Contemporary Theory of Metaphor in Chinese: A
Perspective from Chinese. Amsterdam: John Benjamins.
Meter
METER
Verse is text that is divided into lines (verse lines). One of the
subtypes of verse is metrical verse. In metrical verse, the length
of the lines is controlled by a set of rules (indirectly, all metrical
rules count syllables). The lines in metrical verse are usually
subject to other restrictions as well, most commonly restrictions
on the rhythm of the line (based on stress, syllable weight,
or lexical tone), and/or on a requirement that a syllable in a
specific line-internal location be word-initial or word-final (a
caesura rule). Some meters also include rules about rhyme or
alliteration.
Although verse is probably a universal (see poetic form,
universals of), found in all oral or literary traditions, there are
poetic traditions without metrical verse, of which perhaps the
best known is the Hebrew poetry of the Old Testament, which
is based on syntactic parallelism rather than on counted syllables. Metrical verse is found in European literatures (Greek,
English, the various Celtic, Germanic, Romance, and Slavic literatures, also Finnish), in Arabic and Islamic literatures (e.g.,
Persian, Urdu, Turkish, Hausa), and in literatures less clearly
influenced by Arabic (such as Berber and Somali), in the literatures of South Asia (e.g., Sanskrit, Pali, Hindi, Malayalam, Tamil),
of Southeast Asia (e.g., Thai, Burmese, Vietnamese), and of East
Asia (e.g., Chinese, Korean, Japanese). Metrical verse is reported
to be largely or completely absent in the poetry of ancient Semitic
literatures and of Australia, non-Islamic Africa, the Americas,
and New Guinea, but this may just be because researchers have
not been looking for it (fieldworkers far too rarely ask questions
about the poetics or poetic practice of a culture).
The variety of meters can be illustrated by some examples.
English iambic pentameter requires a 10- or 11-syllable line,
with even-numbered syllables tending to have stress. The French
alexandrin requires a line of 12 or 13 syllables, with the sixth syllable both stressed and word-final. Swahili shairi requires a line
of 16 syllables, with the eighth syllable word-final (and no control over rhythm). Greek iambic trimeter requires a line of 12 syllables, with even-numbered syllables heavy (containing a long
vowel or ending in two consonants), and the third, seventh and
eleventh syllables light (containing a short vowel ending in at
most one consonant). Arabic kamil requires a line of between 12
and 15 syllables with a complex rhythmic control (in the shortest line the third, seventh and eleventh syllables are light and the
others heavy). Sanskrit sardulavikridita requires a line of 19 syllables with the aperiodic rhythm heavy heavy heavy light light
heavy light heavy light light light heavy (word boundary) heavy
heavy light heavy heavy light heavy. Japanese meters require
lines of five or seven light syllables (but permit a heavy syllable
to substitute for two light syllables). A genre of Vietnamese pairs
a six-syllable line with an eight-syllable line, in which the second, and sixth (and eighth) syllables belong to one tonal class
and the fourth to another. Germanic alliterative meter requires
between two and four stressed syllables, at least two of which
must alliterate.
In literary studies, meter is usually discussed as an aid to
interpretation, and less attention has been paid to the theory that underlies the meter than is desirable. The approach to
meter most common in such studies is the foot combination and
495
Meter
substitution approach. In this approach, a meter such as iambic
pentameter is a template made by combining five iambic feet
each of which is composed of a sequence of an unmarked syllable followed by a syllable that is marked. The resulting template
is matched to a line whose syllables are unstressed or stressed,
so that stressed syllables occupy marked positions in the template and unstressed syllables the unmarked positions. For
lines that are not fully periodic (e.g., in an iambic pentameter
line where the rhythm does not involve a uniform repetition of
unstressed-stressed throughout), the template itself is changed
by substituting a foot of a different kind (e.g., a spondee for an
iamb) to match the stress pattern of the variant part of the line.
This approach only describes the actual rhythm of the line, and
though it offers a convenient vocabulary for the literary critic, it
tells us nothing about the organization of the meter of the line
or why some variations are possible in this meter and some not.
Most recent theoretical accounts express strong reservations or
total rejection of this approach.
Recent theoretical approaches to meter are based primarily
on linguistic theory, particularly on the theory of phonology,
following the foundational work of Morris Halle and Samuel Jay
Keyser (1971). For metrical purposes, most such theories adopt
mechanisms that are used in the theory of phonology, particularly the theory of word stress. Following Mark Libermans (1975)
insight that stress is a matter of the relation between syllables, not
a feature of syllables, different approaches explored the use of
trees and grids as representations in accounts both of word stress
and of metrical poetry (Kiparsky 1977; Hayes 1983). The phonological theory of optimality theory has also been adapted for
use in metrical verse (e.g., Golston and Riad 2005). Nigel Fabb
and Halle (2008) develop their account of poetic meter from a
formalism proposed for word stress by William Idsardi (1992); it
groups the syllables with the help of unpaired parentheses both
in phonology (word stress) and in lines (metrical verse). While in
most approaches the metrical representation is a template built
by special rules and then matched to the line, for Fabb and Halle
the metrical representation is generated from the line (much as
in generative syntax the syntactic representation is generated
from the terminal elements, such as words or morphemes).
In metrical verse, as noted, the length of the line is controlled.
In most cases, the basic unit of measurement is the syllable.
However, in many metrical traditions, some syllables are part of
the line but uncounted. In a common convention, a syllable ending in a vowel precedes a syllable beginning in a vowel, but only
one of the two syllables is counted for metrical purposes (though
both are usually pronounced). The latter fact shows something
important; it shows that the grouping and counting of syllables
for metrical purposes is not directly dependent on the phonology
of the lines. It also poses evident problems for other approaches,
such as that of Kristin Hanson and Paul Kiparsky (1996), which
attempt to account for variation in number of syllables by referring to the specific phonology of the language.
In Japanese, some Indian meters, and some other metrical
traditions, morae are counted, a heavy syllable counting as two
morae and a light syllable as one mora. It is often argued that
the heavy syllable actually consists phonologically of two morae,
but this is not necessary for an explanation of the meter, which
can refer just to the syllable as projecting one or two metrical
496
Methodological Solipsism
Methodology
METHODOLOGICAL SOLIPSISM
Methodological solipsism (MS) is the thesis that mental (or
psychological) states are to be individuated solely by referring
to their relationships with other mental states and the physical
state of someones brain, but not by referring to the physical
world outside the individual to whom those states are ascribed.
This phrase was coined by Hilary Putnam (1975a) in an essay
about meaning externalism and internalism, but its
main advocate (Putnam opposes the thesis) is Jerry Fodor, who
defends it as part of the representational and computational theory of mind (Fodor 1980).
The main issue that MS is concerned with is the relationship between mental states and the outside world or, rather, the
absence of such relationships so far as explanation in scientific
psychology is concerned. In Putnams (1975a) Twin Earth
thought experiment, the question is whether Putnam and his
Twin are in the same mental state when thinking about water,
given that on Earth water is H2O, whereas on Twin Earth water
has the chemical formula XYZ (though it behaves otherwise identically to water on Earth). If mental states are to explain behavior,
Fodor argues, it needs to be the case that Putnam and his Twin
are in the same mental state when they are thinking I would like
to take a dive into the deep waters. In an externalist account of
METHODOLOGY
The topic of methodology most generally involves exploring the
range of responses to the following questions that any researcher
in the language sciences must answer: What sort of empirical
data are you collecting, how are you collecting it, and how do
you hope it will bear on the research question(s) you are trying
to answer? Addressing these questions for a particular specialization within the language sciences falls under the purview
of its experts. My goal here is a more general examination of
issues that arise when language scientists assess alternative data
sources and means of data collection and seek to interpret their
data. These high-level choices require more finesse in language
497
Methodology
sciences than in others, perhaps, because (for the purposes of
most researchers) humans are the only creatures that display the
target phenomenon, namely, language. This creates ubiquitous
challenges owing to the great complexity of the human organism (as compared, say, to a fruit fly) and to ethical considerations
that prevent us from carrying out potentially informative procedures that are used with other species.
I structure the discussion around a taxonomy of the ways
in which data can be collected by language scientists. Practical
exigencies limit the discussion to data about (human) language,
though many language scientists need to gather other sorts of
data as well (e.g., computational linguists collect simulation
data, anthropological linguists collect cultural data, dialectologists collect geographical data, etc.). I further narrow the focus
to data collected for research purposes, leaving aside issues particular to clinical (see speech-language pathology), forensic (see forensic linguistics), and other applications. I also
omit discussion of instrumental, statistical, or other formal treatment of data: This is important, but ultimately futile if ones data
do not properly address the questions to be answered.
My taxonomic framework divides empirical methods along
two dimensions. One dimension along which data collection
can be characterized is by the population of speakers/hearers
(hereafter simply speakers, by which I mean also users of sign
languages) from whom one is collecting data: adult native
speakers of a particular language, children growing up in bilingual households, a creole community, the last surviving
speaker of a dying language (see extinction of languages),
the unknown author(s) of an ancient text, the editorial board of a
dictionary, and so on. A second dimension, more or less orthogonal to the first in principle, is how the language data get from
the speaker(s) to the researcher(s). Researchers may observe the
speaker while he or she is doing something involving language,
or they may gather data produced as the result of a prior event
involving language, via an artifact or another person.
As illustrations, I use hypothetical findings that should not
be taken as statements of fact. Because their importance here
lies in clarifying conceptual points, I have not restricted myself
to attested uncontroversial results, though I believe these hypothetical results not to be wildly implausible.
Different Populations
INTRINSIC INTEREST IN SUBGROUPS. Obviously, if ones research
questions are about a particular population (e.g., the language
of autistic children [see autism and language], speakers of
tone languages), then this is an excellent reason for collecting
data from that population, but it is not the only possible reason,
as will be discussed presently. Research on any group other than
the default (healthy adult native speakers) is virtually always
comparative, if only implicitly: In order to know whether one is
really discovering properties of population X, rather than simply
heretofore unknown properties of human language in general,
X must be compared with population Y with regard to the same
properties. For example, finding that some group of bilinguals
has an average vocabulary size of n in their dominant language
would be most interesting in the context of knowing that an
otherwise-comparable group of monolinguals has an average
vocabulary size of, say, 1.75n. Determining what constitutes
498
an otherwise comparable population is one of the major challenges of research. For example, if you want to study specific
language impairment (SLI), you presumably want to compare children with SLI to unaffected children, but which ones?
If you use children of the same chronological age, you will surely
find many differences in their speech, but this will mainly confirm that a speech pathologist was correct in diagnosing the first
group with SLI. More interesting would be to find that younger
unaffected children whose language is similar to that of the
children with SLI in some respects (e.g., mean length of utterance) nonetheless is more advanced in others (e.g. correct use of
inflectional morphology).
There are numerous ways of classifying speakers into groups
that have seemed fruitful: age (see aging and language),
gender (see gender and language), handedness, education,
native versus non-native speaker, mono- versus bi-/multilingual
(see bilingualism and multilingualism), socioeconomic
status (see sociolinguistics), and many more. Interpreting
any correlations one finds between one of these variables and
some language phenomenon is rarely straightforward, however. For example, if we find that increasing age correlates with
increasing frequency of tip-of-the-tongue states, does that implicate a general decline in memory retrieval with age, or rather an
ability to partially retrieve words that younger people could not
retrieve at all, thanks to greater exposure to these words over the
course of a lifetime (Gollan and Brown 2006)?
RELEVANCE OF ATYPICAL SPEAKERS TO THE STUDY OF TYPICAL
SPEAKERS. A second reason for studying a particular population
is to allow us to learn things about typical language that typical
speakers do not. For example, it has been suggested that certain
language disorders represent an otherwise intact language system from which one circumscribed grammatical mechanism has
been removed or rendered inoperative, as in Yosef Grodzinskys
1986 account of agrammatic aphasia, according to which
traces of movement are missing from otherwise normal syntactic representations. No unaffected speakers would provide us
with the opportunity to pose the question What does syntax
look like if you take out just the traces? Similarly, SLI in later
life could, on certain views, allow us to ask how an incompletely
developed morphosyntax behaves when coupled with adultsized open-class vocabulary and general cognitive capacities,
such as working memory (see working memory and language processing) permitting us to test, for example, the
behavior of complex sentences in such circumstances. The
congnitive immaturity of (typically developing) children precludes this kind of test.
There is a major caveat when dealing with atypical populations, however particularly when analyzing them using theories based on typical populations or drawing conclusions about
such populations. We do not know how circumscribed their
deviation from the norm really is. For example, in the case of
focal brain damage, it is not an innocent assumption to posit that
the speakers subsequent use of language is simply the output of
an otherwise normal brain (as it was before the lesion occurred)
minus whatever function(s) used to be performed by the lost
neural structures. Rather, the speakers language use is the product of a damaged brain that has recovered from injury in ways
Methodology
that we do not yet know how to ascertain. Some functions previously performed in the damaged area may have been taken over
by intact areas, which may in turn have lost some of their original
functionality; areas that were inhibited by the damaged area may
now be free to come into play; and so forth.
Another kind of atypical population includes the expert users
of language: authors, comedians, songwriters, journalists, poets,
playwrights, preachers, politicians, and so on. They can be taken
as proof by example of what it is possible for humans to do with
language, but beyond that we know very little about how they
come by their expertise, and so it is hard to say how they might
inform the study of language in nonexpert speakers.
A special reason for choosing particular speakers to study is
because of their genetic relationship to other speakers. This is
most obvious for language disorders suspected to have a heritable component, such as SLI. But genetic relationships, in particular between twins, can be used in language sciences (as in many
sciences) to approach issues concerning the possible contributions of the genotype to aspects of language in the phenotype.
The standard methodology is to compare monozygotic to dizygotic twin pairs, whereby the former share (on average) twice as
much genetic material. Any phenomenon of language where the
monozygotic pairs are more similar is taken to be shaped more
heavily by prewired brain structures (cf. Ganger 1998).
499
Methodology
people who cannot read. Interviews lose fine-grained timing
information (experimenters should still note gross differences
in response times) but gain elsewhere, including the allowance
for open-ended narrative responses and the possibility of asking
follow-up questions contingent thereon. It is important to note
that the presence/absence of a laboratory setting, electronic
equipment, statistical analysis, and so on has no bearing on the
conceptual/epistemological nature of the data collected.
Finally in this category are data from speakers who are actually using language with no extra task imposed on them. It is
surprising prima facie how little time most language scientists
spend actually observing just these events. The reason is largely
practical. Most research necessarily concentrates on a quite specific aspect of some linguistic phenomenon; waiting for it to arise
by chance is too resource intensive. Nonetheless, it is important
to keep in mind that every step away from the real situations we
are interested in will introduce both random errors and systematic distortions for example, in the case of transcripts, due to
imperfect recording quality and the transcribers subconscious
assumptions about what is being said, respectively.
Conclusion
KINDS OF DELAYED DATA COLLECTION. I classify as a delayed data
situation one in which the object of the researchers measurements, observations, and so on is not a person at a time when he
or she is engaging with language, but rather some indication of
what may have happened at such a time: an artifact produced by
that person, or a behavior observed by someone other than the
researcher. There are two major subclasses of such data.
One subtype comprises any instances of written language,
whether created by an original act of writing or representing an
attempt to transcribe or otherwise keep record of language that
was originally spoken. (Although this distinction is important,
all written material, including phonetic transcription, loses
much information found in spoken language.) This includes
documents from now-dead languages, dictionaries and grammars, dialect atlases, poetry, song lyrics, scripts, and so on
(in some of which the writers intent may be to sound unlike
his own or anyone elses natural speech or prose writing), as
well as corpora amassed specifically for academic purposes.
Any text found on the World Wide Web falls into this category
as well. One can, of course, treat textual material as an object of
study unto itself, ignoring how it was created, but if one wants to
use it as evidence bearing on human language in general, then
considering the many differences between writing and talking
becomes paramount. Most significant is the ability to edit written material after initially producing it (in most situations). The
Web, increasingly used as a corpus because of its size, comes
with many special problems: It can be hard to ascertain who
actually wrote any given passage; it is usually impossible to
establish the native language(s), gender, age, and so on of the
author; the intended meaning and discourse function is often
unclear; and so forth.
The second subtype of delayed data is hearsay, that is, reports
given by someone about language phenomena witnessed, told
about, or engaged in personally. For example, elderly speakers might report that their parents used to use some expression
but that they themselves never used it. This is information that
researchers have no way of independently verifying. Likewise, a
500
Metonymy
. 2005. Thinking about what we are asking speakers to do. In
Linguistic Evidence: Empirical, Theoretical, and Computational
Perspectives, ed. Stephan Kepser and Marga Reis, 45784.
Berlin: Mouton de Gruyter.
METONYMY
Metonymy (Greek change of name) is one of the
major figures of speech recognized in classical rhetoric. The
Roman treatise Rhetorica ad Herennium defines metonymy as
a trope that takes its expression from near and close things by
which we can comprehend a word that is not denominated by
its proper word. This ancient characterization already points to
two criterial notions of metonymy, contiguity and substitution,
which still occur in most present-day definitions of metonymy
as the substitution of one word for another with which it is
associated.
Recent studies in cognitive linguistics have shown that
metonymy is not just a matter of words and their substitution
but is part of human thinking and reasoning. The conceptual
nature of metonymy has been demonstrated by George Lakoff
(1987). For example, the term mother makes many people think
of a housewife mother. The relationship between mothers and
housewives is metonymic and operates only on the conceptual
level: The category mother is metonymically associated with the
subcategory housewife mother as one of its members.
Various cognitive linguists have described the conceptual
basis of metonymy using the notion conceptual frame. Frames
are packages of knowledge about coherent segments of experience. The elements of a frame are conceptually contiguous: Any
element evokes the frame as a whole and, concomitantly, other
elements within the frame network. For example, the concept
author establishes a frame that includes literary works, a publisher, biographical information, etc. Since these elements are
conceptually contiguous, they may be exploited by metonymy.
Thus, we may metonymically refer to a book by naming its author,
as in We are reading Shakespeare. Typically, a metonymic interpretation is coerced when there is a conceptual conflict between
expressions belonging to the same frame. In the previous example, the verb read requires an object that denotes a linguistically
coded content, such as a book or a letter. The conceptual conflict
is resolved by understanding Shakespeare as a reference point
that provides mental access to Shakespeares literary work
(Langacker 1993; Radden and Kvecses 1999).
Studies in metonymy have traditionally focused on words.
Standard examples on the synchronic level include The kettle is boiling (container for content) and Jonathan is in the
phone book (person for name). Metonymic processes on the
diachronic level have been long noted by historical linguists and amply demonstrated since the nineteenth century.
Metonymic shifts have been observed cross-linguistically in a
number of conceptual frames (Koch 1999). For example, in the
marriage frame, a preparatory status of being engaged may
stand for the state of being married. Thus, the Latin word sponsus/
sponsa with the meaning fianc/fiance shifted its meaning to
bride/bridegroom and ended up with the meaning husband/
wife, as in Spanish esposo/esposa, French poux/pouse, and
English spouse.
501
Minimalism
Lakoff, George. 1987. Women, Fire, and Dangerous Things: What Categories
Reveal about the Mind. Chicago: University of Chicago Press.
Lakoff, George, and Mark Johnson. 1980. Metaphors We Live By. Chicago
and London: University of Chicago Press.
Langacker, Ronald. 1993. Reference-point constructions. Cognitive
Linguistics 4: 138.
Panther, Klaus-Uwe, and Gnter Radden, eds. 1999. Metonymy in
Language and Thought. Amsterdam and Philadelphia: Benjamins.
Panther, Klaus-Uwe, and Linda L. Thornburg. 1998. A cognitive approach
to inferencing in conversation. Journal of Pragmatics 30: 75569.
Radden, Gnter, and Zoltn Kvecses. 1999. Towards a theory of metonymy. In Panther and Radden 1999, 1759.
MINIMALISM
Minimalism, extending earlier work in transformational
grammar and generative grammar, conjectures that the
computational system central to human language is a perfect solution to the task of relating sound and meaning. Recent
research has investigated the complexities evident in earlier
models and attempted to eliminate them, or to show how they
are only apparent, following from deeper and simpler properties. Major examples of this work include the reduction of the
number of linguistic levels of representation in the model and
the deduction of certain constraints on syntactic derivations
from general considerations of economy and computational
simplicity.
Like earlier versions of generative grammar, the minimalist
program (MP) (Chomsky 1995b, 2000, 2004, 2005) maintains that
linguistic competence is a computational system creating and
manipulating structural representations. MP further proposes
that the derivations and representations conform to economy
criteria, demanding that they be minimal in a sense determined
by the language faculty (perhaps ultimately by general properties of organic systems): no extra steps in derivations, no extra
symbols in representations, and no representations beyond
those that are conceptually necessary.
Reduction of Levels
Minimalism developed out of the government and binding
(GB) or principles and parameters model (Chomsky 1981,
1982; Chomsky and Lasnik 1993). In that model, there are four
significant levels of representation, related by derivation:
(1)
D(eep)-Structure
S(urface)-Structure
PF
(Phonetic Form)
LF
(Logical Form)
Structure Building
Minimalism, in a partial return to the apparatus of pre-1965
transformational theory (Chomsky 1955), has lexical items
502
In this model, there is no one representation following all lexical insertion and preceding all singulary transformations. That
is, there is no D-structure.
Economy, in the form of shortest move, selects (3) over (4) since
the subject is higher than the object, hence, closer to the sentence initial target of wh-movement than the object is.
The simplifying developments in the theory leading toward
the minimalist approach generally led to greater breadth and
depth of understanding of both how human languages are organized (descriptive adequacy) and how they develop in childrens
Minimalism
minds (explanatory adequacy) (see descriptive, observational, and explanatory adequacy). This success
led Noam Chomsky to put forward the audaciously minimalist conjecture that we are now in a position to go even beyond
explanatory adequacy: The human language faculty might be
a computationally perfect solution to the problem of relating
sound and meaning, the minimal computational system given
the boundary conditions provided by other modules of the mind.
This conjecture leads to a general minimalist critique of syntactic
theorizing, including Chomskys own earlier minimalist theorizing. Consider first the leading idea that multiple derivations
from the same initial set of lexical choices are compared. This
introduces considerable complexity into the computation, especially as the number of alternative derivations multiplies. It thus
becomes desirable to develop a model whereby all relevant derivational decisions can be made in strictly Markovian fashion: At
each step, the very next successful step can be determined, and
determined easily. This arguably more tractable local economy
model was suggested by Chomsky (1995a) and developed in
detail by Chris Collins (1997).
(6)
(10)
The assumption is that the position of the verb vis--vis the adverb
indicates whether the verb has raised overtly. For V-raising, the
feature driving the movement is claimed to be one that resides
in Infl. The feature might be strong, forcing overt movement (as
in French), or weak. Similarly, the feature demanding overt whmovement in English is a strong feature of C. The principle procrastinate disallows overt movement except when it is necessary
(i.e., for the satisfaction of a strong feature as in Chomsky 1993;
Lasnik 1999a).
Procrastinate invited a question. Why is delaying an operation
until LF more economical than performing it earlier? Further,
many of the hypothesized instances of covert movement do
not have the semantic effects (with respect to quantifier scope,
anaphora, etc.) that corresponding overt movements have, as
discussed by Lasnik (1999b, Chapters 6 and 8). To address these
questions, Chomsky (2000; 2001) argues for a process of agreement (potentially at a substantial distance) that relates the two
503
Minimalism
items that need to be checked against each other. Many of the
phenomena that had been analyzed as involving covert movement are reanalyzed as involving no movement at all, just the
operation Agree (though Huangs argument indicates that there
are at least some instances of covert movement). Overt phrasal
movement (such as subject raising) is then seen in a different
light: It is not driven by the need for case or agreement features
to be checked (since that could take place via Agree). Instead, it
takes place to satisfy the requirement of certain heads (including Infl) that they have a specifier (in the X-bar theoretic sense).
Such a requirement was already formulated by Chomsky (1981),
and dubbed the extended projection principle (EPP) in Chomsky
(1982). To the extent that long distance A-movement (basically,
movement to a higher subject position) as in (8) proceeds successive-cyclically through each intermediate subject position,
the EPP is motivated, since, as observed earlier, these intermediate positions are not case-checking positions.
An important question at this point is why language has the
seeming imperfection of movement processes at all. We can
distinguish two major types of movement, phrasal movement
and head movement. Chomsky conjectures that phrasal movement is largely to convey topic-comment information (and possibly to make scope relations more transparent), and that the
EPP is the way the computational system formally implements
this. V-movement, on the other hand, is conjectured to have
PF motivation (guaranteeing that the Infl affix will ultimately
be attached to a proper host, V), and may even be a PF process.
Another possibility is that movement is simply a generalization
of the merge operation combining smaller structures into larger
ones. Given that merge is ineliminable, perhaps move is not an
imperfection after all.
Syntactic Interfaces
The connection between syntactic derivation and semantic and
phonological interfaces has long been a central research area. In
minimalism, interpretation could be distributed over many structures in the course of transformational cycles. Already decades
ago, Joan W. Bresnan (1971) argued that the rule responsible for
assigning English sentences their intonation contour applies
following each cycle of transformations, rather than at the end
of the syntactic derivation. Ray Jackendoff (1972) put forward
similar proposals for semantic phenomena involving scope and
anaphora. Chomsky (2000, 2001) argues for a general instantiation of this distributed approach, sometimes called Multiple
Spell-Out, based on Epstein (1999) and Uriagereka (1999).
At the end of each cycle (or phase in Chomskys more recent
work), the syntactic structure thus far created can be encapsulated and sent off to the interface components for phonological
and semantic interpretation. Thus, even the levels of PF and LF
fade away. Samuel D. Epstein argues that such a move represents
a conceptual simplification (in the same way that elimination of
D-structure and S-structure did), and both Juan Uriagereka and
Chomsky provide empirical justification. The role of syntactic
derivation, always important in Chomskian theorizing, becomes
even more central on this view. Epstein reasons that the centrality of (asymmetric) c-command (as opposed to one of a whole
range of other conceivable geometric relations) in syntax is
predicted on this strongly derivational view, but not in a more
504
A
C
B
D
Multiple Spell-Out effectively deals with a range of reconstruction phenomena. For example, an anaphor normally requires an
antecedent that c-commands it:
(14)
(15)
Minimalism
Conclusion
505
[Not to scale]
Human
Wernickes
A rea
Frontal
Lobe
Occipital
Lobe
Brocas
A rea
Homology
Temporal Lobe
Figure 1. A comparative side view of the monkey brain (left) and human brain (right), not to scale. The left view
emphasizes premotor area F5; the right view emphasizes Brocas area and Wernickes area, considered crucial for
language processing. F5 and Brocas area are considered homologous.
firmly, strip off leaves, remove petioles bimanually, fold leaves
over the thumb, pop the bundle into the mouth, and eat. Teaching
is virtually never observed in apes (Caro and Hauser 1992), and
the young seem to look at the food, not at the methods of acquisition (Corp and Byrne 2002). Moreover, chimpanzee mothers seldom, if ever, correct and instruct their young (Tomasello 1999).
The challenge for acquiring such skills is compounded because
the sequence of atomic actions varies greatly from trial to trial.
Byrne (2003) implicates imitation by behavior parsing, a protracted form of statistical learning whereby certain subgoals (e.g.,
nettles folded over the thumb) become evident from repeated
observation as being common to most performances. Apparently,
the young ape, over many months, may acquire the skill by coming to recognize the relevant subgoals and derive action strategies
for achieving subgoals by trial and error.
The ability to learn the overall structure of a specific feeding
behavior over many, many observations, however, is very different
from the human ability to understand any sentence of an openended set as it is heard and to generate another novel sentence
as an appropriate reply. In many cases, humans need just a few
trials to make sense of a relatively complex behavior and can then
repeat it under changing circumstances, if the constituent actions
are familiar and the subgoals these actions must achieve are readily discernible. (The next section places this facility for complex
imitation in an evolutionary and neurological perspective.) It is
interesting to note that even newborn infants can perform certain
acts of imitation, but this capacity for neonatal imitation such
as poking out the tongue on seeing an adult poke out a tongue
(Meltzoff and Moore 1977) is quantitatively different from that
for complex imitation (see communication, prelinguistic).
506
507
Modality
Myowa-Yamakoshi, M., and T. Matsuzawa. 1999. Factors influencing
imitation of manipulatory actions in chimpanzees (Pan troglodytes).
Journal of Comparative Psychology 113: 12836.
Pinker, S., and P. Bloom. 1990. Natural language and natural selection.
Behavioral and Brain Sciences 13: 70784.
Rizzolatti, G., and M. A. Arbib. 1998. Language within our grasp. Trends
in Neuroscience 21.5: 18894.
Rizzolatti, G., L. Fadiga, V. Gallese, and L. Fogassi. 1996. Premotor cortex and the recognition of motor actions. Cognitive Brain Research
3: 13141.
Seyfarth, R. M., D. L. Cheney, and T. J. Bergman. 2005. Primate social
cognition and the origins of language. Trends in Cognitive Sciences
9.6: 2646.
Stokoe, W. C. 2001. Language in Hand: Why Sign Came Before Speech.
Washington, DC: Gallaudet University Press.
Tomasello, M. 1999. The human adaptation for culture. Annual Review
of Anthropology 28: 50929.
Tomasello, M., J. Call, J. Warren, T. Frost, M. Carpenter, and K. Nagell.
1997. The ontogeny of chimpanzee gestural signals. In Evolution
of Communication, ed. S. Wilcox, B. King, and L. Steels, 22459.
Amsterdam and Philadelphia: John Benjamins.
Yoshikawa, Y., M. Asada, K. Hosoda, and J. Koga. 2003. A constructivist
approach to infants vowel acquisition through mother-infant interaction. Connection Science 15: 24558.
The modern approach to the semantics of modal logic, developed by Saul Kripke and others, is a form of truth conditional semantics based on possible worlds. For example,
a sentence of the form S is true, at a given world w, iff S is
true at every accessible world v. Different meanings of and
are represented by establishing different sets of worlds as accessible. For example, suppose we use to represent it is morally
required that; then, at any world w, the accessible worlds are
those that are ideal, from the point of view of morality in w:
MODALITY
(2)
Definition
In its broadest sense, this term encompasses all means by which
we can talk about hypothetical situations. The conception of
modality includes the following, plus more:
(1) Expressions of necessity and possibility, in any sense
of these terms (English examples: necessary, possible, must,
may).
(2) Expressions of knowledge, belief, desire, and so on (know,
believe, want, must).
(3) Expressions used to indicate how strongly the speaker is
committed to what he or she is saying (perhaps, might).
(4) Expressions used to say that some action is obligatory or
permissible (have [to], must, may, allowed, permit).
(5) Conditional sentences (If then ).
There is a range of narrower senses of the term, each used to
describe a grammatical category or set of related categories. For
example, English can express necessity, possibility, obligation,
permissibility, and ability (and other concepts) by means of a
grammatically special set of auxiliary verbs (must, may, should,
can, etc.). So, when studying English, it is reasonable to define
modality as the range of meanings expressed by these verbs. But
other languages do not have this grammatical category, and so
its also reasonable to define modality differently when studying
these languages.
Semantic Theories
MODAL LOGIC. Much research on modality in natural language
has been inspired by modal logic (see Blackburn, de Rejke, and
Venema 2001 for a brief history). Modal logic typically has two
modal operators, (necessarily, must, or obligatory) and
(possibly, may, permissible), which attach to sentences.
508
(1) (the rich give money to the poor) is true at w iff the rich give
money to the poor in every accessible world v.
A world v is accessible if and only if it is a perfect world, from
the point of view of the moral principles holding in w.
Varieties of Modality
SENTENTIAL MODALITY. Most linguists take as the central cases of
modality examples in which some expression combines with a
nonmodal sentence, making it modal. For example, English must
can be analyzed as: must + (the rich give money to the poor). There
are many distinct subtypes of sentential modality, including:
(1) Deontic modality: having to do with rules, including
morality and law (example: Criminals must be punished).
Modality
509
Modularity
very strong political pressure from France. The European Union
has decided that any official language of a member state may be
used. Since this number is very large, and the costs of translating,
for example, Maltese into Finnish are enormous, the result has
been a creeping usage of English as the de facto but not de jure
official language.
In the history of the modern world-system, as Latin fell out
of diplomatic usage, French took its place. Since 1945, given the
hegemony of the United States in the world-system, English has
displaced French. The story in international scientific discourse
is different. In the nineteenth century, German was the favored
lingua franca. After 1918 and especially after 1945, because of
defeats on the battlefield, it lost this status. Before 1939, at an
international scholarly congress, participants felt free to deliver
their papers in English, French, German, and usually Italian as
well. There was normally no translation, and it was assumed that
scholars could understand at least three of the four languages.
After 1945, international scholarly organizations dropped
German and Italian entirely. In the 50 years since then, the use of
French has declined but is still permitted, and Spanish has joined
French as a permitted but seldom-used language. The inclusion
of Spanish is the direct result of the fact that there are 19 states in
which it is an official language.
In commerce, there have always been lingua francas. Anyone
going to a local market in a major center of a country in the global
South will see merchants capable of conducting their business in
multiple relevant languages. If one looks at discussions among
personnel of large corporations, there has been an increasing
tendency to use English. Nonetheless, it is probably still the
case that the ability to use a widely spoken (official) language
other than English is an advantage to persons doing business in
countries outside the English linguistic zone. In commerce, the
decision on linguistic use is less a matter of coercion than of optimizing the ability to engage in profitable business.
Finally, we should notice the consequences for the geoculture of the world-system of the existence of dominant languages.
The widespread use of English in the twenty-first century is very
advantageous for native English speakers. It is not merely convenient but tends to turn English linguistic eyes into world linguistic eyes. It also, however, has its negative side for native English
speakers. They are often the only ones cut off from the internal
communications of other linguistic zones, as well as from the
possibilities of seeing the world through other linguistic eyes.
It is quite possible that the increasing role of the Internet
in communications of all kinds, along with the declining
power of the United States in the world-system, will lead to the
reemergence of a multipolar linguistic situation, with five to
seven world languages that diplomats, scholars, and business
executives will feel the need to master and use.
Immanuel Wallerstein
MODULARITY
Modularity is the claim that human cognition is compartmentalized into a number of discrete components or modules potentially including vision, audition, moral judgement, theory of
mind, and language. These specialized modules contrast with
510
Modularity
Montague Grammar
. 2000. The Mind Doesnt Work That Way: The Scope and Limits of
Computational Psychology. Cambridge MA: MIT Press.
Karmiloff-Smith, A. 1992. Beyond Modularity. Cambridge, MA: MIT
Press.
Plomin, R., and P. Dale 2000. Genetics and early language development: A UK study of twins. In Speech and Language Impairments
in Children: Causes, Characteristics, Intervention and Outcome, ed.
D. Bishop and L. Leonard, 3551. Philadelphia: Psychology Press.
Smith, N. 2003. Dissociation and modularity: Reflections on language
and mind. In Mind, Brain and Language, ed. M. Banich and M. Mack,
87111. Mahwah, NJ: Lawrence Erlbaum. This article treats in greater
depth many of the issues discussed here.
Smith, N., and I.-M. Tsimpli. 1995. The Mind of a Savant: LanguageLearning and Modularity. Oxford, Blackwell.
Sperber, D. 2002. In defense of massive modularity. In Language, Brain
and Cognitive Development: Essays in Honor of Jacques Mehler, ed.
E. Dupoux, 4757. Cambridge, MA: MIT Press.
MONTAGUE GRAMMAR
Montague grammar is a theory of semantics and the syntaxsemantics interface developed by the logician Richard Montague
(193071) and subsequently modified and extended by linguists,
philosophers, and logicians. Classical Montague grammar
had its roots in logic and the philosophy of language; it quickly
became influential in linguistics, and linguists played a large
role in its evolution into contemporary formal semantics.
The most constant features of the theory over time have been
the focus on truth conditional aspects of meaning (see truth
conditional semantics), a model-theoretic conception of
semantics, and the methodological centrality of the principle of
compositionality.
History
Montague was a student of Alfred Tarski (190283), a pioneer in
the model-theoretic semantics of logic. Montague developed an
intensional logic with a rich type theory and a model-theoretic
possible worlds semantics, incorporating certain aspects
of (formal) pragmatics, including the treatment of indexical
words and morphemes like I, you and the present tense. In the
late 1960s, Montague turned to the project of universal grammar,
which for him meant a theory of syntax and semantics encompassing both formal and natural languages.
Montagues idea that a natural language could be formally
described using logicians techniques was radical. Most logicians considered natural languages too unruly for precise
formalization, while most linguists either had no awareness
of model-theoretic techniques in logic or doubted the applicability of logicians methods to natural languages (Chomsky
1955).
At the time of Montagues work, generative grammar was
established, linguists were developing approaches to semantics,
and the relation of semantics to syntax had become central. The
linguistic wars between generative semantics and interpretive semantics were in full swing (Harris 1993). In introducing Montagues work to linguists, Barbara Partee (1973, 1975)
and Richmond Thomason (1974) argued that Montagues work
offered some of the best aspects of both warring approaches,
with added advantages of its own.
511
Montague Grammar
512
Mood
a formal system; Montagues thesis was that English can be
described as an interpreted formal system.
Barbara H. Partee
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bach, Emmon. 1989. Informal Lectures on Formal Semantics. New
York: State University of New York Press.
Chomsky, Noam. 1955. Logical syntax and semantics: Their linguistic
relevance. Language 31: 3645.
Dowty, David, Robert E. Wall, and Stanley Peters, Jr. 1981. Introduction to
Montague Semantics. Dordrecht, the Netherlands: Reidel. The classic
textbook on Montague grammar.
Gamut, L. T. F. 1991. Logic, Language, and Meaning. Vol. 2. Intensional
Logic and Logical Grammar. Chicago: University of Chicago Press. A
good, rigorous introduction to Montague grammar and its logic.
Harris, Randy Allen. 1993. The Linguistics Wars. Oxford: Oxford University
Press.
Lewis, David. 1970. General semantics. Synthese 22: 1867.
Montague, Richard. 1970. Universal Grammar. Theoria 36: 37398.
Repr. in Montague 1974, 22246.
. 1973. The proper treatment of quantification in ordinary English.
In Approaches to Natural Language, ed. K. J. J. Hintikka et al., 22142.
Dordrecht, the Netherlands: Reidel. Repr. in Montague 1974, 24770.
. 1974. Formal Philosophy: Selected Papers of Richard Montague.
Ed. R. Thomason. New Haven, CT: Yale University Press.
Partee, Barbara. 1973. Some transformational extensions of Montague
grammar. Journal of Philosophical Logic 2: 50934.
. 1975. Montague grammar and transformational grammar.
Linguistic Inquiry 6: 203300.
Partee, Barbara H., with Herman L. W. Hendriks. 1997. Montague grammar. In Handbook of Logic and Language, ed. J. van Benthem and A.
ter Meulen, 591. Amsterdam and Cambridge, MA: Elsevier and MIT
Press. A fuller history and explication of Montague grammar and its
impact.
Thomason,
Richmond.
1974.
Introduction.
In
Formal
Philosophy: Selected Papers of Richard Montague, ed. R. Thomason,
169. New Haven, CT: Yale University Press.
MOOD
Mood forms part of the nonspatial setting of an event, alongside
Mood
subjunctive modality typically, a form expressing desire or
uncertainty on a par with moods. This is problematic since
the distinction between moods as speech-acts and clause types
(which include division between main and subordinate clauses,
where conditional forms would be used) is blurred. The introduction of interrogative mood into the system is largely due to
the existence of languages that have an overtly marked verbal
paradigm used for the interrogative speech-act, as in a number
of languages of Amazonia. Further formal distinctions between
moods as clause types involve prosody and constituent order.
Both imperative and interrogative are characterized by a typical intonation contour. Imperatives often have fewer categories than corresponding declaratives. The English imperative is
perhaps the simplest form in the language: It consists of the base
form of the verb without any tense inflection, whose subject
typically, the addressee can be and often is omitted. In contrast, many languages of North and South America and Siberia
distinguish delayed versus immediate imperatives and proximal
versus distal imperatives. The universal property of imperatives
is having the second person as subject, of a transitive or intransitive verb (Dixon 1994, 13142). A prototypical imperative is
agentive, and this is why in numerous languages imperative cannot be formed on passive and stative verbs. Other moods do not
have such restrictions. Imperatives directed at the first person
(e.g., Lets go!), also known as hortatives, are often expressed differently from second person imperatives. Imperatives directed at
the third person (e.g., Long live the king!), also known as jussives,
may share similarities with first person imperatives, or have
properties different from all other imperative forms. Further,
minor moods include exclamative (as in Thats so tacky!) and
expressive types, such as imprecatives (or curses, often cast as
commands but without a command meaning).
Mood interacts with modality, understood as a means used
by the speaker to express his or her opinion or attitude towards
the proposition that the sentence expresses or the situation
that the proposition describes (Lyons 1977, 452). Expressions
of probability, possibility, and belief are epistemic modalities,
and expressions of obligation are deontic modalities. In English,
these meanings are conveyed by modal verbs, e.g., he might
come or he must have come (epistemic), he must come (deontic)
(see Palmer 1986, 51125; Jespersen 1924, 3201). Further modal
distinctions include desiderative (unachievable desire), optative
(achievable desire), conditional, hypothetical, potential, purposive, and apprehensive (lest). Languages with a rich verbal
morphology may have special marking for each distinction.
An alternative (rare) cover term for both mood and modality is
mode (Chung and Timberlake 1985).
Some languages have an affix with a general meaning of
irrealis covering possibility, future, negative statements, and
commands. These languages have the category of reality status,
the grammaticalized expression of an events location either in
the real world or in some hypothetical world (see Elliott 2000, for
its cross-linguistic validity). In Maung, an Australian language,
statements in the present, past, and future are marked with realis suffixes. Potential meanings I can do X are expressed with
irrealis, as are commands. In Manam, an Oceanic language, irrealis covers future, probable, and counterfactual statements, positive commands, and habitual actions. But in Yuman languages
Morpheme
and in Caddo, from North America, realis marks statements and
commands, while irrealis expresses future, possibility, and condition. This shows that the realisirrealis distinction is language
specific and that it is distinct from mood (see Mithun 1999,
17880). Mood, modality, and reality status are distinct from evidentiality (q.v.) whose primary meaning is information source.
Mood is often an obligatory inflectional category of the verb,
marked with affixes (suffixes or prefixes, rarely infixes); it is never
expressed derivationally. In languages of an isolating profile,
mood can be expressed through particles. Modalities are not
obligatory and, thus, do not constitute part of an inflectional paradigm. Modal verbs express modalities rather than moods (this
is the case in English and many other familiar Indo-European
languages).
Forms of mood marking can develop additional meanings
overlapping with modalities. Imperative forms can be used to
express optative and conditional, while indicative forms may
develop overtones of certainty (associated with epistemic modality). Indicative forms for instance, future can be used as command strategies, with differences in illocutionary force.
Alexandra Y. Aikhenvald
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Aikhenvald, Alexandra Y. 2004. Evidentiality. Oxford: Oxford University
Press.
Chung, Sandra, and Alan Timberlake. 1985. Tense, aspect and mood.
In Language Typology and Syntactic Description. Vol. 3: Grammatical
Categories and the Lexicon. Ed. Timothy Shopen, 20258.
Cambridge: Cambridge University Press.
Dixon, R. M. W. 1994. Ergativity. Cambridge: Cambridge University
Press.
Elliott, Jennifer R. 2000. Realis and irrealis: Forms and concepts of the
grammaticalisation of reality. Linguistic Typology 4: 5590.
Fortescue, Michael. 1984. West-Greenlandic. Beckenham, UK: Croom
Helm.
Jespersen, Otto. 1924. The Philosophy of Grammar. London: George Allen
and Unwin.
Lyons, John. 1977. Semantics. Vol. 2. Cambridge: Cambridge University
Press.
Mithun, Marianne. 1999. The Languages of Native North America.
Cambridge: Cambridge University Press.
Palmer, F. R. 1986. Mood and Modality. Cambridge: Cambridge University
Press.
Sadock, Jerrold M., and Arnold M. Zwicky. 1985. Speech act distinctions in
syntax. In Language Typology and Syntactic Description. Vol. 1: Clause
Structure. Ed. Timothy Shopen, 15596. Cambridge: Cambridge
University Press.
MORPHEME
This term has been used in two ways: In Leonard Bloomfields
sense, a morpheme is a minimal meaningful form; in Zellig
Harriss and Charles F. Hocketts later usage, a morpheme is an
abstract unit of analysis realized by a morph (= minimal meaningful form) or by a set of synonymous morphs in complementary distribution. In Bloomfields sense, the plural suffixes -s
and -en are distinct morphemes; in the latter sense, they are
distinct morphs realizing the same morpheme. The term is not
always used consistently in morpheme-based approaches to
513
Morphological Change
morphology. In paradigm-based approaches, no linguistic
principle is assumed to make essential reference to morphemes
as a unified class of elements.
Gregory Stump
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bloomfield, Leonard. 1933. Language. New York: Henry Holt and Co.
Harris, Zellig S. 1942. Morpheme alternants in linguistic analysis.
Language 18: 16980.
Hockett, Charles F. 1947. Problems of morphemic analysis. Language
23: 32143.
MORPHOLOGICAL CHANGE
Morphological change involves alterations made by speakers over
time to the analysis of complex words, or to the relations between
a lexical base and its compounds and derivatives, or to the set
of inflected words that share a common lexical base. The main
mechanisms of morphological change are reanalysis (the reinterpretation of forms) and the extension of patterns to create new
forms. The impetus for reanalysis often comes from changes to
the semantics, phonology, or syntax of the affected forms.
Semantic shift may affect the function of a grammatical element. Thus, in various Australian languages, a subordinating
purposive suffix (in order to VERB), typically -ku, came to be
used in independent clauses as a marker of intentional mood
(may VERB), then further shifted to express future tense (will
VERB). Functional shift in a grammatical affix also took place
in some Karnic (Australian) languages, in which a locative (at)
case suffix -nga came to mark the dative (for) function. Meaning
changes may lead to the reinterpretation of compound words as
simple lexemes (e.g. hlford loaf + ward > lord, shep-herd
sheep+ herder > shepherd), or as a lexical stem plus a derivational affix (king-dom) with downgrading of a bound lexical form
-dom meaning condition to a derivational suffix.
Sound changes often create new allomorphs: For example,
the earlier English plural suffix -z split into three variants -z, -z,
-s, and the stem long developed a variant leng- in length with the
e conditioned by the following i in the former derivational suffix
-ith. As the relationship between words becomes obscured by
the accumulation of sound changes, some phonological differences become morphologized, that is, reinterpreted as partial or
even sole signals of a morphological property. Thus, the vowel e
in slep-t (vs. ee in sleep) helps to mark past tense, and ee in feet
(vs. oo in foot) alone marks plural. Even sound changes that
Pre-Greek
(Doric) Greek
EMGerman
ModGerman
German
leave
leave
give
give
sleep
1Sg
leip
leip
Gib
geb-e
schlafe
2Sg
leiteis
leipeis
gib-st
gib-st
schlfst
3Sg
leitei
leipei
gib-t
gib-t
schlft
1Pl
leipomen
leipomen
geb-en
geb-en
schlafen
2Pl
leitete
leipete
geb-t
geb-t
schlaft
3Pl
leiponti
leiponti
geb-en
geb-en
schlafen
514
Morphological Typology
Morphology
ev
-ler- im
- iz -den
house -PL -POSS. 1-PL -ABL
The third type of language expresses differences in morphosyntactic and lexicosemantic properties through contrasting modifications, or inflections of a words stem. These are inflectional
or fusional languages. The classical languages, Greek, Latin, and
Sanskrit, belong to this type. In Latin you (sg[singular]) loved
is expressed by various modifications of the root am- love to
yield amvist: stem formative -v to express perfect, and -ist to
express perfect (again) + 2d person + singular. Typically, properties are fused in one exponent: Here aspect, person and number agreement are expressed together. Equally, a property can
be expressed by more than one exponent: Here perfect is being
expressed twice.
There has been general unease among modern linguists with
the classical typology. One reason is that languages rarely fall
cleanly into one of these types. For example, Mandarin Chinese
productively uses what looks like a derivational suffix to build
agentive nouns, the word q mechanism: sn-r q cooler,
jin-c q monitor, yng-shng q speaker; compare the
English -er/-or agentive suffix (Hippisley, Cheng, and Ahmad
2005). More importantly, there is some doubt that the typology
offers any theoretical insight, a point argued as far back as Sapir
(1921). Part of the reason is that morphological type is really a
function of other grammatical structures worthy of typological investigation, and is, therefore, epiphenomenal (Anderson
1990).
A more promising approach is to focus on much more narrowly defined word structures and to investigate how they
cross-cut languages that may or may not be genetically or typologically related. The result is then a typology of narrowly defined
structures of words that answer the question What is a possible
word? This is the approach taken by Greville G. Corbett and colleagues, who look at unusual morphology such as suppletion,
deponency, and defectivenesss, recording such structures in a
large number of individual languages and inducing diachronic
and synchronic models of their appearance and use in syntax
(e.g. Corbett 2007; Baerman and Corbett 2007).
Andrew Hippisley
MORPHOLOGICAL TYPOLOGY
typology has its origins in nineteenth-century morphological typology, a method of grouping languages not according
to genetic relatedness but to structural similarity, where the
structure was specifically word structure (see morphology ).
Traditionally, there are three possibilities for phonologically
expressing morphosyntactc (inflectional) and lexicosemantic
(derivational) properties at the level of the word. In an isolating or analytical language, complex words are built from
existing words, free forms. Mandarin Chinese could be viewed
as an isolating language. Productive coining of new terms is
through compounding. The word for Internet is h-lin
wng with h inter + lin related + wng net. In agglutinating languages, the pieces of a complex word map onto
specific meaning elements biuniquely, both at the lexical and
grammatical level. Turkish evlerimizden from our houses is
glossed as
MORPHOLOGY
Morphology and Words
While the lexicon of a language lists basic forms and their content
(meanings and grammatical properties), a languages complex
515
Morphology
words neednt be invariably listed, since their form and content
are often partially or wholly deducible from those of their parts
by means of regular principles. This system of principles is the
languages morphology.
In morphological theory, it is useful to distinguish three senses
of word. In one sense, the form come is the same word in (1) and
(2); in another sense, come in (1) is a different word from come
in (2). This apparent paradox arises because word can be used
to refer to either a phonological or a grammatical unit: In (1) and
(2), come represents the same phonological unit (phonetically
[km]) but two distinct grammatical units: the unmarked infinitive form of the verb come in (1) and the past participial form of
this verb in (2).
(1)
(2)
Grammatical words
b.dog, {plural}
Branches of Morphology
A languages morphology comprises two systems. The inflectional
system defines the phonological realization of the grammatical words in a lexemes paradigm; for instance, the inflectional
system of English specifies that the third-singular (3sg) present
indicative form of the lexeme talk is talks. By contrast, the system of word formation (better: lexeme formation) defines complex lexemes in terms of simpler lexemes. The latter system itself
516
Morphology
kulu should alone suffice to mark this form for class 7 agreement; the appearance of the additional prefix ca- not only seems
unnecessary but actually violates the anti-redundancy principle
(Kiparsky 1982; 136 f) purported to prevent the suffixation of plural -s to English men.
The alternative to an incrementalist theory is a realizational
theory, according to which a words content determines its morphological form (Matthews 1972; Zwicky 1985; Anderson 1992).
In a realizational theory, the paradigm of the verbal lexeme talk
includes the grammatical word in (3a), and the phonological
word that realizes this grammatical word arises from talks root
through the application of any rules associated with the morphosyntactic property set in (3a); there is one such rule, which
realizes the property set {3sg present indicative} through the suffixation of s.
In a realizational theory, the fact that a words form may
underdetermine its content is unproblematic, since content is
not deduced from form in any event. Thus, the fact that the firstsingular aorist form krd-o-x I stole in Bulgarian has no exponent of first-singular subject agreement is simply the effect of a
kind of poverty in the languages verb morpology: It happens not
to have any means of expressing the property 1sg in the realization of the grammatical word steal, {1sg aorist}. Syncretism
is likewise unproblematic: One need only assume that the realization of one word in a lexemes paradigm may pattern after the
realization of a different word in that paradigm. Rules of referral
(Zwicky 1985; Stump 1993) express this kind of relation between
cells in a paradigm; thus, in Sanskrit, a rule of referral specifies
that the realization of a neuter nouns nominative singular cell is
the same as that of its accusative singular cell. Finally, extended
exponence is unproblematic in a realizational theory; in the case
of Nyanja ca-ci-kulu large [class 7], one need only assume that
more than one rule of prefixation participates in the realization
of the grammatical word large, {class 7}.
By (5a), the lexeme walk has the past tense form walked; by
(5b), this lexeme also has walked as its past participle. A central assumption in paradigm-based theories is that rules act as
defaults and are therefore subject to override; in the inflection
of verbal lexemes such as sing, the rules in (5) are overridden
by those in (6). An important concern in such theories is that of
establishing general principles regulating the default/override
relations among rules of morphology. In network morphology,
these relations are regulated by their position in default-inheritance hierarchies; in paradigm function morphology, they are
regulated by the Pinian determinism hypothesis (Stump 2001,
23), according to which Rule A overrides Rule B if and only if A is
narrower in application than B.
(6) Where L belongs to the sing class and has root R,
a. L, {finite past } is realized through the substitution of
[] for [] in R.
b. L, {past participle} is realized through the substitution of
[] for [] in R.
517
Morphology
must be seen as being overridden by a null affix, which by stipulation triggers a rule of internal modification. What emerges is
a widely recurrent coincidence that is never explained: Again
and again, a zero affix is stipulated as the unoverridden override among a set of competing morphemes; over and over, this
unoverridden override by stipulation triggers a rule of internal
modification. There is, of course, no overt class of phonologically
identical affixes in any language that ever shows the kind of syntagmatic and paradigmatic distribution that DM must stipulate
for the artifactual class of phonologically null affixes upon which
this approach depends.
518
Morphology, Acquisition of
Corbett, Greville G., and Norman M. Fraser. 1993. Network morphology: A DATR account of Russian nominal inflection. Journal of
Linguistics 29: 11342.
Embick, David, and Rolf Noyer. 2007. Distributed morphology and the
syntax-morphology interface. In The Oxford Handbook of Linguistic
Interfaces, ed. G. Ramchand and C. Reiss, 289324. Oxford: Oxford
University Press.
Finkel, Raphael, and Gregory Stump. 2009. Principal parts and degrees
of paradigmatic transparency. In Analogy in Grammar, ed. J. Blevins
and J. Blevins, 1353. Oxford: Oxford University Press.
Halle, Morris, and Alec Marantz. 1993. Distributed morphology and
the pieces of inflection. In The View from Building 20, ed. K. Hale and
S. Keyser, 11176. Cambridge, MA: MIT Press.
Kiparsky, Paul. 1982. From cyclic phonology to lexical phonology. In
The Structure of Phonological Representations (Part I), ed. H. van der
Hulst and N. Smith, 13175. Dordrecht, the Netherlands: Foris.
Lieber, Rochelle. 1992. Deconstructing Morphology. Chicago: University
of Chicago Press.
Matthews, P. H. 1972. Inflectional Morphology. Cambridge: Cambridge
University Press.
. 1991. Morphology. 2d ed. Cambridge: Cambridge University
Press.
Selkirk, Elisabeth O. 1982. The Syntax of Words. Cambridge, MA: MIT
Press.
Spencer, Andrew, and Ana Lus. 2005. A paradigm function account of
mesoclisis in European Portuguese (EP). In Yearbook of Morphology
2004, ed. Geert Booij and J. van Marle, 177228. Dordrecht, the
Netherlands: Springer.
Stump, Gregory T. 1993. On rules of referral. Language 69: 44979.
. 2001. Inflectional Morphology. Cambridge: Cambridge University
Press.
. 2006. Heteroclisis and paradigm linkage. Language 82:
279322.
. 2007. A non-canonical pattern of deponency and its implications. In Deponency and Morphological Mismatches, ed. Matthew
Baerman, Greville G. Corbett, Dunstan Brown, and Andrew Hippisley,
7196. Oxford: Oxford University Press.
Zwicky, Arnold M. 1985. How to describe inflection. In Proceedings
of the Eleventh Annual Meeting of the Berkeley Linguistics Society, ed.
M. Niepokuj, M. VanClay, V. Nikiforidou, and D. Feder, 37286.
Berkeley, CA: Berkeley Linguistics Society.
Zwicky, Arnold M., and Geoffrey K. Pullum. 1983. Cliticization vs. inflection: English nt. Language 59: 50213.
MORPHOLOGY, ACQUISITION OF
The acquisition of morphology has played a central role in
exploring both the acquisition of syntax (see syntax, acquisition of) and lexical acquisition. The study of acquisition
of morphology can be distinguished by research in first language
acquisition (FLA) and second language acquisition (SLA)
and by research into inflectional or derivational morphology.
Little research has attempted to link acquisition of inflectional
and derivational morphology. There has been some influence
of work in FLA on that in SLA, in particular in debates about
the order of acquisition of morphemes for second language
learners.
The acquisition of inflection is often examined as the acquisition of morphosyntax, that is, structures that are governed by
both morphological and syntactic rules such as subject-verb
agreement. The majority of literature focuses on children under
the age of seven, as inflectional morphology is acquired during
Inflectional Morphology
FIRST LANGUAGE. One major set of works investigating the
acquisition of morphemes examined the order of acquisition of
inflectional morphemes (Cazden 1968; Brown 1973; de Villiers
and de Villiers 1973), focusing on English. This work found a
consistent (though not identical) order of acquisition among
children. Cross-linguistic work demonstrated that there was no
universal order of morpheme acquisition between languages,
that the order and speed of acquisition depends on the target
language and the morphemes themselves (Slobin 1985; Clark
1998). Work on other languages has also found consistent, but
not identical, orders within a language. Several factors appear to
influence the order of acquisition, including perceptual salience,
complexity of the morpheme either semantically (how many
concepts it encodes) or formally (how variable the affix is, how
many parts it contains), and frequency in the input. Eve V. Clark
(1993) suggested that several principles transparency (how
easily the meaning is derived from the parts), simplicity (how
variable the forms are), and productivity govern the order of
acquisition of both inflectional and derivational morphemes.
Morphemes that are consistent in form (have few allomorphs)
and semantically encode a single feature, such as plural or progressive -ing, tend to be acquired earlier than morphemes that
show more allophonic variation, such as with regular past tense
-ed, and/or encode multiple features, such as third person
singular -s (Brown 1973; Clark 1993). This preference to encode
single features/forms with a single morpheme holds across languages and even for children acquiring more than one language.
For example, Melanija Mikes (1967) discussed the acquisition of
locatives for bilingual Hungarian-Serbo-Croatian children. The
Hungarian locative suffix, which is relatively transparent, was
acquired earlier than the semantically equivalent structure in
Serbo-Croatian, which required locative prepositions + agreement, which varied by gender.
Inflected forms appear early, from the earliest word use, especially in highly or consistently inflected languages (Slobin 1985).
Languages such as English, with fewer inflected forms, have bare
forms appearing first and inflected forms appearing later, often
concurrently with first word combinations. In English, childrens
production of inflectional morphology begins with an initial
period of some variability, gradually becoming more consistent
519
Morphology, Acquisition of
over time. Across languages, some morphemes (such as plurals)
are consistently produced accurately by the age of three or four,
whereas others, such as conditional marking of verbs, are often
not mastered until ages seven or eight. While little research has
been done on inflectional prefixes, it has been suggested that
inflectional prefixes are more difficult and are acquired later
than suffixes (Slobin 1982; Clark 1998).
In all languages, children often go through a period during
which they overregularize irregular forms in their grammar,
for example, breaked or mans. Irregular forms are often acquired
early (broke/men) and then when regular forms are acquired,
overregularized forms coexist with irregular forms (breaked/
broke) for a time until the irregular forms and exceptions are
mastered (Marcus et al. 1992).
The majority of research on inflectional morphology has
focused on the verbal and nominal domains, with relatively
little work on adjectives or adverbs. According to Clark (1998),
the earliest verb forms across languages tend to be imperative,
infinitive verb forms, and third person singular; singular forms
tend to be acquired before plural forms. The characteristics of
individual languages determine which agreement markers
are learned earliest in any given language. In terms of tense
and aspect, present/nonpresent is the first distinction children
make, and distinctions between past, present and future appear
to be in place by age three. In languages that distinguish aspect,
aspect also appears to be acquired around age three. However,
early aspect marking is also associated with the semantic characteristics of the verb. For example, Li and Shirai (2000) argue
that early use of perfective aspect is more likely to occur with
telic (e.g., walked) or resultative verbs (e.g., smashed) than other
types of verbs. Compound or periphrastic tenses such as present
perfect are acquired later, and may not be in place until age five
or older.
Within the nominal domain, number marking occurs early
in nouns and is one of the earliest nominal morphemes to be
seen. gender marking also appears early, just after the first
nouns. Early gender/noun class marking appears to be based
primarily on the phonological shape of the noun and later
becomes associated with individual lexical items. Noun class
marking appears by age three, but adultlike acquisition of gender or noun class marking, which requires attention to both
phonological form and semantics, does not appear until age
four or five in many languages (Demuth 2003). In languages
with classifier systems, such Chinese, Japanese or Thai, general
classifier patterns appear early, and more fine-grained semantic distinctions appear gradually. Case marking occurs early as
well, just after first nouns are learned. In nominative-accusative languages, the first distinction to be acquired appears to
be between nominative and accusative. Dative case is next,
followed by other oblique cases. There is also evidence from
ergative languages that the major case distinctions in these languages are acquired early as well. Languages in which case varies by gender (e.g., Russian) take longer for contrasts to appear
than in languages that simply mark for case (e.g., Hungarian,
or Turkish).
Inflectional morphology is generally mastered by the early
school years for all but the most infrequent or irregular constructions across languages.
520
SECOND LANGUAGE. Early research in second language acquisition of inflectional morphology investigated whether there was
a consistent order of acquisition of morphemes. Initial studies
reported consistent orders of acquisition across both adult and
child second language learners of English (e.g., Dulay and Burt
1974). Later research, however, criticized these early studies for
their methodology, and found variability among learners of different backgrounds that seemed to belie a single order of acquisition for second language learners (e.g., Hakuta 1974; Rosansky
1976). The present consensus seems to be that morpheme language studies provide strong evidence that ILs [interlanguages,
developing grammars] exhibit common accuracy/acquisition
orders (Larsen-Freeman and Long 1991, 92). After this intense
period of debate about the acquisition order of morphemes, later
research in SLA has focused on one of several areas: research on
specific structures such as tense or aspect, the role of context in
the acquisition of morphology, and the implications of missing
morphology for the developing grammars of second language
learners. Research has begun to look again at order of acquisition
in SLA, considering how models of acquisition and/or the functor morphemes themselves can explain the order (Goldschneider
and DeKeyser 2001). Here again, the discussion parallels that of
first language acquisition, with perceptual salience, morphophonological regularity, semantic complexity, and frequency being
posited as factors contributing to the order of acquisition.
Derivational Morphology
FIRST LANGUAGE. Because derived forms are often used to fill
semantic gaps within a lexicon, the acquisition of derivational
morphology is studied within the domain of vocabulary learning. Derivational morphology is acquired somewhat later, and
with a less clear order of development, than inflectional morphology. Highly inflected languages such as Turkish or Finnish
show evidence of early derivational morphology. Studies indicate that the earliest derivations are zero-stem alternations (e.g.,
the noun knife used as a verb) and compounding (e.g., dog-book)
for languages where these processes are productive (Clark 1993).
Agentive suffixes (-er in particular) and diminutives (dogg-y) are
also acquired between ages two and three across many languages.
However, the bulk of derivational morphology is acquired during
middle childhood and adolescence (Tyler and Nagy 1989).
As with inflectional morphology, the course of development for derivational morphology varies from language to language and depends on both the patterns within the language
and the properties of the derivations themselves. For example,
Hebrew-speaking children derive verbs from nouns as young as
age three, while derived nominals are acquired later, after age
eight and continuing into adolescence (Berman 2003; Ravid
and Avidor 1998). Awareness of and ability to decode derivational morphology continues to develop throughout the school
years. Comprehension of derivational morphology is correlated
with reading skills in a number of languages, including English,
Hebrew and Chinese (Tyler and Nagy 1990; Levin, Ravid, and
Rapaport 2001; Kuo and Anderson 2006). As with inflectional
morphology, acquisition appears to be best explained by the
transparency of the morpheme, complexity (semantic and formal), and productivity (Clark 1998). Later-learned morphemes
such as -tion have more allomorphs, make more changes to the
Morphology, Acquisition of
stem they attach to, encode more concepts, and/or are less productive than early learned morphemes such as -er.
SECOND LANGUAGE. As with first language (L1) acquisition, second language acquisition of derivational morphology has been
studied together with general vocabulary growth and knowledge of the lexical semantics of the language (Redouane 2004;
Montrul 2001; Lardiere 1997). Little research has examined this
area of second language acquisition in detail. The research that
exists suggests that beginning learners tend to fill lexical gaps
with word formation strategies from their L1 or by extending
existing second language vocabulary. More advanced learners are more likely to use derivational morphology found in the
target language, although they are still likely to use non-target
forms. Evidence also suggests that derivational morphology patterns that are substantially different from the L1 present considerable problems to second language learners and are learned
only gradually through time.
Current Debates
Major areas of debate in the acquisition of morphology include
the status of missing inflectional morphemes in learners grammars, whether productive use of morphemes reflects rules and
an innate capacity for rule formation, and whether regular and
irregular forms are acquired and represented similarly.
For first language acquisition, one area of strong debate has
been the status of missing or omitted morphology by children.
Within a general nativist framework, explanations range from
maturation of grammatical structures, to prosodic and/or phonological learning, to language-specific lexical learning (e.g.,
Peters and Stromqvist 1996; Santelmann, Berk, and Lust 2000).
More empiricist approaches point to such issues as frequency
of morphemes, vocabulary size, and transparency of the morphemes themselves to explain the course of acquisition (e.g.,
Rowland et al. 2003).
Parallel debates have taken place within the literature on
second language acquisition. Within the search for a consistent
order of morpheme acquisition, debates have focused on the
possible reasons for such an order: Is it due to an underlying universal grammar (Dulay and Burt 1974), characteristics of the
input, or general learning and characteristics of the morphemes
themselves (Goldschneider and DeKeyser 2001). In addition, the
issue concerning why adult learners, unlike child learners, continue to make mistakes with inflectional morphology even after
years of exposure to a language has been a source of significant
debate. Explanations for this phenomenon range from a lack of
access to universal grammar for adult learners to issues with prosodic differences between languages, differences in proficiency,
or other factors.
Within the domains of both inflectional and derivational
morphology, another major theoretical debate concerns the
nature of rules. Do children and adults learn rules, or do they
simply extract regularities from the speech stream? This debate
has been particularly strong in the area of regular versus irregular verbs. One side of the debate argues that children use rules for
regular verbs, and thus have distinct processes for forming and
representing past tense with regular versus irregular forms (e.g.
Marcus 1996). This view argues that irregular forms are stored as
Summary
The acquisition of morphology is central to both morphosyntactic development and lexical development for first and second
language learners. While the acquisition of inflectional morphology is largely complete for children by the time they enter
school, it is often problematic for adult learners of a language.
Derivational morphology, on the other hand, is an ongoing, lifelong process for both first and second language learners. Little
research has attempted to link the two types of morphological
acquisition. More research on both derivational and inflectional
morphology is needed on a wider variety of languages or varying
typologies, in particular on non-Indo-European languages. This
is particularly the case for SLA, where the bulk of the research
has focused on European languages.
Lynn Santelmann
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Akhtar, N., and M. Tomasello. 1997. Young childrens productivity
with word order and verb morphology. Developmental Psychology
33.6: 95265.
Berko, Jean. 1958. The childs learning of English morphology. Word
14: 15077.
Berman, Ruth A. 2003. Childrens lexical innovations: Developmental
perspectives on Hebrew verb structure. In Language Processing and
Acquisition in Languages of Semitic, Root-Based, Morphology, ed.
J. Shimron. Amsterdam: John Benjamins.
Brown, Roger W. 1973. A First Language: The Early Stages.
Cambridge: Harvard University Press.
Cazden, Courtney B. 1968. The acquisition of noun and verb inflections.
Child Development 39: 43348.
Clark, Eve V. 1993. The Lexicon in Acquisition. Cambridge: Cambridge
University Press.
. 1998. Morphology in language acquisition. In The Handbook of
Morphology, ed. A. Spencer and A. Zwicky. Oxford: Blackwell.
Demuth, K. 2003. The acquisition of the Bantu languages. In The Bantu
Languages, ed. D. Nurse and G. Phillipson. Surrey : Curzon.
de Villiers, Jill G., and Peter A. de Villiers. 1973. A crosssectional study of
the acquisition of grammatical morphemes. Journal of Psycholinguistic
Research 2: 26778.
Dulay, Heidi, and Marina Burt. 1974. Natural sequences in child second
language acquisition. Language Learning 24: 3753.
Goldschneider, Julie M., and Robert M. DeKeyser. 2001. Explaining
the natural order of L2 morpheme acquisition. Language Learning
51: 150.
521
Morphology, Acquisition of
Hakuta, Kenji. 1974. A preliminary report on the development of grammatical morphemes in a Japanese girl learning English as a second language. Working Papers on Bilingualism 3: 1838.
Kopcke, K. M. 1998. The acquisition of plural marking in English and
German revisited: Schemata versus rules. Journal of Child Language
25.2: 293319.
Kuo, Li-jen, and Richard C. Anderson. 2006. Morphological awareness
and learning to read: A Cross-language perspective. Educational
Psychologist 41.3: 16180.
Lardiere, Donna. 1997. On the transfer of morphological parameter
values in L2 acquisition. Proceedings of the Annual Boston University
Conference on Language Development, 1997 21.2: 36677.
Larsen-Freeman, Diane, and Michael H. Long. 1991. An Introduction to
Second Language Acquisition Research. New York: Longman.
Levin, Iris, Dorit Ravid, and Sharon Rapaport. 2001. Morphology and
spelling among Hebrew-speaking children: From kindergarten to first
grade. Journal of Child Language 28.3: 74172.
Li, P. and Y. Shirai. 2000. The Acquisition of Lexical and Grammatical
Aspect . Berlin: Mouton de Gruyter.
Marcus, Gary F. 1995. The acquisition of the English past tense in children
and multilayered connectionist networks. Cognition 56.3: 2719.
. 1996. Why do children say breaked? Current Directions in
Psychological Science 5.3: 815.
Marcus, Gary F., Steven Pinker, Michael Ullman, Michelle Hollander, T.
John Rosen, and Fei Xu. 1992. Overregularization in language acquisition. Monographs of the Society for Research in Child Development
57.4: v164.
Mikes, Melanija. 1967. Acquisition des categories grammaticales dans
le langue de lenfant [Acquisition of grammatical categories in the language of the child]. Enfance 3/4: 28998.
Montrul, Silvina. 2001. The acquisition of causative/inchoative verbs in
L2 Turkish. Language Acquisition 9.1: 158.
Peters, Ann M., and Sven Stromqvist. 1996. The role of prosody
in the acquisition of grammatical morphemes. In Signal to
Syntax: Bootstrapping from Speech to Grammar in Early Acquisition,
ed. J. Morgan and K. Demuth. Mahwah, NJ: Lawrence Erlbaum.
Pinker, Steven, and M. T. Ullman. 2002. The past and future of the past
tense. Trends in Cognitive Sciences 6: 45663.
Plunkett, Kim, and Virginia A. Marchman. 1993. From rote learning to
system building: Acquiring verb morphology in children and connectionist nets. Cognition 48.1: 2169.
Ravid, Dorit, and Avraham Avidor. 1998. Acquisition of derived nominals in Hebrew: Developmental and linguistic principles. Journal of
Child Language 25.2: 22966.
Redouane, Rabia. 2004. The acquisition of MSA word formation processes: A case study of English-speaking L2 learners and native speakers. ITL, Review of Applied Linguistics, 145/146: 181217.
Rosansky, E. 1976. Methods and morphemes in second language acquisition. Language Learning 26: 40925.
Rowland, Caroline F., Julian M. Pine, Elena V. Lieven, and Anna L.
Theakston. 2003. Determinants of acquisition order in wh-questions: Re-evaluating the role of caregiver speech. Journal of Child
Language 30.3: 60935.
Santelmann, Lynn, Stephanie Berk, and Barbara Lust. 2000. Assessing
the strong continuity hypothesis in the development of English
inflection: Arguments for the grammatical mapping paradigm.
Proceedings of the XIX West Coast Conference on Formal Linguistics
19: 43952.
Skinner, B. F. 1957. Verbal Behavior. New York: AppletonCentury
Crofts.
Slobin, Dan I. 1982. Universal and particular in the acquisition of language. In Language Acquisition: The State of the Art, ed. E. Wanner
and L. Gleitman, 12872. Cambridge: Cambridge University Press.
522
523
Morphology, Neurobiology of
hypothesi not involve syntax, but could conceivably involve systematic expression of categories such as number, tense, or definiteness.) And, where formal differences involved extra segments
at the beginning or the end of an item, the seeds were sown for
what we now call affixes, arising by a process quite distinct from
grammaticalization.
A variant of Carstairs-McCarthys approach has been developed by Dieter Wunderlich (2006a, 2006b), linking grammatical
evolution with cultural and economic change. The sort of syntax
that many modern languages have, with lavish opportunities for
long-distance syntactic movement, would not have had significant evolutionary advantages (Wunderlich suggests) until after
the emergence of large speech communities whose members
did not all know one another, that is, until the Neolithic period.
Until then, he suggests, that is, as long as all humans were huntergatherers living in small groups, elaborate morphology would
have preponderated over syntax.
It will be seen that, as regards morphological evolution,
widely divergent suggestions have been made about the balance
between cultural and noncultural factors. Carstairs-McCarthys
and Chomskys approaches, though different in many ways,
agree in emphasizing noncultural reasons for the existence of
morphology as a component of grammar. For Comrie, on the
other hand, cultural change is at least as important as biological
or self-organizational factors, while Wunderlich revives the view
that fully modern syntax came late as a cultural by-product of
population expansion and the transition to agriculture, with the
added twist that an elaborate morphological component was
already in existence early. Time will tell which viewpoint prevails or which combination of viewpoints.
Andrew Carstairs-McCarthy
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Carstairs-McCarthy, Andrew. 2010. The Evolution of Morphology. Oxford:
Oxford University Press.
Chomsky, Noam. 2000. New Horizons in the Study of Language and
Mind. Cambridge: Cambridge University Press. Chapter 1 is especially
relevant.
. 2004. Language and mind: Current thoughts on ancient problems. In Variations and Universals in Biolinguistics, ed. L. Jenkins,
379405. Amsterdam: Elsevier.
Comrie, Bernard. 1992. Before complexity. In The Evolution of Human
Languages, ed. J. Hawkins and M. Gell-Mann, 193211. Reading,
MA: Addison-Wesley.
Heine, Bernd, and Tania Kuteva. 2002. World Lexicon of
Grammaticalization. Cambridge: Cambridge University Press.
Hinzen, Wolfram. 2006. Mind Design and Minimal Syntax. Oxford: Oxford
University Press. This book confronts frankly, from a Chomskyan perspective, some of the difficulties that morphology poses for the minimalist program.
Wunderlich, Dieter. 2006a. What forced syntax to emerge? In Between
40 and 60 Puzzles for Krifka, ed. Hans-Martin Grtner, Sigrid Beck,
Regine Eckardt, Renate Musan, and Barbara Stiebels. Berlin: Zentrum
fr Allgemeine Sprachwissenschaft. Available online at: http://www.
zas.gwz-berlin.de/fileadmin/material/40 60-puzzles-for-krifka/
index.html.
. 2006b. Why is there morphology? Abstract of paper presented at
Workshop on Theoretical Morphology, Leipzig, June. Available online
at: http://www.uni-leipzig.de/~jungslav/rmag/Wunderlich.pdf.
524
MORPHOLOGY, NEUROBIOLOGY OF
The nature of word formation and word storage has long been
prominent in the disparate fields of linguistics, psychology, and
neurobiology, and recently the neurobiology of morphology has
emerged as its own distinct subfield of study.
Two prevalent issues in formal linguistic and psycholinguistic
approaches to morphology are the distinctions drawn between
inflectional versus derivational morphology and between regular
versus irregular morphology. Another line of inquiry has sought
to identify the formal structure of basic morphological representations (i.e., words, stems, affixes) and to determine the extent
to which complex words are either composed by a grammatical process or stored as unanalyzed wholes. The issue of compositionality interacts with inflectional/derivational status and
with regularity; for example, various theorists have proposed
that irregular inflection is not compositional (1a), that familiar
inflected forms are not compositional (2a), or that derivational
morphology is not compositional (3a) in the way that they are
stored or processed.
Noncompositional
Compositional
(1)
a. ran = [ran]
b. ran = [run]+[past]
(2)
a. walked = [walked]
b. walked = [walk]+[past]
(3)
a. hopeless = [hopeless]
b. hopeless = [hope]+[less]
Morphology, Neurobiology of
due to a lesion in the inferior frontal gyrus of the frontal lobe,
a lesion/deficit pattern now know as Brocas aphasia. Wernicke
reported that patients with lesions in the left posterior section
of the superior temporal gyrus suffered from a severe comprehension deficit; their speech was fluid and natural sounding but
their word selection seemed divorced from meaning. Numerous
researchers have since argued that lexical storage and retrieval
functions are located in wernickes area, and that morphological grammar functions are housed in brocas area.
A quite different approach to the study of morphology, but one
that is also grounded in neurobiology, is connectionist modeling. Work in connectionist modeling has called into question
some of the most fundamental tenets of morphological theory.
D. E. Rumelhart and J. L. McClelland (1986) presented arguments from modeling against the distinction between regular and
irregular morphology. M. S. Seidenberg and L. M. Gonnerman
(2000) challenged the very existence of morphological representations, arguing instead that what linguists call morphemes are
merely the points of convergence between sound and meaning
codes, and are not a distinct type of entity of their own. These
studies have drawn fierce criticism, stimulated vigorous debate,
and played a major role in shifting the standards of mainstream
morphology to a consideration of brains and simulated brains as
viable data sources.
In each of the disciplines concerned with the neurobiology
of morphology, many central issues remain unresolved and
hotly debated. Furthermore, there had been little occasion until
recently for these methodologically distinct disciplines to communicate, despite their concern with fundamentally the same
topic. However, recent technological advances have provided
each of these disciplinary perspectives with new methods of
study, and thus new insights into perennial questions. These
advances, combined with shifting disciplinary boundaries, have
enabled the neurobiology of morphology to become a largely
coherent line of inquiry into the neural underpinnings of word
storage and processing. We may divide recent approaches into
neuropsychological, hemodynamic, and neurophysiological
methods.
Neuropsychological Methods
Patients with brain damage and resulting impairments have
long been a valuable source of evidence about the neurobiology
of morphology. Modern neuropsychological studies of morphology evaluate patients who suffer impairments that selectively
affect (or selectively spare) morphological functions as the
result of brain damage. Typically, the rationale of these studies
is to identify dissociations between patterns of impaired and
preserved capacities in order to establish whether certain morphological functions are distinct from other components of the
lexical system.
One such dissociation is a morphological deficit that disrupts
the processing of inflectional morphology in the context of a
relatively spared ability to process derivational morphology. The
Italian-speaking patient FS reported in Miceli and Caramazza
(1988) presented with one of the clearest instances of this performance pattern. In spontaneous speech, FS made frequent
errors of agreement between nouns and attributive modifiers
(e.g., the target phrase il mio studio [det.Ms my.Ms office.Ms] was
Hemodynamic Methods
In contrast to neuropsychological methods, hemodynamic
525
Morphology, Neurobiology of
inflectional suffixes, revealing a difference in activation in and
around Brocas area. The researchers interpret these results as
evidence that Brocas area subserves morphological and/or
morphosyntactic functions.
Increasingly, fMRI is the preferred method for hemodynamic
investigations of morphology. Numerous fMRI studies have
pursued a functional localization for morphology by contrasting conditions with and without affixes, but they have often
returned inconclusive results (see, e.g., Davis, Meunier, and
Marslen-Wilson 2004).
A few fMRI studies of morphology and the lexicon, however,
have taken on more refined questions, producing more robust
results. One such study is the investigation by A. Beretta and colleagues (2003) of German regular and irregular inflection. They
find significant overall differences in neural activation between
the processing of regularly inflected and irregularly inflected
nouns and verbs. They interpret this as evidence that regular and
irregular morphological functions are subserved by distinct neural systems, though their findings do not address where irregulars
are processed or where regulars are processed, if such distinct
locations were to exist.
Neurophysiological Methods
In order to examine how specific types of morphologically complex words are processed (e.g., regularly vs. irregularly inflected
words) over the time course of lexical processing, researchers
have increasingly turned to neurophysical recording techniques
whose temporal resolution is well suited to the rapid changes in
brain response to linguistic materials. These imaging methods
include electroencephalography (EEG) also known as eventrelated brain potentials (ERPs) which measures electrical currents caused by neural activity, and magnetoencephalography
(MEG), which measures the magnetic fields that result from this
neuroelectrical activity.
For the most part, neurophysical studies of morphology
exploit well-studied event-related response components under
a variety of stimulus conditions (including lexical priming,
manipulations of lexical properties such as surface or stem frequency, and contextual fit). In EEG/ERP studies, there are two
response components that have been exploited to some advantage: the N400 a negative deflection peaking around 400 ms
that increases in amplitude after a novel or unexpected lexical stimulus and the P600 a positive current shift following
syntactic anomalies (Kutas and Hilyard 1980; Osterhout and
Holcomb 1992; Osterhout and Nicol 1999). In MEG, most morphology studies have focused on the M350 response component believed to reflect some of the currents underlying the
N400 ERP which peaks approximately 350 ms after the presentation of a word stimulus and is sensitive to stimulus factors such
as lexical frequency (Embick et al. 2001).
Several MEG studies have engaged the connectionist literature on morphology, addressing the issue of whether there exists
a distinctly morphological level of representation, one that differs from the representation of meaning and spoken/written
form. In a MEG priming study, L. Stockall and A. Marantz (2006)
found that genuine morphological relatives (e.g., givegave,
teachtaught) pattern differently at the M350 response than
morphologically nonrelated word pairs that are merely similar at
526
the orthographical and semantic levels (e.g., boilbroil, screechscream). The former pairs exhibit a facilitatory priming effect on
the M350 latency, while the latter pairs do not. However, as in
many MEG studies, effects that were visible in the M350 peak
latency were obscured in the behavioral response latency, presumably by a different and opposite effect that arose later in the
time course of processing.
Many other neurophysical studies have found convergent evidence that morphological constituents are actively recognized
in the early stages of lexical processing. Repetition priming has
been found to attenuate the N400 response component to isolated words in lexical decision tasks (Rugg 1985). This effect on
ERP has also been observed with priming by morphological relatives: Regularly inflected primes elicit a weaker N400 response
for uninflected verb targets in comparison to unrelated primes,
whereas irregularly inflected primes do not produce a comparable reduction (Mnte et al. 1999). These contrasting priming
effects have been taken as evidence for morphological parsing
of regularly inflected forms. Other studies have found evidence
that the recognition of regularly inflected words is supported by
morphological decomposition in brain responses to morphologically illegal combinations like bringed (Morris and Holcomb
2005; Lck, Hahne, and Clahsen 2006; see also McKinnon, Allen,
and Osterhout 2003 for evidence of bound-stem parsing).
Evidence for decomposition is also observed in the effects
of regularity and lexical frequency on the P600 response to
inflected words that are ungrammatical for their context. In
a study that manipulated lexical frequency, morphological
regularity, and grammatical fit, high-frequency irregularly
inflected verbs in ungrammatical contexts (e.g., the boy couldnt
*ran / *walked fast enough) showed an earlier onset of the P600
response than did low-frequency irregularly inflected verbs,
in comparison to their grammatical counterparts (e.g., the boy
couldnt run / walk fast enough); but the onset of this response
was unaffected by lexical frequency for regularly inflected verbs
(Allen, Badecker, and Osterhout 2003). The pattern suggests
that for irregular verbs, surface frequency affects the speed with
which lexical recognition mechanisms can gain access to (and
exploit) inflectional content, but that this is not so for regularly
inflected forms.
Recent studies have used MEG to explore how morphology
shapes the recognition and interpretation of compounds and
morphologically derived words (Fiorentino and Poeppel 2007;
Pylkknen et al. 2004). These studies provide further support for
the view that the detection and exploitation of morphological
structure play a major part in the early and subsequent stages of
lexical recognition.
Neural methods provide a distinctive and potentially compelling source of data for the study of morphology. Still, the set of
compelling neural studies of morphology remains quite small,
relative to other research methods, and the coherence among
these studies remains low. This is due in part to the limited availability of costly neuroimaging equipment and to the limited number of scholars who have expertise in both the technical issues of
morphology and the technical methods of cognitive neuroscience. However, as the equipment proliferates and the methods
gain a stronger foothold in the field, we can expect that more rigorous methodological conventions will develop, and that studies
Morphology, Neurobiology of
in this emerging area will have an even greater impact on our
understanding of language and morphology.
Ehren Reilly and William Badecker
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Allen, M., and W. Badecker. 2000. Morphology: The internal structure of words. In What Deficits Reveal about the Human Mind/
Brain: Handbook of Cognitive Neuropsychology, ed. B. Rapp, 21132.
London: Psychology Press.
Allen, M., W. Badecker, and L. Osterhout. 2003. Morphological analysis in sentence processing: An ERP study. Language and Cognitive
Processes 18: 40530.
Baayen, R. H., T. Dijkstra, and R. Schreuder. 1997. Singulars and plurals
in Dutch: Evidence for a parallel dual-route model. Journal of Memory
and Language 37: 94117.
Badecker, W. 1997. Levels of morphological deficit: Indications from
inflectional regularity. Brain and Language 60: 36080.
Badecker, W., and A. Caramazza. 1991. Morphological composition in
the lexical output system. Cognitive Neuropsychology 8: 33567.
Beretta, A., C. Campbell, T. H. Carr, J. Huang, L. M. Schmitt,
K. Christianson, and Y. Cao. 2003. An ER-fMRI investigation of morphological inflection in German reveals that the brain makes a distinction
between regular and irregular forms. Brain and Language 85: 6792.
Davis, M., F. Meunier, and W. Marslen-Wilson. 2004. Neural responses to
morphological, syntactic, and semantic properties of single words: An
fMRI study. Brain and Language 89: 43949.
Embick, D., M. Hackl, J. Schaeffer, M. Kelepir, and A. Marant z. 2001. A
magnetoencephalographic component whose latency reflects lexical
frequency. Cognitive Brain Research 10: 3458.
Fiorentino, R., and D. Poeppel. 2007. Compound words and structure in
the lexicon. Language and Cognitive Processes 22.7: 9531000.
Gnther, T., F. Longoni, O. Sabri, L. Sturz, K. Setani, and W. Huber.
2001. PET study of basic syntax and verb morphology. NeuroImage
13: 538.
Kutas, M., and S. A. Hillyard. 1980. Reading senseless sentences: Brain
potentials reflect semantic incongruity. Science 207: 2035.
Laiacona, M., and A. Caramazza. 2004. The noun/verb dissociation in
language production: Varieties of causes. Cognitive Neuropsychology
21: 10323.
Lck, M., A. Hahne, and H. Clahsen. 2006. Brain potentials to morphologically complex words during listening. Brain Research 1077: 14452.
McKinnon, R., M. Allen, and L. Osterhout. 2003. Morphological decomposition involving non-productive morphemes: ERP evidence.
NeuroReport 14: 8836.
Miceli, G., and A. Caramazza. 1988. Dissociation of inflectional and derivational morphology. Brain and Language 35: 2465.
Miozzo, M. 2003. On the processing of regular and irregular forms of verbs
and nouns: Evidence from neuropsychology. Cognition 87: 10127.
Morris, J., and P. Holcomb. 2005. Event-related potentials to violations
of inflectional verb morphology in English. Cognitive Brain Research
25: 96381.
Mnte, T., S. Tessa, H. Clahsen, K. Schlitz, and M. Kutas. 1999.
Decomposition
of
morphologically
complex
words
in
English: Evidence from event-related brain potentials. Cognitive
Brain Research 7: 24153.
Osterhout, L., and P. J. Holcomb. 1992. Event-related brain potentials
elicited by syntactic anomaly. Journal of Memory and Language
31: 785806.
Osterhout, L., and J. Nicol. 1999. On the distinctiveness, independence,
and time course of brain responses to syntactic and semantic anomalies. Language and Cognitive Processes 14: 283317.
Pinker, S., and M. Ullman. 2002. The past-tense debate: The past and
future of the past tense. Trends in Cognitive Science 6: 45663.
Morphology, Universals of
Pylkknen, L., S. Feintuch, E. Hopkins, and A. Marantz. 2004. Neural
correlates of the effects of morphological family frequency and family
size: An MEG study. Cognition 91: B3545.
Pylkknen, L., and A. Marantz. 2003. Tracking the time course of word
recognition with MEG. Trends in Cognitive Science 7: 1879.
Pylkknen, L., A. Stringfellow, and A. Marantz. 2002. Neuromagnetic
evidence for the timing of lexical activation: An MEG component sensitive to phonotactic probability but not to neighborhood density.
Brain and Language 81: 66678.
Rugg, M. D. 1985. The effects of semantic priming and word repetition
on event-related potentials. Psychophysiology 22: 6427.
Rumelhart, D. E., and J. L. McClelland. 1986. On learning the past tenses
of English verbs. In Parallel Distributed Processing:Explorations in the
Microstructure of Cognition, ed. D. E. Rumelhart, J. L. McClelland, and
the PDP Research Group, 21671. Cambridge, MA: MIT Press.
Seidenberg, M. S., and L. M. Gonnerman. 2000. Explaining derivational
morphology as the convergence of codes. Trends in Cognitive Sciences
4: 35361.
Shapiro, K., and A. Caramazza. 2003. Grammatical processing of nouns
and verbs in the left frontal cortex? Neuropsychologia 41: 118998.
Stockall, L., and A. Marantz. 2006. A single route, full decomposition
model of morphological complexity. Mental Lexicon 1: 85123.
MORPHOLOGY, UNIVERSALS OF
The topic of universals has not been as prominent in the area of
morphology as it has been in some other areas of linguistics.
Many linguists share the impression that morphology is predominantly a domain of the language particular, rather than the
general and universal. Whereas all languages compose words to
make sentences in one way or another (syntax), it is not certain that all languages compose morphemes to make words.
The Chinese languages, for example, have almost no morphology apart from the possibility of compounding two roots to make
a word.
While it is probably too strong to say that some languages have
no morphological system at all, at least it seems clear that there
are no universal morphological categories notions that are
expressed by morphological means in all languages. For example, English requires that past tense be expressed as a suffix on
verbs (stun vs. stunned) and that plural number be expressed as
a suffix on nouns (box vs. boxes), but there are many languages
in which these notions are not expressed morphologically (e.g.,
Yoruba); they are either expressed by syntactic constructions or
not expressed at all. Moreover, morphology is a notorious repository of many historical relics, irregularities, exceptions, and idiosyncrasies. For example, the plural of box in English is boxes, but
the plural of ox is oxen; similarly, the past tense of stun is stunned
but the past tense of run is ran. Irregularities of this sort are tolerated in morphology in a way that they may not be (or not as
much) in other linguistic domains. What universals we can hope
to find in morphology, then, are statistical and implicational
universals, rather than absolute universals (see absolute and
statistical universals).
Despite the inherent noisiness of morphology, some universals of these sorts are discernable. Perhaps the best known
and best understood are universals of markedness. These universals have the form of statements saying that no language will
have an affix that expresses a marked category Y unless it also
has an affix that expresses a less marked category X within the
527
Morphology, Universals of
same semantic domain. For example, many languages (including English) have a plural form for nouns but no dual form that
means exactly two, whereas there are few or no languages that
have morphological marking for the category dual without also
marking the category plural. Similarly, verbs in a given language
do not express the distinction between inclusive first person
plural (we including you) and exclusive first person plural (we
not including you) without also expressing the more basic distinction between first person and second person (we versus
you). Along the same lines, languages do not have special affixes
for remote past and remote future without also having affixes
for simple past and simple future, and if a language makes any
aspect distinctions in its verbal morphology, it will make a distinction between imperfective aspect and perfective aspect.
A plausible reason why universals of this sort hold has to do
with the logic of features. It is assumed that more marked, semantically complex categories are built up out of simpler categories.
For example, the category dual shares a semantic feature [Group]
with the category plural but adds a feature such as [Minimal],
which it shares with singular (Harley and Ritter 2002). It stands
to reason, then, that a language will not have morphemes that
realize a more complex feature bundle like [Group, Minimal]
(dual) without also having morphemes that realize the simpler
feature bundles [Group] (plural) and [Minimal] (singular) that
this bundle properly contains. It seems likely that this vision can
be extended to the full range of nominal and verbal inflectional
categories, although many details remain to be worked out.
One of the most general universals of morphology is that the
order of morphemes in a complex word is almost always rigidly
fixed. Almost all languages allow the words of a sentence to be
rearranged to some degree for stylistic or pragmatic effect.
This is particularly true of a language such as Mohawk, in which
any word order is usually possible. In contrast, no language
allows the morphemes in a complex word to be freely rearranged
in this way. For example, the following Mohawk word consists
of 11 distinct morphemes, but any other ordering of these morphemes is ungrammatical:
(1)
Wa-sha-ko-t-yat-awi-tsher-ahetkv-t-v
FACT-he-her-self-body-put.on-NOML-be.ugly-make-forPUNC
He made the thing you put on the torso [i.e., a shirt or dress]
ugly for her.
528
a. Mikhu-naya-chi-wa-n.
eat-want-make-1sO-3sS
It makes me feel like eating.
b. Mikhu-chi-naya-wa-n.
eat-make-want-1sO-3sS
I feel like making someone eat.
The suffixes chi to make and naya to want can attach to a verb
stem in either order, but there is a systematic difference in meaning. If make attaches before want, the combination means
to want to make someone eat, whereas if want attaches
before make, the combination means to make someone want
to eat. The order in which the affixes attach in Quechua matches
the order in which the words are combined syntactically in the
Morphology, Universals of
corresponding English paraphrases, and this in turn reflects how
the meanings are composed semantically in both languages. Not
all languages that have similar affixes allow both of the orders
shown in (2), but it is generally true that the orders that are used
correspond systematically to the order of interpretation for purposes of syntax and semantics.
Compositional ordering effects of this kind are rather widespread and apply to different types of morphology. One can see
something similar in English in the domain of compounding. For
example, ethics committee proposal refers to a proposal by or for
the ethics committee, whereas ethics proposal committee refers
to a committee in charge of formulating an ethics proposal. As in
Quechua, the different morpheme orders reflect different orders
of semantic composition. The only difference between the two
cases is that chi and naya are bound affixes that must attach to
verb roots, whereas proposal and committee are roots that can be
used as words in their own right in English. The observation that
morpheme orders must directly reflect the order of syntactic/
semantic composition is sometimes called the mirror principle
(Baker 1985); Bybee (1985) refers to a similar idea as the principle
of relevance.
Much the same constraint seems to hold of inflectional morphology as well, except that in this domain, there are very few
cases in which the order of morphemes can be reversed to give a
semantic effect. For example, Joseph Greenberg (1963) showed
that when both number marking and case marking are attached
to a noun root, the number marking almost always attaches
before the case marker does:
(3)
aku-wye-a-y-mi.
(Mapudungun)
arrive-PERF(aspect)-FUT(tense)-IND(mood)-2sS(agreement)
You will have arrived.
Ti-ne:ch-in-tlacua-l-ti:-li-a in
no-pil-hua:ntoto:n.
2sS-1sO-PL-eat-caus-appl the
my-chidren
You made my children eat for me. (*Ti-ne:ch-in-tlacua-li-tia)
529
Morphology, Universals of
Motif
noun and a verb combine to form a single verb. This holds true
regardless of the specific mode of combination, whether it is
the result of syntactic noun incorporation, productive morphological compounding, or idiosyncratic lexical combination. The
order of morphemes in the Mohawk example in (7) is thus typical
in this respect:
(7)
Wa-ke-nakt-a-hninu-
Fact-I-bed-0-buy-PUNC
I bought a bed.
530
MOTIF
Motif is a unit of measurement and content analysis. It is applied
to expressive culture, especially literatures and certain branches
of the arts, such as painting, sculpture, and music. The systematic application of the term was established in the early 1900s
through the work of the Finnish School and its attempts to use
an international folktales history (time) and dispersal in social
and physical space (geography) to reconstruct the tales original form (urform, archetype), place of birth, and other related
matters of diffusion. This approach is also known as the historic-geographic method for its reliance on objective verifiable
criteria, rather than speculative hypotheses. For the Finnish
School, two key concepts became indispensable research instruments: tale-type, a problematic term signifying a full folktale
known cross-culturally, and motif, a smaller unit designating a
detail contributing to the formation of the plot. Motif complex/
cluster/sequence and episode are other related measurement
units. Though arising from the works of the Finnish School,
perceived by many as useful only in comparative studies, and
shackled by problems of name interpretation and linkage to the
currently unfashionable quest for origins, the usefulness of these
terms as tools of data identification and objective analysis transcends these limitations (El-Shamy 1997, 235).
The most salient attributes of a motif are its endurance (continuity in time) and recurrence within a community (continuity
in space). Continuity in time and space are basic requirements
for traditionality. In 1925, Arthur Christensen argued that a motif
persists in tradition according to a psychological law which is
not easily explicable. This empirical characteristic was left unexplained (Bdker 1965, 2013). From the perspective of the bearer
of lore and the author or composer of elite art, certain traditional
themes possess cognitive salience (impressiveness) that make
them stand out and grab a persons attention. This salience (logical or affective) may be due to frequency of occurrence (repetition), meaningfulness, structure, uniqueness, ego involvement,
and so on, properties that make such themes easily perceived,
learned, retained, and recalled (El-Shamy 1997).
The concept of motif is a close parallel to that of theme. The
two terms are often used interchangeably. However, theme has
dominated in the study of elite literature, whereas motif has
been more common in the study of folklore. Research in literary
themes has commonly been pursued because of its interpretive potentialities and its intrinsic congruency with the history
of ideas (Jost 1988, xv). Additionally, literary authorities have
considered it an efficient counteragent against primarily aesthetic movements, such as progressive Universalpoesie. It is
also argued that the motif is intellectual by nature. It expresses
a process of reasoning about mens conduct of life and, as a consequence, does not concern itself with the analysis of individual
characters or extraordinary happenings (Jost 1988, xvii).
The viewpoint outlined here addressing elite literature stands
at some variance with the folkloristic usage of the term motif and
the perceived scope of its applicability. Introducing his MotifIndex, Stith Thompson explained that his system is built around
the interest of students of the traditional narrative and would
address a certain type of character, action, as well as attendant circumstances of the action (1955, I.11). (See, respectively,
Motif
Movement
for examples: W10, Kindness; T72.2.1, Prince marries scornful girl and punishes her; R216, Escape from ship while captors
quarrel).
Influenced by anthropology, Thompsons system may be
compared to the analytical-classificatory devices of culture element, culture complex, and culture institution with culture element being the smallest identifiable component of culture. In
congruence with his companions in the Historic-Geographical
School, Thompson saw folk literature (especially narratives) as
analyzable in terms of motifs, episodes (or motif complexes/
sequences), and full narrative plots constituting tale-types. A
motif, though considerably more intricate, is comparable to
culture element; culture complex is comparable to episode; and
culture institution is comparable to tale-type. For Thompson,
motifs are those details out of which full-fledged narratives are
composed (1955, I.10).
Explaining the rationale for his motif system and its main
objective, Thompson declares that it emulates what the scientists have done with the worldwide phenomena of biology
(1955, I.1011). In this respect, the underlying principle for
motif identification and indexing is comparable to that devised
by anthropologists at Yale for categorizing culture materials
in terms of 78 macrounits (1088) and 629 subdivisions thereof
used to establish The Human Relations Area Files (HRAF);
these files, begun almost contemporaneously with the first publication in the 1930s of Thompsons Motif-Index, may be viewed
as an unprinted index. Twenty-three divisions make up the spectrum of sociocultural materials covered in Thompsons MotifIndex, each treated in an independent chapter (e.g., B. ANIMALS;
C. TABU; F. MARVELS; X. HUMOR). These cardinal themes are
divided into 1,730 subdivisions (El-Shamy 1995).
Because Thompsons Motif-Index seeks global coverage,
numerous geographic regions and national entities did not
receive adequate attention. Consequently, significant fields of
human experience are missing or sketchily presented. Major
expansions are offered in ensuing works, for example (note: the
sign indicates addition to Thompsons system): F70, Ascent
to other planets (worlds) by space ship (flying saucer); J70,
Teaching (training) by cruel example; P610, Homosociality
[]; P770, Markets: buying, selling, trading; X580, Humor
concerning misers and miserliness (El-Shamy 1995; cf. Birkhan,
Lichtblau, and Tuczay 2005).
An offshoot of the motif system is the concept of motifeme,
a hybrid of folklore and linguistics modeled after morphological
analyses of folktales. In this system, an abstract unit of action or
state is labeled motifeme, and its manifestations are motifs. The
variety in which a given motifeme is manifested is termed allomotif (Dundes 1964).
Hasan El-Shamy
WORKS CITED AND SUGGESTED FURTHER READING
Birkhan, Helmut, Karin Lichtblau, and Christa Tuczay. 2005. Motif-Index
of German Secular Narratives from the Beginning to 1400. 6 vols. Berlin
and New York: W. de Gruyter.
Bdker, Laurits. 1965. Folk Literature (Germanic). Vol. 2 of International
Dictionary of Regional European Ethnology and Folklore.
Copenhagen: Rosenkilde and Bagger.
MOVEMENT
Movement is an operation posited in theoretical syntax, in which
words, phrases, and perhaps also morphemes are relocated from
one part of a sentence to another. Movement is generally invoked
in cases in which a phrase has a combination of properties associated with distinct positions in the sentence. Consider (1):
(1)
531
Movement
West Ulster dialect of English in which the questions in (4a) and
(4b) are synonymous:
(4)
a. He is leaving.
b. Is he ___ leaving?
The examples in (9) differ with respect to how far all travels with
what: not at all (9a), only to the intermediate site (9b), or to the
beginning of the clause (9c).
Another controversial type of movement is involved in the
Chinese and Japanese wh-questions in (10a) and (10b), both of
which have the same meaning as the English (10c):
(10)
a. John mai-le
sheme?
John buy Perf what
b. John-wa nani-o
kaimasita ka?
John TOP what ACC bought Q
c. What did John buy?
a.
b.
(6)
(7)
Norvin Richards
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
(8)
One argument for this type of movement comes from West Ulster
English. We saw that this dialect allows all to be stranded in positions formerly occupied by a wh-moved phrase. The all-stranding facts for this dialect in (9) show that wh-movement has an
intermediate landing site:
532
Figure 1.
533
Figure 2.
*
*
**
*
** ** *
Ticonderoga
level
level
level
level
3
2
1
0
Figure 3.
is therefore, in Steeles words, a subjective mental sensation
deriving from a sense of pulsation giv[ing] the mind an idea
of emphasis independent of any actual increment of sound or
even of any sound at all (1775, 117). In recognizing the abstract
character of meter, he anticipated twentieth-century cognitivist approaches that view linguistic stress, along with most other
salient characteristics of language, as mental constructs, phonological rather than phonetic, psychologically real but
only obliquely related to the acoustical or physiological surface
form.
534
L (2)
L (1)
L (0)
a) * Thirteen men
L (2)
L (1)
L (0)
b) Thirteen men
Figure 4.
linguistic computations as independent and self-contained,
Libermans objective was to establish an equivalence between
the underlying representation of song and speech, a connection
that was understood by Liberman to be in some ways a very
deep one (1975, 81). This hypothesis, whatever its ultimate heuristic or conceptual value, has not been influential within the
field of linguistics or in music theory.
Two partial exceptions should, however, be mentioned.
While word-level stress provides evidence for the disassociation
of metrical structure in language and music, higher levels of linguistic structure provide some evidence for a musicalist interpretation of linguistic performance. In particular, phrasal stress,
unlike word stress, is not phonologically austere and requires
for its computation, in addition to morphological, syntactic and pragmatic factors, the quasi-musical considerations
of what is referred to as phrasal euphony. Most conspicuous
among these is the stress clash resulting from stressed syllables
from two words appearing adjacent to each other within the
same phrase. The unacceptable form in a) triggers the application of the rhythm rule, which achieves euphony by retracting
leftward the first of the two syllables involved in the stress clash
to produce the acceptable form b). While not a validation of the
musicalist view, the terminology that is adopted by linguists, as
well as the mechanisms by which this particular phenomenon is
explained, is suggestive of a shared basis underlying the computation of phrasal stress and the assignment of musical meter.
A second point of contact, as noted by Jamie Kassler (2005),
is the musicalist interpretation of metrical structure incorporated into certain approaches to formal prosody and in its
transformational variant, generative metrics. The stated goal
of these approaches is to define the abstract structure of lines
of texts composed in a poetic meter. But even here, the adoption of a musicalist view as representing a significant aspect of
the relevant empirical domain remains controversial. Although
poets routinely invoke the music of poetry and the rhythms
of verse and appeal to explicitly musical terms such as phrasing, staccato, harmony, and so on, it remains an open question
whether poetic rhythm has any relationship to musical rhythm
as musicians understand as the term. The reaction to Derek
Attridges Rhythms of English Poetry (1982) sheds some light on
these questions. Attridge posits a scansion that assigns syllables
to alternating beat and offbeat positions (indicated by b and o,
respectively), some of which can remain vacant, most notably at
the end of lines (see verse line). The general approach and the
representation of empty positions, in particular, is criticized by
Marina Tarlinskaya for failing to recognize that though music
535
N
NARRATIVE, GRAMMAR AND
Analysts of stories have developed two broad strategies for
studying the relations between narrative and grammar. The first
emphasizes grammar of narrative, and the second grammar in
narrative.
Drawing on definitions of grammar as a model of the categories and processes underlying intuitive knowledge of (or competence in) a language, theorists working in fields such as
cognitive psychology, artificial intelligence research, and narratology have proposed story grammars, that is, grammars of narrative. (A key question is whether this work involves
536
Deixis in Narrative
Deictic expressions such as I, here, and now that is, expressions
with interpretations that depend on who utters them to whom
in what communicative circumstances (see pragmatics)
are another part of the grammatical system exploited by narratives in distinctive ways. At the text level, deictic expressions
serve to locate narrators and characters in time and space vis-vis objects, events, and situations in the storyworld, whose
space-time coordinates often do not match those of the current
moment of narration.
Consider, for example, the first two sentences of Ernest
Hemingways 1927 story Hills Like White Elephants: The hills
across the valley of the Ebro were long and white. On this side
there was no shade and no trees and the station was between
two lines of rails in the sun. The preposition across in the first
sentence and the demonstrative pronoun this in the second sentence must both be interpreted in light of the assumed position
of the vantage point from which events are being narrated a
vantage point that here overlaps with that of the characters.
David A. Zubin and Lynne E. Hewitt (1995) propose the notion
of deictic shift to account for such displaced or transposed
modes of deictic reference, which must be anchored in the storyworld evoked by the text, rather than in the world(s) that the
text producer or the text interpreter occupies when producing or
537
Narrative, Neurobiology of
538
NARRATIVE, NEUROBIOLOGY OF
Despite cognitive neurosciences long-standing interest in the
processing of individual words and sentences, most neurobiological enquiry into the comprehension and production of
more holistic and contextual forms of narrative discourse has
come only recently. Chief among the obstacles to such studies
has been a tension between scientific control and ecological
validity: Experimental control demands repeatable and rather
predictable stimuli and constrained responses, whereas freeranging narrative discourse eschews these properties. In addition, an analytical focus on mapping individual brain regions
to individual phonological, morphological, and local
contextual properties has upstaged examination of the ways in
which the construction of meaning may be subserved by network
interactions among many brain regions. With newly broadened
experimental paradigms, though, and with novel analytical methods addressing functional interactions, neuroscience promises
to add a biological dimension to cognitive psychological theories
of narrative discourse.
Central to cognitive neuropsychological explorations of narrative has been the realization of shared structure between the
explicit narratives of literature and the implicit narratives of
everyday mental representations: Information processing
in both these spheres depends on the representational power
conferred by narrative abstraction. In the everyday world no
less than the literary, disconnected percepts gain meaning and
separability from time, place, and action insofar as they become
transformed into representative mental texts, stories whose distinct scenes contain recognizable characters that act in coherent
plots and evince meaningful themes. A neurobiological perspective only reinforces these observations from human thought and
behavior: Neurophysiological study of brain dynamics reveals
that human cognitive architecture may be engineered to
represent its processing in series of discrete frames somewhat
analogous to those of cinema (Freeman 2006), and cognitive
neuroimaging has begun to reinforce a neuropsychological
view of thought as an activity of constructing blended spaces
between narrative schemata. The central role of narrative
scripting in cognition was explored in connection with early
efforts at constructing symbolic systems capable of artificial
intelligence, and was latent in literary criticism even before the
cognitive revolution of the late twentieth century. What is
new about this connection between thought and narrative is its
Narrative, Neurobiology of
explicit elaboration in light of cognitive neuroscience, connecting literature and philosophy with psychobiological information
and constraint.
Neuropsychologically, comprehension has been studied more
completely than production, and this entry focuses on comprehension. Narrative organization is implemented by interacting
and not entirely separable processes of perceptual organization,
attention, and memory. Perceptual organization is the process
that binds separate physical stimuli into coherent higher-order
objects within a scene, replacing, for example, a horizontal plane
and four perpendicular posts with the single entity of a table.
Attention, a group of many subprocesses, focuses processing
on those parts of a scene deemed relevant to the current script
or story schema. Memory encoding, maintenance, and
retrieval, by holding in mind the higher-order representations
of what one has seen before and what one expects to see next,
inform attention and perceptual organization with the context of
this story (Gerrig and McKoon 2001). Although much is known
about the neural substrates of these processes individually, their
significance in narrative processing lies in their interactions. An
understanding of these interactions subserving the comprehension or production of narrative discourse might best begin at the
beginning, with a discussion of contextual integration at the level
of individual words.
The principal physiological index of a words integration into
its context is the N400 (Kutas and Federmeier 2000), a negative
voltage produced by the brain in response to a word, and maximal about 400 milliseconds after the word is presented. The N400
reflects a truly textual process largely independent of the particular sensory mode of representation: Although there are some
more subtle effects of sensory modality, the N400 arises no matter whether the word is read from a page or heard from a speaker.
The N400 is thought to reflect a process or processes of contextual integration since its amplitude varies parametrically with a
words predictability; the canonical method of evoking a large
N400 is to present a word whose semantics violate contextual
expectation, for instance, At breakfast we ate toast with sand.
The initial words in this sentence, and the syntax into which
theyre arranged, prime activations for appropriate breakfast
foods. The ongoing construction of meaning is then challenged
by the non sequitur sand, eliciting a large N400 response. The
N400 arises no matter whether the conflicting context is established by a surrounding sentence, as in this brief example, or
simply by a single adjacent word, or by an entire discourse. For
instance, the following discursive context reduces the N400 in
the example above: We camped on the beach. A stiff wind blew
off the dunes into all our supplies. At breakfast we ate toast with
sand. Thus, words that are semantically related to their surroundings or are otherwise contextually expected evoke small
N400s, whereas words that cannot be predicted from context
and which, by inference, supply new information with which the
context must be updated evoke large N400s.
Anatomically, electromagnetic source localization and functional neuroimaging place the generators of the N400 primarily
in the superior temporal lobe and temporo-parietal junction (Van Petten and Luka 2006). These sources are mainly in the
left hemisphere but have some contribution from the right.
Across the period of the N400 response, they proceed posteriorly
539
Narrative, Neurobiology of
Figure 1.
(caudate nucleus and dorsomedial thalamus) that communicate
with prefrontal cortex (Xu et al. 2005). (See Figure 1.) The involvement of these regions likely reflects discourses demands to imagine scenes perceptually and especially visually, to place scenes and
events in spatial relation and temporal sequence, to take up and to
shift between spatial perspectives and personal points of view,
and to emote and empathize. In particular, the precuneus seems
associated with visual spatial perception and attention, medial
prefrontal cortex and its linked subcortical nuclei with perception
of events in sequence and context, the temporo-parietal junction
with theory of mind, and medial temporal lobe structures with
emotion and memory. Narrative representation can be viewed as
an emergent property of interactions among these and other structures subserving a broad array of cognitive processes.
One of the most discussed capacities involved in narrative
comprehension and production is theory of mind. Theory of
mind was first characterized as the general ability to understand
or to model the thoughts and beliefs of other people. However,
more recent neuroscientific results show that a great deal of
such social attribution can be accomplished using principally
perceptual mechanisms. These social perceptual capacities are
computationally and developmentally prior to theory of mind
and include specialized representations for qualities that typify
agency, such as autonomous movement and direction of gaze.
Such perceptual qualities underlie the attribution of volitional
mental states (she/he/it wants or she/he/it wants to) and
perceptual mental states (she/he/it sees) attributions that
form a ubiquitous shorthand in narrative descriptions even in
the case of plainly mechanical and nonsubjective entities (for
example, the computer doesnt see the network, or the printer
wants attention). Theory of mind in its most specific sense is
essential only for the attribution of belief (she/he/it thinks/
believes/knows), is associated with activation of brain regions
distinct from those subserving more elementary forms of social
attribution, and is distinguished from these more elementary
forms by its appearance at a later stage in child development,
at or near four years of age. This developmental connection
is significant: Theory of mind seems to arise from simpler processes that deal in sensory and especially visual data. Although
early studies associated theory of mind most strongly with the
medial frontal cortex, later work has suggested that this medial
540
frontal activation reflects a more general association with complex social narratives perhaps related to contextual selection of
details that build coherence within a narrative and that engage a
works suggestion structure to make it relevant to personal
experience and self-representation (Ferstl and von Cramon
2002). Such contextual selection may instantiate the prefrontal
cortexs more general involvement in the inhibition of responses
deemed inappropriate to the current behavioral and cognitive
context. In contrast, an experimentally based argument has been
made for a more selective association of temporo-parietal junction with the late-developing, belief-oriented variety of theory of
mind (Saxe and Powell 2006).
It remains an open question in evolutionary psychology as to what extent theory of mind may be a modular cognitive capacity independent of other aspects of cognition, versus to
what extent it may depend on developmental specialization arising in the interaction of earlier-maturing, more general capacities for social perception and executive function; recent views
on genes and language suggest that human cognitive adaptation for narrative discourse may combine these modular and
generalist perspectives, perhaps by putting to novel uses a large
collection of small modules specialized for cognitive processes
that are applicable to language but not necessarily restricted to
language (Bookheimer 2002). Proponents of the modular view
have often pointed to autism and language, or more specifically to autisms impairments in social communication, as
evidence for a modular dysfunction of theory of mind. However,
many people with autism pass theory of mind tests; the dysfunction of medial prefrontal cortex found in imaging studies of
autism seems consistent with a more general deficit in automatically engaging contextual evaluation and self-representation,
and in any case, such abnormalities within specific regions in the
autistic brain may be reflections of a more fundamental disruption in the information transfer and integration between regions.
Behaviorally observed deficits in theory of mind may thus stem
from a more general perturbation of narrative processing, and
may appear especially prominent only because theory of mind is
so frequently applied in everyday social interaction.
In addition to those regions uniquely activated by discourse
processing, most other brain regions involved in language
or higher-order cognition become more heavily recruited by
Narrative, Neurobiology of
discourse than by individual sentences or words. In particular, the middle frontal gyrus, on the dorsolateral surface of the
prefrontal cortex, seems involved in placing sentential or other
discursive elements in temporal, causal, or logical sequence
(Gernsbacher and Kaschak 2003; Mar 2004). This prefrontal sequencing and coordination of ideas seems analogous to
the more concrete executive sequencing and coordination of
body movements implemented in more posterior regions of
the frontal cortex. In a computational sense, therefore, narrative comprehension can be viewed as an elaboration of motor
control or, conversely, motor control itself can be said to
have narrative character, in the sense that it fundamentally
involves sequencing and relating actions in contexts. This relation between language and action gives crucial context to evolutionary biologys efforts to explain the phylogenesis of such
an abstract cognitive capacity, in that the roots of this capacity
may rest in the very concrete domain of motor control. Along
the same lines, the anterior inferior frontal gyrus seems distinguished from neighboring cortex by an involvement in selecting
semantic relations in communication with semantic retrieval
processes in the temporal lobe, whereas posterior regions of the
inferior frontal gyrus are more immediately bound up with the
more concrete, sequencing-related details of syntax and phonology (Bookheimer 2002; Jung-Beeman 2005).
In general, the neural implementation of narrative comprehension seems to take advantage of individual capacities for
movement, sensory perception, and emotion, activating these
systems in an internal simulation of the events evoked by the
narrative. The neural implementation of narrative understanding
thus depends crucially on embodiment. At the most concrete
level, that of simulating movements, this process of comprehension engages the mirror neuron system (see mirror systems,
imitation, and language) in the ventrolateral frontal lobe
and the supplementary motor area in the dorsomedial frontal
lobe (Wilson, Molnar-Szakacs, and Iacoboni 2008).
Cognitively, narrative understanding seems to emerge from
the interaction of many specialized subsystems. Neurally, therefore, a full description of narrative processing must encompass not
only the individual brain regions engaged, but also the dynamics
with which these regions connect and interact over a wide variety of narrative processes and subprocesses. These analyses of
functional connectivity are just beginning (Karunanayaka et al.
2007) via techniques including structural equation modeling,
dynamic causal modeling, and model-free multivariate methods such as partial least squares and independent components
analysis and initial work has demonstrated information flow
from regions of the temporal lobe adjoining the auditory cortex
(often described as wernickes area, though the anatomical
definition of this term has never been precise) to higher-order
processing in the inferior frontal lobe (brocas area) and near
the temporo-parietal junction. From these areas, the network of
interactions becomes more complex, with the inferior frontal lobe
projecting to the dorsolateral prefrontal cortex, temporo-parietal
junction, and medial frontal cortex, implementing wide-ranging effects of core language processing on complex semantics,
sequencing and coordination of ideas, and social attribution.
The data now in hand suggest functional specialization and
subdivision of brain regions beyond the anatomical resolution
that, has thus far been realized in multisubject functional neuroimaging. Contributing to this limited resolution is a potentially
high degree of variation in detailed functional anatomy across
individual subjects. A further challenge is an intersubject variability in information transfer between brain areas that actually
reflects individual differences in cognitive style: For instance, different individuals may make more or less use of working memory in comprehending and producing narrative, and there are
indications that these differences may be reflected in functional
connectivity with cooperating structures in the medial temporal
lobe related to long-term memory.
This connectivity frame holds out the potential for a rapprochement between connectionism and explicitly representational, modularist views of language and narrative processing,
since modular functions may reside not so much in any particular anatomical locus as in the incoming and outgoing links
among these loci: In this sense, the more closely the localization
problem is examined, the more ill-posed it may become. The
aforementioned characterizations in terms of regional functional
mapping may, therefore, be understood as a first approach to a
description in terms of regional functional interaction. It is in this
sense that the new neuroscientific study of discourse processing
is approaching an understanding of narrative connectivity in
terms of neural connectivity.
This entry is current as of 15 August 2007.
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bookheimer, Susan. 2002. Functional MRI of language: New approaches
to understanding the cortical organization of semantic processing.
Annual Review of Neuroscience 25: 15188.
Coulson, Seanna, and Cyma Van Petten. 2007. A special role for the right
hemisphere in metaphor comprehension? ERP evidence from hemifield presentation. Brain Research 1146: 12845.
Ferstl, Evelyn C., and D. Yves von Cramon. 2002. What does the frontomedian cortex contribute to language processing: Coherence or theory
of mind? NeuroImage 17.3: 15991612.
Freeman, Walter J. 2006. A cinematographic hypothesis of cortical
dynamics in perception. International Journal of Psychophysiology
60.2: 14961.
Gernsbacher, Morton Ann, and Michael P. Kaschak. 2003. Neuroimaging
studies of language production and comprehension. Annual Review
of Psychology 54: 91114.
Gerrig, Richard J., and Gail McKoon. 2001. Memory processes and
experiential continuity. Psychological Science 12.1: 815.
Jung-Beeman, Mark. 2005. Bilateral brain processes for comprehending
natural language. Trends in Cognitive Sciences 9: 51218.
Karunanayaka, Prasanna R., Scott K. Holland, Vincent J. Schmithorst,
Ana Solodkin, E. Elinor Chen, Jerzy P. Szaflarski, and Elena Plante.
2007. Age-related connectivity changes in fMRI data from children
listening to stories. NeuroImage 34.1: 34960.
Kutas, Marta, and Kara D. Federmeier. 2000. Electrophysiology reveals
semantic memory use in language comprehension. Trends in
Cognitive Sciences 4: 46370.
Mar, Raymond A. 2004. The neuropsychology of narrative: Story comprehension, story production, and their interrelation. Neuropsychologia
42: 141432.
Saxe, Rebecca, and Lindsey J. Powell. 2006. Its the thought that
counts: Specific brain regions for one component of theory of mind.
Psychological Science 17: 6929.
541
542
Narrative Universals
production and reception of narrative. The mirror neuron system in humans has been consistently reported as being related
to imitation, action observation, intention understanding, and
understanding of the emotional states of others, to mention a
few of the human faculties that are essential for the right perception of fictional narratives and, of course, for the survival and
evolution of the human race. And there seems to be even more
to it. In 2004, Vittorio Gallese, Christian Keysers, and Giacomo
Rizzolatti published a paper in which they explore the possibility that the mirror neuron system, by providing us with an experiential (precognitive) insight into other minds, could provide
the first unifying perspective of the neural basis of social cognition (2004, 401). The mirror neuron system is apparently the
basic mechanism that allows us to grasp the intentions of others
and to experience similar emotions (empathy). As these authors
say in their article, A crucial element of social cognition is the
brains capacity to directly link the first- and third-person experiences of these phenomena (i.e., link < I do and I feel > with < he
does and he feels >) (ibid., 396).
The question arises: Would all art then be nothing but a specific and specialized activity aimed at firing the mirror neurons
in a certain direction? Would the will to style displayed by
authors, composers, film directors, and so on be nothing but the
deliberate use of syntax, semantics, pragmatics, and phonemics
to trigger the mirror neuron system in specific and predetermined ways in order to elicit specific and predetermined insights
and emotions in the reader and the audience?
Frederick Aldama
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Abbott, H. Porter. 2002. The Cambridge Introduction to Narrative.
Cambridge: Cambridge University Press.
Aristotle. 1984. The Rhetoric and Poetics of Aristotle. Trans. Rhys Roberts
and Ingram Bywater. New York: McGraw-Hill.
Carroll, Joseph. 2004. Literary Darwinism: Evolution, Human Nature,
and Literature. New York: Routledge.
Gallese, Vittorio, Christian Keysers, and Giacomo Rizzolatti. 2004. A unifying view of the basis of social cognition. Trends in Cognitive Science
8.9: 396403.
Hogan, Patrick Colm. 2003. The Mind and Its Stories: Narrative Universals
and Human Emotion. Cambridge: Cambridge University Press.
Louwerse, Max, and Don Kuiken, eds. 2004. The Effects of Personal
Involvement in Narrative Discourse Processes. Philadelphia: Lawrence
Erlbaum.
Louwerse, Max, and Willie van Peer, eds. 2002. Thematics: Interdisciplinary
Studies. Amsterdam: John Benjamins.
Miall, David. 2006. Literary Reading: Empirical and Theoretical Studies.
New York: Peter Lang.
Miall, David, and Donald Kuiken. 1994. Foregrounding, defamiliarization, and affect response to literary stories. Poetics 22: 389407.
Propp, Vladimir. 1968. Morphology of the Folk Tale. Austin: University of
Texas Press.
Rizzolati, Giacomo, Vittorio Gallese, Leonardo Fogassi, and Luciano Fadiga.
1996. Action recognition in the premotor cortex. Brain 119: 593609.
Storey, Robert. 1996. Mimesis and the Human Animal: On the Biogenetic
Foundations of Literary Representation. Evanston, IL: Northwestern
University Press.
Tsur, Reuven. 1997. Poetic rhythm: Performance patterns and their
acoustic correlates. Versification 1: 1997. This electronic journal of
NARRATIVE UNIVERSALS
In one Arawak story, a tiger sees a great hunter in the forest,
changes herself into a woman, and marries him. The two have a
happy married life until the good wife suggests that they visit the
hunters family. She warns him that he must not reveal her origin to anyone. The hunter, however, tells his mother the secret.
Feeling ashamed in front of the community, the woman changes
back into a tiger and returns to the forest. The poor husband
would often go into the bush and call his wife, but there never,
never came a reply (Roth 1915, 2034).
At first glance, the story might seem strange and clearly not
universal. A man marries a tiger and must keep it a secret from
his mother? Claude Lvi-Strauss analyzes the story in relation
to a complex of South American myths (see homologies and
transformation sets) that recount the decline from a golden
age. He also connects the tale with cannibalism (1973, 259). LviStrausss analysis does implicitly link the tale with other traditions, simply by referring to a decline from a golden age, for this
topic is found in different cultures. In the Judeo-Christian tradition, the story of the Fall is a case in point. However, in some ways,
this link only makes the story more alien. Pairing Adam and Eve
with a hunter and a tiger seems to highlight cultural difference
and pairing the eating of an apple with cannibalism seems to put
the Arawak at quite a distance from anything Western.
On the other hand, there is something deeply familiar about
the story. It tells of a couple joined by attachment (witness the
husbands pathetic search at the end of the story), suffering conflict due to the husbands divided loyalties to his mother and his
wife; it treats shared secrets, feelings of betrayal and shame, and a
concern about social origins. Moreover, there is nothing in the tale
itself that suggests cannibalism, despite Lvi-Strausss analysis.
In short, it is a story that seems both strange and familiar, both
culturally particular and imbued with cross-cultural concerns.
543
Narrative Universals
propensities are thought to have resulted from adaptations that
are specifically social. While writers in cultural studies tend to see
social practices as quite variable, writers adopting this approach
tend to see societies as manifesting a wide range of universals. Language study has been one area in which universalism
has been prominent, though there has been some disagreement
as to the precise evolutionary origins of language (see
biolinguistics).
narratology has incorporated both tendencies. A common view in cultural studies is that narrative is socially constructed and can vary widely from culture to culture. In contrast,
some researchers drawing on models from linguistics and psychology have argued that there are remarkably consistent narrative patterns across cultures. There are, however, differences
between the study of narrative and the sorts of study that have
occupied linguists for example, the study of syntax. While
writers in cognitive linguistics have viewed syntactic principles as resulting from general cognitive structures and processes, a common view within the field is that there are some
aspects of cognitive architecture that are specially devoted to
syntactic processing (see autonomy of syntax). The case for
an autonomy of narrative is much weaker. It seems much more
likely that narrative results from the interaction of various cognitive structures and processes that are not specially devoted to
narrative. As such, narrative is a less likely candidate for simple
evolutionary analysis. Put differently, if narrative results from
our cognitive abilities to draw causal inferences, to attribute
intentions, to imagine counterfactual or hypothetical situations,
to adopt varying physical points of view, to simulate experiences,
and so on, then it is less likely that there is any single adaptive
function for cross-cultural narrative patterns (comparable to the
commonly posited communicative function for language).
The point is consequential for a number of reasons, relating
to narrative and to other areas of study in the language sciences.
Specifically, if narrative patterns are unlikely to be genetically
coded in any detail, one of two conclusions may be drawn: One
may simply see this as further evidence for the culturalist position, further reason to believe that narrative may vary across
cultures with few limitations. However, the existence or nonexistence of universals is an empirical issue. It cannot be decided
a priori. If one believes that the evidence supports conclusions
of universality, then one is likely to draw a different conclusion
from the indirectness of the relation between narrative and
adaptation. Certainly, some universal patterns will derive more
or less directly from aspects of cognitive architecture (e.g., causal
attribution) that are defined by genetic programs resulting from
selective pressure. Others will derive from commonalities in the
physical environment. But that is not all.
Sticking close to biology, we may note that some cross-cultural patterns are likely to derive from the fact that adaptations
are mechanisms, not functions which means that there are
cases where the mechanism fails. More exactly, the genetic predispositions that serve us so well in daily life do so because they
set out relatively simple procedures that approximate advantageous functions. For example, although evolutionary psychologists often refer to our ability to read minds, we do not
directly know other peoples intentions. Rather, we engage in
complex processes of simulating and inferring those intentions.
544
Narrative Universals
opposition may point either toward in-group hierarchy or toward
in-group/out-group antagonism. Alexandra Aikhenvald reports
that at least among some Arawaks, members of one low-prestige
subgroup are referred to as people of jaguar (2006, 12). Walter
Roth cites an Arawak proverb that identifies tigers with enemies
(1915, 367). Moreover, the story relies on, indeed elaborates on,
the failure of mind reading, which is precisely what allows the
secret and the issue of spousal loyalty to arise in the first place.
So there is certainly commonality here, commonality that makes
cognitive sense. But is that all there is to it? After all, we knew that
there was some commonality already. Is there any greater universality to this story or to narrative patterns more broadly?
Attachment combined with sexual desire (as in pair bonding
or marriage) points toward a set of stories that recur across cultures. Perhaps we will get a better understanding of the issues if
we consider some other stories of this sort, particularly paradigmatic works from other traditions. (In a brief entry, we cannot
consider many stories, or other evidence for narrative universals.
For a range of cases, and for references to other accounts of narrative universality, see Hogan 2004.)
Four Romances
To begin, lets consider what is almost certainly the paradigm of
romantic narrative in the English-speaking world Romeo and
Juliet. Romeo and Juliet fall in love. However, they are prevented
from uniting by the group antagonism of their parents. With the
help of a friar, they are briefly united, but then Romeo is exiled
and Juliet is confined to the home. Juliet is to be married to a
rival by her father. She fakes her death to escape this fate. Romeo
returns, kills his rival, then commits suicide just at the moment
when he might have been united with Juliet. Juliet, too, kills herself, but after their deaths, the families are reunited.
Now consider the Romance of the Western Chamber, Chinas
most popular love comedy, both on stage and in print, beginning in the twelfth century (Idema 2001, 800). Chang and YingYing fall in love. Chang goes off to take the imperial exams.
Meanwhile, a rival comes to marry Ying-Ying with her mothers
approval. Chang succeeds in the examination and returns to
elope with Ying-Ying. He is successful due to the help of a monk.
The rival commits suicide (see Idema 2001, 798800).
The Recognition of akuntal, the most revered work of
Sanskrit drama, begins when Duyanta and akuntal fall in
love. Duyanta worries that they cannot marry due to caste
(thus, an internal group hierarchy). akuntal worries that they
cannot marry due to her fathers disapproval. Both turn out to
be mistaken. They are united. However, akuntal violates her
obligations to a holy man, who curses her with separation from
Duyanta. In consequence, akuntal is exiled, while Duyanta
remains at home, suffering conflict with the demands of an earlier wife (thus, a rival). Duyanta defeats an army of demons
in battle and is subsequently reunited with akuntal and
their son.
In the Arab and Muslim world, few stories have been as popular and influential as that of Layla and Majnun. Layla and Majnun
fall in love, but Laylas father refuses the marriage. Majnuns
father tries to cure Majnun of his love madness through religion,
but Majnun only calls on God to make him worship Layla more.
Majnun wanders the desert, eventually trying to win Layla by
545
Narrative Universals
546
Structural Organization
A narrative is defined here as one way of recounting past events,
in which the order of narrative clauses matches the order of
events as they occurred. Example (1) is a minimal narrative organized in this way:
a.
b.
c.
d.
a.
b.
c.
d.
The end of a narrative is frequently signaled by a coda, a statement that returns the temporal setting to the present, precluding
the question and what happened then?
(4)
a. And you know the man who picked me out of the water?
b. Hes a detective in Union City,
c. and I see him every now and again.
Evaluation
Most adult narratives are more than a simple reporting of events.
A variety of evaluative devices are used to establish the evaluative point of the story (Polanyi 1989). Thus, we find that narratives, which are basically an account of events that happened,
frequently contain irrealis clauses negatives, conditionals,
futures which refer to events that did not happen or might have
happened or had not yet happened:
(5)
And the doctor just says Just that much more, he says, and
youd a been dead.
(6)
Ill tell you if I had ever walloped that dog Id have felt some
bad.
(7)
fact realized. Frequently, such evaluative clauses are concentrated in an evaluation section, suspending the action before
a critical event and establishing that event as the point of the
narrative.
Evaluative clauses vary along a dimension of objectivity. At
one extreme, narrators may interrupt the narrative subjectively
by describing how they felt at the time:
(8)
Narrative Preconstruction
When a narrator has made the decision to tell a narrative, he or
she must solve the fundamental and universal problem: Where
should I begin? The most reportable event, which will be designated henceforth as e0, is most salient, but one cannot begin with
it. Given the marked reportability of e0 and the need to establish its credibility, the narrator must answer the question How
547
How ordinary situations like (10) can give rise to the reportable and violent events that followed is a mystery that narrative
analysis can only contemplate, since they are part and parcel of
the contingent character of history.
548
Narratology
insertions of pseudoevents and removing them, detecting deletions and replacing them, and exchanging excuses for the action
excused. It is then possible to approximate the original chain of
events on which the narrative is based. A useful exercise is to
develop a complementary sub rosa case in the interests of the
antagonist. The comparison of these two constructions deepens
our understanding of how narrative skills are enlisted to transform the social meaning of events without violating our commitment to a faithful rendering of the past.
William Labov
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Goffman, Erving. 1981. Forms of Talk. Oxford: Blackwell.
Labov, William. 2001. The Social Stratification of English in New York City.
2d ed. Cambridge: Cambridge University Press.
. 2003. Uncovering the event structure of narrative. Georgetown
University Round Table 2001. Ed. Deborah Tannen and James Alatis,
6383. Washington, DC: Georgetown University Press. This article
develops the search for the events underlying the narrative.
. 2004. Ordinary events. In Sociolinguistic Variation: Critical
Reflections. ed. C. Fought, 3143. Oxford: Oxford University Press. An
exploration of the evaluative effect of inserting ordinary events into
narrative.
. 2006. Narrative preconstruction. Narrative Inquiry 16: 3745. A
fuller development of this topic.
Labov, William, and Joshua Waletzky. 1967. Narrative analysis. In Essays
on the Verbal and Visual Arts, ed. J. Helm, 1244. Seattle: University
of Washington Press. Repr. Journal of Narrative and Life History 7
(1997): 338.
Maranho, Tulio. 1984. The force of reportive narratives. Papers in
Linguistics 17 (3): 23565.
Norrick, Neal R. 2005. The dark side of tellability. Narrative Inquiry
15: 32344.
Polanyi, Livia. 1989. Telling the American Story. Cambridge, MA: MIT
Press.
Sacks, Harvey. 1992. Lectures on Conversation. Vols. 1 and 2. Ed. Gail
Jefferson. Oxford: Blackwell.
Trabasso, T., and P. van den Broek. 1985. Causal thinking and the
representation of narrative events. Journal of Memory and Language
27: 122.
NARRATOLOGY
The French term narratologie (formed in parallel with biology,
sociology, etc. to denote the study of narrative) was coined by
Tzvetan Todorov in his 1969 book Grammaire du Dcamron.
The early narratologists participated in a broader structuralist
revolution that sought to use Saussurean linguistics as a pilot
science for studying diverse forms of cultural expression, which
structuralist theorists characterized as rule-governed signifying
practices or languages in their own right (see structuralism;
Culler 1975). Likewise, narratologists such as Todorov, Roland
Barthes, Claude Bremond, Grard Genette, and Algirdas Julien
Greimas, adapted Ferdinand de Saussures distinction between
la parole and la langue to construe particular stories as individual narrative messages supported by an underlying semiotic
code (see semiotics). And just as Saussurean linguistics privileged code over message, focusing on the structural constituents
and combinatory principles of the semiotic system of language,
Narratology
rather than on situated uses of that system, structuralist narratologists privileged narrative in general over individual narratives,
emphasizing the general semiotic principles according to which
basic structural units (characters, states, events, actions, etc.) are
combined and transformed to yield specific narrative texts.
In this brief overview, I trace in further detail some of the
developments from which structuralist narratology took rise
and outline key contributions by early theorists. I also review
limitations of the structuralist approach to narrative inquiry
limitations that manifested themselves as story analysts began
to engage more fully with recent research in the language sciences, among other areas of study. To map the evolution of the
field, I draw a distinction between classical and postclassical
approaches to narratological analysis (cf. Herman 1999). Classical
narratology encompasses the tradition of research, rooted in
Russian Formalist literary theory as well as earlier precedents,
that was extended by structuralist narratologists starting in the
mid-1960s and refined and systematized up through the early
1980s by scholars such as Mieke Bal, Seymour Chatman, Wallace
Martin, Gerald Prince, and others. The Anglo-American tradition
of scholarship on fictional narrative can also be included under
the rubric of classical approaches, though for reasons of space,
this discussion focuses mainly on the Formalist-structuralist tradition. Postclassical narratology, meanwhile, designates frameworks for narrative research that build on this classical tradition
but supplement it with concepts and methods that were unavailable to story analysts during the heyday of structuralism. In
developing postclassical approaches, which not only expose the
limits but also exploit the possibilities of older models, theorists
of narrative have drawn on a range of fields, from gender theory,
philosophical ethics, and comparative media studies to sociolinguistics, the philosophy of language, and cognitive science.
Given the focus of the present encyclopedia, I concentrate here
on productive synergies between postclassical narratology and
research in the language sciences.
549
Narratology
constitutes humans narrative COMPETENCE. Hence, having
conferred on linguistics the status of a founding model ([1966]
1977, 82), Barthes identifies for the narratologist the same object
of inquiry that (mutatis mutandis) Saussure had specified for the
linguist: the code or system from which the infinity of narrative
messages derives and on the basis of which they can be understood as stories in the first place.
Narratologists like the early Barthes used structuralist linguistics not just to identify their object of analysis but also to elaborate their method of inquiry. In this connection, the adaptation of
structuralist-linguistic concepts and methods was to prove both
enabling and constraining. On the positive side, the example of
linguistics did provide narratology with a productive vantage
point on stories, affording terms and categories that generated
significant new research questions. For example, the linguistic
paradigm furnished Barthes with what he characterized as the
decisive concept of the level of description (Barthes [1966]
1977: 8588). Imported from grammatical theory, this idea suggests that a narrative is not merely a simple sum of propositions
but, rather, a complex structure that can be analyzed into hierarchical levels in the same way that a natural-language utterance
can be analyzed at the level of its syntactic, its morphological,
or its phonological representation. Barthes himself distinguishes three levels of description. At the lowest or most granular level are basic meaning-bearing elements that he termed
functions, which can be mapped out both distributionally and
in terms of paradigmatic classes; then come characters actions
that collocate to form narrative sequences; and finally there is
the level of narration, or the profile that narrative assumes when
viewed as a communicative process.
Likewise, Genette ([1972] 1980) drew on a broadly grammatical paradigm in using the categories of tense, mood, and
voice to characterize the relations among the story (= the basic
sequence of states, actions, and events recounted), the text on
the basis of which interpreters reconstruct that story, and the
act of narration that produces the text. Indeed, Genettes work
in the area of narrative temporality constitutes one of the truly
outstanding achievements in the field. Developing distinctions
that bear an interesting resemblance to Hans Reichenbachs
(1947) discriminations among event time, reference time, and
speech time, Genette focuses on two kinds of temporal relationships: 1) that between between narration and story and
2) that between text and story. In connection with the first,
Genette distinguishes between simultaneous, retrospective,
prospective, and intercalated modes of narration; in connection with the second, he develops the categories of duration,
order, and frequency. Duration can be computed as a ratio
between the length of time that events take to unfold in the
world of the story and the amount of text devoted to their narration, with speeds ranging from descriptive pause to scene to
summary to ellipsis. Order can be analyzed by matching the
sequence in which events are narrated against the sequence
in which they can be assumed to have occurred, yielding
chronological narration, analepses or flashbacks, and prolepses or flashforwards, together with various subcategories
of these nonchronological modes. Finally, frequency can be
calculated by measuring how many times an event is narrated
against how many times it can be assumed to have occurred in
550
Narratology
than those developed by the classical narratologists. In each of
these domains, story analysts are working to adopt more narrative-appropriate tools from the language sciences that is, tools
that can throw light on how narrative, as a distinctive kind of language use, constitutes a cognitive and communicative resource
by means of which human beings make sense of themselves, one
another, and the world.
Narrative Comprehension: To explore aspects of narrative
processing, story analysts have drawn on a range of theoretical frameworks that were unavailable to the structuralist narratologists, including artificial intelligence research (work on
knowledge representations), accounts of mental models,
cognitive linguistics, and research on text processing.
For example, Catherine Emmott (1997) presents a powerfully
integrative theory of narrative comprehension as a process of
using textual cues to build and update complex mental representations that she terms contextual frames, which contain
information about narrative agents, their situation in time
and space, and their relationships with one another. Mark
Turner (2003), meanwhile, relates story comprehension to
more general cognitive processes that involve conceptual
blending.
Speech and Thought Representation: Narratologists have
drawn on fields ranging from dialectology, pragmatics,
discourse analysis (linguistic), and historical
linguistics to study aspects of speech representation in
narrative, including dialect representations and fictional
portrayals of scenes of conversational interaction. Likewise,
to study representations of characters mental functioning, analysts have begun to work toward a rapprochement
between narratological theory and ideas from cognitive and
social psychology, research on emotion, cognitive linguistics,
and other frameworks for inquiry.
Focalization Theory: Initially given impetus by Genettes
([1972] 1980) attempt to reformulate theories of narrative
perspective or point of view in more rigorous terms, focalization theory has in recent years taken on an increasingly
interdisciplinary profile. Manfred Jahn (1996) has drawn on
the cognitive science of vision to propose a powerful account
of the perspective-marking features of narrative. Meanwhile,
David Herman (2009) uses ideas from cognitive grammar
to propose refinements to Genettes theory.
Quantitative, Corpus-Based Research: Story analysts have
begun to work with large text corpora (see corpus linguistics ) to study whether the distributional facts support
accounts proposed by earlier narratologists on the basis of
their own readerly intuitions. On the one hand, hypothesisdriven approaches to corpus study use a top-down method,
attempting to map assumed categories of structure onto specific texts or corpora to test the validity of prior theories. On
the other hand, bottom-up approaches, seeking to reduce
theoretical presuppositions to a minimum, work to induce
categories and models from surface features that can be
identified through automated analysis of narrative corpora.
It is not just that narratologists have begun to engage more fully
with concepts and methods from the language sciences, however; more than this, the field is currently being revolutionized by
551
552
whole set of ethnolinguistic nationalisms in Europe immediately afterwards. Though the historical differences between the
various social movements cannot be elided, they were inspired
by a number of German post-Kantian idealist thinkers. J. G.
Herders assertion in 1768 that each national language forms
itself in accordance with the ethics and manner of thought of
its people (2002, 50) was an important articulation of the link
between language and nation; by the time that William von
Humboldt gave his definition of a nation in 1836 (a body of men
who form language in a particular way [1988, 153]), the connection appeared almost axiomatic. In 1808, J. G. Fichte spelled out
the political significance of linguistic nationalism by arguing that
wherever a separate language is found, there a separate nation
exists, which has the right to take charge of its independent
affairs and to govern itself (1968, 49). The implications of the
doctrine were realized in the role that it played in national independence campaigns conducted by Greeks, Czechs, Hungarians,
Bulgarians, Ukrainians, Finns, Norwegians, Afrikaners, and
the Irish. Some postcolonial activists today, the Kenyan writer
Ngg wa Thiongo, for example, use the same model of linguistic
nationalism in their contemporary struggle, not so much against
colonialism but in order to counter the legacy of colonial rule.
Andersons account of the nation as an imagined community drew attention to the constructedness of the concept by
pointing to its precise historical origins. Yet the role of language
in the imagining of the community of the nation is also one that
arises at particular moments in history and serves specific functions; it is neither transhistorical nor general. It is also worth
noting that the conception of language underpinning this act of
imagination is one that has been criticized. Thus, M. M. Bakhtin,
in the important essay Discourse in the Novel, points to the fact
that national languages are produced by various types of institutional forces intellectual (linguistic theorizing), educational
(grammars and dictionaries), political (legislation) which act
centripetally in order to create a determinate, fixed, and knowable form. As part of this process, the realities of heteroglossia
(see dialogism and heteroglossia) social difference
inscribed in language by means of variation past and present
have to be banished. Historians, such as E. J. Hobsbawm (1990),
have noted the historical significance of such linguistic selection
and ranking, while linguistic anthropologists have drawn attention to the fact that the homogeneous language of nationalism is
as imaginary as the community that accompanies it (Irvine and
Gal 2000).
The extent to which such insights will have an impact in
political and linguistic thought remains to be seen. It is certainly
the case, however, that the postulated relationship between language and nation is now treated much more skeptically. At the
reactionary edge of forms of linguistic nationalism, there are still
those who argue for the purity of language as a way of guaranteeing the integrity of the nation. But the very fact that the
vast majority of nations past and present have been multilingual
communities including a number of those whose very entrance
into history depended on an emphasis on their supposed monolingualism radically undermines the ideological case for linguistic nationalism.
Tony Crowley
553
Hanoch Ben-Yami
554
Network Theory
Besides NPIs, English also has positive-polarity items
(would rather, sorta) that dont occur in negative polarity contexts; possible polarity items (tell time) that can occur only
within the scope of a possible-type modal; and combinations,
like the impossible polarity item fathom that require both negative scope and a modal.
John M. Lawler
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Atlas, Jay D. 1996. Only noun phrases, pseudo-negative generalized
quantifiers, negative polarity items, and monotonicity. Journal of
Semantics 13.4: 265328. Negative polarity infiltrates logic.
Baker, C. L. 1970. Double negatives. Linguistic Inquiry 1: 16986.
Horn, Laurence R. 1969. A presuppositional analysis of only and even.
Chicago Linguistics Society: CLS 5: 97108.
. 1989. A Natural History of Negation. Chicago: University of
Chicago Press. Horns revision and extension of his 1972 dissertation.
The classical neo-Gricean analysis.
Horn, Laurence R., and Yasuhiko Kato. 2000. Introduction: Negation and
polarity at the millennium. In Studies in Negation and Polarity, ed.
Laurence R. Horn and Yasuhiko Kato, 119. Oxford: Oxford University
Press. An excellent survey.
Israel, Michael. 2004. The pragmatics of polarity. In The Handbook of
Pragmatics, ed. L. Horn and G. Ward, 70123. Oxford: Blackwell.
Klima, Edward S. 1964. Negation in English. In The Structure of
Language, ed. J. Fodor and J. Katz, 246323. Englewood Cliffs,
NJ: Prentice Hall. The first modern syntactic/semantic study.
Ladusaw, William A. 1980. Polarity Sensitivity as Inherent Scope Relations.
New York: Garland. The origins of the downward-entailment theory,
using visual metaphors like scope and focus.
Lakoff, Robin. 1970. Some reasons why there cant be any some-any
rule. Language 45: 60815.
Lawler, John M. 1974. Ample negatives. Chicago Linguistics Society: CLS
10: 35777. After Klima, negative polarity was developed extensively in
the generative semantics tradition (Horn, Lakoff, Ross, Lawler,
and many others), largely using a negative polarity field metaphor,
along with negative triggers and secondary triggering.
Linebarger, Marcia. 1981. The grammar of negative polarity. Ph.D. diss.,
Massachusetts Institute of Technology. A then-orthodox generative
treatment, using rule-based metaphors like NPIs being licensed.
. 1991. Negative polarity as linguistic evidence. Papers from the
Parasession on Negation. Chicago Linguistic Society: CLS 27: 16588.
McCawley, James D. 1993. Everything That Linguists Have Always Wanted
to Know About Logic (But Were Ashamed to Ask). Chicago: University
of Chicago Press. Oxford: Blackwell. McCawleys modern generative
semantic analysis; for example, In natural language, negation is not
truth-functional.
Ross, John R. 1973. Negginess. Paper delivered to the Winter Meeting of
the Linguistic Society of America, San Diego.
van der Wouden, Ton. 1996. Negative Contexts: Collocation, Polarity
and Multiple Negation. London and New York: Routledge. A useful
University of Groningen dissertation.
Zeijlstra, Hedde, and Jan-Philipp Soehn, eds. 2007. Proceedings of the
Workshop on Negation and Polarity. Tbingen: University Collaborative
Research Center. Recent evidence of polarity research expansion into
other languages.
NETWORK THEORY
Network theory concerns itself with the study of elements, called
vertices (e.g., words), and their connections, called edges or links
555
Network Theory
(e.g., two words are connected if one word has been elicited by
the other in a word association experiment; see Figure 1). This
theory has many applications in language sciences and is the
outcome of intersecting work of mathematicians and physicists,
who usually call it graph theory (Bollobs 1998) or complex
network theory (Newman 2003), respectively.
One of the major contributions of physicists has been to
unravel the statistical properties of real networks (Newman
2003), for example, the World Wide Web or protein interaction
networks. Firstly, physicists discovered that practically all real
networks exhibited the small world phenomenon. The term
small world comes from the observation that everyone in the
world can be reached through a short chain of social acquaintances, although the number of people in the whole social network is huge. In the word association network, partially shown
in Figure 1, volcano is reached from ache through a chain of at
least four links, while only one link separates fire from volcano.
Secondly, physicists found that many real networks had a heterogeneous degree distribution. Loosely speaking, this property
means that there are vertices (words) with a disproportionately
large number of connections (the so-called hubs). For instance,
in the network partially shown in Figure 1, the five words with the
highest degrees are food, money, water, car, and good (Steyvers
and Tenenbaum 2005). Finally, another fundamental property of
real networks is clustering; that is, roughly speaking, if two vertices are connected to the same vertex they are likely to be directly
connected as well.
Network theory has contributed to the study of language in
three ways: a) by characterizing the statistical properties of linguistic networks, such as networks of word association (Steyvers
and Tenenbaum 2005), thesauri (Sigman and Cecchi 2002),
and syntactic dependencies (Ferrer i Cancho, Sol, and Khler
2004); b) by modeling the properties of these networks (Steyvers
and Tenenbaum 2005; Motter et al. 2002); and c) by proposing
abstract models that provide a further understanding of the faculty of language (Ferrer i Cancho, Riordan, and Bollobs 2005).
Although the systematic application of network theory to language is a young field (starting in the early twenty-first century)
within quantitative linguistics, it can be concluded that
556
the small world phenomenon, high clustering, and heterogeneous degree distribution are common properties of linguistic
networks (Mehler 2008). Most models are based on the preferential attachment principle proposed by Albert Lszl Barabsi
and Rka Albert (1999): Vertices with many connections are
more likely to become more connected in the future than those
with few connections (Steyvers and Tenenbaum 2001, 2005;
Dorogovtsev and Mendes 2001; Motter et al. 2002).
The challenges of the application of network theory are to
explain the properties of these networks (most studies are merely
descriptive); to incorporate deeper statistical techniques, for
example, degree correlation analysis (Serrano et al. 2007); and
to extend the studies to more languages (most studies are in
English). For these reasons, it is too early to argue that the heterogeneous degree distributions and other statistical patterns
constitute laws of language in the sense of absolute and
statistical universals. When applied to syntactic networks,
network theory has helped to explain the origins of the properties
of the syntactic dependency structure of sentences, for example,
the exceptionality of syntactic dependency crossings (Ferrer
i Cancho 2006) and has provided new tracks for understanding syntax at the large scale of syntactic organization (Ferrer i
Cancho, Sol, and Khler 2004), above the traditional sentence
level (see syntax, universals of).
In their pioneering application of network theory, Mark
Steyvers and Joshua B. Tenenbaum (2001, 2005) studied the
large-scale organization of various kinds of semantic networks
(e.g., word association networks, as in Figure 1) and proposed
a simple model for explaining the small-worldness, high clustering and a heterogeneous degree distribution of semantic
networks. Over time, new vertices (e.g. words) are added and
attached to existing vertices according to two principles: a)
Barabsi and Alberts preferential attachment and b) differentiation. Differentiation means that a new vertex tends to mimic the
connectivity pattern of an existing vertex.
Network theory has shed new light on the evolution of language by defining the necessary conditions for the existence of
language (e.g., word ambiguity) and also by suggesting the
possibility that language could have appeared for free as a side
557
558
Neuroimaging
Nieoullon, A. 2002. Dopamine and the regulation of cognition and attention. Progress in Neurobiology 67: 5283.
Sabe, L., F. Salvarezza, A. Garcia Cuerva, R. Leiguarda, and S. Starkstein.
1995. A randomized, double-blind, placebo-controlled study of bromocriptine in nonfluent aphasia. Neurology 45: 22724.
Tanaka, Y., and D. L. Bachman. 2000. Pharmacotherapy of Aphasia. In
Neurobehavior of Language and Cognition: Studies of Normal Aging
and Brain Damage, ed. and M. Albert, L. Connor, and L. Obler, 15978.
Boston: Kluwer Academic.
Tanaka, Y., M. Miyazaki, and M. Albert. 1997. Effects of cholinergic activity on naming in aphasia. Lancet 350: 11617.
Thierry, A. M., J. P. Tassin, G. Blanc, L. Stinus, B. Scatton, J. Glowinski.
1977. Discovery of the mesocortical dopaminergic system: Some
pharmacological and functional characteristics. Advances in
Biochemistry and Psychopharmacology 16: 512.
Williams, G. V., and P. S. Goldman-Rakic. 1995. Modulation of memory fields by dopamine D1 receptors in prefrontal cortex. Nature
376: 5725.
NEUROIMAGING
Neuroimaging technologies provide a major source of new data
about how the language system is organized in the brain. In particular, activation imaging approaches, in which brain activity is
monitored while subjects perform some language task, allow us
to visualize various aspects of language processing in the normal
brain and test hypotheses about component language functions
or systems. Among the imaging techniques most commonly in
use for understanding language in the brain are structural magnetic resonance imaging (MRI), functional MRI (fMRI), positron
emission tomography (PET), electroencephalography (EEG),
and magnetoencephalography (MEG).
559
Neuroimaging
560
Neuroimaging
language activation studies, where subjects might receive several injections while performing one or more language tasks and
during control tasks (Color Plate 8).
One of the first PET studies to visualize language areas in vivo
was that of S. E. Petersen and colleagues (1988). In this study,
normal volunteers performed a series of language tasks ordered
hierarchically: viewing a crosshair on a screen; seeing printed
words on a screen or hearing words over headphones, reading or
repeating heard words, or generating an action verb corresponding to a visually or auditorally presented noun. By subtracting
lower-level tasks from higher-order language tasks, the authors
isolated areas of the brain involved in word generation, while
removing unwanted effects of sensory stimulation. This subtractive logic forms the basis of activation imaging experiments.
While there are theoretical difficulties with assumptions of hierarchical organization, cognitive subtraction models remain the
mainstay of activation imaging research.
An important disadvantage of PET is the need to expose subjects to radioactivity; a second disadvantage is that subjects must
perform the same task for several minutes continuously to obtain
a single brain image. Also, PET is a relatively noisy methodology, and the signal-to-noise ratio is low enough that scans must
be averaged over a group of subjects. Finally, the spatial resolution of PET is low, usually about 6 mm, so that individual brain
structures cannot be resolved and the areas of significant activation are not easily localized to a specific brain structure. Most
investigators solve this problem by performing an MRI scan for
each subject, mathematically moving or registering the brain
images so they are in the same space, then overlaying the PET
activation regions onto the corresponding MRI scans to localize
the regions of activity.
FUNCTIONAL MRI. In the early 1990s, two groups independently
discovered that blood flow increases during neural activity
could be measured directly with MRI (Kwong et al. 1992; Ogawa
et al. 1992). This is due to the accident that oxygenated blood and
deoxygenated blood have slightly different magnetic properties.
During increased brain activity, an increase in blood flow is not
matched by an increase in oxygen consumption; consequently,
more oxygenated blood spills over to the venous side of the capillary bed. Scans of the brain taken during this state of increased
oxyhemoglobin concentration have slightly higher MRI signals
than those taken in the resting state. Thus, fMRI measures this
blood-oxygen-level dependent, or BOLD, signal when comparing scans taken in different cognitive states. Thanks to the discovery of ultrafast MRI scanning, typically using an approach called
echo-planar imaging or EPI, fMRI takes a complete picture of the
brain as quickly as once per second. By taking complete brain
volumes every few seconds over a period of several minutes or
more, fMRI can track those magnetic changes that correlate with
blood flow during experimental and control tasks that can be
varied with tremendous experimental complexity. Because fMRI
has significantly greater spatial and temporal resolution than
PET, it is often possible to see significant language-related brain
activity within a single individual in a matter of minutes.
The study of language organization in the brain has been
revolutionized by fMRI. Because these studies are relatively easy
and inexpensive to conduct, it is possible to examine language
561
Neuroimaging
Number
562
NUMBER
Number is a grammatical feature that quantifies the denotation
of a linguistic element. It can refer to entities or events, and in
language we find both nominal number (very common, discussed in the following) and verbal number (less common, realized on the verb to indicate the number of events or the number
of participants; also called pluractionality).
Languages vary with regard to the part of their nominal inventory that is involved in the number system. In different languages,
the split into nominals that do and do not express number may
occur at different points of the animacy hierarchy: speaker (first
person pronouns) > addressee (second person pronouns) > third
person pronouns > kin > rational > human > animate > inanimate.
Furthermore, not all nouns are number differentiable. Two types of
noun are traditionally distinguished: count nouns and mass nouns,
the latter regarded as lacking the number distinction. At the level of
semantics, the countmass distinction can be captured with two
semantic features boundedness and internal structure (Jackendoff
1991), which corresponds to the distinction between temporally
bounded and unbounded events in verbal semantics. But countability is really a characteristic of nominal phrases (Allan 1980),
since many nouns can appear in both count and mass syntactic
contexts, for example, Would you like a cake/some cake? We need a
bigger table/ There is not enough table for everyone to sit at.
When nominal number is found expressed on the noun or the
noun phrase as such, it is considered inherent. When found on other
elements of the noun phrase or on the verb, it is contextual. The
Occipital Lobe
expressions of nominal number can involve special number words
(of different syntactic status); syntactic means (i.e., agreement,
found most commonly on demonstratives and verbs but also on
articles, adjectives, pronouns, nouns in possessive constructions,
adverbs, adpositions, and complementizers); a variety of morphological means (inflections, stem changes, zero expressions,
clitics); and lexical means (e.g., suppletion). Number is often
marked in more than one way within one language.
All nominal number systems are built on the primary opposition between singular (expressing the quantity one) and plural
(more than one). Other attested number values are dual (two),
trial (three), and paucal (a few). There may be further divisions
into paucal and greater paucal, plural and greater plural (the last
value may imply an excessive number, or all possible instances of
the referent). No genuine quadrals (four) have been found. The
largest number systems involve five values. In many languages,
the absence of plural marking does not necessarily imply the singular, but the form may be outside the number opposition and
express general number, that is, the meaning of the noun without
reference to number.
Associatives, distributives, and collectives all sometimes
listed as additional values of number are better analyzed as
independent features. Associativity expresses the meaning X
and the group associated with X; distributives indicate that
entities (whether count or mass ones), events, qualities, or locations are to be construed as distinct in space, sort, or time; and
collectives indicate that the members of a group are to be construed together as a unit. Many languages have markers for these
categories in addition to various number markers.
Anna Kibort
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Allan, Keith. 1980. Nouns and countability. Language 56: 54167.
Corbett, Greville G. 2000. Number. Cambridge: Cambridge University
Press.
Jackendoff, Ray. 1991. Parts and boundaries. Cognition 41: 945.
O
OCCIPITAL LOBE
Alcmaeon of Croton, probably the first person to suggest that the
mind is located in the brain and not the heart, also suggested that
the optic nerves are light-bearing paths to the brain. His revolutionary ideas, formulated about 2,500 years ago, were ignored
by Egyptian and Greek scholars alike (most notably by Aristotle).
We now know that visual information is delivered to the occipital lobe mainly via the thalamus. The occipital lobe is the most
posterior of the four lobes of the brain (named after the four skull
bones beneath which they lie).
Even until very recently, it was believed that the occipital lobe
was involved only in the processing of visual information per
se. Language was considered to take place in dedicated areas,
mainly in the frontal and temporal cortex. Thus the relevance of
the occipital lobe to language processing and language research
was only indirect in the sense that its role was limited to providing visual input processed later on by language centers in other
brain regions that contribute to certain aspects of language (such
as reading).
New data are now challenging this view of the occipital lobe
as solely processing visual information. First, several key studies
on the blind have shown the involvement of the occipital lobe
in processing other sensory modalities. In particular, there is
a clear link between occipital lobe processing in the blind and
language and verbal memory functions, where this pattern of
activation is attributed to massive reorganization of the occipital lobe in cases of blindness. Further, the involvement of the
occipital lobe in nonvisual processing has clearly been demonstrated in the sighted (i.e., under normal development of the
occipital lobe), and much of todays research is exploring the
extent to which the occipital lobe is involved in language processing under normal development. These topics are covered in
the following sections.
563
Occipital Lobe
this organization is still subject to debate (for a review, see GrillSpector and Malach 2004 and the following).
2. Visual Pathways. Lesion studies in primates and human
fMRI studies both suggest that there are two processing streams
between early, retinotopic visual areas in the occipital cortex
and higher-order processing centers in the occipito-temporal
and occipito-parietal lobes. These two streams are referred to
as the dorsal and ventral streams, respectively (Ungerleider
and Haxby 1994). Because the dorsal stream is also involved
in visuo-motor transformations, a differentiation is made
between vision for action (dorsal stream) and vision for perception (ventral stream) (Goodale and Milner 1992). The ventral stream contains structures devolved to the fine analysis of
a visual scene such as form and color. Thus, it is also known
as the what pathway. It consists of areas V1V4 in the occipital
lobe and several regions that belong to the lateral and ventral
temporal lobe.
3. Functional Specialization. This principle of division of
labor, which leads to a specialization of function in the various cortical areas, was originally suggested for the visual system (Zeki 1978). Electrophysiological studies in nonhuman
primates have identified organizing principles in addition
to retinotopy, such as selectivity for simple features like spatial orientation in V1 and selectivity for categories of complex
stimuli like faces for spatial layouts in inferior temporal (IT)
cortex. In the last decade or so, the use of noninvasive functional imaging, particularly fMRI, has dramatically increased
our knowledge of the functional organization of the human
visual cortex and its relation to vision, due to its ability to provide a large-scale neuro-anatomical perspective (e.g. Martin
and Chao 2001). An active debate is ongoing about the actual
organization of the ventral stream (Grill-Spector and Malach
2004). One area in the lateral occipital cortex, the lateral occipital complex (LOC; Malach et al. 1995) responds strongly to
pictures of intact objects by contrast to scrambled objects or
nonobject textures. In the ventral occipito-temporal cortex,
specialized areas for faces (fusiform face area, FFA; Kanwisher,
McDermott, and Chun 1997), scenes (Epstein and Kanwisher
1998), and human body parts (Downing et al. 2001) have been
described, as well as for visual word forms (visual word form
area, VWFA; McCandliss, Cohen, and Dehaene 2003) which
has direct implications for the discussion here. Developing a
theoretical framework that captures these specialized regions
continues to be problematic, although the notion of widely
distributed and overlapping cortical object representations
remains a likely principle of organization (Haxby et al. 2001).
The effects of perceptual expertise for certain object categories
(Gauthier et al. 2000) and, more recently, different categoryrelated resolution needs (Grill-Spector and Malach 2004) have
also been put forward as candidate organizational principles of
the human ventral stream.
564
Occipital Lobe
565
Occipital Lobe
Optimality Theory
566
Hubel, D. H., and T. N. Wiesel. 1963. R Shape and arrangement of columns in cats striate cortex. Journal of Neurophysiology 165: 55968.
. 1965. Receptive fields and functional architecture in two nonstriate visual areas (18 and 19) of the cat. Journal of Neurophysiology
28: 22989.
Kanwisher, N., J. McDermott, and M. M. Chun. 1997. The fusiform face
area: A module in human extrastriate cortex specialized for face perception. Journal Neuroscience 17: 430211.
Liu, Y., et al. 2007. Whole brain functional connectivity in the early
blind. Brain 130: 208596.
Malach, R., et al. 1995. Object-related activity revealed by functional
magnetic resonance imaging in human occipital cortex. Proceedings
of the National Academy of Science 92: 81359.
Martin, A. 2006. Shades of Djerine forging a causal link between the
visual word form area and reading. Neuron 50: 1735.
Martin, A., and L. L. Chao. 2001. Semantic memory and the
brain: Structure and processes. Current Opinion in Neurobiology
11: 194201.
McCandliss, B. D., L. Cohen, and S. Dehaene. 2003. The visual word form
area: Expertise for reading in the fusiform gyrus. Trends in Cognitive
Science 7: 2939.
Nobre, A. C., T. Allison, and G. McCarthy. 1994. Word recognition in the
human inferior temporal lobe. Nature 372: 2603.
Noppeney, U., K. J. Friston, and C. J. Price. 2003. Effects of visual deprivation on the organization of the semantic system. Brain 126: 16207.
Ofan, R. H., and E. Zohary. 2006. Visual cortex activation in bilingual
blind individuals during use of native and second language. Cerebral
Cortex 17: 124959.
Pascual-Leone, A., A. Amedi, F. Fregni, and L. B. Merabet. 2005. The plastic human brain cortex. Annual Review of Neuroscience 28: 377401.
Price, C. J., and J. T. Devlin. 2003. The myth of the visual word form area.
Neuroimage 19: 47381.
Rder, B., O. Stock, S. Bien, H. Neville, and F. Rosler. 2002. Speech processing activates visual cortex in congenitally blind humans. European
Journal of Neuroscience 16: 9306.
Sadato N., A. Pascual-Leone, J. Grafman, V. Ibanez, M. P. Deiber, et al.
1996. Activation of the primary visual cortex by Braille reading in
blind subjects. Nature 380: 5268.
Schlaggar, B. L., and B. D. McCandliss. 2007. Development of neural systems for reading. Annual Reviews in Neuroscience 30: 475503.
Serano, S.C., et al. 1995. Borders of multiple visual areas in humans
revealed by functional magnetic resonance imaging. Science 268:
88993.
Ungerleider, L. G., and J. V. Haxby. 1994. What and where in the
human brain. Current Opinion in Neurobiology 4: 15765.
van Atteveldt, N., E. Formisano, R. Goebel, and L. Blomert. 2004.
Integration of letters and speech sounds in the human brain. Neuron
43: 27182.
Zangaladze, A., C. M. Epstein, S. T. Grafton, and K. Sathian. 1999.
Involvement of visual cortex in tactile discrimination of orientation.
Nature 401: 58790.
Zeki, S. M. 1978. Functional specialization in the visual cortex of the rhesus monkey. Nature 274: 4238.
OPTIMALITY THEORY
Optimality Theory (OT; Prince and Smolensky [1993] 2004) is a
formal theory of constraint interaction in grammar that seeks
to explain how and to what extent natural languages may vary.
In addition to this central question of generative grammar,
research in OT addresses questions of the grammars use in performance, its acquisition, and its neural realization.
Optimality Theory
COMPETITION AND CONFLICT. The fact that well-formedness constraints are violable in OT means that two constraints often conflict: Satisfying one requires violating the other. Grammaticality
is therefore not equated with satisfaction of all grammatical
constraints. Grammatical structures are simply structures that
suffer less severe constraint violations than their ungrammatical
counterparts. This means that the evaluation of grammaticality is
inherently comparative. At any level of description (phonological,
syntactic, etc.) the universal set of possible structural descriptions of an input I, Gen(I), forms a candidate set, a collection of
OT in Theoretical Linguistics
Characterizing the set of possible natural languages minimally
requires specifying i) the mental representations that are characteristic of language, ii) the constraints that distinguish possible
from impossible linguistic systems, and iii) the formal mode of
interaction among these constraints. Concerning mental representations, OT imposes no restrictions beyond requiring that
specifications of phonological, syntactic, or semantic structure
be explicit in the sense of generative grammar. This makes OT
compatible with alternative substantive theories of particular grammar components; in syntax, for example, OT versions
of government and binding (Grimshaw 1997; Legendre,
Smolensky, and Wilson 1998), lexical-functional grammar (Bresnan 2000), and the minimalist program (Mller
1997) are flourishing. OTs main contribution concerns constraint interaction (iii) and, as a consequence, the proper formal
characterization of the constraints themselves (ii). Hence, OT
is best characterized as a meta-theory of grammatical structure
compatible with any explicit theory of linguistic representation.
It is therefore applicable to all linguistic levels, and has been
applied to phonology (McCarthy and Prince 1993; Prince and
Smolensky [1993] 2004), syntax (Legendre, Grimshaw and
Vikner 2001), semantics (Hendriks and de Hoop 2001), and
pragmatics (Blutner and Zeevat 2004; Blutner, De Hoop, and
Hendriks 2006).
According to OT, all grammatical constraints are universal
and violable (or soft) a claim that represents a major departure from previous approaches to the characterization of phonological and syntactic knowledge via language-particular rewrite
rules written in a universal notation, or as universal, inviolable
constraints supplemented by additional principles subject to
language-particular parameterization.
567
Optimality Theory
Faithfulness constraints, unique to OT, have been shown to
operate at all levels of linguistic description. In syntax, faithfulness constraints play a crucial role in accounting for languageparticular ineffability, that is, syntactic structures that are simply
impossible in certain languages, for example, multiple wh-questions in Irish (Legendre, Smolensky, and Wilson 1998; Legendre
2009). In most languages, faithfulness to question operators in a
semantic input force the optimal syntactic output to contain multiple wh-phrases. In languages like Irish, however, such faithfulness is ranked below the syntactic constraints violated by clauses
containing multiple wh-phrases, and thus no optimal syntactic
structure contains multiple wh-phrases. Multiple wh-questions
are therefore inexpressible (in a single clause).
VARIATION. OT is naturally extensible to unstable states of language, such as free variation (Anttila 1998), dialectal variation,
and diachronic change (Nagy and Reynolds 1997). Relaxing the
requirement of a complete ranking of the universal constraints, a
language L may be characterized by a single partial ranking P; L
is generated by the set of grammars S full rankings consistent
with P. These generate a set of outputs for each input. According
to Joan Bresnan, Ashwini Deo, and Devyani Sharma (2007),
intraspeaker variation in the British paradigm for be arises from
a partial ranking P among faithfulness constraints, requiring an
output to express input agreement features and markedness
constraints penalizing all features in the output. A total ranking
consistent with P having more highly ranked faithfulness yields a
more highly inflected paradigm.
FURTHER DEVELOPMENTS. OT is an evolving theory; major developments include the following. Faithfulness: Output-output
faithfulness requires identity of a morphemes exponent across
its paradigm (Burzio 1994; Benua 1995); output-output anti-faithfulness constraints achieve morphophonological alternations by
demanding nonidentical surface forms for distinct underlying
forms (Alderete 2001); sympathy theory demands faithfulness
to suboptimal candidates (McCarthy 1999b). Harmonic evaluation: Comparative markedness evaluation distinguishes constraint violations that are shared with the most faithful output
from those that are not (McCarthy 2003); targeted constraints
only compare candidates that differ only in a specified way
(Wilson 2001b). Architecture: Stratal OT assumes differing rankings in a series of lexical levels (Kiparsky 2006); harmonic serialism derives the surface form from a series of small alterations,
each optimal at its point in the derivation (Prince and Smolensky
[1993] 2004; McCarthy 1999a); candidate chain theory evaluates entire derivations (McCarthy 2007); bidirectional optimization adds competition of interpretations/underlying forms to
the competition of expressions/surface forms that is standard in
OT (Smolensky 1996; Blutner 2000; Wilson 2001a). Probabilistic
formulations: In stochastic OT (Boersma 1998), each optimization ranks the constraints according to relative numerical values randomly selected for that optimization from a probability
distribution for each constraint, with a mean value determined
by that constraints strength in the grammar. In the maximum
entropy formulation of OT (Hayes and Wilson 2008), harmony is
numerical: Each constraint has a numerical strength determining the size of the penalties it assesses to violating candidates; the
568
Optimality Theory
neurons maximizes numerical harmony; at a higher level of
description, the same system computes the symbolic structural
description of the input that optimizes the harmony of an OT
grammar. Construed in these terms, the study of grammar is
fully integrated into the contemporary science of the mind/brain
(Smolensky and Legendre 2006).
Geraldine Legendre and Paul Smolensky
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Alderete, John. 2001. Dominance effects as transderivational antifaithfulness. Phonology 18: 20153.
Anttila, Arto. 1998. Deriving variation from grammar. In Variation,
Change, and Phonological Theory, ed. F. Hinskens, R. van Hout, and
W. L. Wetzel, 3568. Amsterdam: Benjamins.
Benua, Laura. 1995. Output-output faithfulness. In University of
Massachusetts Occasional Papers in Linguistics 18: Papers in Optimality
Theory, ed. Jill Beckman, Laura Walsh Dickey, and Suzanne Urbanczyk,
77136. Amherst: University of Massachusetts at Amherst, GLSA.
Blutner, Reinhard. 2000. Some aspects of optimality in natural language
interpretation. Journal of Semantics 17: 189216.
Blutner, Reinhard, Helen De Hoop, and Petra Hendriks. 2006. Optimal
Communication. Stanford, CA: CSLI Publications.
Blutner, Reinhard, and Henk Zeevat, eds. 2004. Pragmatics in Optimality
Theory. London: Palgrave Macmillan.
Boersma, Paul. 1998. Functional Phonology: Formalizing the Interactions
between Articulatory and Perceptual Drives. The Hague: Holland
Academic Graphics.
Boersma, Paul, and Bruce Hayes. 2001. Empirical tests of the gradual
learning algorithm. Linguistic Inquiry 32: 4586.
Bresnan, Joan. 2000. Optimal syntax. In Optimality Theory: Phonology,
Syntax and Acquisition, ed. Joost Dekkers, Frank van der Leeuw, and
Jeroen van de Weijer, 33485. Oxford: Oxford University Press.
Bresnan, Joan, Ashwini Deo, and Devyani Sharma. 2007. Typology in
variation: A probabilistic approach to be and nt in the survey of
English dialects. English Language and Linguistics 11: 30146.
Burzio, Luigi. 1994. Principles of English Stress. Cambridge: Cambridge
University Press.
Davidson, Lisa, Peter W. Jusczyk, and Paul Smolensky. 2006. Optimality
in language acquisition I: The initial and final states of the phonological grammar. In The Harmonic Mind: From Neural Computation
to Optimality-Theoretic Grammar. Vol. 2. Ed. Paul Smolensky and
Graldine Legendre, 793839. Cambridge, MA: MIT Press.
Eisner, J. 1997. Efficient generation in primitive optimality theory. Annual
Meeting of the Association for Computational Linguistics 35: 31320.
Eisner, Jason. 2000. Easy and hard constraint ranking in optimality theory: Algorithms and complexity. In Finite-State Phonology: Proceedings
of the Fifth Workshop of the ACL Special Interest Group in Computational
Phonology (Sigphon), ed. Jason Eisner, Lauri Karttunen and A. Thriault,
2233. Morristown, NJ: Association for Computational Linguistics.
Frank, Robert, and Giorgio Satta. 1998. Optimality theory and the generative complexity of constraint violability. Computational Linguistics
24: 30715.
Gibson, Edward, and Kevin Broihier. 1998. Optimality theory and human
sentence processing. In Is the Best Good Enough? Optimality and
Competition in Syntax, ed. Pilar Barbosa, Danny Fox, Paul Hagstrom,
Martha McGinnis, and David Pesetsky, 15791. MIT Working Papers
in Linguistics. Cambridge, MA: MIT Press.
Grimshaw, Jane. 1997. Projection, heads, and optimality. Linguistic
Inquiry 28: 373422.
Hayes, Bruce. 2004. Phonological acquisition in optimality theory: The
early stages. In Constraints in Phonological Acquisition, ed. Ren
569
Oral Composition
MA: Blackwell. Original version was a technical report, Rutgers
University and University of Colorado at Boulder..
Prince, Alan, and Bruce B. Tesar. 2004. Learning phonotactic distributions. In Constraints in Phonological Acquisition, ed. Ren Kager, Joe
Pater, and Wim Zonneveld. Cambridge: Cambridge University Press.
Rumelhart, David E., James L. McClelland, and the PDP Research Group.
1986. Parallel Distributed Processing: Explorations in the Microstructure
of Cognition. 2 vols. Cambridge, MA: MIT Press.
Smolensky, Paul. 1986. Information processing in dynamical systems: Foundations of harmony theory. In David E. Rumelhart, James
L. McClelland, and the PDP Research Group, Parallel Distributed
Processing: Explorations in the Microstructure of Cognition, I: 194281.
Cambridge, MA: MIT Press.
. 1988. On the proper treatment of connectionism. Behavioral
and Brain Sciences 11: 174.
. 1996. On the comprehension/production dilemma in child language. Linguistic Inquiry 27: 72031.
Smolensky, Paul, and Graldine Legendre. 2006. The Harmonic
Mind: From Neural Computation to Optimality-Theoretic Grammar.
Vol. 1. Cognitive Architecture. Vol. 2. Linguistic and Philosophical
Implications. Cambridge, MA: MIT Press.
Speas, Margaret. 1997. Optimality theory and syntax: Null pronouns and
control. In Optimality Theory: An Overview, ed. Diana Archangeli and
D.Terrence Langendoen, 17199. Malden, MA: Blackwell.
Stevenson, Suzanne, and Paul Smolensky. 2006. Optimality in sentence processing. In The Harmonic Mind: From Neural Computation
to Optimality-Theoretic Grammar. Vol. 2. Ed. Paul Smolensky and
Graldine Legendre, 30738. Cambridge, MA: MIT Press.
Tesar, Bruce B. 1996. Computing optimal descriptions for optimality
theory grammars with context-free position structures. Proceedings of
the Thirty-Fourth Annual Meeting of the Association for Computational
Linguistics, 1017. Morristown, NJ: Association for Computational
Linguistics.
. 1998. Error-driven learning in optimality theory via the efficient computation of optimal forms. In Is the Best Good Enough?
Optimality and Competition in Syntax, ed. Pilar Barbosa, Danny Fox,
Paul Hagstrom, Martha McGinnis, and David Pesetsky, 42135. MIT
Working Papers in Linguistics. Cambridge, MA: MIT Press.
Tesar, Bruce B., and Paul Smolensky. 1998. Learnability in optimality
theory. Linguistic Inquiry 29: 22968.
Wilson, Colin. 2001a. Bidirectional Optimization and the Theory of
Anaphora. In Optimality-Theoretic Syntax, eds. Geraldine Legendre,
Sten Vikner and Jane Grimshaw. Cambridge, MA: MIT Press.
. 2001b. Consonant cluster neutralisation and targeted constraints. Phonology 18: 14797.
Yip, Moira. 1993. Cantonese loanword phonology and optimality theory. Journal of East Asian Linguistics 2: 26191.
ORAL COMPOSITION
Oral composition broadly refers to the creation of organized verbal formulations without reliance on writing. Though in essence
a familiar process in everyday speech, it has become a quasitechnical and debated term, applied especially to bringing into
being relatively sustained examples of entextualized verbal art,
both ancient and recent. It has thus been of interest to linguists,
anthropologists, folklorists, psychologists, historians, and specialists in specific languages and cultures, also linking to work
on oral culture, performance, story, literacy, and
memory.
The central issues have been, first and most directly, how
lengthy oral poems, narratives, and other sustained verbal forms
570
can come into being without writing a puzzle for those steeped
in literate traditions that assume the centrality of the written
word; and second, how this relates to performance (for performance is arguably how an oral creation exists).
Earlier approaches focused largely on nonliterate settings,
especially those characterized as primitive or traditional. One
model was of spontaneous improvization by the unself-conscious child of nature, unfettered (and unhelped) by recognized artistic conventions. Another was of unchanging tradition
from the far-distant past, not composed by living creators but
stored in the communal tribal memory.
These models were largely superseded by the influential oralformulaic approach (also known as the oral theory) which came
to the fore in the mid-twentieth century. The concept of oral
composition acquired a specific meaning and became a key term
of analysis and explanation. Its classic statement in Albert Lords
seminal The Singer of Tales (1960) used fieldwork in the 1930s in
Yugoslavia to demonstrate how lengthy oral poems were composed during performance: The singers drew on a traditional
store of formulaic phrases and themes, which enabled them,
without writing or verbatim memorization, to pour forth long
epic songs in uninterrupted flow. Variations around such formulaic phrases as, for example, By Allah, he said, and mounted his
white horse recur throughout the poems, providing a parallel to
the Homeric epithets like fleet-footed Achilles found in early,
putatively oral, Greek epics. This special technique of composition (Lord 1960, 17) relied not on preplanned, memorized texts
but on composition-in-performance. Contrary to literate expectations, there was no fixed correct version: each performance was
authentic in its own right, a unique product composed and performed on one occasion. Oral-formulaic composing was linked
to a traditional, oral mindset incompatible with literacy and the
literate mind, and once singers became literate, it was posited,
they lost the power to compose orally.
The oral-formulaic theory was enormously influential
throughout much of the later twentieth century and across a
wide span of disciplines, providing, as it apparently did, an
answer to the puzzle of verbal composition without writing.
Examples of comparable formulaic expression and hence, it
seemed, of oral composition were identified throughout the
globe, from early Greek epic, Old English texts, or the Bible to
living examples recorded from the field, soon also extending to
the full range of poetic genres and to prose-like forms, such as
sermons or storytelling.
Though still regarded as a classic approach, oral-formulaic
theory has been both modified and challenged, especially during the last two decades. First, it has become apparent that not
all genres of unwritten verbal art follow the oral-formulaic composition-in-performance mode, nor, as implied by the classic
oral-formulaic analysts, is oral composition a single identifiable
process. Their often somewhat generalized conclusions have not
been fully supported by the empirical evidence, for oral forms
turn out to be created in diverse ways. Some are composed before
and separated from performance. Some do, after all, involve
memorization. One much-quoted case is of the Somali poets
who spend hours, sometimes days, composing elaborate genres
of oral poetry, later delivered word for word either by themselves
or by reciters who are able to memorize poems and, without
Oral Composition
writing, store large and exactly reproducible repertoires in their
memory over many years. Elsewhere, too, prior composition is
sometimes a long-drawn-out and carefully considered procedure, in some cases involving multiple authors and/or rehearsals before being performed. Certain womens personal songs in
mid-twentieth-century Zambia, for example, were thought out
by one woman, elaborated with her friends, worked over for days
by an expert composer, then rehearsed and memorized before
final performance. In other cases, a composer may speak aloud
words of rapid inspiration designed for later performance, to
be captured by listeners on the spot through memorizing, tape
recording, or writing (further details and discussion in Finnegan
1992, 5287; 2007, 96113, 179200). Contrary to the classic oralformulaic model, oral composing varies in different cultures,
genres, and circumstances.
Second, the assumption that literacy and orality are mutually
incompatible has been extensively challenged. By now, many
empirical examples of their interaction in both historical and
more recent times have been noted and investigated. At a more
theoretical level, there are also the current transdisciplinary critiques of the West-centered binary dichotomizing between primitive/civilized, non-Western/Western, traditional/modern, and,
alongside these, oral/literate, together with parallel challenges
to the arguably ethnocentric and ideological presuppositions of
a simple and necessary link between literacy and modernity. In
practice, it appears, there are multiple forms of literacy, interacting, therefore, in multiple ways with oral modes.
Despite challenges to some of its central presuppositions, the
legacy of the oral-formulaic school lives on. It rightly unsettled
the (literate) concept of fixed correct text, highlighted the significance of performance and audience, and, if in the (arguably)
somewhat elusive terminology of formulae, pointed up the
importance for composition of conventionalized verbal formulations in generic settings. Scholars identifying themselves with
that tradition have continued their (largely textual) examinations of oral and oral-derived texts while also reconfiguring
their approaches by attention to the specificities of aesthetic and
cultural traditions, interacting fruitfully with trends elsewhere to
produce sophisticated analyses of the complex interrelations of
oral with written composition (Amodio 2005; Foley 2002).
Although there is currently no one dominant approach to
complement the earlier oral theory, the topic of composition
without writing (or anyway, without central reliance on writing)
has continued to attract interdisciplinary interest. The focus is
now less on attempting to delineate oral composition as a single
process, or as pertaining to some special kind of culture or mentality, and more on complexity and plurality.
Oral composition is thus no longer conceptualized as primarily confined to traditional, historic, or non-Western settings
but as also including such examples as contemporary popular
songs or the spoken oratory of modern statesmen and publicists. It has also been noted how readily some long-established
oral genres are exploited in new settings, like the South African
praise poems now composed for Nelson Mandela, the national
football team, or university graduation ceremonies, and circulated not only in live performance but in writing and on radio,
CD-ROMs, and the Web. The relation between oral and literate is
now more often envisaged as continuum than as opposition or,
571
Oral Culture
investigations it has stimulated have enabled a firmer grasp on
the complexity of the processes by which, without much or any
direct recourse to writing, people can and do produce verbal
formulations both lengthy and short, aesthetically marked and
everyday. Further, all of this has helped to challenge traditional
models of language as realized preeminently either, on the one
hand, in stable written texts or, on the other, in relatively unconstrained and perhaps trivial everyday speech. A consideration of
oral composing highlights the sustained and creative marshaling of language in situations where writing does not necessarily
lie at the core: verbal genres that are by no means outdated or
peculiar but have had a wide spread in the world, both yesterday
and today.
Ruth Finnegan
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Amodio, Mark C., ed. 2005. New Directions in Oral Theory. Tempe: Arizona
Center for Medieval and Renaissance Studies.
Banti, G., and F. Giannatasio. 2004. Poetry. In A Companion to Linguistic
Anthropology, ed. Alessandro Duranti, 290320. Oxford: Blackwell.
Barber, Karin. 2007. Texts, Persons and Publics in Africa and Beyond.
Cambridge: Cambridge University Press.
Carruthers, Mary. 1990. The Book of Memory. Cambridge: Cambridge
University Press.
Finnegan, Ruth. 1992. Oral Poetry: Its Nature, Significance and Social
Context. 2d ed. Bloomington: Indiana University Press.
. 2007. The Oral and Beyond: Doing Things with Words in Africa.
Oxford: James Currey. Chicago: Chicago University Press.
Foley, John Miles. 1988. The Theory of Oral Composition: History and
Methodology. Bloomington: Indiana University Press.
. 2002. How to Read an Oral Poem. Urbana: University of Illinois
Press.
Kendon, Adam. 2004 Gesture: Visible Action as Utterance.
Cambridge: Cambridge University Press.
Lord, Albert B. 1960. The Singer of Tales. Cambridge: Harvard University
Press.
McNeill, David, ed. 2000. Language and Gesture. Cambridge: Cambridge
University Press.
Rubin, David C. 1995. Memory in Oral Traditions: The Cognitive
Psychology of Epic, Ballads, and Counting-Out Rhymes. Oxford: Oxford
University Press.
Silverstein, Michael, and Greg Urban, eds. 1996. Natural Histories of
Discourse. Chicago: Chicago University Press.
ORAL CULTURE
Oral culture is a conceptual construct associated primarily with
the work of Walter J. Ong, S.J., Marshall McLuhan, and Eric A.
Havelock, whereas the term oral tradition, which they also use,
is more often associated with the work of Milman Parry, Albert B.
Lord, and their many followers (see Foley 1985). Ong, McLuhan,
and Havelock use the term oral culture to refer primarily to preliterate cultures but also to characterize the thought and expression that carry over into manuscript culture and even into print
culture. Moreover, oral culture, which Ong also refers to as
primary oral culture, endures in the sense that people continue
to talk with one another. The later subsequent cultural developments in manuscript culture and print culture may be seen
as cultural overlays that influence and transform the base oral
572
Oral Culture
the linear accounts of time in the Bible with literacy; he likes
to style linear conceptions of history as evolutionary thought,
thereby rooting later forms of evolutionary thought in Darwin
and others within the biblical cultural tradition in Western tradition. Independently of Ong, Donald L. Fixico, who is himself
of American Indian descent, works comfortably with these contrasts in The American Indian Mind in a Linear World: American
Indian Studies and Traditional Knowledge (2003; also see Lee
1987, 10520).
In Manliness, Harvey C. Mansfield does not happen to refer
explicitly to oral culture, but he refers to Achilles frequently to
illustrate certain points regarding manliness (2006, 558, 601),
an ambivalent quality that he sees as needing to be disciplined
toward socially constructive ends. Male puberty rites, for example, have long been used in oral cultures to help discipline and
orient young men in socially constructive ways (see van Gennep
1960; Ong 1971, 11341). The kind of socially constructive warrior manliness that Achilles and Agamemnon and Hector and
Odysseus represent is a necessity in oral cultures: The entire
enterprise of modernity, however, could be understood as a project to keep manliness unemployed (Mansfield 2006, 230). In
David Riesmans terminology, oral culture is tradition directed,
whereas modernity is dominated by inner-directedness (1950).
(For further recent studies of the historical development of innerdirectedness, see Williams 1993; Brakke 2006; Cary 2000; van t
Spijker 2004; Renevey 2001; Low 2003; Connor 2006; Bloom 1998;
Kahler 1973). In Honor and the Epic Hero, Maurice B. McNamee
(1960) shows that concepts about heroic and great-spirited persons have shifted from time to time. Even though the concepts of
a heroic and magnanimous person in oral culture no longer work
for modernity, we do need to formulate some concepts of heroic
and magnanimous persons that will work for modernity.
Before concluding, we should note the critique that some
authors have made of Ongs work and related work regarding
oral culture. The critique alleges that Ong has set forth a great
divide theory in which there is a great divide with oral culture
when literacy emerges (see, for example, Daniell 1986; but also
see Ongs 1987 letter about her article). Beth Daniell and others
who advance this critique do not accurately summarize what
Ong has said, and so their supposed critique amounts to little
more than knocking down a straw man named by them Ong.
(For a more detailed response to this alleged line of critique, see
Farrell 2000, 1626, 2004).
In conclusion, in oral culture, people are culturally conditioned so that they tend to favor cyclic patterns of thought and
expression, to have a world-as-event sense of life, to put manliness to work in socially constructive ways, to use oral stories of
heroes as ways to help orient and put manliness to work, and to
use ritual process very effectively to promote and support socially
constructive behavior.
Thomas J. Farrell
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Ballew, Lynne. 1979. Straight and Circular: A Study of Imagery in Greek
Philosophy. Assen, the Netherlands: Van Gorcum.
Bloom, Harold. 1998. Shakespeare: The Invention of the Human. New
York: Riverhead Books.
Brakke, David. 2006. Demons and the Making of the Monk: Spiritual
Combat in Early Christianity. Cambridge: Harvard University Press.
Cary, Phillip. 2000. Augustines Invention of the Inner Self: The Legacy of a
Christian Platonist. New York: Oxford University Press.
Connor, James L. 2006. The Dynamism of Desire: Bernard J. F.
Lonergan, S.J., on The Spiritual Exercises of Saint Ignatius of Loyola.
St. Louis, MO: Institute of Jesuit Sources.
Daniell, Beth. 1986. Against the great leap theory of literacy. Pre/Text
7.3/4: 18193. Also see Ong 1987.
Draper, Jonathan A. 2004. Orality, Literacy, and Colonialism in Antiquity.
Leiden: Brill.
Draper, Jonathan A., ed. 2003. Orality, Literacy, and Colonialism in
Southern Africa. Leiden: Brill.
Eliade, Mircea. [1949] 2005. The Myth of the Eternal Return. 2d ed. Trans.
Willard R. Trask, new introduction by Jonathan Z. Smith. Princeton,
NJ: Princeton University Press.
Farrell, Thomas J. 2000. Walter Ongs Contributions to Cultural
Studies: The Phenomenology of the Word and I-Thou Communication.
Cresskill, NJ: Hampton.
Fixico, Donald L. 2003. The American Indian Mind in a Linear
World: American Indian Studies and Traditional Knowledge. New
York: Routledge.
Foley, John Miles. 1999. Homers Traditional Art. University
Park: Pennsylvania State University Press.
Foley, John Miles, ed. 1985. Oral-Formulaic Theory and Research: An
Introduction and Annotated Bibliography. New York: Garland.
Havelock, Eric A. 1963. Preface to Plato. Cambridge: Belknap Press/
Harvard University Press.
. 1978. The Greek Concept of Justice: From Its Shadow in Homer to Its
Substance in Plato. Cambridge: Harvard University Press.
. 1982. The Literate Revolution in Greece and Its Cultural
Consequences. Princeton, NJ: Princeton University Press.
Horsley, Richard A., Jonathan A. Draper, and John Miles Foley, eds.
2006. Performing the Gospel: Orality, Memory, and Mark. Minneapolis,
MN: Fortress.
Jousse, Marcel. 1990. The Oral Style. Trans. Edgard Sienaert and Richard
Whitaker. New York: Garland.
Kahler, Erich. 1973. The Inward Turn of Narrative. Trans. Richard
Winston and Clara Winston, foreword by Joseph Frank. Princeton,
NJ: Princeton University Press.
Kelber, Werner H. 1997. The Oral and the Written Gospel: The Hermeneutics
of Speaking and Writing in the Synoptic Tradition: Mark, Paul, and Q.
2d ed. Foreword by Walter J. Ong, S.J., new introduction by Werner H.
Kelber. Bloomington: Indiana University Press.
Lee, Dorothy. 1987. Freedom and Culture. Prospect Heights,
IL: Waveland.
Lord, Albert B. 1960. The Singer of Tales. Cambridge: Harvard University
Press.
Low, Anthony. 2003. Aspects of Subjectivity: Society and Individuality from
the Middle Ages to Shakespeare and Milton. Pittsburgh, PA: Duquesne
University Press.
Mansfield, Harvey C. 2006. Manliness. New Haven, CT: Yale University
Press.
McLuhan, Marshall. 1962. The Gutenberg Galaxy: The Making of
Typographic Man. Toronto: University of Toronto Press.
McNamee, Maurice B. 1960. Honor and the Epic Hero: A Study of the
Shifting Concept of Magnanimity in Philosophy and Epic Poetry.
New York: Holt, Rinehart and Winston.
Morris, Jan, and Barry Powell, eds. 1997. A New Companion to Homer.
Leiden: Brill.
Nightingale, Andrea. 2004. Spectacles of Truth in Classical Greek
Philosophy: Theoria in Its Cultural Context. Cambridge: Cambridge
University Press.
573
574
575
576
Origins of Language
words and sentences in abstraction from their ordinary uses. For
Wittgenstein, philosophys proper task is simply the assembling
of a series of reminders of actual uses, with the purpose of dispelling confusions that arise in specific contexts.
That conception of philosophy is one that Strawson shares up
to a point, but he also finds that an appropriate philosophy of language provides the basis for a descriptive metaphysics, one that
is content to give an account of the actual structure of the world
of our experience. This is set in contrast to a revisionary metaphysics that vainly strives to do better. Descartes, Leibniz, and
George Berkeley are revisionary; Kant and Aristotle are descriptive. That contrast in many ways mirrors the distinction between
ordinary language philosophy and those formalist attempts that
only mar whats well.
Clifford Brown
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Austin, J. L. 1962 . How to Do Things with Words. Oxford: Clarendon.
Brown, C. 2006. Peter Strawson. Montreal: McGill-Queens University
Press.
Quine, W. V. O. 1972. Methods of Logic. New York: Holt, Rinehart and
Winston.
Russell, B. 1905. On denoting. Mind 14.4: 47993.
Strawson, P. 1950. On referring. Mind 59: 2152.
. 1952. Introduction to Logical Theory. London: Methuen.
. 1954. Wittgensteins Philosophical Investigations. Mind, n.s.,
63.249: 7099.
. 1974. Subjects and Predicates in Logic and Grammar. London:
Methuen.
ORIGINS OF LANGUAGE
This term and language evolution are sometimes used interchangeably. Here, origins of language will mean the earliest
emergence of a system structurally distinct from the communication systems of other animals and having at least some of the
attributes of human language; the term will not include further
evolutionary developments leading to the emergence of fully
modern language or the subsequent diversification of human
languages (see historical linguistics). Although the topic
is one that has engaged the human imagination throughout history, an orgy of armchair speculation following Darwin caused
it to fall into disrepute. However, now that advances in a variety of sciences have made possible more informed (if still inevitably speculative) approaches, the attention devoted to it has
increased annually, with perhaps an overly ebullient proliferation of theories.
The issues may be more sharply defined by considering separately three major questions to which the topic gives rise:
(1) Was language directly selected for, or an emergent product of other faculties?
(2) If it was selected for, what pressure(s) selected for it?
(3) What form did its earliest emergence take?
Other issues involve the timing of language emergence, the
modality it originally employed, and whether or not language
evolved directly from prior means of animal communication.
Selective Pressures
Among those who see language as an adaptation, explanations
for the selective pressure involved have changed over time. Until
the 1980s, it was widely assumed that language arose for purposes of tool making and/or cooperative hunting. However, ecological studies revealed complex cooperative hunting patterns in
nonhuman species, while anthropological studies showed that
preliterate peoples made tools and taught tool making largely
without using words. Moreover, ethological studies of ape species showed highly complex societies in which individuals competed with and sought to deceive and outwit one another (Byrne
and Whiten 1992). An influential essay (Humphrey 1976) had
already suggested that higher cognitive faculties, including language, had most likely been generated through intense withingroup competition.
The view that language arose from social intelligence is
nowadays shared by a majority, but it has problems. Social
competitiveness is far from unique to humans; so why has no
form of language, however rudimentary, evolved in other primate species? A unique adaptation suggests a unique pressure. Furthermore, there must surely have been a stage when
577
Origins of Language
language was limited to a handful of symbols with which it
would have been impossible to express any socially significant
meaning. What, in such a situation, would have reinforced language use?
Advocates of some form of social adaptive pressure whether
for gossip (see grooming, gossip, and language), sexual
display, or social manipulation have so far failed to address
such problems adequately. An alternative proposal is that some
primitive form of language developed for exchanging information about food sources among small groups of extractive foragers (Bickerton 2002). Carcasses of megafauna, in particular,
would have required the rapid recruitment of significant numbers for efficient exploitation. Nobody doubts that language,
once it had emerged, would have been used for a variety of social
functions; such functions, in turn, would have expanded language. The real, and still unanswered, question is exactly what
led to its initial emergence.
The issue is rendered still more problematic by the fact that
words are cheap tokens (Zahavi 1975). Since they take so little
effort to produce, and since primate species constantly engage in
deception, why would anyone have believed them, and if no one
believed them, who would have persevered in their use?
Initial Structure
While some (as noted) believe that language has always possessed its present structure, most researchers would probably
agree that some simpler form developed first a stage generally
termed protolanguage (Bickerton 1990) and subsequently grew
more complex. Until recently, it was assumed that protolanguage,
like early-stage pidgins, consisted of a small quantity of units
(roughly equivalent in semantic coverage to modern words)
that could be concatenated, without any consistent grammatical
structure, to form brief propositions; in other words, protolanguage was compositional. This view is now challenged by
the proposal that protolanguage was synthetic, with holophrastic
units (like the units of animal communication systems) roughly
the semantic equivalents of complete propositions and not divisible into smaller meaningful units the whole thing means the
whole thing (Wray 2002, 118).
Defenders of a synthetic system note that (in contrast with
a compositional system) there would nowhere be any break in
continuity between language and the prelinguistic communication system of hominids (assumed to be similar to those of other
primates; see primate vocalizations), which it would at first
resemble except for productivity (holophrastic units could be
multiplied indefinitely). At a subsequent stage, chance phonetic
similarities between portions of holophrases would cause the latter to be reanalyzed into wordlike segments; these could then be
recombined to form a modern, compositional language.
It is claimed that a synthetic protolanguage would be less subject to ambiguities than a compositional one and would be better adapted for manipulation of other group members. Support
has come from computational linguists, many of whose
simulations of language evolution begin with units that represent propositional rather than lexical units (Briscoe 2002). Those
who, following Darwin and Otto Jespersen, assume a common
origin for language and music are more or less obliged to adopt
some form of the synthetic hypothesis.
578
Other Issues
A further controversy revolves around whether language was originally spoken or signed. Given that sign languages develop as
naturally among the deaf as do spoken ones among the hearing,
and that the hands of our closest primate relatives are more agile
and under more volitional control than their vocal organs, the
notion of a signed protolanguage is not unreasonable and has
been vigorously defended (Corballis 2002). However, even if the
original modality could be determined (and for all we know, protolanguage could originally have mixed signs and vocalizations
indiscriminately), this would not answer the questions discussed
here or tell us how language came to acquire the properties that
distinguish it from other modes of communication.
Another unresolved issue concerns the timing of emergence.
None of the evidence from the fossil record is unambiguous.
Endocasts of Homo habilis suggest a developed brocas area,
and this has been taken to indicate an early (~2.5 million years
ago) beginning for language. But since, even today, Brocas area
subserves both linguistic and nonlinguistic functions, we cannot know what functions it performed in antecedent species.
Symbolic artifacts are sometimes used to date language origins,
but while these indicate that language already existed, they cannot tell how long before their appearance it began. Absent reliable evidence, estimates of when language originated tend to
be determined by researchers positions on other issues. For
instance, those who believe that language emerged abruptly
more or less in its present state favor a recent date coincidental with the emergence of anatomically modern humans (~140
thousand years ago), or even later. Conversely, those who take
an adaptationist approach argue for a much earlier date, anything up to a few million years ago. The origin of language is
probably associated with some speciation event, but this issue,
like most others, is unlikely to be resolved without new sources
of evidence.
The question of continuity with prelinguistic systems is
somewhat clearer. That language evolved from some prior communicative system was, to Darwin, an article of faith, and some
subsequent authors have assumed that a commitment to gradual evolution entails such continuity, discounting the possible
capacity of mutations, changes in function, and interactions
between different faculties to produce evolutionary novelties.
But the only plausible continuist scenario is the holophrastic,
Overregularizations
synthetic model of Wray, discussed previously. If objections to
this are overcome, the case for continuity could be maintained;
otherwise, the differences between language or even protolanguage and any nonlinguistic system suggest a sharp discontinuity between the two.
Derek Bickerton
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bickerton, Derek. 1990. Language and Species. Chicago: University of
Chicago Press.
. 2002. Foraging versus social intelligence in the evolution of protolanguage. In The Transition to Language, ed. Alison Wray, 20726.
Oxford: Oxford University Press.
Bowie, Jill. 2006. The evolution of meaningful combinatoriality. Paper
presented at the Sixth International Conference on the Evolution of
Language, Rome.
Briscoe, Ted,
ed. 2002. Linguistic Evolution through Language
Acquisition: Formal and Computational Models. Cambridge: Cambridge
University Press.
Byrne, Frank, and Andrew Whiten. 1992. Cognitive evolution in primates. Man 27: 60927.
Corballis, Michael. 2002. From Hand to Mouth: The Origins of Language.
Princeton, NJ: Princeton University Press.
Donald, Merlin. 1991. Origins of the Modern Mind. Cambridge: Harvard
University Press.
Hauser, Marc, Noam Chomsky, and Tecumseh Fitch. 2002. The faculty of language: What is it, who has it, and how did it evolve? Science
298: 156979.
Humphrey, Nicholas K. 1976. The social function of intellect. In
Growing Points in Ethology, ed. P. P. G. Bateson and R. A. Hinde,
30317. Cambridge: Cambridge University Press.
Jenkins, Lyle. 2000. Biolinguistics: Exploring the Biology of Language.
Cambridge: Cambridge University Press.
Kirby, Simon, and Morton H. Christiansen, eds. 2003. Language Evolution.
Oxford: Oxford University Press. A collection of position papers by
leading scholars in the field.
Pinker, Steven, and Paul Bloom. 1990. Natural language and natural
selection. Behavioral and Brain Sciences 13: 70726.
Tallerman, Maggie. 2007. Did our ancestors speak a holistic protolanguage? Lingua 117: 579604.
Wray, Alison. 2002. Dual processing in protolanguage: Performance
without competence. In The Transition to Language, ed. A. Wray,
11337. Oxford: Oxford University Press.
Zahavi, Amotz. 1975. Mate selection a selection for a handicap.
Journal of Theoretical Biology 53: 20514.
OVERREGULARIZATIONS
Overregularizations like runned and mans have played a major
role in the language development literature for more than 40 years
(see also childrens grammatical errors; morphology,
acquisition of; syntax, acquisition of). Once brought
into focus, overregularizations comprised prime examples of
the way in which childrens use of general grammatical rule
knowledge (the regular past tense rule of adding -ed) could productively overwhelm the word-specific knowledge they gained
form actual input, a prime example of using rules to go beyond
the input. Less attention was paid to the way in which children
would get rid of overregularizations, but results indicated that
there is a verb-by-verb competition between the regular rule and
579
Overregularizations
Parable
580
P
PARABLE
Standard definitions, such as the one given in the Oxford English
Dictionary, conceive of parable as a literary term; it is said to be
the expression of one story through another. Literary historians
have modified this conception by placing limits on the kind of
story that counts as parable, attempting to distinguish it from, for
example, fable or allegory.
There are, however, even among literary scholars, some who
see parable as a much larger phenomenon, belonging not merely
to expression and not exclusively to historical genres but, rather,
as C. S. Lewis (1936, 44) observed, to mind in general. (See also
Louis MacNeices discussion of literary critical perspectives on
parable in MacNiece [1963] 1965, 5.)
For the language sciences, parable is not only, or even chiefly,
a kind of story; it is not an expression at all but, rather, a mental faculty that allows the human mind to integrate two conceptual stories or narratives into a third story, thereby creating a
conceptual blending network that has emergent meaning.
Straight history, or the observation of human interaction,
often can serve as the material for such parabolic blending. For
example, Sun Tzus The Art of War treats 13 aspects of warfare.
It has been studied in the West by military strategists since the
eighteenth century. Written in the sixth century b.c. in China, it
precedes by a couple of millennia the origin of modern business
management. But in the 1980s, it underwent extensive parabolic rendering in numerous books and articles for the purpose
of offering guidance to twentieth-century graduate students of
business and investment on how to conduct their professional
lives.
Parable frequently blends two stories that have strong conflicts in their content. It is a scientific riddle why human beings
should be able to activate two conflicting stories simultaneously,
given the evident risks of mental confusion, distraction, and
error. Yet, uniquely among species, human beings can evidently
not only activate fundamentally conflicting stories simultaneously and construct connections between them but also blend
them to create emergent meaning. This ability to blend two conceptual arrays with strong conflicts in their framing structure is
central to higher-order human cognition and is a hallmark of the
cognitively modern human mind. It is known as double-scope
blending (Fauconnier and Turner 2002).
Consider a parable from the Fourth Gospel. In John 10:1118,
Jesus presents Himself as the good shepherd, who lays down
His life for the sheep, in contrast to the hired hand, who does
not care for the sheep and flees in the face of the wolf. He says
the Father loves Him because He lays down his life and that no
one takes it from Him. Rather, He has the power to lay it down
Parable
Paralanguage
and take it up again. The clash between the story of the shepherd and the blend Jesus proposes is astonishing. It is quite
implausible that a shepherd would choose to die defending the
sheep, because then the sheep would be without a defender.
Yet this consequence is not projected to the blend: The actual
shepherd cannot return after being killed to look out for the
flock, but in Jesuss blend, He can. The emergent structure in the blend is crucial: Jesuss narrative blends dying
with physical manipulation of an object. (Physical manipulation is at the root of human understanding. See Chapter 4 of
Turner 1996, Actors Are Movers and Manipulators.) In the
story of manipulation, we can lay down an object and pick it
back up. Blending manipulation of a physical object with the
state of being alive or dead, Jesus achieves the remarkable
ability of self-revival.
As discussed in Chapter 4 (Analogy) of Cognitive Dimensions
of Social Science (Turner 2001), almost all the mental achievements analyzed by analogy theorists as analogy involve considerable unrecognized blending. In general, analogy involves
dynamic forging of mental spaces, construction of connections between them, and blending of the mental spaces to create
a conceptual integration network of spreading coherence, whose
final version contains a set of what are recognized, after the fact
in the rearview mirror, as systematic, even obvious analogical
connections. But those analogical connections are more often
the outcome of conceptual blending than its preconditions. Put
differently, what is commonly discussed as analogy manifests
the faculty for parable.
It is also important to recognize that a parable is not, in general, a conceptual metaphor for understanding one conceptual domain in terms of another. Consider 2 Samuel 12. The
prophet Nathan creates an elaborate blend in which a rich man
is blended with King David, a poor man is blended with Uriah
the Hittite, Uriahs wife Bathsheba is blended with a favored ewe
lamb, and there is a traveler who comes to dinner. The point of
the complex blend is that David has done wrong. The source
and target are complicated, drawing on many conceptual
domains, and the principal connection is that in both of them,
one man abuses another and deserves punishment. No general
conceptual metaphor provides this set of cross-space connections. Most of them are not metaphoric.
Parable as a form of literary expression might be of interest
to historians, anthropologists, and critics. But parable as a species-specific mental faculty that can activate, connect, and blend
sharply conflicting stories to produce new emergent meaning is
a far larger and more fundamental topic, posing one of the central riddles of the cognitive and language sciences.
Mark Turner
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Fauconnier, Gilles, and Mark Turner. 2002. The Way We Think: Conceptual
Blending and the Minds Hidden Complexities. New York: Basic
Books.
Lewis, C. S. 1936. The Allegory of Love. Oxford: Oxford University Press.
MacNeice, Louis. [1963] 1965. The Varieties of Parable [The Clark
Lectures]. Cambridge: Cambridge University Press.
Turner, Mark. 1996. The Literary Mind: The Origins of Thought and
Language. New York: Oxford University Press.
PARALANGUAGE
Nuances, connotations, and innuendos, which are integral characteristics of verbal communication, are given the vague term
paralanguage. These meanings arise from sources both within
and outside of standard linguistic structure. Linguistic elements
words, word order, semantics, grammar can be utilized
for paralinguistic communication. These combine with variations
of speech melody in ways that often defy structural description.
Paralanguage (as the term implies) both draws on and lies over
the known and describable ortholinguistic levels of phonetics,
phonology, morphology and lexicon, syntax, and semantics. All of these elements can be harnessed for paralinguistic
communication, as is well known from baby talk, from connotational meaning differences in terms such as skinny, slim, slender,
and from word-order choices such as Herman Melvilles That
inscrutable thing is chiefly what I hate. In addition, emotion,
attitude, intention, mood, psychological state, personality,
and personal identity can be communicated without referring
to words. Because of the power of the intonational contribution to paralanguage, the notion of two channels in the speech
signal has been invoked, but their intimate interplay has been
emphasized (Bolinger 1964). Words can communicate emotions,
but when prosody does so using a different channel, the paralinguistic intent overrides the ortholinguistic content, as in Im not
angry! spoken with increased pitch, amplitude, and rate.
Much of paralanguage is carried over longer stretches of
utterance than the phonetic element or the single word. Lexical
and syntactic choices may interact with intonational features
with a cumulative effect. Formulaic and nonliteral expressions
may be called into play. It has come to my attention that your
stonewalling is holding up the works contains conventional and
metaphoric utterances that build to a message more fraught
with paralinguistic content than Ive learned that your hesitation is contributing to a delay. Although subtle contrasts can be
conveyed on short utterances (see Nine ways of saying yes in
Crystal 1995), paralanguage prefers a larger canvas. Repetition of
words (Shakespeares a little, little grave) may have a powerful
paralinguistic effect. Movement from low to high pitch across an
intonational unit displays surprise or amazement; temporal units
are stretched to express sadness or disappointment; increased
intensity signals aggression or thematic emphasis; voice quality
becomes creaky to communicate victimization or breathy to
signal excitement.
Prosody, a major vehicle of paralanguage, can be decomposed into measurable elements: timing, pitch, amplitude, and
voice quality. These measures combine into complex patterns,
such that associating acoustic cues with paralinguistic meanings
is far from straightforward. John didnt drive the car can be
intoned with sadness, happiness, fear, or disgust, and may enfold
attitudes such as incredulity, relief, perplexity, or amusement.
Contradiction or denial, and conversational presumptions, such
as sincerity and truthfulness, are carried by phrasal intonation.
Take a common paralinguistic trope, sarcasm (see irony), in
581
Paralanguage
Parameters
That was a good effort. We know sarcasm when we hear it, but
exactly what in the signal conveys that the speaker is intending
to communicate the opposite of the usual lexical meanings is difficult to specify. In one version, the sarcastic utterance utilizes
higher pitch and greater amplitude on the first word followed
by falling intonation, pharyngeal voice quality, tensed vocal
tract, and spread lips. While morphological, lexical, and syntactic meanings can be structurally analyzed using units, features,
and rules, paralinguistic meanings constitute a brew of unstable,
fleeting, and subjective qualities. These paralinguistic qualities
shade into one another, and they impinge on purely linguistic
uses of prosodic contrasts, as in question and statement intonation. The auditory-acoustic cues that comprise paralanguage are
graded, in that they are not perceptually allocated by the listener
into discrete, contrastive categories as are the acoustic signals
for phonetic and lexical elements. Using deft combinations and
placements of prosodic cues, a speaker can communicate more
or less fear, gradations of perplexity, and degrees of denial.
The development of the pragmatics of communication,
a branch of linguistics that studies language use in conversation
(see conversation analysis), jokes (see verbal humor),
and storytelling, has advanced understanding of paralanguage.
Communicative elements such as turn-taking, inference, and
theme (topic of the discourse), and how they are signaled by the
speaker and comprehended by the listener, are investigated. The
fields of prosody and pragmatics have provided another valuable
impetus for the productive study of paralanguage: investigation
of the communicative competence of right hemisphere language processing in humans. While it has long been known
that the left hemisphere modulates language processing,
studies of pragmatics and prosody indicate involvement of the
right hemisphere in processing emotions and attitudes, inference and theme. The notion of two channels, ortholinguistic and
paralinguistic, is supported by the model that allocates processing to left and right hemispheres, respectively. Paralinguistic
nuances are intimately woven into the propositional message, so
much so that synthesized speech is often judged as unpleasant.
A goal of speech synthesis is to produce more natural-sounding
speech, which means infusing paralanguage, a challenging but
worthy goal.
Diana Van Lancker Sidtis
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bolinger, Dwight. 1964. Around the edge of language: Intonation. In
Intonation, ed. D. Bolinger. Harmondsworth, UK: Penguin Books.
Crystal, David. 1995. Nine ways of saying yes. The Cambridge
Encyclopedia of the English Language, ed. David Crystal, 248.
Cambridge: Cambridge University Press.
Kreiman, Jody, Diana Van Lancker Sidtis, and Bruce Gerratt. 2005.
Perception of voice quality. In Handbook of Speech Perception, ed.
David Pisoni and Robert Remez, 33862. Maldon, MA: Blackwell.
Van Lancker Sidtis, Diana. 2007. The relation of human language to
human emotion. In Handbook of the Neuroscience of Language,
ed. Brigitte Stemmer and Harry Whitaker. San Diego, CA: Academic
Press.
Williams, C., and K. Stevens. 1972. Emotions and speech: Some
acoustical correlates. Journal of the Acoustical Society of America
52.4B: 123850.
582
PARAMETERS
The term parameter is used in linguistics on analogy with its
usage in mathematics and engineering. In mathematics, the
parameters of a function are those aspects of the function that
are held constant when defining a particular function, but which
can vary in a larger context so as to characterize a family of similar functions. For example, the function for a line in analytic
geometry is f(x) = mx + b, with x the variable and m and b parameters of the function (the slope and the y-intercept). In the definition of any one line, the parameters m and b are held constant,
while the value of x varies, giving the different points on the same
line. In a broader context, however, the parameters m and b can
vary so as to define a family of similar functions: the set of all
lines. In the same way, parameters in linguistics are properties
of a grammatical system that are held constant when characterizing one particular human language, but which are allowed to
take different values in a broader context so as to characterize
a whole family of possible human languages. The idea that the
observed variation in human languages can be understood as
the fixing of certain parameters within an otherwise innate and
invariant system of principles (universal grammar) is most
commonly associated with the Chomskyan approach to formal
generative linguistics (see generative grammar). As a result,
this approach is sometimes called the principles and parameters theory. The idea is, however, a very general one, and it
can also be used in the context of other views about the nature of
the human language faculty.
This notion of linguistic parameters was introduced into linguistic theory by Chomsky (1981) and Rizzi (1982), during the
government and binding period. The paradigmatic case
was the pro-drop parameter (or null subject parameter). It was
observed that languages like Spanish and Italian differ from
French and English in several ways that appear to be interrelated. First, Spanish and Italian allow the subject pronoun of a
finite clause to be omitted, whereas French and English do not:
(1)
Second, Spanish and Italian allow the subject to come after the
verb as well as before it, whereas French and English generally
do not:
(2)
a. Chi credi
che verr? (Italian)
Who you-think that will-come
b. *Who do you think that came? (English)
Although these are clearly three distinct properties of the languages in question, they have a common theme: Informally put,
French and English require that there be an overt noun phrase
in the canonical subject position immediately before the finite
Parameters
verb, whereas Spanish and Italian do not. This difference in the
syntax of subjects was also related to a morphological difference: The agreement morphology on the finite verb is rich
enough to uniquely identify which pronoun would be in the
subject position in Spanish and Italian, whereas in French and
English it is not. The universal syntactic condition, then, is that
finite clauses require subjects (the extended projection principle); the parameter concerns exactly what kind of subject is
necessary to fulfill this condition. In Italian and Spanish, the rich
agreement on the verb means that null or displaced subjects are
permissible because (roughly) much of the information concerning the sort of subject it was is locally available on the finite
verb. In French and English, the agreement on the verb is of little
help, and so an overt subject in the canonical subject position
is required. A parameter, then, is a way of attributing a unified
theoretical account of the systematic differences that distinguish
one class of languages from another.
While the pro-drop parameter was the first important parameter to be proposed, it is by now not considered the best case. A
look at a wider range of languages both nonstandard dialects
of the Romance languages and languages from other families
quickly showed that the properties in (1)(3) do not correlate
with one another as closely as was thought (Jaeggli and Safir
1989). This implies that the pro-drop parameter as it was originally conceived is either false or highly oversimplified.
That does not mean that the idea of a parameter was ill-conceived, however. The current paradigmatic example is what is
sometimes called the head directionality parameter (terminologies vary). This can be stated as an open factor in the principles
of phrase structure (see x-bar theory). Roughly put,
when a word-level category x merges with a phrase Y to create
a phrase of type X, there are two ways that the elements can be
ordered: The order can be X-Y within XP, or it can be Y-X. Setting
the parameter in the first way gives head-initial languages like
English, in which complementizers come before embedded
clauses, tense particles come before verb phrases, verbs come
before their objects, prepositions come before their objects, and
so on:
(4)
Setting the parameter in the second way gives head-final languages like Japanese, in which complementizers come after
embedded clauses, tense particles come after verb phrases, verbs
come after their objects, prepositions come after their objects,
and so on:
(5)
parameters that are intended to account for the large-scale differences among the major classes of languages discovered by
typology. The head directionality parameter is a parameter of
this sort. Another early example was Ken Hales (1983) nonconfigurationality parameter, which was designed to explain why
Australian languages like Warlpiri tolerate free word order and
discontinuous phrases, whereas languages like English do not.
Similarly, parameters have been proposed to capture the difference between ergative languages (like Basque and Eskimo), in
which the object of a transitive clause is treated in some respects
like the subject of an intransitive clause, and accusative languages (like English and most Indo-European languages), in
which all subjects are treated similarly. These proposals range
from radical differences in how syntactic structure is initially
constructed (Marantz 1984) to relatively minor differences in
how case and agreement morphology are assigned in a simple
sentence (Bittner and Hale 1996). Mark Baker (1996) proposes
a polysynthesis parameter that attempts to give a unified characterization of the difference between many native American languages, in which a large part of the expressive burden is placed
on verbal morphology, and languages like English, in which the
primary expressive burden is borne by syntactic combination.
Taken together, some set of parameters such as these might
characterize the major linguistic types we observe.
Other parameters operate on a smaller scale, defining the differences between historically-related languages or dialects.
The pro-drop parameter was a parameter of this sort, distinguishing French from Italian. Another example is the parameter that
determines whether the subject of a clause moves from its original position inside the verb phrase to the highest position in the
clause or not; this accounts for the difference between English,
which has subject-finite verb-object word order, and Celtic languages like Welsh, which have finite verb-subject-object word
order (Koopman and Sportiche 1991). Jean-Yves Pollock (1989)
argues that there is a parameter that says that verbs move to a
higher position in French than they do in English; this accounts
for a cluster of subtle word-order differences having to do with the
placement of verbs, negation, and adverbs in the two languages
(e.g., John kisses often Mary is normal French but bad English).
A third example is Jonathan Bobaljik and Dianne Jonass (1996)
proposal that some Germanic languages have an extra position available for subjects that other Germanic languages dont
have; this makes sentences like There have some trolls eaten
Christmas pudding possible in some Germanic languages but
not others, among other things. (See Baker 2001 for a general
overview of these parameters and several others.)
In the early days of parametric theory, it was thought that
virtually any syntactic principle could be parametrized, and
parameters were proposed that were relevant not only to X-bar
theory but also to movement, the theory of binding, and even
the projection principle. On that view, there would be a modest number of parameters (dozens or perhaps hundreds), each
of which would have a relatively large impact on the language
generated. But this view has been questioned in more recent
work. Hagit Borer (1984) proposed almost immediately that the
syntactic principles themselves are invariant, and what is parameterized is the features associated with individual lexical items.
Rather than saying that the syntax of French is different from
583
Parameters
Parietal Lobe
584
Jaeggli, Osvaldo, and Kenneth Safir. 1989. The null subject parameter
and parametric theory. In The Null Subject Parameter, ed. Osvaldo
Jaeggli and Kenneth Safir, 144. Dordrecht, the Netherlands: Kluwer.
Kayne, Richard. 2005. Some notes on comparative syntax, with special reference to English and French. In The Oxford Handbook of
Comparative Syntax, ed. Guglielmo Cinque and Richard Kayne, 369.
New York: Oxford University Press. A detailed discussion of general
considerations and very small-scale parameters.
Koopman, Hilda, and Dominique Sportiche. 1991. The position of subjects. Lingua 85: 21158.
Marantz, Alec. 1984. On the Nature of Grammatical Relations. Cambridge,
MA: MIT Press.
Pollock, Jean-Yves. 1989. Verb movement, universal grammar, and the
structure of IP. Linguistic Inquiry 20: 365424.
Rizzi, Luigi. 1982. Issues in Italian Syntax. Dordrecht, the
Netherlands: Foris.
PARIETAL LOBE
Anatomy
The parietal lobe is situated superior to the occipital lobe and
posterior to the frontal lobe. More specifically, it extends from the
central sulcus anteriorly, to the imaginary boundary of the parietal-occipital fissure posteriorly, to the sylvian fissure (perisylvian cortex) inferiorly. The parietal lobe(s) can be further
subdivided into three main areas. These include: 1) the somatosensory strip, also known as the postcentral gyrus (Brodmanns
area [BA] 1, 2, 3, 43), 2) the superior parietal lobule (BA 5), and 3)
inferior parietal lobule (includes BA 39-angular gyrus and 40-supramarginal gyrus). The latter two areas are separated by the
intraparietal sulcus (see Figure 1). Medially, the parietal lobe(s)
comprises the postcentral gyrus extension of the paracentral lobule, the precuneus, and part of the cingulate gyrus (see Figure 2).
Physiology
There are two parietal lobes, one in each hemisphere, which are
divided functionally on the basis of dominance. The dominant
lobe is typically the left one and the nondominant the right.
There are many different non-language functions performed
by the parietal lobe, for example, perception and localization of
touch, pressure, pain, and temperature on the opposite side of
the body, and visuospatial processing. The variety of languagerelated functions associated with the parietal lobe will be especially highlighted in the context of non-language functions.
The dominant parietal lobe is involved primarily in integrating sensory information to create a particular perception. The
inferior portion of this lobe, particularly the supramarginal gyrus
and angular gyrus, is involved in structuring information for
reading and writing (see writing and reading, neurobiology of), performing mathematical calculations, and perceiving
objects normally.
Damage to the dominant lobe can result in apraxia (motor
planning deficit) aphasia (language disorder), agnosia (abnormal
perception of objects), and sensory impairment (e.g., touch, pain).
Lesions to the inferior portion of the dominant lobe involving the
angular gyrus can result in Gerstmanns syndrome, which is characterized by leftright confusion, difficulty pointing to named
fingers (finger agnosia), impaired writing ability (agraphia), and
inability to perform mathematical calculations (acalculia).
Parietal Lobe
Figure 1.
Figure 2.
The nondominant parietal lobe, however, is involved in a different set of functions that are mostly non-language related. In
particular, this region is responsible for visuospatial functions as
it receives and integrates input from the visual system (occipital
lobe) to make sense of the spatial order of the world around us.
M. A. Eckert and colleagues (2005) found that Williams syndrome,
whose phenotype (visuospatial deficits) and genotype (deletion
on chromosome 7) are well characterized, is linked to superior
parietal impairment. Williams syndrome, thus, may provide a
valuable system for understanding parietal lobe function.
Damage to the right parietal lobe can result in a constellation
of deficits involving spatial and body relations. Bilateral lesions
may result in Balints syndrome, which affects both visual attention and motor skills. If both the parietal and temporal lobes are
damaged, memory impairments and personality changes may
result. Specifically, if this damage occurs on the dominant (left)
side, it may result in verbal memory deficits and difficulty in the
retrieval of strings of numbers. If the damage is on the right side,
it will affect nonverbal memory functions and will significantly
impair personality.
History
For more than a century, the exact role of the parietal lobe has
been debated by neuroanatomists and psychologists, with much
585
Parietal Lobe
and form, though it is noteworthy that differences have been
identified, such as larger parietal cortex, asymmetry of the lobes,
and more neural subdivisions in humans (Kolb and Whishaw
1990).
Language
Continuing the classical connectionist tradition of Hugo
Liepmann, Norman Geschwind (1965) championed the simplified yet controversial position that the parietal lobe acts as the
association area of association areas. Neural tissue damage to
this area often results in the classical disconnection syndromes,
for example, apraxia and others.
Aleksandr Romanovich Luria (1973) considered the parietal
cortex one piece in his two-part model of mental activity, stating that it was important for understanding reception, analysis,
and storage of information. Lesions to the left parietal lobe were
understood to result in afferent motor aphasias (difficulty in
finding the correct articulatory positions for specific phonemes),
particularly lesioned primary and secondary sensory areas
affecting speech motor control and lesioned tertiary sensory
area resulting in aphasia (the loss of speech production and/or
comprehension).
Recent research, such as that of Gregory Hickok (2000), suggests that the inferior parietal lobe serves as the connection
between phonological representations and motor control for
those representations, that is, the auditory-motor interface,
which is part of a larger network of interfaces and systems
subserving language function. Marco Catani, D. Jones, and
H. Dominic (2005) in a significant paper, confirm the analysis
that includes the inferior parietal lobe in the use and possibly
the acquisition of language via a new circuit connecting the
traditional language areas of broca and wernicke. It has
been labeled Geschwinds territory in honor of Geschwinds
original proposal that the parietal lobe is critical to language
function.
In sum, the left parietal cortex has particular areas that are
responsible for various linguistic functions. However, there are
other extralinguistic processes that the parietal lobe is known for
as well.
Extralinguistic Processes
ATTENTION. The function of the parietal lobe in attention mechanisms has been discussed over a period of time. Michael Posner
and Steven E. Peterson (1990) outlined the different subsystems
of attention: a) orientation to sensory events (not conscious), b)
signal detection for focal processing (conscious), and c) maintenance of a vigilant state (conscious). From the available neurocognitve evidence, the researchers assert that the posterior
parietal lobe plays an important role in attention mechanisms,
specifically in orientation and signal detection that are essential
for linguistic processing. Earlier, Luria (1973) identified this parietal region that mediates attention as an involuntary orienting
system. However the posterior parietal attentional mechanisms
are greatly impacted by the frontal regions that subserve alerting
mechanisms as well.
MEMORY. Traditionally, episodic memory, or declarative memory, has been attributed to the medial temporal lobe (MTL);
586
Parsing, Human
Parsing, Machine
PARSING, HUMAN
In general, parsing refers to breaking something into its
constituent parts. Thus, machines (see parsing, machine) and
humans can decompose a message (such as print or spoken language) into phrases, words, and morphemes. Most commonly
human parsing has been considered in the context of sentence
processing, particularly its syntactic and semantic aspects.
Language spoken, written, and signed can also be described
in terms of smaller functional units, including syllables, phonemes, features, and gestures.
An understanding of grammatical constraints has guided the
development of descriptive representations (e.g., sentence diagrams) and formal systems of language structure and use. Parsing
models also have been influenced by linguistic, psycholinguistic,
and cognitive theory and by techniques used in computational
linguistics, natural language processing, and speech recognition (Chomsky 1965; Bresnan 1982; Jurafsky and Martin 2000).
Representative approaches include linguistic, statistical, connectionist, and dynamical systems models (Charniak 1993; Hale
2006; McClelland and St. John 1989; Steedman 1999; Tabor and
Tanenhaus 1998; see also self-organizing systems).
The scale at which we can break the signal into pieces depends
upon both our attention to detail and to our descriptive goals, as
can be seen in numerous psychological studies that range from
ambiguity resolution (Frazier and Fodor 1978; Frazier 1987) to
assessment of our ability to perceive, produce, and use information at various levels of description. Parsing linguistic information
is not restricted to sound and print but can include a consideration of the gestures underlying the production of language by
voice (the coordinated movement of speech articulators, such as
the tongue body, tongue tip, jaw, and lips; see speech production) and sign (manual, facial, and body orientation) (Battison
1978; Browman and Goldstein 1990; Fowler and Brown 2000).
Philip Rubin
PARSING, MACHINE
The query Over which strait in North Wales did Thomas Telford
build a suspension bridge? illustrates the fact that natural languages have complex syntactic structures. Comparison of the
question with the answer He built a suspension bridge over the
Menai Strait reveals that the phrases including strait, occur in
different positions in the two utterances, and that the verb positions are quite different, leading linguists to propose a constituent structure like (1) for the question:
(1)
S NP VP
Sq PP Sinv
Sinv Vaux NP PP
PP Preposition NP
NP Determiner N1
etc.
587
Passing Theories
group the words into phrases and phrases into clauses, according
to the grammar rules, keeping track of multiple possible analyses
because of the extensive ambiguity of natural languages. Topdown parsers, though perhaps less intuitive, are more frequently
used: They essentially work by attempting to generate the input
sentence.
Standard parsing algorithms for analyzing any artificial
language (such as programming languages) have been developed and can be used with any context-free grammar. Thus,
linguists can write the grammar rules: They do not need to be
programmers. But linguists grammars of natural languages
often make use of additional devices, such as agreement or
subcategory features (as with Sinv and Vaux in (1) and (2), denoting inverted sentences and auxiliary verbs). generative
grammars, therefore, usually augment constituent structure
with additional information: There may be labels to uniquely
identify individuals or additional levels of information, such as
meanings. Work on feature-based frameworks such as lexicalfunctional grammar and head-driven phrase structure grammar has gone hand in hand with the development
of complementary parsing methods.
John Coleman
SUGGESTION FOR FURTHER READING
Jurafsky, Daniel, and James H. Martin. 2000. Speech and Language
Processing. Upper Saddle River, NJ: Prentice Hall.
PASSING THEORIES
Passing theories are utterance-specific formal semantic theories; they specify the correct interpretation, or literal meaning,
of particular linguistic utterances: sentences uttered by particular speakers at particular times. The expression passing
theory was coined by Donald Davidson in his 1986 paper A
Nice Derangement of Epitaphs, which was part of his attack on
accounts of linguistic communication essentially involving conventionally determined, shared meanings ([1986] 2005).
According to Davidson, expressions like language, meaning,
or sentence are theoretical terms used for describing, or explaining, successful linguistic communication (cf. [1992] 2001, 108f).
For communicative success, regular or conventional use of
words is not necessary; what is necessary is only that the hearer
understand what the speaker intends to mean. For instance, if by
the words a nice derangement of epitaphs the speaker intends
to mean a nice arrangement of epithets and the hearer understands that, we have a case of successful linguistic communication. Davidson suggests characterizing communicative success
in terms of the semantic intentions of the speaker. These he construes as intentions to be interpreted in a particular way on a particular occasion and by a particular hearer. Moreover, they are of
a Gricean, self-referential form (see communicative intention): A semantic intention is an intention to achieve the end
of being interpreted in a certain way by means of the intentions
being recognized by the hearer (Davidson [1986] 2005, 92 f).
Any utterance is made with a number of intentions that can be
ordered in terms of means to ends; the first intention in such
a sequence (as ordered by in order to) specifies its literal, or
588
Performance
for Davidsons account of successful linguistic communication
in terms of the complicated semantic intentions of the speaker.
Kathrin Gler
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bar-On, Dorit, and M. Risjord. 1992. Is there such a thing as a language?
Canadian Journal of Philosophy 22: 16390.
Davidson, Donald. [1982] 1984. Communication and convention.
Inquiries into Truth and Interpretation, 26580. Oxford: Clarendon.
. [1986] 2005. A nice derangement of epitaphs. In Truth, Language,
and History, 89108. Oxford: Clarendon.
. [1992] 2001. The second person. In Subjective, Intersubjective,
Objective: 10722. Oxford: Clarendon.
. [1994] 2005. The social aspect of language. In Truth, Language,
and History, 10926. Oxford: Clarendon.
Dummett, Michael. A nice derangement of epitaphs: Some comments
on Davidson and Hacking. In Truth and Interpretation: Perspectives
on the Philosophy of Donald Davidson, ed. E. Lepore, 45976.
Oxford: Blackwell.
Gler, Kathrin. 2001. Dreams and nightmares: Conventions, Norms,
and Meaning in Davidsons Philosophy of Language. In Interpreting
Davidson, ed. P. Kotatko, P. Pagin, and G. Segal, 5374. Stanford,
CA: CSLI.
Pietroski, Paul. 1994. A defense of derangement. Canadian Journal of
Philosophy 24: 95118.
PERFORMANCE
The study of performance investigates communicative practices
in their sociocultural contexts from three perspectives. First, it
foregrounds the performativity of communicative forms and
practices as modes of action or means of accomplishing social
ends. Second, it directs attention to the poetics of communicative practice or to the forms of verbal artistry through which communicative acts are crafted and communicative skill is displayed.
Third, it focuses attention on performances as a special class of
events, such as rituals, spectacles, festivals, or fairs, in which a
societys symbols and values are publicly displayed, interpreted,
and transformed. Within language study, the first and second
perspectives have been foregrounded.
The contemporary focus on the poetics and performance of
communicative practice emerged in the subdiscipline of linguistic anthropology from a line of inquiry called the ethnography of
speaking. Developed by Dell Hymes and his students during the
1960s and 1970s, the ethnography of speaking highlights performance in two linked ways: as speaking practice and as artfully
marked ways of speaking (Bauman and Sherzer 1975). Its centerpiece is what Hymes called the speech event or communicative event, a framework that allowed scholars to analyze multiple
components of language in use, including setting, participants,
ends (goals, purposes), act sequences, key (tone, tenor), instrumentalities (channel, code), norms, and genres (the SPEAKING
acronym provides a mnemonic) (Hymes 1967). The interest was
not simply in cataloging these components but, rather, in understanding how speakers use language within the conduct of social
life. In highlighting the emergent and creative nature of speech
performance, the ethnography of speaking focused attention
on linguistic forms as resources for living, in Kenneth Burkes
([1941] 1973) sense. Further, it proposed a new unit of study, the
589
Performance
Richard Bauman highlights the dimensions of risk, responsibility, and accountability in what has become a classic definition of performance: Performance as a mode of spoken verbal
communication consists in the assumption of responsibility to
an audience for a display of communicative competence (1977,
11). Inspired by Hymes, Bauman has been particularly interested in the forms of verbal artistry through which communicative skill is put on display. His work generated a pivotal shift in
folklore studies from a classificatory concern with texts independent of their contexts of use to an interest in the performance of
verbal art as a constitutive ingredient of social life. Performance
in this sense may range from sustained, full performance to a
fleeting breakthrough into performance, with hedged or negotiated performance lying somewhere in between (Bauman 2004,
110; the phrase breakthrough into performance comes from
Hymes 1981). Both Bauman and the interactional sociologist
Erving Goffman have been interested in how performances are
framed or keyed, but whereas Goffmans approach is dramaturgical, highlighting how social actors move from back stage
regions to perform the face work associated with an array of
social roles (Goffman 1959), Baumans interest lies in poetics,
voice, and genre as verbal resources for the accomplishment of
social ends.
Thus far, performance has been considered from two related
vantage points, each grounded in particular disciplinary perspectives: Performance as speaking practice has been a focus
of linguistic anthropology and sociolinguistics; performance
as verbal art has been highlighted in folkloristics and linguistic
anthropology. A third approach views performance as a special
class of marked events in which a societys symbols are displayed for commentary, interpretation, or transformation. This
approach, pioneered by Victor Turner (1967, 1969), is less concerned with language per se. Through its focus on collective representations, cultural symbolism, and collective effervescence,
or communitas, it is located in a Durkheimian paradigm, with
inspiration from Arnold Van Genneps work on rites of passage.
Increasingly, however, scholars are drawing on aspects of all
three approaches. One example of how the three approaches
may be productively considered together is Jane Goodmans
analysis of a childrens performance in the Kabyle Berber region
of Algeria (2005).
The performance in question took place at a wedding,
understood as a festive occasion in which villagers suspended
interpersonal or political conflicts and came together to collectively celebrate the new union. The wedding was set apart from
everyday life by various formal markers: location (an outdoor
public square), timing (late evening), dress, music (traditional
band), and activities (dance). Special forms of verbal art also
marked the occasion: A hired poet recited a poem after henna
was applied to the grooms hand; older village women sang
traditional songs to mark transitions. Wedding guests danced
to show support for the new couple. In this village, men and
women shared the same dancing space but typically danced
sequentially rather than concurrently; in no case did they
dance as couples. One summer, however, village youth active
in the national Berber Cultural Movement formed a mixedgender childrens chorus as a way of changing gender relations
in the community and, more broadly, fostering a commitment
590
591
592
Perisylvian Cortex
ethnomethodological paradigm of conversation analysis.
Consider the following exchange:
S: Another glass of wine would hit the spot.
H: I dont think so mate, youve had enough.
Ss utterance, despite being a declarative, is clearly not interpreted as simply stating a fact by H; rather, Hs response (a
refusal to comply) shows that it was interpreted as a request
for another glass of wine. The basis for interpretation here lies
in the conversational sequencing of the two contributions: They
are conditionally reliant upon each other, by virtue of being two
parts of a request-refusal adjacency pair. The question as to
the intended illocutionary force of Ss turn becomes moot in this
approach; what matters is that H has clearly interpreted it as
request-like, having provided an appropriate second part to the
adjacency pair. Of course, H might provide an incorrect interpretation, but if this is the case, it will become appararent in the subsequent interaction. Such an inductive approach avoids some of
the pitfalls inherent in attempts to classify speech-acts according
to the nonobservable, and therefore unfalsifiable, intentions of
the speaker.
Ronald Geluykens
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Austin, J. L. 1962. How to Do Things with Words. Oxford: Oxford University
Press.
Levinson, S. C. 1983. Pragmatics. Cambridge: Cambridge University
Press.
Searle, J. R. 1969. Speech Acts: An Essay in the Philosophy of Language.
Cambridge: Cambridge University Press.
. 1979. Expression and Meaning: Studies in the Theory of Speech Acts.
Cambridge: Cambridge University Press.
PERISYLVIAN CORTEX
Beginning in the late nineteenth century, the application of deficit-lesion correlations based on autopsy material to the problem
of the regional specialization of the brain for language yielded
the fact that human language requires parts of the association
cortex in the lateral portion of one cerebral hemisphere (Broca
Brocas area
593
Perisylvian Cortex
Despite its widespread clinical use, however, this model has
serious limitations. It deals only with words, not other levels of
the language code. From a linguistic and psycholinguistic
point of view, the syndromes are all composed of many processing deficits, which are different in different patients. The syndromes themselves do not provide a guide to the localization of
more specific components of the language processing system.
As reviewed in the following, Geschwinds critical contribution regarding the role of the parietal lobe receives no empirical
support.
594
Perisylvian Cortex
Perlocution
Overview
The left perisylvian association cortex appears to be the most
important brain region supporting human language. However,
it is not the sole area involved in these abilities. How this area
and other brain regions act to support particular language operations is not yet understood. There is evidence for both localization of some functions in subparts of this region and other
brain areas, and for either multifocal or distributed involvement of brain areas in other language functions. It may be that
some higher-level principles are operative in this domain. For
instance, content-addressable activation and associative operations such as those that underlie phoneme recognition, lexical
access, and lexical semantic activation, may be invariantly
localized, while combinatorial computational operations such
as those that constitute the syntax of natural language may not
be. However, many aspects of these topics remain to be studied
with tools of modern cognitive neuroscience.
David Caplan
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Ben-Shachar, M., T. Hendler, I. Kahn, D. Ben-Bashat, and Y. Grodzinsky.
2003. The neural reality of syntactic transformations: Evidence from
fMRI. Psychology Science 14: 43340.
Ben-Shachar M., D. Palti, and Y. Grodzinsky. 2004. The neural correlates
of syntactic movement: Converging evidence from two fMRI experiments. Neuroimage 21: 132036.
Bornkessel I., C. Fiebach, and A. Friederici. 2005. On the cost of syntactic
ambiguity in human language comprehension: An individual differences approach. Cognitive Brain Research 21: 1121.
Broca, P. 1861. Remarques sur le sige de la facult du parole articul,
suivis dune observation daphmie (perte de parole). Bulletin de l
Socit dAnatomie (Paris) 36: 33057.
Broca P. 1865. Sur le sige de la facult du langage articul. Bulletin de
la Socit danthropologie 6: 33793.
Caplan, D., E. Chen, and G. Waters. 2008. Task-dependent and taskindependent neurovascular responses to syntactic processing. Cortex
44: 25775.
Caplan, D., G. DeDe, and J. Michaud. 2006. Task-independent and taskspecific syntactic deficits in aphasic comprehension. Aphasiology
20: 893920.
Caplan, D., G. Waters, G. DeDe, J. Michaud, and A. Reddy. 2007. A study
of syntactic processing in Aphasia I: Behavioral (psycholinguistic)
aspects. Brain and Language 101: 10350.
Caplan, D., G. Waters, D. Kennedy, N. Alpert, N. Makris, G. DeDe,
J. Michaud, and A. Reddy. 2007. A study of syntactic processing in
aphasia II: Neurological aspects. Brain and Language 101: 15177.
PERLOCUTION
In pragmatics, perlocution refers to the effect speech-acts
have on the hearer (H). J. L. Austin (1962) distinguishes three
types of act that utterances perform simultaneously: locution
(roughly equivalent to the meaning in a propositional sense),
illocution (the intended force of the speech-act), and perlocution. Austin characterizes perlocution as follows: Saying something will often, or even normally, produce certain consequential
effects upon the feelings, thoughts, or actions or actions of the
audience, or of the speaker, or of other persons: and it may be
done with the design, intention, or purpose of producing them
(1962, 101). Hs reaction to an illocutionary act might be verbal
(e.g., asking a question might prompt an answer), or nonverbal
595
Person
(e.g., an insult may result in a slap in the face), but also an internal
psychological or emotional state (e.g., a threat might result in H
being frightened or angry).
Although Austin intended perlocution to be an integral part
of a speech-act, later developments of speech-act theory have
focused almost exclusively on illocution, that is, the speakers
mental state or intention (e.g., Searle 1969). As a result, the term
speech-act has become virtually synonomous with illocutionary force. This is perhaps unsurprising, given that perlocutions
do not always consist of observable behavior (and might therefore
be argued to fall outside a linguistic theory of pragmatics; but see
Gu 1993). Moreover, perlocutions are hard to classify: Not only
do certain illocutions allow for a range of possible perlocutions
(a request, for instance, may result in either compliance or rejection by the hearer); there is often no way of knowing whether the
achieved perlocution is actually the one the speaker (S) intended
to achieve (a warning, say, may be intended to make the hearer
(H) take evasive action but may only result in frightening him/
her). Nevertheless, it is easy to demonstrate that perlocutions are
intrinsic parts of speech-acts since their successful performance
often depends on them. As Austin points out, an utterance such
as I bet you 10 dollars the Knicks will win by 5 points is felicitous
only if it receives uptake, that is, if H acknowledges and accepts
the bet (see felicity conditions).
conversation analysis (Sacks 1992) offers a potential
alternative, inductive approach to (verbal) perlocutions based on
local sequential organization. Consider the following exchange:
S: Have a cookie
H: ehm no thanks Ive just had dinner
In this exchange, Ss contribution can be labeled an offer by virtue of its being the first part of an offer-refusal (or offer-acceptance) adjacency pair. If H recognizes Ss utterance as such,
he/she will have to provide a sequentially appropriate response
(or perlocution). The second part is thus conditionally reliant on
the first part.
Ronald Geluykens
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Austin, J. L. 1962. How to Do Things with Words. Oxford: Clarendon
Press.
Gu, Yueguo. 1993. The impasse of perlocution. Journal of Pragmatics
20: 40532.
Sacks, Harvey. 1992. Lectures on Conversation. Oxford: Blackwell.
Searle, J. R. 1969. Speech Acts: An Essay in the Philosophy of Language.
Cambridge: Cambridge University Press.
PERSON
Person is a morphosyntactic property of nominal phrases (nouns
and pronouns) used to indicate the discourse role of their referent. English personal pronouns show three person distinctions: first person, indicating speakers (I, we); second person,
indicating addressees (you); and third person, indicating discourse nonparticipants (he, she, it, they). Some languages also
distinguish inclusive and exclusive we: Ojibwa has kiinawint for
groups including speakers and addressees, and niinawint for
596
Person
597
598
verbal tense and aspect. All these advances were authoritatively compiled by Dionysius Thrax in his Treatise on Grammar
(ca. 100 b.c.), which was so influential that it was often called
simply The Manual (and thereby probably subject to extensive
later revision by others). Some of this work, such as the eightfold
division of parts of speech and the treatment of Greek nominal
and verbal systems, still appears in twentieth-century textbooks.
In Rome, the Greek grammatical heritage was appropriated by
writers from Varro (On the Latin Language, first cent. b.c., only
partially preserved) down to Priscian (fifth to sixth cent. a.d.),
whose exhaustive Principles of Grammar (ca. 500), fortuitously
designed to assist the Greek speakers of the longer-lived eastern
Roman empire, would become the ultimate authority for learning Latin throughout medieval Europe. Although (as is so often
the case) much Latin grammatical theory slavishly followed
Greek models, it was impossible to ignore obvious differences
between the two languages (e.g., Latins lack of an article, one
past tense fewer, and one additional case).
Since the dominant unit of linguistic analysis of the time was
the word, and less so the sentence, the primary achievements
of classical language science lay in its descriptive and pragmatic
dimensions, particularly in linguistic pedagogy and the accurate
preservation, understanding, and annotation of written texts. For
instance, we have the Hellenistic era to thank for the invention of
such scholarly staples as footnotes, commentaries, critical editions, dictionaries, encyclopedias, and library catalogs. On the
other hand, investigations of phonetics and syntax, though
found in some early classical theorists, remained rudimentary.
And sadly, despite the story of King Mithridates of Pontus (or
Mithradates VI, 12063 b.c.), who was fluent in all 22 languages
of his subjects, there was almost no formal ethnographic study
of the many other now-extinct languages of the Mediterranean
region; non-Greek speakers were simply barbarians (barbaroi, babblers). Lexicographical work was driven by the need
to translate Greek and Latin, as well as to comprehend archaic
texts (e.g., Homer), and many word lists have been preserved as
hermeneumata translations and lists of glosses (glossaries,
from glossai, unfamiliar words). Although prodigious effort
from Socrates on was invested in etymology (the pursuit of a
words etymon, truth), this was almost a complete failure since
ancient philologists did not yet grasp how important phonology and rules of sound change are for tracing the historical roots
of words. The results ranged from the fanciful to the ridiculous.
Thus, Latin lignum wood hid potential ignis fire; lepus hare
was light-foot (compounding levis + pes); and words could
stem from their opposites: bellum war was so named for being
not at all bellum beautiful. Much of this dubious heritage was
compiled by Isidore of Seville (sixth to seventh cent. a.d.), whose
Etymologiae remained influential throughout the medieval
period. Many such classical and medieval compilations remain
secondarily valuable, however, because they often preserve the
sole remaining fragments of hundreds of ancient texts.
As a time of consolidation and preservation of the GrecoRoman heritage, the Middle Ages made relatively few significant
contributions to the study of language, as for many centuries
the Latin culture of Europe lagged behind the Greek learning of
the eastern Roman or Byzantine empire and the Arabic scholarship of Moorish Spain. Based on Varros lost writings on the
599
600
Phonetics
Another great paradigm split was marked by the publication
of the one-time philologist Ferdinand de Saussures Course in
General Linguistics (1916). Perhaps a victim of its own success,
diachronic philology, which so carefully traced the evolution of
parole, eventually yielded its disciplinary headship of language
study to Saussures synchronic langue (see synchrony and
diachonry and structuralism).
Presently partitioned among various university disciplines,
philology and hermeneutics still govern the fields of medieval
and classical studies, historical linguistics, literary theory and
criticism, textual editing, lexicography, prosody and metrics,
and many others (see Cerquiglini [1989] 1999; Gumbrecht 2003).
Today, though the normal science of language emphasizes
such synchronic contexts as society, psychology, and the brain,
there is little doubt that philology and hermeneutics will persist
and reappear, like Hermes and Mercury, in many new guises in
the future.
Christopher M. Kuipers
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Boeckh, August. [1886] 1968. On Interpretation and Criticism. Ed. and
trans. John Paul Pritchard. Norman: University of Oklahoma Press.
Cerquiglini, Bernard. [1989] 1999. In Praise of the Variant: A Critical
History of Philology. Trans. Betsy Wing. Baltimore: Johns Hopkins
University Press.
Gumbrecht, Hans Ulrich. 2003. The Powers of Philology: Dynamics of
Textual Scholarship. Urbana: University of Illinois Press.
Mueller-Volmer, Kurt, ed. 1985. The Hermeneutics Reader: Texts of the
German Tradition from the Enlightenment to the Present. New York:
Continuum.
Ormiston, Gayle L., and Alan D. Shrift, eds. 1990. The Hermeneutic
Tradition: From Ast to Ricoeur. Albany: State University of New York
Press.
Palmer, Richard E. 1969. Hermeneutics: Interpretation Theory in
Schleiermacher, Dilthey, Heidegger, and Gadamer. Evanston,
IL: Northwestern University Press. Still the standard introduction.
Robins, R. H. 1997. A Short History of Linguistics. 4th ed. London:
Longman.
PHONEME
The phoneme is the smallest unit of speech that discriminates
one word from another in a particular language. Phonemes
are represented by symbols between slashes thus /p/ or /b/.
Phonemes may have alternate forms, called allophones. For
example, in English, the same phoneme /p/ is produced differently in pit and spit. Minimal pairs are used to determine whether
two speech sounds are allophones or separate phonemes. For
example, in English, the phonemes /p/ and /b/ distinguish the
word pull from bull, and /t/ and /d/ distinguish the word bat
from bad.
Miwako Hisagi
PHONETICS
What Is Phonetics?
Phonetics is the area of language science research that studies
the articulation, acoustic properties, and auditory perception of
Phonetics
speech units (see speech production, acoustic phonetics, articulatory phonetics , and speech perception ,
respectively). More specifically, phonetics can be understood
as linguistically informed speech science, and research phoneticians are generally trained linguists who bring to bear their
knowledge of the structural properties of language. Rather than
focusing only on one particular language or on universal anatomical properties of hearing or articulation, a phonetician
has a special interest in understanding the full range of distinct possibilities in human speech or signed communication.
Because of the important role that linguistics plays in phonetic
study, most often phonetics finds itself housed academically
as a linguistic discipline, though sometimes it finds its home
in engineering, psychology, or in a language-specific setting.
The most prominent textbook used in educating phoneticians
is P. Ladefogeds A Course in Phonetics (2006), now in its fifth
edition.
Within linguistics, phonetics is related to the field of
phonology, another area of theoretical linguistic research.
Linguists vary in their opinions regarding the degree of distinctness and areas of overlap between the phenomena considered to be the objects of phonetic versus phonological
research. Both are concerned with the component speech units
or building blocks into which words can be divided. However,
the general view is that phonetics investigates measurable,
physical properties of these speech units, such as the precise
articulation of speech units, their detailed and contextually
dependent acoustic properties, and cross-linguistic variation
in these physical properties. Phonology, in contrast, is generally concerned with how these speech units are combined or
organized into acceptable word forms within a language (e.g.,
allowable sequences) and with the underlying principles of
organization shared across languages (see phonology, universals of ). On analogy to chemistry, phonetics investigates
subatomic structure, and phonology studies the formation of
molecules out of basic atoms. Traditionally, the phonological
structure has been viewed as cognitive or grammatical, while
the phonetic structure has been viewed as purely physical and
implementational. However, the dividing line between cognitive and physical has blurred or dissolved over the years (e.g.,
Browman and Goldstein 1995).
601
Phonetics
Figure 1. The IPA Chart. Reprinted with permission from the International Phonetic Association. Copyright 2005 by
International Phonetic Association.
602
height [high-mid-low], related to the lowest resonant frequency (the first formant) of the vowel; and
backness [front-central-back], related to the distance between
the first and second resonant frequencies (formants) of a
vowel.
Rounding or lip protrusion or compression is also encoded in
the symbol choice itself.
In addition, in order to adequately describe speech units, the
mechanism by which the air moves in the vocal tract must be
Phonetics
6000
Frequency (Hz)
5000
4000
3000
2000
1000
1450
Time (ms)
603
Phonetics
604
Neurobiological Underpinnings
The physiology (or function) of phonological processing is
described in terms of the structures (anatomy) activated in
processing and the function of these structures. A comprehensive understanding of the physiology requires explication
at the micro- and macrolevels of processing. The microlevel
describes the microstructures and their processing (neuron,
axon, synaptic potential) and are general to brain function,
whereas the macrolevel focuses on larger-scale structures and
processes specific to a particular motor, sensory, or cognitive
process (e.g., phonetic processing). Several points concerning
the microlevel are necessary for understanding how neurobiological methods are used to examine phonetics and phonology
(see Kandel, Schwartz, and Jessell 2000; Shafer and GarridoNag 2007, for greater detail).
First, brain function is in terms of electrochemical messages between neurons. neuroimaging methods index different aspects of these processes and the metabolic processes
that support these. Electrophysiological methods (electroencephalogram [EEG], magnetoencephalogram [MEG]) record
changes in electrical potential at the scalp. These changes are the
result of the synchronous firing of large assemblies of neurons.
Functional magnetic resonance imaging (fMRI) and positron
emission tomography (PET) measure changes in the metabolism of oxygen, and PET can also measure changes in the chemical aspect of the electrochemical signals sent between neurons.
These changes in electrochemical and metabolic measures are
used to make inferences about timing and localization of neural
activity related to some stimulus or event.
A second point is that different brain regions have distinctive
structure in terms of neurons and connectivity and that these
distinctions are the basis of Korbinian Brodmanns classification
system. For example, primary auditory cortex (Brodmanns area
[BA] 41) has a thick layer of neurons specialized to receive information from the peripheral auditory system. These neurons then
send signals to other cortical regions but not directly back to the
periphery for motor responses. Ultimately, phonological functioning will need to be described in terms of connectivity at this
neural level for a complete understanding of the brainbehavior
relationship.
At the macrolevel, neurobiology of phonetics/phonology is
described in terms of the activated brain regions and the timing
of activation of these regions in perception or production (see
speech perception and speech production). These brain
regions are referred to by Brodmanns areas, by names describing
function (e.g., primary auditory cortex), by the scientist involved
in identifying the regions (e.g., brocas area), or by some term
describing an attribute of the regions (e.g., Greek hippocampus
for a region that is shaped like a seahorse).
The principal brain structures involved in phonetic/phonological perception are found in the perisylvian cortex and
include primary (BA 41) and secondary (BA 42) auditory cortex
for processing the acoustic-phonetic aspects of speech (Scott and
Wise 2004) (see Color Plate 10). Sound in general (e.g., noise)
activates bilateral regions of the dorsal plane of the superior
temporal gyrus (STG) and regions of the lateral STG. In contrast
with noise, temporally complex signals, including speech, more
strongly activate the dorsal region of STG, and the lateral STG
activation extends more ventrally. Auditory information identified as speech compared to non-speech leads to increased activation of regions of the STG and superior temporal sulcus (STS)
that are more anterior and ventral (inferior). The left STS appears
to be active in mapping speech onto lexical-semantic representation. In contrast, the right STS shows sensitivity to melodic
features. The left planum temporale (PT, in superior posterior
temporal cortex) is believed to have a special role in phonetic/
phonological processing and appears to support a motor/sensory interface for acoustic information. A left-greater-than-right
asymmetry is generally stronger for speech than non-speech (see
605
606
Alternatively, listeners may have difficulty refocusing their attention to the relevant cues needed for rapid processing of the second language (Strange and Shafer 2008).
Neurophysiological data can address this question by
examining where in the nervous system differences in processing are found for first and second language learners. The current research has not shown differences earlier than the MMN
response. Furthermore, a recent study from our laboratory suggests that attention plays a role in loss of ability to learn novel
categories. Specifically, listeners learn to automatically attend to
relevant cues in their first language and can only overcome these
weightings with great attentional effort. This result suggests that
the loss of sensitivity in adjusting to novel phonological categories by second language learners is not directly due to a closure
of a critical period for changing the sensitivity or resolution of the
primary and auditory cortex; rather, it is due, at least in part, to
attentional issues (Hisagi 2007).
These findings do not answer all the questions regarding
critical and sensitive periods for setting up phonetic and phonological categories since second language learners acquired
categories for a first language early in life. Recent research examining the neurophysiological and behavioral consequences of
deprivation of hearing, which is reversed by cochlear implants,
will have much to contribute for addressing this question. Recent
advances have led to implantation at earlier ages, which is allowing researchers to compare the quality of phonological processing across different ages of first exposure to speech information.
Improvements in these implanted devices will also allow examination of how the quality of auditory-speech input impacts phonetic and phonological systems. This emerging area of research
is likely to provide less ambiguous evidence regarding a critical
or sensitive period for speech.
607
Phonological Awareness
Conclusion
This entry illustrated the importance of neurobiological data in
addressing significant questions concerning phonetic and phonological processing. In particular, an understanding of the neurobiology supporting phonetic and phonological processing will
allow researchers to construct better models of processing and to
address questions related to first and second language learning
and disorders (such as dyslexia and aphasia) attributable to
deficits in phonological processing.
Valerie Shafer
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Dehaene-Lambertz, G., and T. Gigla. 2004. Common neural basis for
phoneme processing in infants and adults. Journal of Cognitive
Neuroscience 16: 137587.
Hisagi, M. 2007. Perception of Japanese temporally-cued phonetic contrasts by Japanese and American English listeners: Behavioral and electrophysiological measures. Ph.D. diss., City University of New York.
Kandel, E. , J. Schwartz, and T. Jessell. 2000. Principles of Neural Science.
New York: William Heinemann and Harvard University Press.
Kujala, A., K. Alho, E. Service, R. J. Ilmoniemi, and J. F. Connolly. 2004.
Activation in the anterior left auditory cortex associated with phonological analysis of speech input: Localization of the phonological
mismatch negativity response with MEG. Cognitive Brain Research
21: 10613.
Magnuson, J. S., and H. C. Nusbaum. 2007. Acoustic differences, listener expectations, and the perceptual accommodation of talker variability. Journal of Experimental Psychology, Human Perception and
Performance 33: 391409.
Ntnen, Risto 2001. The perception of speech sounds by the human
brain as reflected by the mismatch negativity (MMN) and its magnetic
equivalent (MMNm). Psychophysiology 38: 121.
Poeppel, David, and Gregory Hickok. 2004. Towards a new functional
anatomy of language. Cognition 92: 112.
Scott, S., and R. Wise. 2004. The functional neuroanatomy of prelexical
processing in speech perception. Cognition 92: 1345.
Shafer, V. L., and K. Garrido-Nag. 2007. The neurodevelopmental bases
of language. In The Handbook of Language Development, ed. M. Shatz
and E. Hoff, 2145. Oxford: Blackwell.
Shafer, V. L., R. G. Schwartz, and D. Kurtzberg. 2004, Language-specific
memory traces of consonants in the brain. Cognitive Brain Research
18: 24254.
608
PHONOLOGICAL AWARENESS
Phonological awareness encompasses the broad class of abilities that enable one to attend to, isolate, identify, and manipulate
the speech sounds in spoken words. The domain of phonological awareness abilities can be subdivided into two levels. The
first, phonological sensitivity, pertains to conscious awareness of larger, more salient sound structures within words,
including rhymes and syllable structures (i.e., syllables and
subsyllabic units) (Scarborough and Brady 2002). (Rhymes,
defined at the word level, consist of the stressed vowel and
what follows [e.g., be/we; feather/weather]; subsyllabic units
include onsets, i.e., the portion of each syllable preceding the
vowel [e.g., be; spot; magnet], and rimes, i.e., the remaining portion [e.g., be; spot; magnet]). The second level of phonological
awareness, phoneme awareness, refers to explicit awareness of
the individual phonemes making up words. Generally, children
acquire at least some degree of phonological sensitivity prior
to phoneme awareness (see phonology, acquisition of ).
However, questions remain as to whether attainment of phonological sensitivity is a necessary prerequisite for the development of phoneme awareness (Gillon 2005). When children
begin to acquire phoneme awareness, they usually first are able
to isolate and identify the external phonemes (i.e., the beginning and/or final phonemes in words). Ultimately, proficiency
in phoneme awareness entails the ability to segment, identify,
and blend all of the individual phonemes, including those
within consonant clusters (e.g., in words such as blast).
The significance of phoneme awareness stems from its role
in reading acquisition (see writing and reading, acquisition of). Understanding that spoken words are made up of
individual speech sounds provides a conceptual foundation for
understanding the alphabetic principle (i.e., that letters correspond with phonemes). This awareness, in turn, facilitates
learning to read and spell. The relationship between phoneme
awareness and literacy development is reciprocal: With
some emergent awareness of phonemes, the student can start
to acquire lettersound knowledge. In turn, awareness of phonemes is heightened by experience with print.
Since the concept of phoneme awareness was established in
the 1970s (e.g., Liberman 1971), evidence for the significance of
phoneme awareness for reading achievement has accrued from
correlational, prediction, and training studies. At all ages, including adulthood, less-skilled readers demonstrate weaker performance on phoneme awareness measures than better-reading
peers, whether the same age or younger reading-age controls.
Prediction studies with kindergarten students document that
phoneme awareness performance is one of the strongest predictors of their subsequent reading achievement, particularly
for decoding and word recognition skills, but also for reading
Phonology
comprehension (see teaching reading). Most compelling,
intervention studies confirm a causal link between instruction in
phoneme awareness and increased success at learning to read,
with greater benefits when discovery of phonemes is linked with
letter knowledge (Ehri et al. 2001).
Susan A. Brady
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Ehri, L. C., S. R. Nunes, D. M. Willows, B. Schuster, Z. Yaghoub-Zadeh,
and T. Shanahan. 2001. Phonemic awareness instruction helps
children learn to read: Evidence from the National Reading Panels
meta-analysis. Reading Research Quarterly 36.3: 25087.
Gillon, G. 2005. Phonological Awareness: From Research to Practice.
New York: Guilford.
Liberman, I. Y. 1971. Basic research in speech and lateralization of
language: Some implications for reading disability. Bulletin of the
Orton Society 21: 7187.
Scarborough, H. S., and S. A. Brady. 2002. Toward a common terminology for talking about speech and reading: A glossary of the phon words
and some related terms. Journal of Literacy Research 34: 299334.
PHONOLOGY
As opposed to phonetics, which deals with the properties
of sounds from a language-independent point of view, phonology constitutes the study of the sound structure of units
(morphemes, words, phrases, utterances) within individual
languages. Its goal is to elucidate the system of distinctions in
sound that differentiate such units within a particular language,
and the range of realizations of a given units sound structure as
a function of the shape of other units in its context. These two
goals the study of invariants of sound structure and of the variation shown by these elements in combination are obviously
closely related, but attention has tended to shift between them
over time.
Late nineteenth- and early twentieth-century study of sound
structure focused on the details of sound production. As these
studies (in both articulatory phonetics and acoustic
phonetics) became more sophisticated, however, it was
increasingly apparent that the resulting explosion of data about
sound properties was obscuring, rather than enhancing, scholars understanding of the way sound is organized for linguistic
purposes. Much that is measurable in the speech signal is predictable, internal to the system of a given language, even though
exactly comparable properties may serve to distinguish items
from one another in a different language.
Vowels in English, for example, are relatively longer before
certain consonants than before others, but the difference in the
vowels of, for example, cod and cot is entirely predictable from
this principle alone. By contrast, an exactly parallel difference
between the vowels of kaade dip and kade envious in Finnish
serves as the sole difference between these words. A focus on
phonetic features alone fails to reveal the role played by sound
properties within a language.
The result of this insight was the development within various theories of structuralism of attempts (Anderson 1985)
to define the phoneme, a presumed minimal unit of contrast
within the sound system of a single language. While there is
609
Phonology
1970s, attempts to describe the phonology of tonal systems led
to important changes in assumptions about representations and
a concurrent shift of attention on the part of phonologists.
The classical theory had assumed that phonological (and
phonetic) representations were given in the form of a simple
matrix, where each row represented a phonological distinctive
feature and the columns represented successive segments. Such
a representation is based on the assumption that there is a oneto-one relation between the specifications for any given feature
and those for all other features, since each column contains
exactly one specification for each feature.
Tonal phenomena, however, made it clear that features need
not be synchronized in this way: A given feature specification
might take as its scope either more or less than a single segment.
A classic example of this, offered by W. Leben, is found in Mende,
where each word bears one of a limited set of tonal patterns,
regardless of the number of syllables on which this pattern is
realized. Thus, the tone pattern high-low appears on a single
syllable in mb (and thus the low has scope over only the last half
of the vowel), on two in ngl, and on three in flm (where the
single low of the pattern takes scope over two vowels). This led
to the development of autosegmental representations, in which
feature specifications were linked by lines of association (subject
to specific constraints), rather than all being aligned into segments. The extension of this insight to other phenomena, and its
consolidation, essentially displaced the earlier concerns of rule
notation and ordering in phonologists attention.
METRICAL PHONOLOGY. A similar development took place in
the analysis of stress and the study of the syllable. The analysis
in Chomsky and Halle (1968) treated stress as simply one more
phonological feature, with a value assigned to some (but not all)
of the segments in the representation of a word. This account was
forced to attribute a number of basic properties to the feature
[Stress], however, that had no obvious correlates in the behavior
of other features.
It became possible to rationalize these properties by viewing
stress not as a segmental feature but as a relational property of
the organization of syllables into larger structures. This, in turn,
required the recognition of syllables as significant structural
units: a notion that was explicitly rejected in the earlier theory
in favor of an attempt to reformulate all apparent syllable-based
generalizations in terms of segmental structure alone. The organization of segments into syllables and these, in turn, into larger
units called feet, which themselves are organized into phonological words (and phrases, etc.), allows for the elimination of the
anomalous character of segmentalized stress. The study within
metrical phonology of these units, their internal organization,
and their relation to one another completed the enrichment of
the notion of phonological representation begun within autosegmental phonology.
FEATURE GEOMETRY. A standard theme of classical generative
phonology was that of natural classes of phonological segments,
groups of segments that function in some parallel fashion in phonological rules to the exclusion of others. It was originally hoped
that the analysis of segments into distinctive features would provide the solution to this issue: Segments sharing a feature (or set
610
Phonology
of the next, and, from then on, no further elements from the initial stratum can be added. This process continues (perhaps vacuously) through all of the strata of the lexicon, yielding a potential
surface word. All of the words in a given syntactic structure are
then subject to adjustment by another set of postlexical phonological processes.
There are a number of further points that characterize this
view, including proposed differences in the properties of lexical
and postlexical rules and the relations between rules on one level
and those on the others. The central point for a broader theory
of grammar, however, is probably the replacement of a syntaxbased (but purely phonological) notion of cyclic rule application
by a repeated cycle of morphological addition and phonological adjustment. This results, for example, in the possibility that
a phonologically derived property (on one cycle) can be relevant
to the conditioning of a morphological operation (on a following
cycle), a possibility that has been shown to be quite real.
OPTIMALITY THEORY. In the early 1990s, a much more radical
challenge to the classical model was presented by the development of optimality theory (OT), a view of phonology based
on a system of ranked, violable constraints on surface shape,
as opposed to a system of ordered rules deriving the phonetics
from an underlying phonological representation. These constraints govern (in the standard formulation) a one-step relation
between underlying and surface representations (cf. underlying structure and surface structure), with no intermediate stages of the sort produced in a rule-based description. The
constraints can be divided into general classes: a) markedness
constraints, which express universally preferred configurations,
and b) faithfulness constraints, requiring that contrasts present
in the phonological representation be preserved in the surface
form. In general, these are in conflict, and the ranking of the constraints governs the resolution of those conflicts, in conformity
with general principles of grammar.
Initially, OT seemed to offer its greatest promise in the analysis of stress, syllable structure, and related phenomena, but subsequent development has encompassed a full range of segmental
and other facts. Descriptions in constraint-based terms are at
least superficially very different from those couched in terms of
traditional rules, and theoretical discussion in phonology since
their introduction has been largely dominated by comparisons
of the two frameworks.
example, when languages accommodate loan words to the surface patterns of other words of the language, the adjustments
needed to achieve this may include changes that do not correspond to any rule of the phonology of native forms. Constraints
accomplish this directly and without further stipulation, whereas
a system of rules may have to be arbitrarily extended to account
for loanword adaptation.
On the other hand, some of the same issues that rule-based
phonology dealt with (and at least largely resolved) have resurfaced as serious challenges to the architecture of grammar generally assumed in constraint-based theories. Most important
among these is the problem of opaque generalizations. The
standard model of OT assumes that its constraints apply directly
to surface forms and govern a single-stage mapping between
these and underlying phonological representations, and so
has no place for generalizations that crucially apply to any sort
of intermediate level. Nonetheless, a number of compelling
examples of such phenomena have been demonstrated, and
some sort of accommodation of these facts must be provided by
an adequate phonological theory.
Some responses to this challenge have attempted to maintain
the standard OT model by introducing new sorts of constraints.
Mechanisms such as output-output constraints or sympathy theory, however, have not generally succeeded in dealing with all of
the relevant phenomena and have been shown to produce new
difficulties of their own.
One approach that seems promising is that of stratal OT,
an architecture that grafts a constraint-based account onto the
standard model of lexical phonology. The result is a framework
in which the phonological mapping at each stage is a one-step
process governed by a constraint system. Since the model is
built on a cyclic interaction of phonology and morphology, however, it also provides for multiple successive stages in the overall
derivation, thus accommodating opacity to the extent it can be
related to morphological structure (as in the best-established
examples).
Examples also seem to exist in which the specific changes
through which a language achieves conformity with a general constraint on surface forms do not follow directly from
the content of the constraint (together with other interacting
generalizations). In such a case, something like a rewriting
rule might be necessary, as a supplement to the constraint
system a notion that is clearly antithetical to the basic philosophy of OT.
A quite different problem concerns the very nature of the
universals of phonological structure (see phonology, universals of). Phonological theorizing has generally accepted
the premise that generalizations that are true of phonological
systems in general result from the cognitive organization of the
human language faculty and, thus, must be incorporated in some
way into the architecture of phonological theory. Recently, however, it has been argued that at least some such typological regularities result not from the content of a universal grammar
constraining synchronic systems but, rather, from the universals
of language change (see language change, universals of)
governing the diachronic developments resulting in the systems
we observe. To the extent that this is true, it requires investigators to examine closely the arguments for incorporating any
611
Phonology, Acquisition of
particular regularity into phonological theory per se, as opposed
to seeking its basis elsewhere.
understand the system that relates childrens stored representations to their productions, and to formalize the developmental
paths that children follow.
Conclusion
While there have, of course, been other trends not covered here,
it seems fair to say that the bulk of the theoretical discussion in
phonology from the 1960s to the present has been devoted to
the elaboration and refinement of the generative program of
Chomsky and Halle (1968). The most recent developments in
that tradition, involving the wholesale replacement of rules by
constraints as the mechanism for expressing regularities of a languages sound pattern, have shown great promise but cannot yet
be considered wholly consolidated. Apparently, some appropriate synthesis of the classical and OT models remains to be found,
and it is that search that dominates discussion today.
Stephen R. Anderson
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Anderson, Stephen R. 1985. Phonology in the Twentieth Century: Theories
of Rules and Theories of Representations. Chicago: University of Chicago
Press. Describes the development of phonological theory, from its origins through the classical period of generative phonology.
Chomsky, Noam, and Morris Halle. 1968. The Sound Pattern of English.
New York: Harper and Row.
Gussenhovern, Carlos, and Haike Jacobs, 2005. Understanding Phonology.
2d ed. New York: Oxford University Press. Lucid elementary introduction to current phonology.
Kager, Ren 1999. Optimality Theory: A Textbook. Cambridge: Cambridge
University Press. Introduces the main ideas of optimality theory in
phonology and their implementation.
Kenstowicz, Michael 1994. Phonology in Generative Grammar.
Cambridge, MA: Blackwell. Provides a comprehensive description of
the principal themes in phonology up to the introduction of optimality
theory.
PHONOLOGY, ACQUISITION OF
A diversity of issues informs work in the field of phonological
acquisition, as it encompasses both first (L1) and second (L2)
language acquisition examined by researchers in linguistics,
psychology, speechlanguage pathology, and language education. In L1, there are questions such as how the acquisition
of phonology interfaces with perceptual and motor development (Locke 1993), and how an examination of disordered
development can illuminate the normally developing grammar (Bernhardt and Stemberger 1998; Dinnsen 1999). In L2,
there are questions as to whether the acquisition process is
fundamentally like L1 acquisition (Flege 1995), or whether L2
grammars are in some sense impaired due to, for example, L1
constraints that impede native-like attainment (see Brown 1998
on perception).
Due to space constraints, this entry focuses on L1, although
many of the same issues arise for L2. The acquisition of phonology is examined from the perspective of generative grammar;
thus, a principal theme is to examine how acquisition research
has used linguistic theory to inform development. This theme
considers the starting hypothesis to be that childrens productions are largely system driven: Acquisition research strives to
612
Markedness
Although children take different paths to the adult grammar, early
phonologies are also strikingly similar (Jakobson [1941] 1968).
As Roman Jakobson emphasizes, these similarities reflect crosslinguistically unmarked properties. markedness constrains the
shapes of linguistic systems such that less complex properties are
favored. For example, there is a well-documented preference for
consonant+vowel (CV) syllables among children (Ingram 1978;
cf. Grijzenhout and Joppen-Hellwig 2002); this is also a syllable
shape that no end-state grammar forbids (Jakobson 1962). Since
unmarked patterns are systematically observed across learners,
one might reasonably infer that they reflect early grammatical
organization. However, markedness has not always been well
integrated into the theory of grammar (as part of the theory of
representations or formulation of rules/constraints). This begs
the question of whether markedness should instead be part of
the theory of acquisition, which interfaces with, but is independent of, the theory of grammar.
Phonology, Acquisition of
Table 1.
Ambient
form:
Stage
Grammar:
Stored form:
Grammar:
[wei]
M >> F-perc, F
/wei/
M >> F-perc, F
F-perc, F >> M
[wei]
/wei/
Produced
form:
F-perc, F >> M
[wei]
[gep] escape
[an] banana
613
Phonology, Acquisition of
for developing and target systems. To capture truncation, Neil
Smith (1973) provides the rules below, neither of which operates
in the adult grammar:
(2)
R14 deletes initial vowels in words like escape. For consonantinitial forms like banana, the result is [bnan], which then undergoes R16, yielding [ban].
Since SPE employed linear representations, the theory did
not offer any insight into why pretonic rather than posttonic syllables delete (escape [gep], but tiger [aig], *[ai()]).
The development of nonlinear phonology (see Goldsmith 1985
for an overview), notably the move to highly articulated prosodic
representations, led to significant breakthroughs in understanding this asymmetry. In trochaic languages, where the foot (the
rhythmic unit in which stress is assigned) is left-headed (stressinitial), escape cannot form a single foot, [s(kip)Ft]Wd, whereas
tiger can, [(tig)Ft]Wd.
Much work in nonlinear phonology has explored the idea
that prosodically defined templates constrain output shape
(McCarthy and Prince 1995). Paula Fikkert (1994) proposes that
templates, which at early developmental stages reflect what
is unmarked, are responsible for truncation. If the childs productions are limited to one foot, circumscribed from the adult
output, this template will determine which material is preserved
from the adult form and which is deleted:
Wd
A dult output:
Ft
s
e
Child output:
keip
Ft
Wd
In contrast to SPE, nonlinear phonology reveals the relationship between target and truncated forms, and the role that
markedness plays in shaping outputs. The material inside the
foot survives, as syllables organized by feet ([keip]) are less
marked than those linking directly to the word ([s]). One problem with the templatic approach, however, is that it is too rigid: If
the segments predicted to survive are precisely those delimited
by the constituent that serves to organize them in the adult form,
it becomes difficult to capture the observation that material from
the truncated syllable can also survive. For example, in Amahls
pronunciation of banana in (1), onset selection favors [b], replacing [n] from the stressed syllable; that is, his production is [an]
not *[nan] as expected from adult [b(nan)Ft]Wd (see Kehoe
and Stoel-Gammon 1997 for other problems with the templatic
approach).
This problem is rectified in OT. First, there are no templates;
templatic effects arise from the interaction of markedness constraints. Second, segmental content (e.g., labial preservation)
is the responsibility of faithfulness constraints. Finally, all constraints are interranked; thus, the co-occurrence of truncation
and onset selection is not unexpected (see Pater 1997).
614
Table 2.
ParseSyll
a.
[b(nan)Ft]Wd
b.
[(nan)Ft]Wd
c. [(ban)Ft]Wd
Max[lab]-IO
Max-IO
*!
**
I-Contig
*!
**
To illustrate, concerning truncation in Table 2, the constraint ParseSyllable (syllables are parsed into feet), along
with other markedness constraints, must be satisfied at the
expense of the lower-ranked faithfulness constraint Max-IO
(every segment in the input has a correspondent in the output). Fully faithful (a) is thus eliminated because the initial syllable is unfooted. Concerning onset selection, Max[labial]-IO
(every [labial] in the input has a correspondent in the output) must be ranked over I-Contiguity (the portion of the
input standing in correspondence forms a contiguous string).
Preservation of [labial] in banana will thus be favored, (c),
even though the result violates I-Contig through morphemeinternal segment deletion.
OT has had a major impact on acquisition research.
Phonological processes are now generally expressed through
constraints, rather than rules, as this provides a better conceptualization of the observation that markedness shapes early
grammars. As discussed, childrens productions become more
target-like when markedness constraints are demoted below
faithfulness. A similar idea, that development is best viewed as
the gradual relaxing of constraints, had been proposed earlier
(Stampe 1969; Menn 1980), but it was difficult to formally implement it in the rule-based frameworks of the time.
OT seems to provide an appealing view of the initial state and
of development; researchers can address important questions,
such as how the theory may restrict what a possible developing grammar is, and how, in turn, data from development may
inform the theory. However, this is not to say that OT has solved
all problems in phonological acquisition. One understudied
problem is rogue behavior. We have been assuming that childrens grammars are possible grammars, thereby ignoring the
fact that some commonly attested processes, notably consonant harmony (CH), have no adult analogs (Drachman 1978). In
CH, consonants share place over vowels of any quality (Vihman
1978), as seen in (3) for Amahl, age 2.60 (Smith 1973):
(3)
[aig] tiger
[ok] stroke
Some recent accounts of CH (Goad 1997; Rose 2000) incorrectly predict that the process should be attested in adult grammars; others (Pater 1997) appeal to child-specific constraints,
thereby challenging the notion that childrens grammars are possible grammars. Neither of these approaches questions whether
CH is truly grammar-driven nor addresses, more generally, the
criteria that should factor into the determination concerning
what is grammar-driven and what is not. I leave these questions
to future work.
Heather Goad
Phonology, Acquisition of
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bernhardt, Barbara, and Joe Stemberger. 1998. Handbook of Phonological
Development from the Perspective of Constraint-Based Nonlinear
Phonology. San Diego, CA: Academic Press.
Best, Catherine, Gerald McRoberts, and Nomathemba Sithole. 1988.
Examination of perceptual reorganization for nonnative speech
contrasts: Zulu click discrimination by English-speaking adults and
infants. Journal of Experimental Psychology: Human Perception and
Performance 14: 34560.
Brown, Cindy. 1998. The role of the L1 grammar in the L2 acquisition of
segmental structure. Second Language Research 14: 13693.
Brown, Cindy, and John Matthews. 1997. The role of feature geometry
in the development of phonetic contrasts. In Focus on Phonological
Acquisition, ed. S. J. Hannahs and Martha Young-Scholten, 67112.
Amsterdam: Benjamins.
Chomsky, Noam, and Morris Halle. 1968. The Sound Pattern of English.
New York: Harper & Row.
Demuth, Katherine. 1995. Markedness and the development of prosodic
structure. Proceedings of NELS 25: 1325.
Dinnsen, Daniel. 1992. Variation in developing and fully developed phonetic inventories. In Phonological Development: Models, Research,
Implications, ed. Charles Ferguson, Lise Menn, and Carol StoelGammon, 191210. Timonium, MD: York.
. 1999. Some empirical and theoretical issues in disordered child
phonology. In Handbook of Child Language Acquisition, ed. William
Ritchie and Tej Bhatia, 647704. San Diego, CA: Academic Press.
Drachman, Gaberell. 1978. Child language and language change: A conjecture and some refutations. In Recent Developments in Historical
Phonology, ed. Jacek Fisiak, 12344. The Hague: Mouton.
Edwards, Mary Louise. 1974. Perception and production in child phonology: The testing of four hypotheses. Journal of Child Language
1: 20519.
Eimas, Peter, Einar Siqueland, Peter Jusczyk, and James Vigorito. 1971.
Speech perception in infants. Science 171: 3036.
Ferguson, Charles, and Carol Farwell. 1975. Words and sounds in early
language acquisition. Language 51: 41939.
Fikkert, Paula. 1994. On the Acquisition of Prosodic Structure. The
Hague: Holland Academic Graphics.
Flege, James. 1995. Second language speech learning: Theory,
findings, and problems. In Speech Perception and Linguistic
Experience: Theoretical and Methodological Issues, ed. Winifred
Strange, 23377. Timonium, MD: York.
Gnanadesikan, Amalia. [1995] 2004. Markedness and faithfulness
constraints in child phonology. In Constraints in Phonological
Development, ed. Ren Kager, Joe Pater, and Wim Zonneveld, 73108.
Cambridge: Cambridge University Press.
Goad, Heather. 1997. Consonant harmony in child language: An
optimality-theoretic account. In Focus on Phonological
Acquisition, ed. S. J. Hannahs and Martha Young-Scholten, 11342.
Amsterdam: Benjamins.
Goad, Heather, and Yvan Rose. 2004. Input elaboration, head faithfulness and evidence for representation in the acquisition of leftedge clusters in West Germanic. In Constraints in Phonological
Development, ed. Ren Kager, Joe Pater, and Wim Zonneveld, 10957.
Cambridge: Cambridge University Press.
Goldsmith, John, ed. 1995. The Handbook of Phonological Theory.
Oxford: Blackwell.
Grijzenhout, Janet, and Sandra Joppen-Hellwig. 2002. The lack of onsets
in German child phonology. In The Process of Language Acquisition,
ed. Ingeborg Lasser, 31939. Frankfurt am Main: Peter Lang.
Hale, Mark, and Charles Reiss. 1998. Formal and empirical arguments
concerning phonological acquisition. Linguistic Inquiry 29: 65683.
615
Phonology, Evolution of
. 1996. Phonological Development: The Origins of Language in the
Child. Oxford: Blackwell.
Werker, Janet, John Gilbert, Keith Humphrey, and Richard Tees. 1981.
Developmental aspects of cross-language speech perception. Child
Development 52: 34953.
Werker, Janet, and Richard Tees. 1984. Cross-language speech perception: Evidence for perceptual reorganization during the first year of
life. Infant Behavior and Development 7: 4963.
White, Lydia. 1982. Grammatical Theory and Language Acquisition.
Dordrecht, the Netherlands: Foris.
PHONOLOGY, EVOLUTION OF
True Phonology: How Did It Evolve?
phonology is the study of how languages use segmental and
prosodic categories to build spoken words and signal their differences in meaning. Many animals that communicate vocally
use distinct sound patterns to signal different meanings, but
their repertoires are typically small and closed sets. By contrast,
humans have large vocabularies and learn new words all of
their lives. From infancy to adolescence, children acquire lexical entries at a remarkably fast rate. The vocabulary size of high
school students has been conservatively estimated at 60,000 root
forms a number that implies an average acquisition rate of
more than 10 words per day (Miller 1991).
This difference is linked to the uniquely human method of
coding information: the combinatorial use of discrete entities.
Combinatorial structure, the hallmark of true language, creates
the conditions for open-ended lexical and syntactic systems
that provide the foundation for the singular expressive power of
human languages. How did it evolve?
We focus on two areas of empirical research. One is the study
of human cognition. The other is the investigation of the phonetic
signal space from which all phonological patterns are drawn.
The first theme highlights mans rich semantic abilities. The
second looks for phenomena that presage combinatorial sound
structure.
Cognitive Growth
The virtually infinite set of meanings encodable by language
raises the question of how mans cognitive capacity evolved
from skills not unlike those of present-day apes. How do we picture the transition from a nonhuman to a human primate mind?
What was the role of language?
According to Merlin Donalds synthesis of neurobiological, psychological, archeological, and anthropological evidence
(1991), our ancestors broke away from the stimulus-driven
behavior of apes in two steps. First, during the period of Homo
erectus (from 1.5 million years ago), an adaptation called mimesis
occurred, a communicative culture allowing individuals to share
mental states and begin to represent reality in new and expanding ways.
Mimetic behavior is an ability to voluntarily access and
retrieve motor memories and to rehearse and model them for
communication with others. The whole body is used as a representational device, as in imitating vocal, manual, and postural
movements for a communicative purpose. Mimesis involved
major changes of motor and memory mechanisms based on
existing capacities.
616
Phonology, Evolution of
PHYLOGENY OF THE SYLLABLE. The frame/content theory
(MacNeilage 2008) offers an evolutionary account of the syllable.
Syllables are universally associated with openclose alternations
of the mandible, vowels being open and consonants closed articulations. This movement has a parallel in childrens babbles,
which resemble consonant-vowel sequences such as [bababa],
but are in no way organized in terms of discrete segments. Rather,
their syllabic and segmental character arises fortuitously from
adding phonation to the openclose jaw motion. This rhythmically repeated up-and-down movement is also found in so-called
lipsmacks, a facio-visual behavior in primates often combined
with phonation during grooming.
Accordingly, the evolutionary path to the syllable began in
deep prehistoric time when mammal biomechanics evolved for
feeding. A second stage was the use of this machinery in primate
communication. In a third step, this primate mechanism was coopted for speech by scaffolding early phonology on its pseudosyllables and pseudo-segments.
QUANTAL THEORY. The acoustic consequences of a continuous
articulatory movement are often noncontinuous, as illustrated by the pseudo-segmental character of babbling. In the
babble example, the jaw moves continuously but the acoustics
shows an abrupt change from a vowel-like to a stoplike pattern.
This quantal jump illustrates a general fact about the phonetic
space. The mapping of articulation onto acoustic parameters
creates a set of acoustic patterns that forms a number of disjoint
subspaces, rather than a single continuous, coherent space.
Within each such subregion, sound quality is homogeneous.
Voiced and voiceless sounds, as well as different manners of
articulation (e.g., stops, nasals, fricatives, trills), exemplify such
distinct subspaces (Stevens 1989).
USER-BASED CONSTRAINTS: ON-LINE SPEECH. The human voice
is an expressive instrument that undergoes moment-to-moment
retuning by many nonlinguistic factors. Consequently, the phonetic patterns conveying linguistically the same utterance exhibit
great variability. However, the need for messages to be both intelligible and pronounceable imposes a systematic distribution on
phonetic variations, placing them between clear hyperforms and
reduced hypoforms. This view portrays speakerlistener interactions as a tug-of-war between the listeners need for comprehension and the speakers tendency to simplify. There is a great deal of
experimental evidence for this view of speech (Lindblom 1990).
USER-BASED CONSTRAINTS: PHONOLOGY. These user-based constraints also leave their mark on phonology, as is evident from
typological data on strengthening and weakening processes in
phonological rules and sound changes (Kiparsky 1988) and from
attempts to simulate segment inventories. These studies indicate
that systemic selections have been favored that simultaneously
optimize distinctiveness and articulatory ease. An example of
the effect of these conditions is the size principle: The larger the
system, the greater the proportion of articulatorily complex segments (Lindblom and Maddieson 1988)
SELF-ORGANIZATION. These user-based constraints in conjunction with the quantal nature of the signal space help explain
why phonologies do not recruit more of a humans total soundmaking capabilities (e.g., mouth sounds and other non-speech
vocalizations; Catford 1982) but prefer practically the same small
set of phonetic properties. However, the study of these constraints
only partially illuminates the roots of combinatorial coding. This
topic has been explicitly addressed in computer modeling experiments. One such study shows how discrete phonetic targets and
reuse can emerge from a dynamic systems network of agents
(speaker/listener models) whose vocalizations initially randomly distributed in phonetic space tend to converge (driven
by a magnet-like dominance of the patterns heard most often) on
a few targets (Oudeyer 2006).
TARGETS AND MOTOR EQUIVALENCE. Traditionally, the basic units
of speech have been assumed to be targets, the intertarget transitions being primarily determined by the response characteristics
of the production system. Speech, like other movements, exhibits motor equivalence: the ability of motor systems to compensate
and reach a given goal irrespective of initial conditions. This view
implies that the end state of phonetic learning is a set of contextindependent targets and a system capable of motor equivalence.
It moreover suggests that once a target has been learned in one
context, it can immediately be reused in other contexts, since the
motor equivalence capability handles the new trajectory. Also, it
means that, developmentally, discrete segments derive from the
emergent targets and recombination from motor equivalence.
A further relevant observation on the target hypothesis is that
linguistic systems with phonemically coded vocabularies would
be learned faster, more easily, and in an open-ended manner
than repertoires based on holistic forms (Lindblom 2007).
Conclusion
Where does combinatorial structure come from? From prespecifications in our genetic endowment? Or from a modality-independent principle shared by sign and speech and perhaps also
operating in genetics and chemistry (cf. the particulate principle
[Abler 1989])? Or from a mutually reinforcing interplay between
cognitive growth and a suite of conditions entailed by communicating by vocal sounds?
In view of the materials reviewed here, a positive treatment of
the last possibility appears within reach. More lexical inventions
imply an increasing number of soundmeaning pairs. The linking of phonetic shapes with distinct meanings would be subject
to numerous user-based constraints and processes shaping the
instrinsic content of lexical entries, fractionating them into discrete units and facilitating unit recombination. Sound structure
could, thus, plausibly have evolved in response to the expressive
needs associated with growing semantic abilities and as a process of phonetically biased scaling, self-organizing without
any formal a priori or modality-independent blueprint.
Bjrn Lindblom
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Abler, William L. 1989. On the particulate principle of self-diversifying
systems. Journal of Social and Biological Structures 12: 113.
Arbib, Michael. 2005. The mirror system hypothesis: How did protolanguage evolve? In Language Origins: Perspectives on Evolution, ed.
Maggie Tallerman, 2147. New York: Oxford University Press.
617
Phonology, Universals of
Bellugi, U., And M. Studdert-Kennedy. 1980. Signed and Spoken
Language: Constraints on Linguistic Form. (Dahlem Konferenzen.)
Weinheim, Germany: Verlag Chemie GmbH.
Carr, Ren, B. Lindblom, and P. MacNeilage. 1995. Rle de lacoustique
dans lvolution du conduit vocal humain. Comptes Rendus de
lAcadmie des Sciences (Paris) t 30, srie Iib: 47176.
Catford, John C. 1982. Fundamental Problems in Phonetics.
Bloomington: Indiana University Press.
Donald, Merlin. 1991. Origins of the Modern Mind. Cambridge: Harvard
University Press
Fitch, W. Tecumseh. 2000. The evolution of speech: A comparative
review. Trends in Cognitive Science 4.3: 25867.
Hewitt, G., A. MacLarnon, and K. E. Jones. 2002. The functions of laryngeal air sacs in primates: A new hypothesis. Folia Primatol 73: 7094.
Hurley, Susan, and Nick Chater. 2005. Perspectives on Imitation: From
Neuroscience to Social Science. Vols. 1, 2. Cambridge, MA: MIT Press.
Kent, Ray D. 2004. Development, pathology and remediation of
speech. In From Sound to Sense: 50+ Years of Discoveries in Speech
Communication, ed J. Slifka et al. Cambridge, MA: Research
Laboratories of Electronics, MIT.
Kiparsky, Paul. 1988. Phonological change. In Linguistics: The Cambridge
Survey. Vol. 1. Ed. F. J. Newmeyer, 363415. Cambridge: Cambridge
University Press.
Lieberman, Philip. 1991. Uniquely Human. Cambridge: Harvard
University Press.
Lindblom, Bjrn. 1990. Explaining phonetic variation: A sketch of the H&H
theory. In Speech Production and Speech Modeling, ed. W. Hardcastle
and A. Marchal, 40339. Dordrecht, the Netherlands: Kluwer.
. 2007. The target hypothesis, dynamic specification and segmental independence. In Syllable Development: The Frame/Content
Theory and Beyond, ed. B. Davis and K. Zajd. Hillsdale, NJ: Lawrence
Erlbaum.
Lindblom, B., and I. Maddieson. 1988. Phonetic universals in consonant systems. Language, Speech and Mind, ed. Larry M. Hyman and
C. N. Li, 6278. London and New York: Routledge.
MacNeilage, Peter F. 2008. The Origin of Speech. New York: Oxford
University Press.
Miller, George A. 1991. The Science of Words. New York: Freeman.
Oudeyer, Pierre-Yves. 2006. Self-Organization in the Evolution of Speech.
New York: Oxford University Press.
Stevens, Kenneth N. 1989. On the quantal nature of speech. J Phonetics
17: 346.
Studdert-Kennedy, Michael. 2005. How did language go discrete? In
Language Origins: Perspectives on Evolution, ed. Maggie Tallerman,
4867. New York: Oxford University Press.
PHONOLOGY, UNIVERSALS OF
Phonological universals are those aspects of languages sound
system that are found either in every or most human languages or in diverse languages where their presence cannot be
accounted for by inheritance from a common parent language,
geographical proximity, or borrowing. They are often referred
to as unmarked or default conditions in languages phonologies
when these terms imply reference to common cross-language
patterns. There are universal patterns in languages 1) sound
inventories, including their prosodies, 2) sequential constraints
(how sounds are sequenced), and 3) sound changes and the
phonological alternations they create within a given language.
They are of interest because they give insight into the physical
factors that shape human speech, help to elucidate mechanisms
of sound change, and, perhaps, suggest something about the
618
Table 1
Noun stem
Stem + and
English gloss
/fab/
/fab:n/
Father
/sgd/
/sgt:n/
Scorpion
/ka/
/ka: n/
Donkey
/mg/
/mk:n/
Dog
Phonology, Universals of
Table 3.
Table 2.
Language
Voiced only
a. Sundanese:
Awadi
i, u, e
a, o
to wet
to be rich
Campa
o, e, a
bhr
Chatino
i, u
o, e, a
hkn
to inform
Dagur
i, u, e
o, a
msih
to love
Huichol
i, , e
u, a
Serbo-Croatian
i, u
e, o, a
Tadjik
i, u, a
e, o, u
Tunica
i, e, , a, , o
Uzbek
i, u
e, , o, a
h
and lieutenant [lwtnnt] ~ [lftnnt] and truck [t k]~[th k]
(and similar patterns in many other languages). The same factors
frequently lead to the affrication of stops before high close vowels or glides as in Japanese, for example, or the sound change that
converted Benjamin Franklins natural [ntjul] to the modern
pronunciation [n t l].
Aerodynamic factors also explain patterns of nasal prosody
in languages as diverse as Sundanese (spoken in the Indonesian
archipelago) and Tereno (spoken in the Mato Grosso, Brazil).
As shown in Table 3, in these (and other) languages, the presence of a nasal consonant induces nasalization on all vowels and
b. Tereno:
1st person
3rd person
piho
mbiho
I/he went
ahjaao
anaao
I/he desire(s)
iso
nzo
I/he hoed
owoku
wgu
my/his house
ajo
jo
my/his brother
emou
my/his word
iha
nza
my/his name
Sources: For Sudanese: Robins 1957; for Tereno: Bendor-Samuel 1960, 1966.
619
Phonology, Universals of
Table 4.
Kpelle: [w] patterns with velars in nasal assimilation:
Indefinte
Definite
`mi
wax
lu
`nui
fog, mist
`il
dog
we
`wei
white clay
Phonotactics
The conventional view of common cross-language sound
sequencing or phonotactics is couched in terms of whats
called the sonority hierarchy (attributed to E. Sievers and O.
Jespersen), whereby the favored pattern at syllable onset shows
sounds sequenced in the following order (where omissions
are possible): stop + fricative + nasal + liquid (i.e., non-nasal
620
Sound Changes
Of the thousands of regular sound changes that have been
identified using the comparative method in historical phonology, certain ones are recognized as showing independent
cross-language incidence. One such is velar palatalization, k >
t, t, s, /__i (j), (and similar changes involving the voiced velar
/g/), for example, English cheese [thiz]] from Latin caseus (cf.
Dutch kaas); Ikalanga [ti-ledu] chin < Proto-Bantu *ki-dedu.
Traditionally, the causes of sound change were attributed to two
opposite tendencies: speakers striving for ease of articulation,
which would lead to assimilations and reductions, and speakers striving to speak more clearly, which would lead to exaggeration of articulation and augmentation of pronunciation.
There is no doubt that speakers do alter their pronunciation
in these manners, but it may be seriously questioned whether
Phonology, Universals of
Phrase Structure
PHRASE STRUCTURE
It is an ancient observation that natural language syntax is hierarchically organized. As can be seen from a variety of diagnostics,
the words comprising a sentence do not behave as beads on
a string but group into successively larger units, or constituents.
Phrase structure (PS) is a formal representation of this constituent structure. PS is typically depicted as a tree-structured
graph (Figure 1), which encodes three sorts of structural information: i) dominance, specifying the words and constituents that a
constituent contains within it (e.g., as shown by vertical placement
in the figure, prepositional phrase (PP) dominates on and television); ii) precedence, specifying the temporal orderings among the
words and constituents (e.g., as shown by horizontal position, the
constituent most fans precedes the constituent watched the game
on television); and iii) labeling, specifying the grammatical category of each word and constituent (e.g., the constituent the game
is a noun phrase (NP)). In PS-based approaches, this structural
information plays an important role in defining the conditions
under which grammatical dependencies may obtain (see agreement, anaphora, binding, and case), and PS is often taken to
be the input to transformational operations (see movement and
transformational grammar). Further, PS representations
serve as the interface between syntax and semantics, as they provide the structural information necessary for interpretation (see
compositionality, thematic roles, and logical form).
A fundamental question concerns how the range of possible
PS is specified in a grammar. The earliest answer comes from
Noam Chomsky (1957), who suggests that PS is generated by a
set of phrase structure rules, like the following:
1. S NP VP
2. NP N
3. NP Det N
4. VP V NP
5. VP VP PP
6. PP P NP
621
Pidgins
Figure 1. Phrase structure representation for Most fans watched the game
on television.
PIDGINS
Pidgins are the worlds only non-native languages. They are
typically acquired by adults, after the critical period for language
acquisition has passed. They normally arise wherever sufficient
speakers of mutually incomprehensible languages must interact with one another. Some pidgins arose through and for trade;
the most plausible derivation offered for the origin of the name
pidgin attributes it to the Chinese pronunciation of business
622
Pidgins
Pitch
PITCH
When an object vibrates, its movement produces changes in
air pressure that radiate like waves from the source. If the frequencies of the vibrations are roughly between 20 and 20,000
cycles per second, or Hertz (Hz), ideally they can be heard by a
young, healthy human listener. The range of hearing frequencies declines with age. The physical characteristic of the vibrating body, frequency, produces a psychological experience called
pitch. In general, a low frequency produces the sensation of a low
pitch (for example, the 60 Hz hum produced by electrical power
in a poorly grounded radio), with the pitch increasing as the frequency increases (a male voice at 100 Hz, a female voice at 200
Hz, a childs voice at 300 Hz). Because there is a close correspondence between frequency and pitch, people frequently use
the terms interchangeably. However, in addition to frequency,
the sensation of pitch is influenced by an interaction between
the amplitude of the vibration and the range of the frequency.
Pitch is also influenced by the complexity of the vibration and its
corresponding wave form.
A vibrating body oscillates as a single entity, producing a frequency referred to as the fundamental frequency. So, when the
key for A above middle C is played on a piano, a string vibrates
at 440 Hz. Vibrating bodies are not perfectly rigid, though, and
the string also vibrates in parts as if it is two strings (producing
a frequency of 880 Hz), and three strings (1320 Hz), etc. Thus, a
vibrating body produces a series of frequencies beginning with
the fundamental frequency (f0) and including its harmonics,
which are multiples of the f0. The distribution of acoustic energy
623
Pitch
across the harmonic series contributes to the quality or timbre
of the sound. In addition to sound quality, the harmonics contribute significantly to the perception of pitch. The fundamental
frequency of a harmonic series can be artificially removed without changing the pitch, a demonstration referred to as the missing fundamental. In speech, the harmonic series is a function of
the complex wave produced by the glottal source. Formants are
bands of resonance that concentrate the acoustic energy produced by the glottal source as a function of the vocal tract configuration and have center frequencies that reflect the vocal tract,
rather than the harmonic series.
Pitch can be experienced from pure tones (f0 alone), which
do not occur naturally, as well as from complex tones (f0 + harmonic series), but the pitch of complex tones is a stronger percept, allowing finer discriminations of f0 frequency differences.
The processing of pitch from a determination of the f0 versus
the pattern recognition of a harmonic series relies on different neurological systems. Simple frequency determination can
occur at multiple levels of the nervous system, but complex pitch
processing occurs in auditory areas in the right cerebral hemisphere that complement speech and language areas in the left
cerebral hemisphere (Sidtis 1980; see also right hemisphere
language processing and left hemisphere language
processing).
For practical purposes, pitch in language can be viewed as a
direct function of f0. Unlike the pitch distinctions made in music,
linguistic pitch distinctions are comparatively coarse. Whereas
a musical octave can be divided into 12 semitones, and vibrato
in experienced singers can be consistently less than a semitone,
most linguistic communicative situations only require distinctions of three semitones or more. Further, the pitch distinctions
in language are relative, allowing men, women, and children to
make the same linguistic and paralinguistic distinctions despite
different vocal f0s (see paralanguage), whereas pitch distinctions in music reference specific frequencies (e.g., a musical
scale tuned to 440 Hz).
A number of linguistic and paralinguistic phenomena are
provided by pitch. At the suprasegmental level, pitch produces
the melodic line of an utterance to convey linguistic intonation (e.g., declination effect: falling pitch anticipating the end
of a statement, rising pitch indicating a question), sociolinguistic information (e.g., uptalk, rising pitch at the end of a
statement, falling pitch as a cue for turn-taking), and paralinguistic information (e.g., emotion, attitude). Pitch can also be
used with loudness to provide syllable accent at the segmental level.
Pitch also has a lexical role in tone and pitch accent languages. Tone languages may have many tone patterns (estimates vary, but the numbers are fewer than the 12 notes in the
musical octave), and they tend to fall in relative categories like
high, medium, and low, further distinguished by rising and falling patterns. Because such distinctions are relative, the listener is
required to perform a tone normalization to identify a speakers
lexical tones. Just as simple and complex pitch perception rely
on different brain mechanisms, the processing of pitch for linguistic and nonlinguistic purposes engage different neurological
systems, principally the temporal lobes in the left and right
cerebral hemispheres (Van Lancker and Fromkin 1973).
624
625
626
627
628
to normal subjects, and from behavioral tests to technology-assisted observations. In the 1970s and 1980s, RHD subjects performed poorly on connotative word meaning tests, while LHD
subjects experienced problems with denotation (Gardner and
Denes 1973; Brownell, Potter, and Michelow 1984; Drews 1987).
One might have assumed that the left hemisphere processed
denotation and the right, connotation, but over time, a more
complex picture emerged. Christine Chiarello and others established, using visual-field testing, that primary and subordinate
word meanings are initially activated in both hemispheres, but
that subordinate meanings are quickly suppressed in the left
hemisphere, resulting in a more efficient processing time for the
dominant meaning not unlike Gioras graded salience model,
where the most salient meaning of an expression, metaphoric or
not, gets processed first and faster than a less commonly occurring meaning (Chiarello and Maxfield 1995).
Finally, the neurobiology of poets may play a significant role
in the neurobiology of poetic language. Poets are known to suffer
from affective disorders in particular, hypomania and bipolar
illness at rates far exceeding those of the general population
or other categories of writers (Andreasen 1987; Jamison 1989;
Ludwig 1994; Post 1996). Feeling negative emotion strongly,
being introspective, and spending time alone are traits associated with expressive writing as well as mental dysfunction, and
mentally ill persons may feel drawn to express their anguish in
writing; James Kaufman and John Baer (2002) propose these
and other behavioral explanations for the poetry/affective disorder connection. Taking a neurobiological approach, Felix
Post (1996) suggests that the intensive intellectual and emotional effort involved in writing poetry may trigger overactivation
of neural networks and, thus, cause mental illness. Julie Kane
(2004) suggests the opposite, that overactivation may precede
poetic output: Pointing to substantial evidence that handedness and dominance for language can shift temporarily from the
left to right hemisphere during manic episodes, she proposes
that abnormal mood elevation may activate right-brain regions
involved in processing poetic language. Recently, too, Dawn
Blasko and Victoria Kazmerski (2006) have shown that poets and
nonpoets differ in the brain regions that they activate while reading poems.
There is a vast amount of territory yet to be covered in exploring the elephant of poetic language, complicated by the fact
that new research findings often seem to challenge the old. But
as neuroimaging techniques become more precise and less invasive, illuminating features that could only be guessed at before,
one thing becomes increasingly clear: The neurobiology of poetic
language is not the same animal as the neurobiology of ordinary
language.
Julie Kane
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Andreasen, Nancy. 1987. Creativity and mental illness: Prevalence
rates in writers and their first-degree relatives. American Journal of
Psychiatry 144: 128892.
Blasko, Dawn, and Victoria Kazmerski. 2006. ERP correlates of individual
differences in the comprehension of nonliteral language. Metaphor
and Symbol 21.4: 26784.
Poetic Metaphor
Post, Felix. 1996. Verbal creativity, depression and alcoholism: An investigation of one hundred American and British writers. British Journal
of Psychiatry 168: 54555.
Sotillo, Maria, Luis Carreti, Jos Hinojosa, Manuel Tapia, Francisco
Mercado, Sara Lpez-Mrtin, and Jacobo Albert. 2005. Neural activity associated with metaphor comprehension: Spatial analysis.
Neuroscience Letters 373: 59.
Turner, Frederick, and Ernst Pppel. 1989. The neural lyre: Poetic meter,
the brain, and time. In Expansive Poetry: Essays on the New Narrative
and the New Formalism, ed. Frederick Feinstein, 20954. Santa Cruz,
CA: Story Line.
Winner, Ellen, and Howard Gardner. 1977. The comprehension of metaphor in brain-damaged patients. Brain 100: 71729.
POETIC METAPHOR
Since Aristotles first articulation of a comparative theory of
metaphor, metaphor studies in literary and ordinary language
have proceeded without interruption in philosophy, rhetoric,
linguistics, and literary criticism. Two traditions have emerged
in metaphor theory: conceptual and linguistic traditions. The
conceptual view emphasizes metaphors fundamental role in
everyday thought and language; the linguistic tradition limits the range of metaphor to local pragmatic and aesthetic
functions (Ortony 1993). The range of accounts within both
traditions is variegated and well beyond the scope of this entry.
Nonetheless, for the language sciences, it seems that the conceptualist tradition has dominated in recent years. The present
discussion assumes a conceptual view of metaphor as understood through the frameworks of conceptual metaphor
theory (CMT) and conceptual blending theory (CBT),
where poetic metaphor is regarded as a special case of these
underlying conceptual operations. At present, poetic or literary
metaphor cannot be easily extracted from the central questions
of metaphor theory in general, namely: What is metaphor, and
what is metaphor for? The present discussion merely touches on
the first question, in favor of a more elaborate treatment of the
second question.
The question of what metaphor is and how poetic metaphor
can help language sciences understand the everyday mind and
language is addressed in the first section, where I compare and
contrast the two models of metaphor. The question of what metaphor is for is addressed in the second section.
What Is Metaphor?
CMT purports to unearth the systematic correlations of experience and meaning. Meaning arises from everyday experience.
Abstract notions such as time, causation, states, change, and
purposes depend on a rich system of metaphors. Metaphor is
the name given to the process of conceptual mappings from
source to target domains (see source and target). The latest
incarnation of CMT (Lakoff and Johnson 1999) builds on Joseph
Gradys (1997) theory of primary metaphor, in which the ontogenetically basic process of domain correlation constitutes the
experiential basis of conceptual metaphors. A primary metaphor
is a correlation of subjective experience with a more abstract
concept. For instance, MORE IS UP is a primary metaphor, based
on the tight ontological correlation between the accumulation of
the same entities and vertical height.
629
Poetic Metaphor
CBT, while not a theory of metaphor, accounts for metaphor
as a species of conceptual blending that often involves the integration of concepts that do not normally go together. CBT takes
a decidedly usage-based perspective to metaphor and other
phenomena, in which systematic correlations arise from conceptual blending itself, the process of constructing new scenes
and scenarios with specific emergent properties from multiple
mental models. The aim is to see how metaphors arise on the
fly as we think and talk. CMT has as its basic unit of cognitive
structure the conceptual domain. CBT has as its basic unit of
organization a mental space, or scenes and scenarios set up as
we think, talk, and otherwise interact. CBT models the dynamic
unfolding of a language users representations. In this respect,
CBT has developed analytic routines and modeling techniques
that capture constitutive principles and governing constraints of
blending. (See Fauconnier and Turner 2002, 30952.)
630
Ship
Leader
Ships captain
National policies/actions
Ships course
National success/improvement
National failures/problems
Sailing mishaps
Sea conditions
All these metaphoric mappings derive from the basic primary metaphoric couplings of ACTION-AS-SELF-MOTION,
COURSES-OF-ACTION-AS-PATHS, SOCIAL-RELATIONSHIPS-ASDEGREES-OF-PHYSICAL-PROXIMITY,andCIRCUMSTANCES-ASWEATHER. These experiential correlations (and perhaps others)
interact in a way that motivates the framing of a nation and its history as a ship gliding through water.
As George Lakoff and Mark Turner (1989, 6772) argue, the
power of poetic metaphor, in particular, issues from the extension of these mappings for local expressive purposes. A conventionalized metaphor never gives you all you need, and poetic
thought is marked by its ability to stretch or extend conventional
metaphors. Notice that with lines 67, DeMond extends the typical range of mappings to include ship building and the role
shipwright.
Poems also employ expressions in which the schemas and
domains underlying the metaphor can be elaborated in unusual
or novel ways. Lines 6 and 7, when understood against the context of the whole speech, take on unusual significance. The implication of line 6 is that the master shipwright is God, while the
workman is identified with the Negro, echoing a consistent
theme of the speech the hard labor of the Negro race in building America.
Poetic Metaphor
A blending analysis helps account for ways in which the
NATION-AS-SHIP metaphor is not a simple and obvious mapping between two conceptual domains. While conceptual
domains name large depositories of knowledge about the physical and social world, mental spaces comprise on-line scenes and
scenarios; they are specific and sensitive to pressures from local
context.
Levins (1993) account of poetic metaphor is more completely captured in CBT, a theoretical framework in which preternatural scenes are constructed to reveal how to reason and
draw inferences about something else. Conceptual blends are
often richly counterfactual, but rarely do they exist for their
own ends. In this case, the blended scenario extends and elaborates the conventional metaphor for local rhetorical purposes. A
basic blending analysis of DeMonds introductory sentence and
the first five lines of the poem proper would include a discourse
ground specifying the participants, the situation, and setting, a
mental space for Seafaring, a mental space for Nation, and the
initial blended space for Nation-as-Ship, each of which is set up
in the very first line of the poem.
Let us assume the analysis from the perspective of a worshiper
sitting in the Dexter Avenue Baptist Church in 1900. Under these
conditions, the ground includes the identities of the churchgoers, the speaker, and the setting. Let us further assume that the
discourse participants are African Americans and that the poetic
persona represents them. Initially, the Seafaring and Nation
spaces project conceptual structure into the blend under the
influence of the cross-space mappings as specified here. In the
blend, America is a ship, Negro citizens are among its passengers, the ocean is time, and the twentieth century is an unspecified landmark on open water. The blend allows the audience
to imagine temporally, causally, and spatially diffuse political
events as attaining, for the moment, the look and feel of primary
experience.
Once composed, the blend and the network of mental spaces
permit the addition of new information and relations. A noteworthy contribution of the blending framework here is that it offers
precise ways of accounting for the elements of the Nation-asShip image that have no specific counterparts in the target space
of nations and politics. Once the network is up and running,
readers can combine concepts fluidly. For instance, line 2 commands, Sail on, O Union, strong and great, wherein the poet
fuses elements from different mental spaces into tight syntactic
units. Thus, in the blend it is perfectly natural and logical for a
union to sail. What is more, it is perfectly natural for the ship to
plot a straight course. Once the image is created, many other elements of ships become mentally accessible. For instance, ships
must be made of particular materials in order to be seaworthy.
The phrase, ribs of steel, in line 6 satisfies local formal and conceptual imperatives in 1) providing completion for the couplet
with line 5, and 2) suggesting that the nation is made of sturdy
material and (opportunistically) made from the very material
that the Negro worker has been responsible for manufacturing.
Importantly, the mapping between Shipwright and Creator is
responsible for all aspects of the nation.
The goal of using this conventional metaphor is to construct
a view of social reality for the Negro race, focusing on communal
activities and on achieving collective goals. The ship of state has
Poetics
been conventionalized for just that purpose because the image
potential associated with building, operating, and navigating is
of rich social activities. In the blended space, however, the choice
to sail on is framed as an all-or-nothing proposition. If the ship of
state does not sail, it ceases to exist. In the sailing space, however,
the ship, once built, exists whether or not the crew sails; in the
sailing space, a captain and crew can choose when and when not
to sail, and the crew can still be referred to as sailors whether on
land or on sea. In the blend, a refusal to board and sail is tantamount to renouncing ones citizenship. By exploiting elements
of the shipwright (a collective activity) and by attributing that
activity to a divine creator, DeMonds version of the ship of state
takes on the voice of a divine decree.
As suggested, DeMond takes considerable risk in quoting a
poem that makes extensive use of this metaphor, for members of
the congregation may generate a metaphoric mapping in which
the cross-domain counterpart of American Negro is not passenger but cargo, destroying the political legitimacy of the image.
DeMond, however, assiduously avoids focusing any attention
on the circumstances that brought them to America. Instead, he
picks up the story at their arrival and tells of the Negro race as
those who built the nation.
The present analysis presents CMT and CBT as complementary analytic frameworks, wherein the first focuses solely on conventionalized mappings, while the latter is much more interested
in how these mappings operate in local rhetorical contexts, and
thus can point scholars in the direction of a usage-based theory
of poetic metaphor.
Todd Oakley
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
DeMond, A. L. 1900. The Negro element in American life: An oration.
Available online at: http://lcweb2.loc.gov/cgi-bin/query/r?ammem/
murray:@field(DOCID+@lit(lcrbmrpt0e10div2).
Fauconnier, Gilles, and Mark Turner. 2002. The Way We Think. New
York: Basic Books.
Grady, Joseph. 1997. Foundations of meaning: Primary metaphors and
primary scenes. Ph.D. diss., University of California, Berkeley.
Grady, Joseph, Todd Oakley, and Seana Coulson. 1999. Blending and
metaphor. In Metaphor in Cognitive Linguistics, ed. F. Gibbs and G.
Steen, 10124. Amsterdam: Benjamins.
Lakoff, George, and Mark Johnson. 1999. Philosophy in the Flesh. New
York: Basic Books.
Lakoff, George, and Mark Turner. 1989. More Than Cool Reason: A Field
Guide to Poetic Metaphor. Chicago: University of Chicago Press.
Levin, Samuel. 1993. Language, concepts, and worlds. In Metaphor
and Thought. 2d ed. Ed. A. Ortony, 11223. Cambridge: Cambridge
University Press.
Ortony, Andrew. 1993. Metaphor, language, and thought. In Metaphor
and Thought. 2d ed. Ed. A. Ortony, 116. Cambridge: Cambridge
University Press.
POETICS
In ancient Greece, as Aristotle pointed out, there was no common name for all the different poetic genres (Poetics, 47b),
including epic, tragic drama, dialogue, elegy, and poems written in various meters. Poetry in the sense of making or creation
became the general name for literary expressions in diverse
631
Poetics
forms, and Poetics, the term used by Aristotle for his treatise on
tragedy and epic, thus represented the kind of critical and analytical treatment of poetry that would be called in later times literary criticism or literary theory.
Aristotles Poetics offers an important model in Western literary criticism, but it was not widely known in Europe in antiquity
or in medieval times, and it did not become a classic until the
latter half of the sixteenth century. During the time that it was lost
in medieval Europe, however, the Poetics, along with some other
works by Aristotle, was being studied by Arabic scholars, notably
Ibn Rushd, known in the West as Averros. But once it was rediscovered and commented on by such influential Renaissance critics as Lodovico Castelvetro (150571) and Francesco Robertello
(151667), the Poetics quickly became one of the most influential
works in Western literary criticism. Epic and tragedy discussed
therein became the two major classical genres before the rise
of the modern novel and, after Dante, poets of every European
nation tried to create an epic in the vernacular to mark the maturity of a modern language and the establishment of a national
literary tradition. Aristotles philosophical treatment of plot, language, and rhetoric of the tragic drama provides a model of critical analysis, and many basic concepts used in the Poetics, such
as imitation, recognition, the reversal of fortune, tragic hubris,
and the catharsis of pity and fear, have all had a tremendous
influence on later criticism. In our own time, Aristotles Poetics
remains a major classic and continues to be discussed and commented on by important critics and theoreticians from various
perspectives.
As the aforementioned Arabic commentaries suggest, the
systematic study of the literary art is by no means confined to
the European tradition. There are, for example, well-established
traditions of sophisticated literary criticism or poetics in South
and East Asia. The earliest treatise on dance and dramatic art in
ancient India, Bharatamunis Ntyastra (ca. second cent. b.c.),
offers a comprehensive discussion of Sanskrit drama in terms of
taste and emotions (rasa) and of language and bodily gestures
that give expression to various emotions. In the seventh century,
Sanskrit poetics was fully established by such important theorists
as Bhmaha and Dandin. In the ninth century, nandavardhana
made significant contributions to its further development with
discussions of the theoretical notions of rasa and dhvani, while
Abhinavagupta and Kuntaka in the tenth century explored new
areas by debating on the issue of indirect and suggestive expressions (vakrokti) in poetic language. Indeed, as an Indian scholar
remarks, A study of Sanskrit poetics from Bharata (5th century
b.c.) to Panditarja Jaganntha (17th century a.d.) will bear witness to the existence of a highly developed poetics in ancient
India, with a rigorous scientific method for description and analysis of literature (Pathak 1998, 345).
In China, the Great Preface to the Mao edition of the Book
of Poetry (second century b.c.) articulated the Confucian ideas
about poetry and its functions, and laid the foundation of a poetics that both acknowledges the release of emotions as the origin
of poetry and the efficacy of moral teaching as its ultimate justification. Lu Ji (261303), with his Rhyme-Prose on Literature added
to the critical tradition a more focused attention on the importance of emotions (qing), and he argued for the necessity to learn
both from nature and from the ancients. Liu Xies (465?520?)
632
Point of View
Literary Mind and the Carving of Dragons is deservedly famous
as the most systematic study of the literary art in the Chinese
critical tradition. This substantial work of Chinese poetics relates
literature to the cosmic tao and the exemplary classics of ancient
sages, thereby elevating literature to a position of high social and
moral values. Its focus, however, is on the art of literature. The
Literary Mind first formulates some basic principles of the idea of
wen or literature, gives a survey of all the literary genres in classical Chinese literature, commenting on their origin and development, and then presents a highly developed theory of literary
creation, making contributions to the important issues of the
relationship between poetry and reality, the style and characteristics of a literary work, the effect of imagery and poetic imagination, and the regulations of metric composition. Since the eighth
century in Tang China, and particularly the eleventh century in
the Song Dynasty, there have been numerous works in a critical
genre known as remarks on poetry (shihua), which often contain
valuable insights into the nature of poetry, the techniques of the
literary art, and the principles of aesthetic appreciation. Like the
aforementioned Indian example, the Chinese critical tradition
also offers an alternative form of poetics outside the Aristotelian
and European tradition.
In a broad sense, then, poetics can be understood as a critical, theoretical, and more or less systematic treatment of poetry
or literature in general. In such an expanded usage, what the
term signals is a theoretical discourse on a subject in arts or literature, covering a considerable range of oeuvre, and offering
some philosophical insights into the nature of the subject under
discussion. Poetics, therefore, becomes a general term for a sustained argument or a long essay in literary and art criticism.
Zhang Longxi
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Aristotle. 1987. Poetics with the Tractatus Coislinianus, Reconstruction
of Poetics II, and the Fragments of the On Poets. Trans. Richard Janko.
Indianapolis: Hackett.
Averros. 1977. Three Short Commentaries on Aristotles Topics,
Rhetoric, and Poetics. Ed. and trans. Charles E. Butterworth.
Albany: State University of New York Press.
De, S. K. 1988. History of Sanskrit Poetics. Calcutta: Firma KLM Pvt. Ltd.
Liu, James J. Y. 1975. Chinese Theories of Literature. Chicago: University
of Chicago Press.
. 1988. Language-Paradox-Poetics: A Chinese Perspective. Ed.
Richard John Lynn. Princeton, NJ: Princeton University Press.
Miner, Earl. 1990. Comparative Poetics: An Intercultural Essay on Theories
of Literature. Princeton, NJ: Princeton University Press.
Pathak, R. S. 1998. Comparative Poetics. New Delhi: Creative Books.
Rajendran, C. 2001. Studies in Comparative Poetics. Delhi: New Bharatiya
Book Co.
POINT OF VIEW
In narrative studies (see narratology), this term, also perspective or focalization, refers to textual strategies that provide
the reader with the illusion of seeing things through the eyes
of a character. These strategies are mostly linguistic in nature,
ranging from deictic positioning in the characters mental here
and now (see deixis) to lexical choices linking up with the
Point of View
characters worldview and ways of thinking and perceiving the
world. Point of view, from a linguistic perspective, is therefore an
important aspect of linguistic pragmatics.
633
Point of View
stylistic and lexical markers relating to the characters social
position, age, gender, and so on. For instance, when in Charles
Dickenss Our Mutual Friend Mrs. Veneering remarks that
these social mysteries make one afraid of leaving Baby ([18645]
1952, 414; emphasis added), the word Baby relates to the motherchild relationship of the reported speaker and represents her
point of view. At the same time, the phrase these mysteries and
the pronoun one underline Mrs. Veneerings upper-class status.
Addressee-oriented expressions like forms of address (Maam,
Sir, Your Excellency, etc.) also invoke social position by linguistic
means (cf. Fillmore 1983, 1997).
Most basically, deictics serve the function of positioning
speakers and, hence, creating point of view. For example, in
Bleak House Mr. Bucket (still grave) inquires if to-morrow morning, now, would suit ([18523] 1962, 720; emphasis added),
in which the futurity of to-morrow relates to Mr. Buckets
moment of utterance. Among linguists, Charles Fillmores work
on deixis (1983, 1997) needs to be credited with incisive insights
into the generation of point of view by means of deixis. From a
linguistic perspective, these signals are expressivity markers,
implying a speaking or thinking consciousness, a deictic center
(Bhlers [1934] origo) from which the world is being viewed. In
the widest possible sense, such expressivity markers are indicative of ideation and emotion, the latter capable of being textually
suggested by syntactic means, such as intensifying repetitions
besides merely lexical intensifiers and emphatic vocabulary.
Evaluative point of view can be illustrated in sentences like Do
talk to the poor dear. Incomplete sentences (indicating hesitation or derangement), sentence modifiers (in any case, sure
enough), clause-initial adjuncts (oh, well), interjections (good
grief), negative inversion (Never will he forget) or left and right
dislocation are among the most common strategies used (cp.
Fludernik 1993, 22779). In oral discourse, moreover, expressivity shows up in intonation and the echoing of idiosyncratic
pronunciation (imitated in writing: she sho was happy).
In medieval literature, point of view is often signaled by interjections like alas or by means of repetition. Such signals of point
of view occur intermittently in medieval literature and early
modern English prose but do not constitute a continuous representation of a characters perspective as in the Gothic novel and
the later stream of consciousness novel.
Like the study of narrative discourse markers, the focus on
expressivity signals can help to emphasize the specifically narrative uses of point of view for the linguist. Point of view markers not
only establish free indirect discourse; they are, moreover, crucial
to text beginnings, where they help distinguish between narratives with a prominent speaker (= narrator) function and others
in which the reader is eased into the story by means of a protagonists perspective. (Roland Harweg [1968] has contrasted these
as emic and etic text beginnings, respectively.) Peculiarities of
thought and worldview are also constitutive of M. A. K. Hallidays
mind-style (1971) as the distinctive linguistic representation of
individual self (Shen 2005, 312). Ultimately, an analysis of point
of view as expressivity links up with the linguistic enquiry into
individual style.
It should be noted that all of these signals of expressivity are
clichs and cannot directly claim mimetic relevance (Fludernik
1993, 43464). On the contrary, they depend on typical recurrent
634
Politeness
Uspensky, Boris. [1973] 1983. A Poetics of Composition: The Structure of the
Artistic Text and Typology of a Compositional Form. Trans. Valentina
Zavarin and Susan Wittig. Berkeley: University of California Press.
POLITENESS
Politeness is essentially a matter of taking into account the feelings of others as to how they should be interactionally treated,
including behaving in a way that demonstrates appropriate
concern for interactors social status and their social relationship. In this broad sense of speech oriented to an interactors
social persona or face, politeness is ubiquitous in language use.
Since taking account of peoples feelings generally involves
saying things in a less straightforward or more elaborate manner than when one is not considering such feelings, ways of
being polite provide a major source of indirectness, reasons for
not saying exactly what one means, in how people frame their
utterances.
There are many folk notions for these kinds of attention to
feelings, captured in terms like courtesy, tact, deference, sensibility, poise, rapport, and urbanity, as well as terms for the contrasting behaviors rudeness, gaucheness, social gaffes and their
consequences, embarrassment or humiliation. Such terms attest
both to the pervasiveness of notions of politeness and to their
cultural framing.
Peoples face is invested in their social status and in their
relationships with one another, and so indexing this relationship
appropriately is necessary for maintaining face expectations. In
addition, one often has interactional goals that potentially contravene face, and the expression of such communicative intentions (e.g., requests, offers, disagreements, complaints) tends to
be mitigated by attention to face.
Politeness is crucial to the construction and maintenance
of social relationships; indeed, it is probably a precondition for
human cooperation in general. Politeness phenomena have,
therefore, attracted interest in a wide range of social sciences,
particularly linguistics, anthropology, psychology, sociology,
and communication. Work in these disparate fields can be characterized in terms of three main classes of theoretical approach.
635
Politeness
Politics of Language
3) in any culture there are norms and values affecting the degree
of imposition or unwelcomeness of an utterance, and one tends
to be more polite for more serious impositions. The linguistic
structures for conveying particular kinds of politeness are also
underlyingly similar across languages, with the politeness of
solidarity (positive politeness) characterized by expressions of
interest in the addressee, exaggerated expressions of approval,
use of in-group identity markers and address forms, seeking
of agreement, and avoidance of disagreement, whereas avoidance-based politeness (negative politeness) is characterized
by self-effacement, formality, restraint, deference, hedges, and
impersonalizing mechanisms like nominalization or passive
constructions.
To explain these kinds of detailed parallels across languages
and cultures in the minutiae of linguistic expression in socially
analogous contexts, Brown and Levinson proposed an abstract
model of politeness as strategic attention to face, deriving strategies for constructing polite utterances in different contexts on
the basis of assessments of three social factors: the relative power
(P) of speaker and addressee, their social distance (D), and the
intrinsic ranking (R) of the face-threateningness of an imposition.
In contrast with rule-based approaches, Brown and Levinson
argued that politeness inheres not in words or sentences per se;
politeness is an implicature that may be conveyed by utterances
spoken in context, by virtue of successful communication of a
polite attitude or intention.
Politeness continues to be a major focus for research in many
disciplines concerned with social interaction, and the topic
now has its own professional journal, the Journal of Politeness
Research. Over the past 30 years, empirical descriptions of particular politeness phenomena from many different parts of the
world have accumulated, with the research emphasis largely
on cross-cultural differences. There has been much theoretical
controversy over whether, indeed, there are any universal principles of politeness and if so, what form they take. The recent
trend seems to be toward emphasizing emic rather than etic
approaches (cf. Watts 2003; Eelen 2005). But the importance of
politeness goes far beyond the ps and qs of appropriate behavior and speech in a particular cultural setting. Its wider significance is in the interactional, communicative, day-to-day basis of
social life and the conduct of social relationships. Recent developments in the theory of social interaction that take account
of our common human nature (e.g., Goody 1995; Enfield and
Levinson 2005; see also universal pragmatics) offer hope
for theoretical progress in this field.
Penelope Brown
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Brown, Penelope, and Stephen C. Levinson. [1978] 1987. Politeness: Some
Universals in Language Use. Cambridge: Cambridge University Press.
Eelen, Gino. 2005. A Critique of Politeness Theories. Manchester: St.
Jerome.
Enfield, Nick, and Stephen C. Levinson, eds. 2005. Roots of Sociality.
Cambridge: Cambridge University Press.
Goffman, Erving. 1967. The nature of deference and demeanor. In
Interaction Ritual, ed. Erving Goffman, 4795. New York: Anchor
Books.
. 1971. Relations in Public. New York: Harper Colophon Books.
636
POLITICS OF LANGUAGE
Politics of language is not a domain or subdiscipline. It is an idea
that puts the study of language in a perspective: the idea that
language is a politically invested object and that people, consequently, act politically in, through, and on language. In that
sense, the term covers an enormous range of phenomena and
cuts across numerous disciplinarily organized practices. The
issue is one of function, and the politics of language suggests
that political meanings and effects are among the functions of
language. In fact, some would emphasize that there are no nonpolitical meanings.
Such political functions are metapragmatic: They operate
through meta-discourses on language, on things people say
about language in language. Thus, the politics of language is a
language-ideological phenomenon. Clear instances of it are
widespread utterances such as English is the language of business or Xhosa is a language for community interaction. In
both instances, a particular language is defined as a language
that operates with a specific load, a specific set of social, cultural, economic political attributes, all of them implicitly
articulated: Whenever I use English, my language use will be
framed as business, and I will speak like a businessman. The
politics of language has to do with the way in which we associate particular varieties of language (forms) with particular normative complexes, genres and topical domains, and identities
(functions). The relationship between forms and functions, thus
defined and seen as relatively stable (stable enough to generate
shared meanings), is usually defined as ideology, and authors
explicitly addressing the politics of language often focus on ideology, hegemony, and ideological naturalization.
In what follows, I first give a brief overview of some key notions
and authors, then engage in a brief survey of some recent work
and focus on language ideologies as a frame for understanding such political functions. I conclude with an appraisal of this
work.
Politics of Language
Rather, I would suggest we read history backwards, starting from
the current approaches to the politics of language and looking into those authors who are seen as formative now. From
that vantage point, two groups of authors stand out: authors
who developed a political view of language and authors whose
political-analytic work provides tools for scholars in the field of
language. The first category is dominated by such scholars as M.
M. Bakhtin, V. N. Voloshinov, Roland Barthes, Michel Foucault,
and Pierre Bourdieu (see habitus); in the latter, Karl Marx (see
marxism) and Antonio Gramsci stand out.
This collection of authors and insights, it must be realized,
can only be discussed in a more or less coherent way when a
number of conditions are met. In particular, two presuppositions
are required:
(i) It is clear that reflections of this kind are predicated on a
view of language as a social object (not a mental object); such
reflections belong to the realm of a social theory of language.
(ii) They also are predicated on a view in which language
displays intricate connections with social structure: Either
language mirrors social structure (especially structures of
inequality) or it can become an instrument for changing
social structure.
These presuppositions ensure that the authors mentioned can
become interlocutors for current practitioners in the field, and such
practitioners would then be clustered in applied fields, such as
discourse analysis (both linguistic and foucaultian),
sociolinguistics, and linguistic anthropology.
The work of Bakhtin and Voloshinov has been influential in
its emphasis on the social and political dimensions of a key feature of real language: its heteroglossic nature (see dialogism
and heteroglossia). Heteroglossia stands for the presence
of multiple voices in an act of communication, and such
voices are intricately related to social formations and interests.
Whenever we communicate, thus, we engage with existing complexes of social (and cultural) meaning, we insert ourselves in an
intertextual tradition in which such complexes make sense,
and we articulate interests, not only (neutral, self-contained)
meanings. In addition, the articulation of such interests is not
a unilateral and linear event. Bakhtin (1981) emphasizes the
importance of evaluative uptake in interaction his dialogical
principle in which every act of communication requires ratification by the other in order to be valid, that is, in order to be
meaningful. This process of ratification is evaluative: It is done
from within ordered complexes of forms-and-meanings in
which appropriateness, social roles, fluency, and other quality
attributes are specified. Thus, even if I think I produce a cogent
story, my interlocutor may judge it to be off the mark because
what I say and how I say it do not qualify as good enough in his/
her evaluative framework. And evidently, such evaluative frameworks are reflections and instruments of the social and political
order (Voloshinov 1973).
This social and political order penetrates language at a fundamental level: It shapes discourses. Discourses are complexes
of communicative forms (genres, styles) mapped onto thematic
and social domains, and what the social and political order does
is to create spaces in which particular discourses operate while it
eliminates other such spaces. This idea is central to the work of
State of Affairs
The political load of language is one of the central concerns
for critical discourse analysis (CDA), an approach to discourse analysis that, especially since the 1990s, explicitly focuses
on the ways in which discourse reflects power and social structure and constructs it (Fairclough 1989, 1992; Blommaert 2005).
It is CDAs stated goal to analyze opaque as well as transparent
structural relationships of dominance, discrimination, power
and control as manifested in language (Wodak 1995, 204) a
paradigmatic choice that is reflected in numerous studies on racism, sexism, media, and political discourse and advertisements.
In all of these, linguistic and textual patterns are analyzed as conduits for hegemony and power abuse, and CDA has been influential in identifying registers and genres of power and control.
CDA clearly subscribes to a view of language as loaded (cf.
Bolinger 1980) and as invested with social and political interests
that steer discourse into particular, structural (i.e., nonarbitrary)
patterns of use and abuse. The influence of Foucault, Gramsci,
Bourdieu, and other critical theorists is explicitly acknowledged
in much CDA work.
The same paradigmatic choice underlies work in what
could be called critical sociolinguistics: an approach in which
the distribution of language in society is also seen as a reflection of power processes, often crystallized in normative (standards) discourses and invariably entailing judgments of users
through judgments of language use (e.g., Milroy and Milroy
1985; Cameron 1995; see standardization). Variation in language speaks to variation in society, and such forms of variation
637
Politics of Language
are evaluated given different value. Institutionalization, such
as, for example, in the education system (Rampton 1995) or in
bureaucracy (Sarangi and Slembrouck 1996) can stabilize and
reify such evaluative patterns and use them as normative, exclusive, and excluding instruments of power and control. Sensitive
social identities, such as gender and immigrant identities, can
be especially vulnerable to exclusion or marginalization in such
reified normative structures.
Both CDA and critical sociolinguistics seek an integration of
the linguistic or discourse-analytic method with social theory,
thus reversing the tendency toward autonomy and disciplinary
recognizability that characterized earlier phases in the development of these disciplines (e.g., Cameron 1992; Chouliaraki and
Fairclough 1999). This move is aimed at strengthening the fundamental theoretical assumption: that language and social structure stand in an intricate relationship to each other and that one
cannot be understood without an understanding of the other.
From another theoretical angle, linguistic anthropology has
significantly contributed to the study of the politics of language.
In contrast to the previous schools, linguistic anthropology has its
roots in an integrated science of human behavior. The anthropological notion of language, consequently, appears easier to integrate into a mature social-theoretical framework than notions of
language that have their feet in twentieth-century linguistic traditions. The fact that language forms and structures need to be
seen as reflective and constructive of sociocultural and political
realities was central to Edward Sapirs work (1921), and the post
World War II reemergence of the ethnography of communication
(Gumperz and Hymes 1972; Gumperz 1982; Hymes 1996) started
from the assumption that there is no occurrence of language that
is not drenched in social, cultural, historical, and political contexts and that, consequently, can be understood without attention to these contexts (Duranti 1997). It is from within linguistic
anthropology that the paradigm of language ideologies developed (Schieffelin, Woolard, and Kroskrity 1998; Blommaert 1999;
Kroskrity 2000; Bauman and Briggs 2003).
Language Ideologies
Language ideologies are beliefs, ideas, views, and perceptions
about language and communication. Such ideational complexes
pertain to every aspect of communication: about linguistic forms
and functions, as well as about the wider behavioral frames (often
called nonlinguistic) in which they occur. Thus, in the field of
language ideologies, people are seen to perform meanings, and
language in the narrow sense of the term is seen as just one mode
of meaning production. People produce semiosis (meaningful
symbolic behavior) as performance, and they do so within a
regimented field in which language ideologies produce stability
and recognizability. Seen from that perspective, language ideologies are of course not just ideational; they are practical in the
sense of Bourdieu, referring to the Marxian praxis, rather than to
the Mannheimian or Durkheimian notion of ideology.
The study of language ideologies emerged out of the Whorfian
concern with connections between language form and world
view (Hill and Mannheim 1992). To recap Benjamin Whorfs
basic idea, he argued that grammatical categories encoded
and thus revealed aspects of collective perceptions of reality;
as such, grammatical organization was not random, logical, or
638
Evaluation
The idea that language is a politically invested object and that
people act politically in, through, and on language is by now a
well-established theoretical frame, the legitimacy of which no
longer requires debate. One reason is the fact that the different
approaches discussed here all have very strong empirical inclinations and that studies documenting the politics of language
often manage to transcend the slogans of a committed social
science and bring theoretical and methodological innovation to
the field. CDA has done much to sensitize discourse analysts at
large about the fact that discourse matters to people because it
Politics of Language
is invested with power and social capital; critical sociolinguistics has likewise drawn attention to the fact that sociolinguistic
distribution is not just a horizontal phenomenon but also a vertical one: Sociolinguistic difference is complemented by sociolinguistic inequality. And from within linguistic anthropology,
we have witnessed the emergence of a powerful ethnographic
paradigm that recovers the holistic and rich agenda developed
earlier by the likes of Sapir and Whorf and applies these insights
to an expanding field of fundamental and applied topics of language in society. The language-ideological approach appears to
be the most promising one because of its compelling theoretical coherence and empirical applicability, and it would benefit
adjacent disciplines if the central insight that every pragmatics of language is accompanied by a metapragmatics would be
adopted.
Jan Blommaert
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bakhtin, M. M. 1981. The Dialogic Imagination: Four Essays.
Austin: University of Texas Press.
Barthes, R. 1957. Mythologies. Paris: Seuil.
Bauman, R., and C. Briggs. 2003. Voices of Modernity: Language Ideologies
and the Politics of Inequality. Cambridge: Cambridge University Press.
Blommaert,
J.
2005.
Discourse:
A
Critical
Introduction.
Cambridge: Cambridge University Press.
Blommaert, J., ed. 1999. Language Ideological Debates. Berlin: Mouton de
Gruyter.
Bolinger, D. 1980. Language: The Loaded Weapon. London: Longman.
Bourdieu, P. 1991. Language and Symbolic Power. Cambridge,
UK: Polity.
Cameron, D. 1992. Feminism and Linguistic Theory. London: Macmillan.
. 1995. Verbal Hygiene. London: Routledge.
Chouliaraki, L., and N. Fairclough. 1999. Discourse in
Late
Modernity:
Rethinking
Critical
Discourse
Analysis.
Edinburgh: Edinburgh University Press.
Duranti, A. 1997. Linguistic Anthropology. Cambridge: Cambridge
University Press.
Fairclough, N. 1989. Language and Power. London: Longman.
. 1992. Discourse and Social Change. Cambridge, UK: Polity.
Foucault, M. 1984. The order of discourse. In Language and Politics, ed.
M. Shapiro, 10838. London: Blackwell.
Gumperz, J. 1982. Discourse Strategies. Cambridge: Cambridge University
Press
Gumperz, J., and D. Hymes, eds. 1972. Directions in Sociolinguistics: The
Ethnography of Communication. New York: Holt, Rinehart and
Winston
Hill, J., and B. Mannheim. 1992. Language and world view. Annual
Review of Anthropology 21: 381406.
Hymes, D. 1996. Ethnography, Linguistics, Narrative Inequality: Toward
an Understanding of Voice. London: Taylor and Francis.
Kroskrity, P., ed. 2000. Regimes of Language. Santa Fe, NM: SAR.
Milroy, J., and L. Milroy. 1985. Authority in Language: Investigating
Language Prescription and Standardisation. London: Routledge.
Rampton, B. 1995. Crossing: Language and Ethnicity among Adolescents.
London: Longman.
Sapir, E. 1921. Language: An Introduction to the Study of Speech. Orlando,
FL: Harcourt Brace.
Sarangi, S., and S. Slembrouck. 1996. Language, Bureaucracy and Social
Control. London: Longman.
Schieffelin, B., K. Woolard, and P. Kroskrity, eds. 1998. Language
Ideologies: Practice and Theory. New York: Oxford University Press.
639
640
We cannot stay on the first-order level here without the independence indicator.
This does not clear up all interpretational problems, however.
We can still ask: What are the relevant possible worlds in different applications? This question is connected with the question as
to what kinds of modalities and other intensional notions there
are.
The characterization of possible worlds as represented by
maximal consistent classes of sentences of a given language has
encouraged the idea that what is intended by possible worlds are
indeed worlds in the sense of entire universes. However, a comparison with probability theory shows that such grandiose interpretations are neither unavoidable nor even preferable. In most
applications of probability theory, the possible worlds (sample
space points) are not worlds in any ordinary sense of the word.
They usually are what might be called scenarios, namely, courses of
events involving a small region of space-time, for example, tosses
of a die. Some probability theorists speak of small worlds, and
practically all realistic applications of possible worlds semantics
are to such small worlds. In some of his work, Montague, in fact,
operates with contexts of use, rather than possible worlds.
There remains the question of different modalities. Are they
all viable in the light of possible world semantics? There are no
major unsolved conceptual problems about epistemic or doxastic modalities or other similar intensional concepts. The class
of alternative epistemic worlds has a clear meaning or at least
as clear a meaning as our language has. Logical (conceptual)
modalities are interpretable only if we look at the structure of
the possible worlds that they involve. If we begin to speak of all
possibilities of individual existence, the alternatives to a given
world do not form a viable class anymore than the set of all sets
in set theory. The idea of natural possibility has a clear sense if it
is taken to mean conformity with natural laws (nomic necessity).
But when it is claimed that there exists a metaphysical necessity
separate from nomic (physical) and conceptual (logical) necessities, it is hard to see what is being meant. It is not enough to
claim that we have intuitions about them, for the very notion of
intuition in its current philosophical use is highly suspect.
Jaakko Hintikka and Risto Hilpinen
WORKS CITED AND SUGGESTIONS FUR FURTHER READING
Carnap, Rudolf. 1947. Meaning and Necessity: A Study in Semantics of
Modal Logic. Chicago: University of Chicago Press. An enlarged edition appeared in 1956.
Copeland, B. Jack. 2002. The genesis of possible worlds semantics.
Journal of Philosophical Logic 31: 99137.
Hintikka, Jaakko. 1957a. Quantifiers in Deontic Logic. Helsinki: Societas
Scientiarum Fennica (Commentationes Humanarum Litterarum, Vol.
13.4).
. 1957b. Modality as referential multiplicity. Ajatus 20: 4964.
. 2003. A second generation epistemic logic and its general significance. In Knowledge Contributors, ed. by V. Hendricks, K. F. Jrgensen,
and S. A. Pedersen, 3356. Dordrecht, the Netherlands: Kluwer
Academic.
Hintikka, Jaakko, and John Symons. 2003. Systems of visual identification and neuroscience: Lessons from epistemic logic. Philosophy of
Science 70: 89104.
Kanger, Stig. 1957. Provability in Logic. Stockholm Studies in Philosophy.
Vol. 1. Stockholm: Almqvist and Wicksell.
Kripke, Saul A. 1959. A completeness theorem in modal logic. Journal
of Symbolic Logic 24: 114.
Montague, Richard. 1974. Formal Philosophy: Selected Papers by
Richard Montague, ed. by Richmond Thomason. New Haven, CT: Yale
University Press.
Partee, Barbara H. 1989. Possible worlds in model-theoretic semantics: A linguistic perspective. In Possible Worlds in Humanities, Arts
and Sciences: Proceedings of Nobel Symposium, ed. by S. Alln, 93123.
Berlin and New York: Walter de Gruyter.
Partee, Barbara H., ed. 1976. Montague Grammar. New York: Academic
Press.
Vaina, Lucia. 1990. What and Where in the Human Visual System: Two
Hierarchies of Visual Modules. Synthese 83: 4991.
641
642
643
644
the interlocutor, rather than for its ability to seek reasons and
motivations, which at that stage are beyond the childs level of
cognitive development.
Not long after children acquire the ability to produce individual communicative acts, they begin to combine them into larger
language acts and events, such as stories, explanations, and
complex observations. Narratives range from the simple statement of a problematic event in the past followed by a solution
(The baby cried; the mummy picked it up) to heavily scaffolded
productions such as the following:
Ross (aged 2;6) and his mother are sitting together eating a
snack.
R: Sometimes Ross come out bed bed come out night.
M: What are you talking about? What about that bed at night?
Sometimes you what the bed at night?
R: Mmm.
M: What did you say?
R: In the dark.
M: In the dark!
R: Ross, erm, Ross runs in the dark.
M: Run in the dark?!
R: Ross runs.
M: You get out of the bed in the night did you and ran around
in the dark. That sounds a daft thing to do! (Foster-Cohen
1990)
Here, Rosss mother helps him get his story out piece by piece
and puts it together for him.
As children develop their story skills, we start seeing complex
depictions with recognizable phases and characteristic packaging of information (Labov and Walestsky 1967). The literature on
childrens narratives has been enhanced by several large-scale
studies, such as that by Ruth Berman and Dan Slobin, whose collection of stories told in response to wordless books about a small
boy and a frog (known, naturally enough, as The Frog Stories)
has provided a cross-linguistic, cross-cultural view of how children develop the ability to tell a story (Berman and Slobin 1994;
Strmqvist and Verhoeven 2004).
Stories and other large discourse units are held together via
the coherence of their informational structure and by the
markers of cohesion that link individual utterances to each other.
The presentation of new information in relation to assumed or
known information is one key aspect of coherence, and requires
children to be able to infer what their interlocutor knows and
to structure the information provided accordingly. As such, the
development of coherence in childrens narratives, and in their
language use generally, is dependent on the evolving understanding of other minds (see theory of mind and language
acquisition). When there is different and conflicting information held by the child narrator and by a protagonist in the story,
it places considerable strain on the young childs pragmatic
competence. An example occurs in the Frog, where are you?
story. At one point, the boy in the story grabs what he believes to
be branches but the narrator knows to be the antlers of a deer.
Children struggle with how to represent this conflicting information, as the following representative samples from Berman and
Slobins (1994) work suggest:
tages and those clinching business deals know, they may hone
their pragmatic skills for the rest of their lives.
645
Pragmatics
646
PRAGMATICS
Pragmatics refers to the study of meaning in context. Consider,
for example, the following exchange between two close friends:
Harvey: Are you going to the big party tonight?
Molly: Didnt you hear that Jason would be there?
Pragmatics
Expressive: The speaker expresses an attitude toward the
propositional content of an utterance, such as Im sad your
wallet was stolen.
Declarative: The speaker performs an action just representing
himself or herself as performing that action, such as We find
the defendant guilty of murder in the first degree.
Philosophers have explored the various social and institutional facts that must hold for an utterance to be faithfully seen
as an example of any of these speech-acts, such as whether an
individual must be capable of fulfilling the act represented in
an utterance for it to be seen as a sincere promise (i.e., that the
speaker actually has $5 to loan and can give this money to the
listener).
Pragmatic theory, however, has paid greater attention in the
last 40 years to the process by which listeners infer what speakers
mean by what they say. Recall the conversation between Harvey
and Molly. Understanding that Mollys comment is meant as
a particular answer to Harveys question requires that Harvey
go through a chain of reasoning regarding Mollys intentions
because her answer does not logically follow from his question. The philosopher H. Paul Grice called the intended message
behind Mollys utterance a conversational implicature,
which is a natural outcome of speakers and listeners tacit
adherence to the cooperative principle. This states that a
speaker must make your conversational contribution such as
is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged
(Grice, 1975: 45). The cooperative principle carries with it four
maxims:
Maxim of Quantity: Make your contribution as informative as
is required, but not more so, for the current purposes of the
exchange.
Maxim of Quality: Do not say anything you believe to be false
or for which you lack adequate evidence.
Maxim of Relation: Say only what is relevant for the current
purposes of the conversation.
Maxim of Manner: Be brief, but avoid ambiguity and obscurity of expression.
Grice noted that speakers do not always uphold these maxims. So long as speakers generally adhere to the overall cooperative principle, they can flout any of these maxims to produce
certain implicatures. For example, Mollys response to Harveys
question flouts the maxim of manner to implicate that she is not
going to the party because of Jason. According to Grices analysis, Harvey would not consider Mollys response to be uncooperative. Instead, Harvey would continue to assume that Mollys
rhetorical response was cooperative and would seek an interpretation given what he assumes about Molly, and what he believes
Molly assumes about him, in order to derive an acceptable and
authorized interpretation.
One place where speakers flout conversational maxims is in
their use of figurative language, such as metaphor (e.g.,Lawyers
are sharks) and irony (e.g., A fine friend you are!). Grices
theory assumes that figurative language is understood in a series
of steps (1975). First, listeners analyze the literal meaning of the
entire expression. Second, they assess whether this literal interpretation is appropriate for the specific context. Third, if the literal
meaning is contextually inappropriate, as is the case for figurative
language, listeners must then derive the intended figurative (e.g.,
metaphorical, ironic) meaning via the cooperative principle. This
view suggests, then, that figurative language should be more difficult to comprehend than corresponding literal speech, because
figurative speech requires an additional processing step in which
the literal meanings are rejected and the intended figurative
meanings are subsequently inferred.
Many pragmatic theories, especially in philosophy, embrace
all or some of the Gricean view of conversational implicature and
his specific proposals on understanding indirect and figurative
language. Indeed, much of the focus in philosophical and linguistic studies on pragmatics is devoted to demonstrating how classic semantic phenomena, such as reference, indexicals, and
demonstratives, can be explained in terms of an understanding of
the specific facts about the speaker, time, and location of an utterance (Kaplan 1989; Stalnaker 1999). But psychological experiments
have raised important questions about Grices theory. Although
there is considerable evidence showing that speakers generally
aim to be cooperative, with talk being primarily organized around
the recovery of speakers pragmatic intentions (Clark 1996; Gibbs
1999), it is less clear that meaning is processed in the serial manner
that Grice and other pragmatists assume. For instance, numerous
psycholinguistic studies indicate that many kinds of figurative language, including novel metaphors, can be understood as quickly
as literal speech when these expressions are encountered in rich
linguistic contexts (Gibbs 1994). Thus, pragmatic knowledge may
be immediately accessed and applied in order to understand what
speakers imply by what they say, without listeners first having to
analyze the literal, semantic meaning of utterances.
A different proposal on the pragmatics of utterance interpretation assumes that speakers aim to be optimally relevant
in saying what they do. Optimizing relevance is a fundamental
tenet of relevance theory (Sperber and Wilson 1995). Under
this optimally relevant view, every act of ostensive behavior
communicates a presumption of its own optimal relevance,
that is, a presumption that it will be relevant enough to warrant the addressees attention and as relevant as compatible
with the communicators own goals and preferences (the communicative principle of relevance). Speakers design their utterances to maximize the number of cognitive effects that listeners
infer, while minimizing the amount of cognitive effort to do so.
Listeners understand speakers communicative intentions via
the relevance-theoretic comprehension procedure (Sperber
and Wilson 2002), by following a path of least effort in computing
cognitive effects. They do this by testing interpretive hypotheses
(e.g., disambiguations, reference resolutions, implicatures) in
order of accessibility, and then stopping when their expectations
of relevance are satisfied.
For example, consider the following exchange between two
university professors (Sperber and Wilson 2002, 19):
Peter: Can we trust John to do as we tell him and defend the
interests of the Linguistics Department in the University
Council?
Mary: John is a soldier!
647
Pragmatics
How does Peter understand Marys metaphorical assertion
about John? Peters mentally represented concept of a soldier
includes many ideas that may be attributed to John. Among
these are a) John is devoted to his duty, b) John willingly follows
orders, c) John does not question authority, d) John identifies
with the goals of his team, e) John is a patriot, f) John earns a
soldiers pay, and g) John is a member of the military. Each of
these ideas may possibly be activated to some degree by Marys
use of soldier in relation to John. However, certain of these
attributes may be particularly accessible given Peters preceding question where he alludes to trust, doing as one is told, and
defending interests. Following the relevance-theoretic comprehension procedure, Peter considers these implications in
order of accessibility, arrives at an interpretation that satisfies
his expectations of relevance at d, and stops there. He does not
even consider further possible implications, such as eg, let
alone evaluate and reject them. In particular, Peter does not
consider g, the literal interpretation of Marys utterance, contrary to what is advanced by the Gricean view, and consistent
with the psychological evidence on inferring metaphorical
meaning.
Relevance theory has also advanced the idea that significant aspects of what speakers say, and not just what they totally
communicate, are deeply dependent upon enriched pragmatic
knowledge. Essentially, the same sorts of inferential processes
used to determine conversational implicatures also enter into
determining what speakers say (Carston 2002; Recanati 2004;
Sperber and Wilson 1995). Consider a case where a speaker says
to you I havent eaten in response to a question about whether
she found time for breakfast that morning. Once the indexical references and the time of the utterance are fixed, the literal meaning of the sentence determines a definite proposition, with a
definite truth condition, which can be expressed as The speaker
has not eaten prior to the time of the utterance. This paraphrase
reflects the minimal proposition expressed by I havent eaten.
However, a speaker of I havent eaten is likely to be communicating not a minimal proposition but some pragmatic expansion
of it, such as I havent eaten today. This possibility suggests
that significant pragmatic knowledge plays a role in enabling
listeners to expand upon the minimal proposition expressed in
order to recover an enriched pragmatic understanding of what
a speaker says. Several experimental studies indicate that pragmatics plays a major role in peoples intuitions of what speakers say (Gibbs and Moise 1997). Thus, the distinction between
what speakers say and imply may possibly be orthogonal to any
distinction between semantics and pragmatics, contrary to the
traditional Gricean view.
The vast number of studies on a wide assortment of linguistic
and nonlinguistic phenomena conducted within the relevance
theory framework makes it the most salient model of pragmatics and utterance interpretation available today. At the very least,
part of relevance theorys significant appeal in interdisciplinary
language studies is its explicit aim to situate pragmatics within
broader concerns of human cognition and communication,
through its embrace of the principles of relevance. Not surprisingly, relevance theory has its critics, ranging from scholars,
primarily in linguistics, who assume that utterance meaning is
determined by heuristics of default or preferred interpretations
648
Pragmatics
cognitive and linguistic processes that underlie their understanding of others pragmatic intentions, unless the attempt to
coordinate fails and leads to misunderstandings.
Not all psychologists agree that speakers and listeners always
aim to be cooperative in conversation by taking the other persons perspective into account during speaking and listening.
Some experiments show, for example, that speakers and listeners can each adopt an egocentric bias as they speak and comprehend, particularly when they experience additional cognitive
load or stress (Horton and Keysar 1996). Speakers also sometimes overestimate how effective they are communicating their
messages to listeners, with listeners also sometimes assuming
that they correctly understood speakers when in fact they did not
(Keysar and Henly 2002). These studies show how there are at
least some systematic sources of misunderstanding attributable
to what might be best characterized as an egocentric bias in communication effectiveness.
The study of pragmatics will undoubtedly continue to have a
strong interdisciplinary flavor in the future. Scholarly intuitions
about how knowledge of the world and context shape utterance
interpretation must be supplemented by experimental studies that examine fast, unconscious cognitive and linguistic processes operating when speaker meaning is understood. We need
to understand not only what pragmatic information shapes contextual meaning but also when that knowledge in recruited in the
psychology of ordinary language interpretation.
Raymond W. Gibbs, Jr. and Gregory A. Bryant
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Austin, John. 1962. How to Do Things with Words. Oxford: Clarendon.
Carston, Robyn. 2002. Thoughts and Utterances: The Pragmatics of Explicit
Communication. Oxford: Blackwell.
Clark, Herbert. 1996. Using Language. New York: Cambridge University
Press.
Clark, Herbert, and Edward Schaeffer. 1987. Concealing meaning from
overhearers. Journal of Memory and Language 26: 20925.
Clark, Herbert, and Deanna Wilkes-Gibbs. 1986. Referring as a collaborative act. Cognition 22: 139.
Giora, Rachel. 1997. Discourse coherence and theory of relevance: Stumbling blocks in search of a unified theory. Journal of
Pragmatics 27: 1734.
Gibbs, Raymond. 1994. The Poetics of Mind: Figurative Thought, Language,
and Understanding. New York: Cambridge University Press.
. 1999. Intentions in the Experience of Meaning. New York: Cambridge
University Press.
Gibbs, Raymond, and Jessica Moise. 1997. Pragmatics in understanding
what is said. Cognition 62: 5174.
Grice, H. Paul. 1975. Logic and conversation. In Syntax and Semantics.
Vol 3: Speech Acts. Ed. Peter Cole and Jerald Morgan, 4158. New
York: Academic Press.
Horn, Larry. 2004. Implicature. In The Handbook of Pragmatics, ed.
Larry Horn and Gregory Ward, 328. Oxford: Blackwell.
Horton, William, and Boaz Keysar. 1996. When do speakers take into
account common ground? Cognition 59: 91117.
Kaplan, David. 1989. Demonstratives. In Themes from Kaplan, ed.
Joseph Almog, John Perry, and Howard Weinstein, 481563. New
York: Oxford University Press.
Keysar, Boaz, and Anne Henly. 2002. Speakers overestimation of their
effectiveness. Psychological Science 13: 20712.
649
650
(laughs) Oh my God.
Oh no. Weird.
Chad:
Peter:
Chad:
No.
Pragmatics, Neuroscience of
PRAGMATICS, NEUROSCIENCE OF
What Is Pragmatics and What Is the Neuroscience of
Pragmatics?
The neuroscience of pragmatics has not been extensively studied. This is understandable, inasmuch as pragmatics came
into focus fairly late as an area of study in linguistics. The field
is not dominated by one specific theoretical framework. There
are controversies concerning its delimitation with respect to,
for example, semantics and nonlinguistic behavior. Finally,
since pragmatic phenomena are crucially related to connected
discourse and communicative interaction, they do not lend
themselves easily to investigation by established experimental
approaches or to such methods as neuroimaging, electroencephalography (EEG), and so on that focus on limited, often
decontextualized, linguistic units. This entry is an attempt to
summarize some typical approaches in relating pragmatic
phenomena to neural processing, and what they have found to
date.
The term pragmatics is used here in accordance with C. W.
Morris (1938), who posited a framework in which syntax deals
with the formal relations among signs, semantics adds the relations of signs to objects, and pragmatics further adds the relations of signs to the interpreter. Pragmatics is about language use
or communication, in a broad sense and in context. It is assumed
here that there is no clear sense in which pragmatics can be separated from semantics. The focus here, however, is on phenomena that are typically considered to be part of pragmatics.
In terms of the relation of pragmatics to neuroscience, different types of approaches have to be considered, since this is far
from a uniform field of study. One set of approaches constitutes
experimental studies of phenomena that are considered to be
651
Pragmatics, Neuroscience of
important in pragmatics. This may involve trying to isolate such
phenomena in an experimental setting that allows for neuroimaging techniques, such as fMRI (functional magnetic resonance
imaging), or EEG/ERP (event-related potentials) to be used
to measure brain activity. Lesion data can be handled either
by experimental group studies, comparing, for example, individuals with left hemisphere damage (LHD), right hemisphere
damage (RHD) and no brain lesion, or by case studies comparing specific phenomena in one individual or a number of
individuals with lesions and communication disorders. Lesion
studies in the area of pragmatics often also include a focus on
communicative interaction in context, involving persons with
brain damage, and use such methods as videorecording of faceto-face interaction, transcription, and coding and microanalysis of sequences and patterns of interaction. These studies also
include more social constructivist empirical approaches, such
as conversation analysis. Lesion studies can also be done with
brain activity measurements, but this is more difficult, especially when homogeneous groups and many repeated events
are needed and when an interaction between two persons is
being studied.
The field of pragmatics primarily needs models of connected
speech, for example, of story structure and topic flow, and models of linguistic communicative interaction, that is, models
involving two (or more) participants and the interactive flow,
co-construction, activation, and so on between them, including
different levels of conscious control. These models can vary considerably in degree of detail and specificity. Typically in studies
of pragmatics, communicative, cognitive, and emotive factors
are included. Multimodality, such as body communication and
prosody in speech, is considered important. Data often consist
of sequences longer than words and sentences monologue and
dialogue samples, for example. Overall structure and the course
of communication are studied, and interactive phenomena are
often in focus. Specific pragmatic phenomena that can be studied from a neuroscience perspective are listed in Table 1.
652
inhibition
selection and ordering of speech, behavior, and logic
formation and execution of plans of action
memory processes (working memory, episodic memory
retrieval, emotional modulation of memory processes and
executive processing, and cues for long-term memory)
theory of mind (ToM), mental inferencing, attribution of
mental states, simulation for comprehension, visuospatial
imagery, abstraction
TYPICAL STUDIES I EEG/ERP. Studies of brain activity using
EEG/ERP involve the correlation of specific temporal components of brain activity with performance. Conversation is not
easily studied with this type of method. One of the most frequently used components is the N400, a negative ERP response
to semantic anomaly.
One of the findings from studies of the N400 is that there
is rapid incremental processing all the way through; in other
words, listeners start early to respond to unfolding words as
influenced by topic, how the speech is produced, and by whom.
Listeners also use discourse information to automatically make
predictions about semantics, syntax, phonology, and referents;
discourse information can overrule local constraints. No evidence of context-free sentence-internal interpretation has been
found, and the conclusion is that the brain does not engage in
a two-step interpretation of semantics and pragmatics. What
seems to be happening is that an initial quick and superficial
Pragmatics, Neuroscience of
Table 1. Possible and typical (x) combinations of pragmatic objects of study and neuroscientific methods
Pragmatic phenomena/methods
Empirical studies of
behavior in context
Experimental studies
EEG/ERP
Inference
Cognitive semantics
Emotion in communication
Conversational principles
Body communication,embodiment,
multimodality
653
Pragmatics, Universals in
It should also be stressed that when fine-grained methods are
used, pragmatic deficits stemming from LH aphasia, traumatic
brain injury, and other brain damage conditions are also found.
Concerning the analysis of face-to-face interaction, the generalizability of results tends to be fairly low. Still, these types of analysis are extremely important to the neuroscience of pragmatics, as
they provide studies of important pragmatic phenomena, which
can also serve as input for further development of theories and
methods. Most of the theoretical claims made on the basis of
brain activity studies today in the area of pragmatics were already
made much earlier on the basis of empirical studies of behavior
following brain damage.
Elisabeth Ahlsn
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Brownell, H., and O. Friedman. 2001. Discourse ability in patients with
unilateral left and right hemisphere brain damage. In Handbook
of Neuropsychology. 2d ed. Vol. 3. Ed. R. S. Berndt, 189203.
Amsterdam: Elsevier.
Gernsbacher, M. A., and M. P. Kashak. 2003. Neuroimaging studies of language production and comprehension. Annual Review of
Psychology 54: 91114.
Mar, R. A. 2004. The neuropsychology of narrative: Story comprehension, story production and their interrelation. Neuropsychologia 42:
141434.
Morris, C. W. 1938. Foundations of the theory of signs. Chicago: Chicago
University Press.
Robertson, D.A., M. Gernsbacher, S. Guidotti, R. Robertson, W. Irwin,
B. Mock, and M. Campana. 2000. Functional neuroanatomy of the
cognitive process of mapping during discourse comprehension.
Psychological Science 11.3: 25560.
van Berkum, J. J. A. 2005. The electrophysiology of discourse and conversation. In The Cambridge Handbook of Psycholinguistics, ed. M.
Spivey, M. Joanisse, and K. McRae. Cambridge: Cambridge University
Press.
Deixis
PRAGMATICS, UNIVERSALS IN
Changing Prospects for Universals in Pragmatics
The term pragmatics has come to denote the study of
general principles of language use. It is usually understood
to contrast with semantics , the study of encoded meaning, and also, by some authors, to contrast with sociolinguistics and the ethnography of speaking, which are more
concerned with local sociocultural practices. Given that pragmaticists come from disciplines as varied as philosophy, sociology, linguistics, communication studies, psychology, and
anthropology, it is not surprising that definitions of pragmatics vary. Nevertheless, most authors agree on a list of topics
that come under the rubric, including deixis, presupposition , implicature (see conversational implicature),
speech-acts, and conversational organization (see conversational analysis ). Here, we can use this extensional
definition as a starting point (Levinson 1988; Huang 2007).
With the rise of generative grammar, and the insistence on universals of grammar (see universal grammar),
anthropologists began to emphasize the diversity of language
use, implicitly accepting the underlying uniformity of grammar
654
The fundamental use of language is in face-to-face conversation, where participants take turns at speaking. Aspects of this
context are built into languages in many detailed ways. All spoken languages have a grammatical category of person, that is,
a grammatical reflection of the different roles that participants
(and nonparticipants) have in an utterance (speaker, addressee,
third party), which is likely to be reflected in personal pronouns, verbal inflections, imperatives, vocatives (as in address
forms), and so forth. Likewise, all languages have at least one
demonstrative, a special form for indicating entities in the context typically, there are contrastive forms (like this and that)
associated with pointing. They also have ways to distinguish the
time and place of speaking (they may not have tense, but they
will have forms denoting now, today, here, etc.). These
aspects of language structure are pragmatic in the sense that
they refer to aspects of the context of utterance, and their interpretation is relative to that context. The peculiarity of these systems is that as speakers alternate, the reference of these terms
also alternates (my I is your you, and my this may be your that),
a fact that children can find difficult when learning a language.
Since artificial languages (logics, programming languages) successfully purge their structures of such items, it is clear that
Pragmatics, Universals in
natural languages could be different and, thus, that deictic organization constitutes a nontrivial universal aspect of language
built for interactive use.
Presupposition
Languages have various ways to foreground and background
information, and this is crucial if the speakers current point
is to be identified. Information that is presumed in the context
(either because it has already been mentioned or is taken for
granted) is typically not asserted but presupposed, and this is
reflected in language structure. The contrast between definite
and indefinite articles, in those languages that have them, is a
simple example: Both The ninth planet has a peculiar orbit and
The ninth planet does not have a peculiar orbit presuppose that
there is a ninth planet. This constancy under negation is often
taken to be a defining property of presupposition it shows
that the presupposed content is not what is being asserted.
Note that unlike what is asserted, presuppositions are defeasible (fall away) in certain contexts, as in If there is one, the ninth
planet must have a peculiar orbit. Many structures have been
identified that signal this presuppositional property: factive
verbs like regret in he regrets publishing it (which presupposes
he did publish it), cleft-sentences like It was the police who hid
the crime (which presupposes that someone hid the crime), or
comparatives like Hes a better golfer than Tiger (which presupposes that Tiger is a golfer). Although this might seem
to be purely a matter of the arbitrary conventions of a single
language, in fact structures with similar semantics also tend
to carry similar presuppositions in other unrelated languages
(Levinson and Annamalai 1992), suggesting that it is properties of the semantic representation that trigger the presuppositional inferences. It is thus possible to make an inventory of
types of structure that tend to universally signal presuppositional content.
Implicature
A conversational implicature is an inference that comes about
by virtue of background assumptions about language use,
interacting closely with the form of what has been said. H.
Paul Grice (1975, 1989) outlined a cooperative principle
instantiated in four such background maxims of use: Speak
the truth (quality), provide enough but not too much information (quantity), be relevant (relevance), and be perspicuous (manner). For example, if A says Have you seen Henk?
and B says His office door is open, we read Bs utterance
as a partial answer (by relevance), which B chooses because
he hasnt seen Henk but wishes to provide information that
is both true (quality) and relevant, and sufficient to be useful (quantity) and clear enough (manner). By virtue of the
assumption that B is following these maxims, Bs utterance
can suggest, or conversationally implicate, in Grices terminology, that Henk is somewhere close by. Despite the fact
that we often have reasons or cultural conventions for being
obscure or economical with the truth (Sacks 1975; Ochs 1976),
such indirect answers seem to be universal, suggesting that
the background assumption of cooperation holds right across
the cultures of the world.
Speech-Acts
The speech acts of questioning, requesting, and stating are found
in conversation in any language, and they have grammatical
repercussions in all language systems for example, in interrogative, imperative, and declarative syntax (Sadock and Zwicky
1985). Languages differ, of course, in how, and the extent to
which, these acts are grammatically coded, but they always are
at least partially reflected in grammar. John Searle (1976) suggested that there are five major kinds of speech-acts: directives (a
class including questions and requests), representatives (including statements), commissives (promising, threatening, offering),
expressives (thanking, apologizing, congratulating, etc.), and
declarations (declaring war, christening, firing, excommunicating, etc.). The types are individuated by different preconditions
and intended effects, known as their felicity conditions.
The broad taxonomy offers plausible universal classes, while
subsuming culture-specific actions like declarations, such as
divorce by announcement in Moslem societies or magical spells
in a Melanesian society.
Despite the fact that there is an association between, for
example, interrogative form and questioning, the link between
form and action performed is often complex. In English, for
example, requests are rarely done in the imperative, but typically in the interrogative, as in Can you help me get this suitcase
down? It has been noticed that if a distinctive felicity condition for a successful request is stated or requested, this will itself
serve as a request (the addressee being able to get the suitcase
down being a precondition to a felicitous request). This seems
to have general cross-linguistic application, suggesting that the
action performed is in fact implicated by what is said (Brown
and Levinson 1987, 136 ff). However, in many cases, less regular strategies link what is said to the actions performed, and the
mapping from utterances to actions remains a serious theoretical problem in pragmatics.
Conversation Structure
The organization of conversation seems likely to provide some
of the most robust pragmatic universals. As far as we know,
in all societies the most informal type of talk involves rapid
655
Pragmatics, Universals in
alternation of speaking roles (Sacks, Schegloff, and Jefferson
1974). This turn-taking, of course, motivates the deictic system
already mentioned. Such informal talk is also characterized by
the immediacy of conversational repair; that is, if addressees do not hear or understand what is said, they may query either
the whole or part, getting immediate feedback in the next turn
(Schegloff, Jefferson, and Sacks 1977). Such talk is structured
locally in terms of sequences (Schegloff 2006) in the simplest
case, adjacency pairs, that is, pairs of utterances performing actions like question-answer, offer-acceptance, requestcompliance, greeting-greeting, and so forth. Sequences can be
embedded, as in A: Do you have Marlboros? B: You want
20s? A: Yes. B: Ah sorry, no. We do have 10s. They can also
be extended over more turns, for example by adding a presequence as in: A: Do you mind if I ask you something? B: No.
A: Why did you give up that amazing job? B: Burnout. Given
the general expectation for rapid turn-taking, any participant
wishing to have an extended turn at talk is likely to negotiate this, for example, through a prestory of the kind Have you
heard what happened to Bonny? During such an extended turn
at talk, feedback of restricted types (mmhm, uhuh, etc.) may be
expected. In addition to these local levels of organization, conversations also generally have overall structures for example,
they are likely to be initiated by greetings and ended with partings, each with its distinctive structure.
All of this detailed structure seems entirely general across
cultures and languages, although there may be constraints of
many local kinds about who can talk to whom and where in
this informal way. Ethnographic reports to the contrary do not
seem to stand the test of close examination. There are, though,
many aspects of cultural patterning that can be very distinctive.
For example, although in all cultures conversation makes use
of multimodal signals (gaze, gesture, facial expression, etc.) in
face-to-face interaction, the details can differ strikingly, whereas
Tzeltal speakers avoid gaze and the signals that would be thus
made available, Rossel Islanders presume mutual gaze and so
can systematically signal responses like yes, no, amazing!
and so on by facial expression.
In addition to these general observations about conversational universals, there seem to be very detailed generalizations
about specific actions. For instance, in a wide sample of languages, it seems that reference to persons follows a precise set
of expectations about the form of reference expressions, as well
as the procedures to follow when the expression proves inadequate (Stivers and Enfield 2007). Thus, utterances of the following kind, where specific components are added incrementally
and in order until recognition is signaled, can be expected
in any language: John (.) Wilkins (.) The man you met at the
party.
656
Pragmatics, Universals in
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Brown, Penelope, and Stephen C. Levinson. 1987. Politeness: Universals
in Language Usage. Studies in Interactional Sociolinguistics 4.
Cambridge: Cambridge University Press.
Enfield, Nick J., and Stephen C. Levinson, eds. 2006. Roots of Human
Sociality: Culture, Cognition and Human Interaction. Oxford:
Berg.
Goldin-Meadow, Susan. 2003. The Resilience of Language: What Gesture
Creation in Deaf Children Can Tell Us About How All Children Learn
Language. New York: Psychology Press.
Grice, H. P. 1957. Meaning. Philosophical Review 67: 37788.
. 1975. Logic and conversation. In Syntax and Semantics. Vol.
3: Speech Acts. Ed. P. Cole and J. Morgan, 4158. New York: Academic
Press.
. 1989. Studies in the Way of Words. Cambridge: Harvard University
Press
Haspelmath, Martin, Matthew Dryer, David Gil, and Bernard Comrie,
eds. 2005. The World Atlas of Language Structures. Oxford: Oxford
University Press.
Hockett, C. F. 1960. The origin of speech. Scientific American
203: 8996.
Horn, Laurence. 1984. Toward a new taxonomy for pragmatic inference: Q- and R-based implicature. In Meaning, Form and Use in
Context, ed. Deborah Shiffrin, 1142. Washington, DC: Georgetown
University Press.
Huang, Yan. 2007. Pragmatics. Oxford: Oxford University Press.
Hymes, Dell. 1982. Models of the interaction of language and social life.
In Directions in Sociolinguistics, ed. John J. Gumperz and Dell Hymes,
3571. New York: Holt, Rinehart and Winston.
Jackendoff, Ray. 2002. Foundations of Language. Oxford: Oxford
University Press.
Levinson, Stephen. 1983. Pragmatics. Cambridge: Cambridge University
Press.
. 2000. Presumptive Meanings: The Theory of Generalized
Conversational Implicature. Cambridge, MA: MIT Press.
. 2006. On the human interaction engine. In Roots of Human
Sociality, ed. N. Enfield and S. Levinson, 3969. Oxford: Berg.
Levinson, Stephen C., and E. Annamalai 1992. Why presuppositions
arent conventional. In Language and Text: Studies in Honour of Ashok
R. Kelkar, ed. R. N. Srivastava, Suresh Kumar, K. K. Goswami, and
R. V. Dhongde, 22742. Dehli: Kalinga Publications.
Newmeyer, Fredrick. 2004. Typological evidence and universal grammar. Studies in Language 28: 52748.
Ochs, Elinor. 1976. The universality of conversational postulates.
Language in Society 5.1/3: 6780.
Sacks, Harvey. 1975. Everyone has to lie. In Sociocultural Dimensions
of Language Use, ed. M. Sanches and B. Blount, 5780. New
York: Academic Press.
Sacks, Harvey, Emanuel Schegloff, and Gail Jefferson. 1974. A simplest systematics for the organization of turn-taking in conversation.
Language 50: 696735.
Sadock, Jerrold, and Arnold Zwicky. 1985. Speech act distinctions in
syntax. In Language Typology and Syntactic Description. Vol. l: Clause
Structure. Ed. Timothy Shopen, 15596. Cambridge: Cambridge
University Press.
Schegloff, Emanuel. 2006. Sequence Organization in Interaction: A
Primer in Conversation Analysis. Cambridge: Cambridge University
Press.
Schegloff, Emanuel, Gail Jefferson, and Harvey Sacks. 1977. The preference for selfcorrection in the organization of repair in conversation.
Language 53: 36182.
Searle, John. 1976. Speech Acts: An Essay in the Philosophy of Language.
Cambridge: Cambridge University Press.
657
658
Preference Rules
PREFERENCE RULES
Preference rule systems constitute a form of rule interaction
related to default logic and harmonic grammar (Smolensky and
Legendre 2006). They are introduced in semantic theory in
Jackendoff (1983) and in generative music theory in Lerdahl and
Jackendoff (1983) and argued to be ubiquitous in cognition.
A standard example is the meaning of the verb climb. A stereotypical case such as John climbed for hours is interpreted as John a)
moving upwards on a surface, b) with an effortful clambering manner of motion. Both conditions are violable. John climbed down the
mountain and John climbed across the cliff do not involve upward
motion; the airplane climbed steadily entails upward motion but
not clambering. However, both conditions cannot be violated at
once: *The airplane climbed down 5,000 feet.
These examples make it impossible to analyze the meaning
of climb in terms of necessary and sufficient conditions,
as assumed in the philosophical and formal logic traditions.
Neither condition is necessary, but either one is sufficient for
an action to count as climbing. At first blush, this suggests that
the conditions are simply disjunctive. However, there is a further
wrinkle: Satisfying both constraints results in a more stereotypical use of climb, and in cases where there is no evidence to the
contrary, both conditions are assumed by default. Thus, preference rule systems provide a formal characterization of Ludwig
Wittgensteins (1953) and E. Rosch and C. Merviss (1975) notion
of categories displaying a family resemblance: There is no
single criterial condition for members of the category, stereotypical members satisfy all or most conditions, and marginal members satisfy fewer conditions.
Preference rule systems differ from optimality-theoretic
rule systems in that the constraints, though violable, are not
ranked: Under proper conditions, either rule can dominate
the other. A classical example comes from gestalt principles of
visual grouping (Wertheimer [1924] 1938), where grouping of
units can be based either on their relative distance (1a) or their
relative similarity (1b). Thus, either condition is sufficient for
grouping:
(1) a. x x x x x x x x x
b. x x xX X Xx x x
Anat Ninio
Ray Jackendoff
(2)
a. x x x X X X x x x
b. x x X X X x x x x
x xxx
c. x x
X XX
[stronger judgment]
[distance overrules size]
[size overrules distance]
659
Prestige
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Jackendoff, R. 1983. Semantics and Cognition. Cambridge, MA: MIT
Press.
Lerdahl, F., and R. Jackendoff. 1983. A Generative Theory of Tonal Music.
Cambridge, MA: MIT Press.
Rosch, E., and C. Mervis. 1975. Family resemblances: Studies in the
internal structure of categories. Cognitive Psychology 7: 573605.
Smolensky, P., and G. Legendre. 2006. The Harmonic Mind. Cambridge,
MA: MIT Press.
Wertheimer, M. [1924] 1938. Laws of organization in perceptual
forms. In A Source Book of Gestalt Psychology, ed. W. D. Ellis, 7188.
London: Routledge and Kegan Paul.
Wittgenstein, L. 1953. Philosophical Investigations. Oxford: Blackwell.
Presupposition
or what people think its record to have been (1989, 4), not only
describing the status quo but also including a diachronic dimension (see synchrony and diachrony). While prestige is
mostly associated with speaker evaluations, its role as a catalyst
for other characteristics, such as functional specialization, literary heritage, and standardization, has also been acknowledged. In a dynamic model of language prestige and prestige
change, Susanne Mhleisen (2002) looks at the interaction
among societal, institutional and interactional, and sociopsychological dimensions of prestige. The dynamics of this interaction may result in various types and directions of changes in
language prestige.
Susanne Mhleisen
PRESTIGE
660
PRESUPPOSITION
A presupposition is a precondition of a sentence such that the
sentence cannot be uttered meaningfully unless the presupposition is satisfied. The concept of a presupposition originated
with Gottlob Frege (1892), but the English term was coined by
Peter F. Strawson (1950). Presupposition theory is an area of
active research at the semantics/pragmatics interface. A
related term is conventional implicature. H. Paul Grice
(1975) distinguished between presuppositions and conventional implicatures, however, it is still under debate whether
such a distinction is necessary (cf. Potts 2007 and ensuing
discussion).
definite descriptions have played a major role in the
development of presupposition theory and are still generally
analyzed as introducing a presupposition. Consider for example
(1). No entity that satisfies the description biggest natural number exists. What is the status of (1)? Is it true or false?
(1)
Presupposition
existence of an individual that satisfies NP in other words,
definite descriptions carry an existence presupposition.
Presupposition failure describes the case wherein a presupposition is not fulfilled like (1). Presupposition failures are analyzed as being neither true nor false, but as being truth value
gaps. Presupposition theory, therefore, relies on a distinction
among three possible truth values a sentence may have: true,
false, and undefined. One important argument in support of
a third truth value has been the interaction between negation
and presuppositions: A presupposition failure in many cases
remains a presupposition failure even when the sentence is
negated:
(2)
It follows that (2), like (1), is a presupposition failure just if negation does not change the conditions under which a sentence
has a truth value. Negation can be used in this way as a presupposition test: A presupposition follows from a sentence and its
negation. The assertion, on the other hand, only follows from the
sentence itself, and not from its negation.
Just as the existence presupposition of the sentences in (1)
and (2) is triggered by the definite article the, many other words
trigger presuppositions. Stephen C. Levinsons (1983) textbook
lists several pages of presupposition triggers in English. A particularly interesting paradigm is that in (3) (cf. Abusch 2005): (3a)
has no relevant lexically triggered presupposition, whereas (3b)
presupposes that it is actually raining outside and asserts that
Bill thinks so, too. Finally, (3c) presupposes that Bill thinks that
it is raining outside, and asserts that it actually is raining outside.
It is particularly interesting that be right and know have the same
truth conditions, but differ on which part of them is presupposed. Paradigm (3) shows that part of our specific knowledge
about think, know, and be right is whether they trigger a presupposition and which one.
(3)
Some presuppositions are not lexically triggered. For example, (3a) cannot be used if it is known that it really is raining
outside. This presupposition, however, has been analyzed as an
implicated presupposition (Sauerland 2008). It can be derived in
a similar way to conversational implicatures as arising from the
avoidance of a presupposition trigger and a principle of presupposition maximization (Heim 1991).
One central problem of presupposition theory is the question
of how to predict the presuppositions of complex sentences the
problem of presupposition projection. Lauri Karttunen and S.
Peters (1979) show that while negation does not affect presuppositions, in other complex sentences presupposition triggers can
occur, but the presupposition may not project to the entire sentence: In example (4), the conditional clause blocks projection of
the existence presupposition of the biggest natural number.
(4)
661
662
663
664
Priming, Semantic
Romanski, L. M., J. F. Bates, and P. S. Goldman-Rakic. 1999. Auditory
belt and parabelt projections to the prefrontal cortex in the rhesus
monkey. J Comp Neurol 403.2: 14157.
Seyfarth, R. M., D. L. Cheney, and P. Marler. 1980. Monkey responses
to three different alarm calls: Evidence of predator classification and
semantic communication. Science 210.4471: 8013.
Tian, B., D. Reser, A. Durham, A. Kustov, and J. P. Rauschecker. 2001.
Functional specialization in rhesus monkey auditory cortex. Science
292: 2903.
Wang, X., M. M. Merzenich, R. E. Beitel, and C. E. Schreiner. 1995.
Representation of a species-specific vocalization in the primary auditory cortex of the common marmoset: Temporal and spectral characteristics. J Neurophysiol 74.6: 26852706.
Waser, P. M. 1977. Individual recognition, intragroup cohesion, and
intergroup spacing: Evidence from sound playback to forest monkeys.
Behaviour 60: 2874.
. 1982. The evolution of male loud calls among mangabeys
and baboons. In Primate communication, ed. C. T. Snowdon,
C. H. Brown, and M. R. Petersen. New York: Cambridge University
Press.
Zuberbuhler, K. 2000. Referential labelling in Diana monkeys. Anim
Behav 59: 91727.
. 2002. A syntactic rule in forest monkey communication. Anim
Behav 63: 2939.
PRIMING, SEMANTIC
priming is used to describe a situation with two words (or other
entities) that are related, whereby an encounter with the one will
either facilitate or inhibit recovery of the other, either as regards
the speed or the accuracy with which it is recovered (see also
spreading activation). There are various kinds of priming,
of which semantic priming is perhaps the most discussed in the
psycholinguistic literature. An example of semantic priming
is that the word dog will be processed more quickly if the word
bark has just been encountered. If the two words are unrelated
or an encounter with a non-word precedes the encounter with
a word, an encounter with the former will have no effect on the
speed with which the latter is processed; thus, the processing
of dog will be unaffected if either teaching or priff has just been
encountered. This is automatic and independent of any intention or task-related motivation. When processing is affected,
the affected word is known as the target; the word or entity that
affects the process is known as the prime. Primes do not have to
be words; they can be groups of words or complete sentences, or
they can be pictorial or aural. The prime is assumed to partially
activate circuits that include the target.
The assumption is that the degree of priming is proportional
to the semantic relatedness of the items (though this is not accurate for all types of priming) and that priming is an automatic
process. Another assumption, which gives rise to the assumption of proportionality, is that related words are stored closer
in semantic space to one another than are nonrelated words
(non-words are, of course, stored nowhere prior to encounter).
Indeed, one of the major values of priming studies is that they
permit investigation of the interconnections of the mental lexicon (Rumelhart and Norman 1985). Interpreting the results of
such studies, though, is not unproblematic; Kenneth I. Forster
(1976) shows that the facilitation of processing is affected in
Priming, Semantic
complex ways by whether the prime and target are low frequency
or high frequency (or combinations of these).
Different kinds of priming have different kinds of effect. If the
prime is formally similar but semantically unrelated to the target
(e.g., from and frog), the prime inhibits recognition of the target,
presumably because one is momentarily mistaken for the other
in the recognition process (Colombo 1986). Where the prime
and the target are drawn from the same semantic set (e.g., dog
and cat) or are otherwise related semantically, as in the dog-bark
example, the target is retrieved more quickly. This phenomenon
was first demonstrated (though not first noticed) by D. E. Meyer
and R. W. Schvaneveldt (1971) and has become a cornerstone of
psycholinguistic methodology.
Semantic priming automatically accelerates the recognition process if the time between sight of the prime and sight
of the target is short, but if the gap is greater, other factors
may impact. In particular, even if the prime and the target
are semantically related, there may be inhibition rather than
acceleration if the informant has been led to expect something else to occur (Neely 1977), though this inhibitory effect
does not occur if there is a very short time between prime and
target display (technically known as the stimulus-onset asynchrony, or SOA). This leads one to conclude that there are two
types of priming, one of which is automatic and short-lived
and the other of which attentional priming is longer lasting and more available for conscious inspection. The former
can be interfered with but will persist in the speaker on subsequent occasions; the latter can be created and equally can
disappear. So in an influential experiment, Neely (1977) told
informants that whenever they were primed with body, they
should expect a word associated with buildings to be the target. He then sometimes gave a word associated with the body
as target instead. He found that with an extremely short SOA,
both heart and door were processed equally quickly after use
of body as a prime. With a slightly longer SOA, however, the
processing of the unexpected heart actually took longer than
the (in the circumstances) expected door. Outside the context
of the experiment, though, there are no reasons to suppose
that the body-heart connection would be affected, and equally
it seems unlikely that the attentional priming would persist
much beyond the end of the experiment.
Successful processing has an effect here, too, and further
confirms the existence of attentional priming as a separate kind
of priming from automatic priming. If informants find that the
primes they are given relate regularly and reliably to the targets,
the effect of the priming gets increasingly strong. If, on the other
hand, they find little connection between the primes and the targets, the priming effect gets weaker (Den Heyer 1985)
It is important to note that priming of either the automatic
or attentional kind presupposes a causeeffect relationship: The
prime affects the target. However, it is implicit in this causal relationship that the word that is the target has some prior mental
connection with the word (or other linguistic, or indeed nonlinguistic, entity) that is chosen as prime, and it is this prior connection that the prime activates, whether the connection takes
the form of the words being stored near to each other because of
their semantic or pragmatic proximity or because they regularly
co-occur for some reason. An exploitation of the latter kind of
665
Priming, Semantic
of the set in conjunction with bark. Thus, the mind stores several instances of the Alsatian barked and of the poodle barked, as
well of course the dog barked, and from this creates an association of bark with all types of dog. This set remains sound until
conflicting evidence is encountered. So Chihuahua may never
be encountered with barked but with yelped. The conflict persists
until the speaker modifies the original priming or treats the new
priming as either an exception or an anomaly. So, for instance,
bark might be placed in a semantic set of nonverbal noises, which
would include growl as well as bark and yelp. According to Hoey,
the same processes result in the establishment for the speaker
of quasi-grammatical relations associated with the lexical item
(colligational relations; Hoey 2005; Sinclair 2004), and grammar
is argued to be an output from the primings, rather than having
an existence independent of them.
Hoey (2005) also claims that speakers are primed to associate
words with textual position and cohesive patterning. All primings are assumed to be genre and domain specific, and this is
particularly noticeable of the textual primings. Thus, for example, British newspaper readers are primed by British news stories
to associate yesterday with text-initial sentences and to associate
Mr with first-word position in paragraph-initial sentences. The
words Blair and Bush are primed for most readers to be followed
cohesively by pro-forms (he, his), while Pluto is primed to be followed by co-hyponyms (Neptune, Saturn, etc.).
The term priming is used by Hoey to describe both the process and the product of that process. The association created
by priming may itself be subject to priming (which Hoey terms
nesting). Thus, we are primed to collocate Bush with George, and
then to collocate the nested pair George Bush (with or without W)
with President. As a further step, we are then primed to associate
(President) (George)(W) Bush as a combination with pronominal
cohesion. No assumptions are made, however, about the order
in which the primings may occur. Primings necessarily vary from
individual to individual.
The accumulation of collocational, colligational, and semantic relationships may explain linguistically the efficacy of semantic primings, though experimental evidence has not yet been
offered in support of this claim. The phenomenon of semantic
priming would appear to have no effect in giving rise to the lexical primings as described by Hoey, but the existence of semantic priming is confirmation of the efficacy of the prior lexical
primings.
Michael Hoey
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Colombo, Lucia. 1986. Activation and inhibition with orthographically
similar words. Journal of Experimental Psychology: Human Perception
and Performance 12: 22634.
Den Heyer, Ken. 1985. On the nature of the proportion effect in semantic
priming. Acta Psychologica 60: 2538.
Forster, Kenneth I. 1976. Accessing the mental lexicon. In New
Approaches to Language Mechanisms, ed. R. J. Wales and E. Walker,
25787. Amsterdam: North-Holland.
Harley, Trevor. 2001. The Psychology of Language: From Data to Theory.
2d ed. Hove and New York: Psychology Press. Although the information on priming is spread around the text, this is an ideal and accessible
introduction to the different kinds of priming.
666
the lexically specified theta-roles of a predicate and its argument expressions in the syntactic representation. As for structures derived by transformations, movement rules are reduced
to a single and maximally general operation Move that can
move anything anywhere. A representational filter that limits the application of Move is the empty category principle
(ECP), which demands that traces of movement be licensed
under a local structural relation called government. Apart from
the ECP and the bounding theory, which places an upper
bound on how far movement can take an element, various other
modules of UG, not narrowly geared to cut down the overgeneration of structures resulting from Move , act to filter the output representations produced by movements. case theory
requires that (phonetically overt) noun phrases (NPs) occupy a
position at surface structure where they are assigned a case. The
three principles of binding theory (which constrain the distribution of anaphors, personal pronouns, and referential NPs,
respectively, relative to potential antecedents they can/cannot
be coreferential with) are sensitive to the binary [ anaphoric]
and [ pronominal] features of NP categories generally, including phonetically empty NPs like various types of traces and null
pronouns.
The modular organization itself, that is, the dissociation of
various aspects of syntactic phenomena for the purposes of the
grammar, is what makes it possible to keep principles of UG maximally simple. The cohesion of each module is supplied by some
notion and/or formal relation on which its principles are centered. The whole of the grammatical system is also characterized
by unifying concepts, most notably the notion of government,
which plays a key role in a variety of modules. The components
interact in complex ways to restrict the massive overgeneration
of syntactic expressions that would otherwise result from the
fundamental freedom of possible basic phrase structures and
transformations applied to them, which ultimately yields the
actual set of well-formed expressions.
The modularity of the different (sets of) principles is due
not only to the dissociation of the properties relevant to them
but also to the stipulation of distinctions with regard to where
in the grammar they apply. According to GB theory, each sentence corresponds to a sequence of representations, starting
from D-structure (or deep structure, DS), proceeding through
S-structure (or surface structure, SS) to the final representation
called logical form (LF), where adjacent representations are
related by transformations. The derivation from DS to SS feeds
phonetic realization, in particular the mapping from SS to phonetic form (PF) (it is overt), whereas the derivation from SS to LF
does not (it is covert). A principle can apply to transformations
(like bounding theory), or to one or more of the three syntactic representational levels DS, SS, and LF (these constraints are
filters), though not to any intermediate representation. Figure 1
depicts this so-called Y- or T-model of GB, tagged to indicate
where the most prominent modules apply.
Modularity
In the government and binding model of the P&P approach
(Chomsky 1981), principles of UG are organized into modules,
or subtheories. Such modules include x-bar theory, which
constrains possible phrase structure configurations, and
theta theory, which determines a bi-unique mapping between
Parameters
UG, as a model of language competence, includes the principles
along with the locus of their application, as well as the primitive syntactic objects (e.g., labels distinguishing full phrases,
heads of phrases, and intermediate-level categories), relations
667
Theta Theory
X - bar Theory
Lexicon
DS
Bounding Theory
overt transformations
Theta Theory
Case Theory
Theta Theory
ECP
Binding Theory
SS
LF
covert transformations
Figure 1.
PF
668
Conceptual interpretation
Full Interpretation
principles of computational economy
Lexicon
Full Interpretation
Phonetic interpretation
Figure 2.
(applying them as soon as their respective trigger is merged in)
and that has multiple transfers. Derivational sequences between
two transfer points are called phases. The basic architecture is
shown in Figure 2.
Finally, grammatical components are reduced as well. First of
all, there are no distinct phrase structure and transformational
components, as both basic phrase structure and movements are
brought about by the operation merge: While basic structure
building involves merging two distinct elements, movement
involves (re)merging an element with a constituent that contains it. In addition, the burden of description carried by modules of GB is partly reallocated to syntax-external components,
and is partly redistributed among the residual factors that can
enter syntactic explanation: the principal constraint imposed by
the interface components (full interpretation), the character of
the syntactic derivation (multiple transfers, principles of computational economy, the nature of basic syntactic operations,
etc.), and the properties of lexical items. For instance, much of
the binding theory of UG is reduced to movement operations
and rules of interpretation, case theory is recast in more general
terms and is subsumed in a broader account of triggers for movements (called checking theory), and bounding theory is deduced
from the multiple transfers nature of the derivation.
Conclusion
The fundamental question pursued by the P&P framework is
whether it is possible to construct an explanatorily adequate
theory of natural language grammar based on general principles. Two further ambitions of P&P, gaining prominence with
the advent of its minimalist research program, are to find out
whether the primitive notions and principles of such a model are
characterized by a certain degree of naturalness, simplicity, and
nonredundancy, and concurrently, whether some properties of
the language faculty can be explained in terms of design considerations pertaining to computational cognitive subsystems in
general, or even more broadly, in terms of laws of nature. Should
it turn out that these questions are answered in the affirmative (as
some initial results suggest), that would be a surprising empirical discovery about an apparently complex biological subsystem
(cf. biolinguistics): in the case at hand, the human language
faculty. The exploration of the ways in which general laws of
nature might enter linguistic explanation has barely begun.
Clearly, most of the work lies ahead.
Balzs Surnyi
669
670
a child acquire the verb-particle construction significantly earlier than compounding. This prediction received strong support
from a longitudinal study of 10 children.
This example illustrates how the investigation of language
acquisition and the investigation of mature grammars can
be mutually reinforcing activities within the P&P framework.
Another example is provided by the work of Diane Lillo-Martin
and Ronice Mller de Quadros (2005), who considered the parametric prerequisites for the different types of wh-questions in
American Sign Language (ASL), according to two competing
syntactic analyses. The two analyses yielded distinct predictions
about the time course of acquisition, which were then successfully tested against longitudinal data from children acquiring
ASL.
Areas of Debate
We mention here two areas of debate within the P&P approach
to child language acquisition, and of course there are others.
1) What types of parameters, exactly, is the child required to set?
2) What are the observable consequences of an unset or misset
parameter?
One point of disagreement in the P&P literature quite generally, including the acquisition literature, concerns the proper
conception of parameters. A classic conception, which Noam
Chomsky (1986, 146) attributes to James Higginbotham, is the
switchbox metaphor: Each parameter is like an electrical switch,
with a small number of possible settings.
Yet this is only one of many possible ways that parameters
could work. A radically different conception is found in optimality theory, which posits a universal set of violable constraints.
Instead of choosing particular settings for switches in a switchbox, the learner has to rank the constraints correctly. The result
is a narrowly restricted set of options for the target grammar,
as required by the P&P framework. (Indeed, on the mathematical equivalence of a constraint ranking to a set of switchboxstyle dominance parameters, see Tesar and Smolensky 2005,
456.)
Still another approach to parameters is to connect them to the
lexicon. (See lexical learning hypothesis.) This is conceptually attractive because the lexicon is independently needed as
a repository of information that varies across languages. Exactly
what it means to connect parameters to the lexicon, however,
has been open to interpretation.
One idea is to connect points of abstract grammatical (e.g.,
syntactic) variation to the paradigms of inflectional morphology.
The idea is that paradigmatic morphology has to be stored in the
lexicon anyway, and might provide a way to encode parametric
choices. This approach can be found in Borer (1984) and LilloMartin (1991), for example. A related idea is to encode parametric
choices in the morphology of closed-class lexical items. A good
example is Picas (1984) proposal to derive cross-linguistic variation in the binding domain of a reflexive pronoun from the pronouns morphological shape. A variant of Pierre Picas approach
is to encode parametric choices as abstract (rather than morphologically overt) properties of individual lexical items. This is the
lexical parameterization hypothesis of Wexler and Rita Manzini
(1987), who took this approach to cross-linguistic variation in the
binding domain for both reflexives and pronominals.
671
672
Conclusion
Language acquisition is a rich source of evidence about both the
principles and the parameters of the human language faculty.
For this reason, research on language acquisition plays a central
role in the P&P framework.
William Snyder and Diane Lillo-Martin
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Borer, Hagit. 1984. Parametric Syntax: Case Studies in Semitic and
Romance Languages. Dordrecht, the Netherlands: Foris.
Borer, Hagit, and Ken Wexler. 1992. Bi-unique relations and the maturation of grammatical principles. Natural Language and Linguistic
Theory 10: 14789.
Brown, Roger, and Camille Hanlon. 1970. Derivational complexity and
order of acquisition in child speech. In Cognition and the Development
of Language, ed. John R. Hayes, 155207. New York: Wiley.
Chomsky, Noam. 1981. Lectures on Government and Binding. Dordrecht,
the Netherlands: Foris.
. 1986. Knowledge of Language: Its Nature, Origin, and Use. New
York: Praeger.
. 1995. The Minimalist Program. Cambridge, MA: MIT Press.
. 2001. Derivation by phase. In Ken Hale: A Life in Language, ed.
Michael Kenstowicz, 152. Cambridge, MA: MIT Press.
Crain, Stephen, and Diane Lillo-Martin. 1999. Linguistic Theory and
Language Acquisition. Cambridge, MA: Blackwell.
Crain, Stephen, and Mineharu Nakayama. 1987. Structure dependency
in grammar formation. Language 63: 52243
Gibson, Edward, and Kenneth Wexler. 1994. Triggers. Linguistic Inquiry
25: 355407.
Hyams, Nina. 1986. Language Acquisition and the Theory of Parameters.
Dordrecht, the Netherlands: Reidel.
Kaye, Jonathan D., Jean Lowenstamm, and Jean-Roger Vergnaud. 1990.
Constituent structure and government in phonology. Phonology
7: 193231.
Lenneberg, Eric H. 1967. Biological Foundations of Language. New
York: Wiley.
Print Culture
Lillo-Martin, Diane C. 1991. Universal Grammar and American Sign
Language: Setting the Null Argument Parameters. Dordrecht, the
Netherlands: Kluwer.
Lillo-Martin, Diane, and Ronice Mller de Quadros. 2005. The acquisition of focus constructions in American Sign Language and Lngua
de Sinais Brasileira. In BUCLD 29 Proceedings, ed. Alejna Brugos,
Manuella R. Clark-Cotton, and Seungwan Ha, 36575. Somerville,
MA: Cascadilla Press.
Marcus, Gary F. 1993. Negative evidence in language acquisition.
Cognition 46: 5385.
Pica, Pierre. 1984. On the distinction between argumental and nonargumental anaphors. In Sentential Complementation, ed. Wim de Geest
and Yvan Putseys, 18594. Dordrecht, the Netherlands: Foris.
Prince, Alan, and Paul Smolensky. 2004. Optimality Theory: Constraint
Interaction in Generative Grammar. Malden, MA: Blackwell.
Snyder, William. 2001. On the nature of syntactic variation: Evidence
from complex predicates and complex word-formation. Language
77: 32442.
. 2007. Child Language: The Parametric Approach. Oxford: Oxford
University Press.
Tesar, Bruce, and Paul Smolensky. 2000. Learnability in Optimality
Theory. Cambridge, MA: MIT Press.
Thornton, Rosalind, and Stephen Crain. 1994. Successful cyclic movement. In Language Acquisition Studies in Generative Grammar, ed.
Teun Hoekstra and Bonnie D. Schwartz, 21553. Amsterdam: John
Benjamins.
Valian, Virginia. 1991. Syntactic subjects in the early speech of American
and Italian children. Cognition 40: 2181.
Wexler, Kenneth, and Rita Manzini. 1987. Parameters and learnability in
binding theory. In Parameter Setting, ed. Thomas Roeper and Edwin
Williams, 4176. Dordrecht, the Netherlands: Reidel.
Yang, Charles D. 2002. Knowledge and Learning in Natural Language.
Oxford: Oxford University Press.
PRINT CULTURE
Print (or typographic) culture designates all the activities entailed
in producing, distributing, collecting, and reading printed materials and engraved images. As a historical construct, it usually
refers to the literary environment that first emerged in Western
Europe during the second half of the fifteenth century. Diverse
developments elsewhere (such as the use of xylography in China
and of movable type in Korea, prohibitions against Arabic printing by Ottoman rulers, and the sluggish pace of Russian printing)
lend themselves to comparative study but cannot be covered
here.
In Western Europe (unlike other areas), the printing arts, once
introduced, spread with remarkable rapidity. Between the 1460s
and 1490s, printing shops were established in all of the major
political and commercial centers. New occupations (typefounding and presswork) were introduced; bookmaking arts were
reorganized; trade networks were extended; book fairs inaugurated. By 1500, the use of movable type had become the dominant mode for duplicating texts, and xylography had replaced
hand illumination for replicating images.
Though often classified under the heading of book history,
print culture encompasses a vast variety of nonbook materials,
such as advertisements, almanacs, calendars, horoscopes, proclamations, tickets, and timetables. It also entails the provision
of visual aids (such as maps, charts, tables, graphs, and detailed
drawings) that are especially difficult to duplicate in large
673
Print Culture
more strenuous effort was required to cope with the burden of
the past (Bate 1970).
There are synchronic as well as diachronic aspects to print
culture (see synchrony and diachrony). Unlike handcopied texts, printed copies get issued not consecutively but
simultaneously. The distribution of printed copies was relatively
slow before the development of modern transport systems.
Nevertheless, the age of the wooden handpress saw a marked
improvement in the coordination of diverse activities, such as
checking the path of a comet against diverse predictions, incorporating new findings and corrections in successive editions of
reference works, or mobilizing a protest movement in different
parts of a given realm. We made the thirteen clocks strike as
one, commented an American revolutionary.
Simultaneity went together with standardization, as is illustrated in an anecdote about Napoleons minister of education,
who looked at his watch and announced that at that moment all
French schoolboys of a certain age were turning the same page
of Caesars Gallic Wars. The output of the handpress fell short
of achieving the degree of standardization that marks modern
editions. Yet early modern readers were able to argue, in both
scholarly tomes and polemical pamphlets, about identical passages on identically numbered pages.
Simultaneity is of special significance in conjunction with
journalism. Especially after the introduction of iron presses
harnessed to steam and wire services that made use of the telegraph, the newspaper press would restructure the way readers
experienced the flow of time. Simultaneity is nicely illustrated by
the front page layout of a modern newspaper, which has been
described by Marshall McLuhan as a mosaic of unrelated scraps
in a field unified only by a dateline (1964, 249). Given the juxtapositions and discontinuities presented by the front page, it is
a mistake to associate print only with linear sequential modes
of thought. Although books and newspapers are now filed separately in libraries and archives, they are intertwined manifestations of print culture, beginning with early newsbooks and going
on to later serialized novels.
Even in the age of the handpress, newsprint altered the way
readers learned about affairs of state. It created a forum outside
parliaments and assembly halls and invited ordinary subjects
to participate in debates (by contributing letters to editors). It
provided ambitious journalists (from Jean-Paul Marat to Benito
Mussolini) with new pathways to political power. It served to
knit together the inhabitants of large cities for whom the daily
newspaper provided a kind of surrogate community. According
to Benedict Anderson (1983, 3740), newsprint served a similar
function for millions of compatriots who lived within the boundaries of a given nation-state.
The reception of news via print rather than voice points to a
facet of print culture that has given rise to much debate. It centers on a contrast not with scribal but with oral culture. To an
Enlightenment philosopher such as the Marquis de Condorcet
(see Baker 1982, 268), who was impressed by advances in mathematics, the use of print held the promise of introducing rationality into political affairs. Whereas speech was ephemeral,
Condorcet argued, a printed account lent itself to rereading and
careful consideration. By means of rhetorical devices, orators
could persuade their audiences to perform ill-considered acts.
674
Private Language
nineteenth century and after the adoption of other new technologies (such as lithography, photography and the shift from hot
to cold type). Nineteenth-century observers believed that the
advent of newspapers signaled the end of the book. Late twentieth-century commentators believed that radio, television, and
other electronic media were going to supersede print. At present,
the movement of texts onto screens has persuaded some observers that supersession is finally at hand. In my view, continued
coexistence seems more likely, especially since the preservative
powers of print are still uncontested. An ever-growing shortage of space on library shelves and an unending concern about
information overload suggest that print culture is still exerting a
cumulative effect and will continue to do so for the foreseeable
future.
Elizabeth L. Eisenstein
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Anderson, Benedict. 1983. Imagined Communities. London: Verso.
Baker, Keith M. 1982. Condorcet. Chicago: University of Chicago Press.
Bate, W. Jackson. 1970. The Burden of the Past and the English Poet.
Cambridge: Harvard University Press.
Burke, Edmund. 1790. Reflections on the Revolution in France.
Available online at: http://www.constitution.org/eb/rev_fran.htm.
Chartier, Roger. 1986. Texts, printing, readings. In The New Cultural
History, ed. L.Hunt, 15471. Berkeley: University of California Press.
. 1987. The Culture of Print. Princeton, NJ: Princeton University
Press.
Cormack, Bradin, and Carla Mazzio. 2005. Book Use, Book Theory, 1500
1700. Chicago: University of Chicago Press. See Part III on The How-to
Book.
Eisenstein. Elizabeth L. 1997. From the printed word to the moving
image. Social Research 64 (Fall): 104966.
. 2005. The Printing Revolution in Early Modern Europe. New
York: Cambridge University Press.
Martin, H. J., and Lucien Febvre. 1976. The Coming of the Book. Trans.
D. Gerard. London: NLB. See pages 716 for section on Chinese
precedents.
McKenzie, D. F. 2002. Speech manuscript print. In Making Meaning:
Printers of the Mind and Other Essays, ed. P. MacDonald and
M. Suarez, 23759. Amherst: University of Massachusetts Press.
McLuhan, Marshall. 1962. The Gutenberg Galaxy. Toronto: University of
Toronto Press.
. 1964. Understanding Media. New York: McGraw-Hill [Signet
Books].
Mumford, Lewis. 1934. Technics and Civilization. New York: Harcourt
Brace.
Ong, Walter. 1982. Orality and Literacy. London: Routledge.
PRIVATE LANGUAGE
Ludwig Wittgenstein (18891951) is considered one of the most
influential philosophers of the twentieth century. While his contributions to philosophy are wide-ranging, one of his most widely
discussed and influential contributions is taken to follow from the
sections of his Philosophical Investigations that explore the possibility of a logically private language ([1953] 1958, 243315).
These remarks have come to be called Wittgensteins private language argument. The label is not, however, without controversy;
we explore why in this entry.
675
Private Language
order debate such readers take, they all (if they take Wittgenstein
to have been successful in his alleged aim) hold that a logically
private language, as described in the final paragraph of PI 243,
is shown by Wittgenstein to fail to fulfill the relevant criteria for
being a language. It is this, such readers claim, that he demonstrates argues in the 72 or so remarks that follow it.
The alternative way of reading these remarks is as follows: Wittgenstein, in PI, asks us to imagine such a language
([1953] 1958, 243), that is, to (try to) entertain the thought of the
possibility of a language which describes my inner experiences
and which only I can understand (256). The remarks that follow
work on the reader to the extent that he or she sees that however
one tries to give sense to such a (putative) language, we never
arrive at a position where our desire to see it as such is satisfied.
Read aright, the remarks serve to dispel the desire to attempt to
give sense to the locution [a] language which describes my inner
experiences and which only I can understand or a logically private language. On this reading, it is not that there is something
akin to a misuse according to the rules of grammar of the
concepts private and language, such that such locutions are
nonsense. It is not something that the philosopher wants to say
but cannot owing to the configuration of grammatical rules. It is,
rather, that when we try to imagine a private language, we realize that there is no determinable thing a private language to
imagine. The very notion of a private language dissipates as we
try to grasp it.
On the latter reading, therefore, Wittgenstein does not advance
a theory as to the nature of first person, present tense, psychological utterances but merely offers suggestions as to how it might be
possible to understand these as learned replacements for (say) a
cry of pain, rather than as, for example, a description or report of
an inner state, such as a sensation. He offers such suggestions, as
it were, as prophylactics. To accept such suggestions as possible
weans one off of the assumption that such utterances must be
descriptions of inner states (sensations, for example) and feeds
into weaning one away from the assumed need and the desire to
give sense to the locution private language.
The debate between the two readings, therefore, hinges
on how one should understand Wittgensteins philosophical
method (or metaphilosophy). Those who take him to offer a refutation of a logically private language and to be, in the course of
doing so, advancing positive claims as to (say) the nature of first
person, present tense, psychological utterances, do so, as their
opponents suggest, by underplaying his remarks on philosophical method (especially PI 109 and 126 through 133), where he
lays out his therapeutic vision of philosophy. Here, the practice
of philosophy is undertaken as a therapeutic dialogue between
the Wittgensteinian philosopher and his or her interlocutor
indeed, the therapist and interlocutor might be conflicting tendencies in oneself. The task of the philosopher-as-therapist is to
facilitate the interlocutors free realization that he or she is in the
grip of a particular picture of the way things must be that leads
him or her to be committed to certain nonobligatory philosophical positions.
What one takes to be done by Wittgenstein in these 72 (or
so) remarks has more than merely exegetical significance. If one
understands him to have refuted the possibility of a logically private language and, in so doing, to have advanced, for example,
676
Projectibility of Predicates
the expressive theory for the meaning of first person, present
tense, psychological utterances, then one will be led to argue
that such utterances are not cannot be reports or descriptions
of inner states but are must be rather, expressions or avowals
of judgments or evaluations/appraisals: So, for example, many
cognitivist philosophers (e.g., Lyons 1980; Nash 1989) and
psychologists (e.g., Lazarus 1982) of emotion advance this view
(some, such as Kenny 1963, even drawing on Wittgenstein as
chief influence).
If one understands Wittgenstein as not having advanced such
views and, rather, takes his remarks to be designed to work on
one so as to facilitate the realization that there is nothing we
would wish to hold onto answering to the name private language, we are not led to a philosophical commitment to any
view on psychological language. We might then, rather, engage
with those who claim or assume that first person, present tense,
psychological utterances must be reports or descriptions of inner
states or sensations Jamesian accounts of emotion in general
and cognitive neuroscience in particular (e.g., Damasio 1994)
and with those who claim or assume that they must be expressions or avowals as both being driven by prejudice. Jamesian/
neoJamesian and cognitivist theories of emotion can be seen
to rest on prejudice about our use of psychological language (see
also emotion words).
Wittgenstein, read aright, can provide much help in our
attempts to dissolve such prejudice in the human sciences.
Phil Hutchinson
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Baker, Gordon. 2004. Wittgensteins Method: Neglected Aspects.
Oxford: Blackwell. See Part I, Section B, Chapters 5, 6, and 7. The book
contains a number of excellent chapters on the (so-called) private language argument and its interpretation.
Damasio, A. R. 1994. Descartes Error: Emotion, Reason, and the Human
Brain. New York: Grosset/Putnam.
Kenny, A. 1963. Action, Emotion and Will. London: Routledge.
Lazarus, R. S. 1982. Thoughts on the relations between emotion and cognition. American Psychologist 37: 101924.
Lyons, W. 1980. Emotion. Cambridge: Cambridge University Press.
Mulhall, Stephen. 2006. Wittgensteins Private Language: Grammar,
Nonsense and Imagination in Philosophical Investigations, 243315.
Oxford: Oxford University Press. Nuanced reading.
Nash, R. A. 1989. Cognitive theories of emotion. Nos 23: 482504.
Wittgenstein,
Ludwig.
1922.
Tractatus
Logico-Philosophicus.
London: Routledge.
. [1953] 1958. Philosophical Investigations. Oxford: Blackwell.
PROJECTIBILITY OF PREDICATES
The distinction between projectible and nonprojectible predicates arises in analyses of inductive inference. David Hume
([1748] 2000) showed that it is not possible to justify the belief
that past empirical regularities will continue into the future.
Nelson Goodmans 1953 new riddle of induction raised a further problem. Goodman showed that there are always an unlimited number of hypotheses that encompass all of the evidence,
yet conflict in their predictions. For example, if we introduce
the predicate grue (grue = examined before some future time
Projectibility of Predicates
Projection Principle
Proposition
properties of lexical items (that is, both their features and selectional requirements) must be represented at every syntactic level of representation. Its key effect is to allow structure at
every level of representation to be directly determined by lexical
properties, thereby largely eliminating the need for languageparticular rules. For this reason, the projection principle has
played a key role in the transition from early rule-based transformational grammar to principles and parameters theory.
Let us briefly review two major consequences of the projection principle. The first concerns the elimination of languageparticular base rules (the rules that generate deep structures; see
underlying structure and surface structure) in transformational grammar. Such rules would state, for example, that
in English a verb phrase may consist of a verb followed by either
a noun phrase or a prepositional phrase or a noun phrase and a
prepositional phrase, and so on. In addition, the lexical entries of
verbs specify the syntactic environment in which they can occur.
Thus, the entry for the verb put states that it must be followed by a
noun phrase and a prepositional phrase. This arrangement gives
rise to redundancy: In the case at hand, the fact that there are
English verbs that can be followed by a noun phrase and a prepositional phrase is expressed twice, namely, in the base rules and
in the lexical entry of put (and other verbs). Since lexical properties are idiosyncratic, elimination of this redundancy requires a
simplification of the base component, and this is precisely what
the projection principle makes possible. As it requires syntactic
representations to be projections of lexical properties, it allows
the base component to be reduced to a small universal skeleton,
known as x-bar theory, and relegates language-specific properties of deep structures to the lexicon and to a set of word-order
parameters.
The projection principle also implies that syntactic representations that are the result of movement must contain traces
that function as place holders for categories that have undergone
this operation. It is easy to see why this should be so. As stated
earlier, the lexical entry for the verb put must state that this verb is
followed by a noun phrase and a prepositional phrase. However,
in the question in (1a), the selectional requirements of this verb
appear not to be met, since the noun phrase that usually follows
it has undergone movement. Satisfaction of the projection principle, therefore, requires that movement leave behind a trace
shown as tNP in (1b) so that the selectional properties of put are
also expressed in the structure that results from movement.
(1)
a. I wonder [[NP what] Jack has [VP [V put] [PP in the oven]]]
b. I wonder [[NP what] Jack has [VP [V put] tNP [PP in the oven]]]
PROJECTION PRINCIPLE
PROPOSITION
678
Propositional Attitudes
existence of propositions is that they make sense of the fact that
from
Terry believes that cats are cute.
PROPOSITIONAL ATTITUDES
Propositional attitudes are mental relations believing, desiring,
hoping, and so on between an individual and a proposition.
The sentence Amie believes that 1 + 1 = 2 is a propositional
attitude report, where Amie stands in the belief relation to the
proposition that 1 + 1 = 2. The proposition specifies the content of
the attitude (see intentionality). Propositional attitudes play
a central role in ordinary psychological explanations. We explain
why Elia presses the doorbell by saying that he hopes someone
will answer the door, and he believes that this will happen if he
presses the bell.
679
Prototypes
among a subject, a proposition, and a way of believing. A way
of believing is supposed to be similar to a Fregean sense, but it
is contextually specified and does not correspond to any explicit
referring item in the report.
Joe Lau
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Burge, Tyler. 1979. Individualism and the mental. In Midwest Studies in
Philosophy. Vol. 4. Ed. P. French, T. Uehling, and H. Wettstein, 73121.
Minneapolis: University of Minnesota Press.
Carnap, Rudolf. 1958. Meaning and Necessity: A Study in Semantics and
Modal Logic. 2d ed. Chicago: University of Chicago Press.
Crimmins, Mark. 1992. Talk about Beliefs. Cambridge, MA: MIT Press.
Davidson, Donald. 1968. On saying that. Synthese 19: 13046.
Frege, Gottlob. [1892] 1948. Sense and reference. Philosophical Review
57: 20930.
Larson, Richard, and P. Ludlow. 1993. Interpreted logical forms.
Synthese 95: 30555.
McGinn, Colin. 1977. Charity, interpretation, and belief. Journal of
Philosophy 74: 52135.
McKay, Thomas, and Michael Nelson. 2008. Propositional attitude
reports. In The Stanford Encyclopedia of Philosophy (Fall), ed. Edward
N. Zalta. Available online at: http://plato.stanford.edu/archives/
fall2008/entries/prop-attitude-reports/. This is a comprehensive survey on the semantics of propositional attitude reports.
Quine, Willard Van Orman. 1956. Quantifiers and propositional attitudes. Journal of Philosophy 53: 17787.
Richard, Mark. 1990. Propositional Attitudes: An Essay on Thoughts and
How We Ascribe Them. Cambridge: Cambridge University Press.
Salmon, Nathan. 1986. Freges Puzzle. Cambridge, MA: MIT Press.
Searle, John. 1983. Intentionality. An Essay in the Philosophy of Mind.
Cambridge: Cambridge University Press.
PROTOTYPES
Prototypes refers to one of the ways in which psychology has
attempted to account for the nature of concepts and conceptual categories (see categorization). In the prototypes view,
conceptual categories form around and/or are represented in the
mind by salient, information-rich ideas and images that become
prototypes for the category. Other members of the conceptual
category are judged in relation to these prototypes, thus forming
gradients of category membership.
The importance of prototype theory needs to be seen in its
historical context. Prior to this work, concepts and conceptual
categories were assumed from philosophy to be arbitrary logical
sets with defining features and clear-cut boundaries. All members of the conceptual category were considered equivalent with
respect to membership. This is now called the classical view.
Psychological research on concept learning used artificial sets
of stimuli, structured into microworlds in which these assumed
characteristics of categories were already built in. Mainstream
linguistics was constructed on similar assumptions; phonology, semantics, and syntax all sought to decompose the
subject matter of their domains (speech sounds, word meaning, and grammar) into sets of abstract binary defining features,
a procedure called componential analysis. In contrast, in the prototype view, there need be no defining attributes that all conceptual category members have in common; category boundaries
680
Prototypes
6. Probability Judgments: Representativeness strongly influences probability judgments (Kahneman, Slovic, and Tversky
1982), which is important because probability is thought by
many philosophers to be the basis of inductive inference and,
thus, of the way in which we learn about the world.
7. Natural Language Indicators of Graded Structure: Natural
languages themselves contain various devices that acknowledge
and point to graded structure, such as hedge words like technically and really (see also Lakoff 1987; Taylor 2003).
8. Judgment of Similarity: Less good examples of categories
are judged more similar to good examples than vice versa. This
violates the way similarity is treated in logic, where similarity
relations are symmetrical and reversible (Tversky 1977).
What determines the items that will be prototypical of categories in the first place? Some are based on statistical frequencies,
such as the means or modes (or family resemblance structures) for various attributes; others appear to be ideals made
salient by such factors as physiology (good colors, good forms),
social structure (president, teacher), culture (saints), goals
(ideal foods to eat on a diet), formal structure (multiples of 10
in the decimal system), causal theories (sequences that look
random), and individual experiences (the first learned or most
recently encountered items or items made particularly salient
because they are emotionally charged, vivid, concrete, meaningful, or interesting). Note that particular exemplars can be prototypes if they serve as reference points for the category; thus,
prototype and exemplar theories are not necessarily contradictory. Note also that it is a misapprehension to take prototypes to
mean only one kind of prototype and to critique prototype theory
on that basis.
The prototype view has spread beyond psychology to many
fields, including linguistics and narratology. Gradients of
exemplariness are ubiquitous in linguistic phenomena, even
in phonology where actual speech is less clear-cut than would
appear in an abstract componential analysis. In semantic and
syntactic analyses (particularly in cognitive grammar and
the understanding of metaphor), prototype effects, as well as
providing specific case studies are often used as evidence that
formal analysis is insufficient of itself and that world knowledge
must be part of ones theory (Lakoff 1987; Langacker 1990; Taylor
2003).
There are societal implications of prototype theory. For example, social stereotypes are a type of prototype; it is called a
stereotype when it applies to a group of people and has social
consequences. Another example: Anglo-American case law is
based on prototypes; precedent cases provide the reference
points in arguing present cases (see legal interpretation).
Further examples: Political arguments are often about conflicting
prototypes (e.g., different images of welfare mothers) even when
goals (care for children) may be quite similar, while scholarly
debates are frequently about attempts to draw clear-cut boundaries where there are none when the participants actually agree
on central prototypes (e.g., agreement on clear cases of what we
mean by language or religion but disagreement about borderline
phenomena and about criteria). Many such issues could be clarified by understanding the principles of prototyping.
Although prototype effects are now acknowledged as empirically established, there have been many criticisms of prototypes
as an account of concepts and categorization. The main objections fall into two camps: In the first, prototypes and graded
structure violate the classical requirement that the real meaning
of a concept (that to which it refers) must be the identifiable necessary attributes of a classical definition. One argument for this
view is that prototype and graded structure effects can be found
for conceptual categories that have a formal classical definition, such as odd number (Armstrong, Gleitman, and Gleitman
1983), another that prototypes do not form componential combinations as do the elements of classical definitions (e.g., a good
example of pet fish is neither a prototypical pet nor a prototypical
fish; Osherson and Smith 1981). Both findings are taken to indicate that prototypes are something other than and irrelevant to a
concepts meaning. One solution is a dual model in which prototypes are assigned the function of rapid recognition of conceptual
referents, whereas the true meaning is provided by a classically
defined core (Osherson and Smith 1981; Smith, Shoben, and Rips
1974). In the second camp, prototypes change with context for
example, prototypical animals in the context of a zoo differ from
those in the context of a farm (Barsalou 1987). Such findings are
taken to indicate that it is theories that determine concept meaning (Medin 1989). (For a review of more philosophical objections
to prototypes and all other accounts of concepts see Fodor
1998 and Laurence and Margolis 1999.)
In conclusion, prototype and graded structure effects are well
established; the debate concerns what they mean for our understanding of concepts, categorization, thinking, decision making,
the meaning of words and word combinations, and innumerable
other aspects of human functioning. The prototype view is not
necessarily incompatible with other approaches if the different
views are seen at a level deep enough that their complementarity can be appreciated. Prototype effects do not deny that under
some circumstances we explicitly think in terms of necessary
and sufficient conditions, nor does the essentialist intuition that there are nonobvious realities behind outer appearances (Gelman and Wellman 1991) exclusively require the form
of classical definitions for its implementation. Furthermore, prototypes, like any other mental or cultural process, only function
within the ever-shifting contexts of partially organized forms of
encyclopedic knowledge and belief the sort of understanding
toward which the theories view points. What prototype research
uniquely points out is that there is a level of organization of the
mind in which concepts appear to be represented by informationrich, imagistic, sensory-based, often emotion-linked wholes that
are used in thinking and communicating without reference to
definitions, category boundaries, or even truth conditions.
Eleanor Rosch
WORKS CITED AND SUGGESTIONS FOR FURTHER REFERENCES
Armstrong, Sharon, Lila Gleitman, and Henry Gleitman. 1983. What
some concepts might not be. Cognition 13: 263308.
Barsalou, Lawrence. 1987. The instability of graded structure: Implications for the nature of concepts. In Concepts and
Conceptual Development: Ecological and Intellectual Factors in
Categorization, ed. Ulric Neisser, 10140. Cambridge: Cambridge
University Press.
Fodor, Jerry. 1998. Concepts: Where Cognitive Science Went Wrong. New
York: Oxford University Press.
681
Proverbs
Gelman, Susan, and Henry Wellman, 1991. Insides and essences: Early
understanding of the non-obvious. Cognition 38: 21344.
Kahneman, Daniel, Paul Slovic, and Amos Tversky, eds. 1982. Judgment
under Uncertainty: Heuristics and Biases. New York: Cambridge
University Press.
Lakoff, George. 1987. Women, Fire, and Dangerous Things: What
Categories Reveal about the Mind. Chicago: University of Chicago
Press.
Langacker, Ronald. 1990. Concept, Image, and Symbol: The Cognitive
Basis of Grammar. Berlin: Mouton de Gruyter.
Laurence, Stephen, and Eric Margolis. 1999. Concepts and cognitive
science. In Concepts: Core Readings, ed. Eric Margolis and Stephen
Laurence, 381. Cambridge, MA: MIT Press.
Markman, Ellen. 1989. Categorization and Naming in Children.
Cambridge, MA: MIT Press.
Medin, Douglas. 1989. Concepts and conceptual structure. American
Psychologist 44: 146981.
Mervis, Carolyn. 1980. Category structure and the development of
categorization. In Theoretical Issues in Reading Comprehension, ed.
William Brewer, Bertram Bruce, and Rand Spiro, 279308. Hillsdale,
NJ: Erlbaum.
Osherson, Daniel, and Edward Smith. 1981. On the adequacy of prototype theory as a theory of concepts. Cognition 9: 3558.
Rosch, Eleanor. 1973. Natural categories. Cognitive Psychology
4: 32850.
. 1975. Cognitive representations of semantic categories. Journal
of Experimental Psychology: General 104: 192233.
. 1978. Principles of categorization. In Cognition and
Categorization, ed. Eleanor Rosch and Barbara Lloyd, 2748. Hillsdale,
NJ: Lawrence Erlbaum.
. 1999. Reclaiming concepts. Journal of Consciousness Studies
6.11/12: 6177.
Rosch, Eleanor, and Carolyn Mervis. 1975. Family resemblances: Studies
in the internal structure of categories. Cognitive Psychology
7: 573605.
Rosch, Eleanor, Carolyn Mervis, Wayne Gray, David Johnson, and
Penelope Boyes-Braem. 1976. Basic objects in natural categories.
Cognitive Psychology 8: 382439.
Smith, Edward, and Douglas Medin. 1981. Categories and Concepts.
Cambridge: Harvard University Press.
Smith, Edward, Edward Shoben, and Lance Rips. 1974. Structure and
process in semantic memory: A featural model for semantic decisions. Psychological Review 81: 21441.
Taylor, John. 2003. Linguistic Categorization. Oxford: Oxford University
Press.
Tversky, Amos. 1977. Features of similarity. Psychological Review
84: 32752.
PROVERBS
Proverbs in the Humanities and Cognitive Science
The humanities treat proverbs as repositories of wisdom about
everyday life. This premise motivates religious, literary, practical,
and culturalfolklore approaches that capitalize on the fact that
proverbs can be pithy, express a moral or precept, sound authoritative, perform various pragmatic functions (e.g., exhortation),
and serve as indirect speech-acts, in which what is intended
encompasses more than what is actually said.
The cognitive science perspective examines how proverbs
both illuminate and are illuminated by knowledge about the
mind. Ultimately, the humanities and cognitive science views
are complementary. The former is represented in Mieder (1994)
682
683
684
impulses. But the generation of ordinary speech is no less determinate, and it is not determined solely by ones self-conscious
purposes. At any given moment, there are many ways in which
I might phrase what I wish to say. However, I say things in one
way only. For example, in speaking of a particular object, I might
refer to it as a p.c., a computer, a laptop, and so on. But I choose
one. In a connectionist model, we would say that a range of
factors has given the greatest activation to a particular phrasing.
In some cases, this is just the basic level term. In other cases,
there is some contextual or other reason for the choice. It might
be a matter of repetition (I just saw a copy of PC Computing,
so I say p.c.); alternatively, it might involve a more complex
lexical relation (someone just mentioned desktops, so I
say laptop). Psychoanalysis adds a series of circuit activations
from dynamically unconscious motivational contents. For example, Ernst Lanzer (Freuds Rat Man) tells Freud about someone who flattered and deceived him. He identifies the person as
a medical student. There are many ways in which Lanzer could
have introduced this person. What caused the specifically medical detail to be activated? One possibility is that Lanzer had a
transferential relation to Freud in which he feared that Freud
would flatter and deceive him in the treatment. Circuits connecting Freud with this student (and presumably with parental
figures from the period of primary repression; see Hogan 1996,
14850) would then have shaped this speech specifically, the
choice of medicine as an identifying characteristic separately
from Lanzers self-conscious intentions.
CODE. Various features of a linguistic code particularities of the
lexicon, morphology, grammar may allow the expression of
unconscious ideas and impulses beyond the straightforward
semantics of lexical selection. This may occur most obviously
through ambiguity or puns, a point stressed particularly by
Lacan. For example, James Gorney (1990) reports a case in which
someone reported conscious anxiety over a bill and thereby
expressed unconscious anxiety over someone named Bill. This
is most readily understood in terms of left hemisphere versus
right hemisphere processes. Neurocognitive research shows
us that the right hemisphere generates multiple meanings, some
of which are inhibited by the left hemisphere, which selects the
contextually relevant meanings (see Chiarello 1998, 145; Faust
1998, 180). For example, in the semantic context of references to
a restaurant, and the linguistic context of following a determiner,
left hemisphere interpretive processes will limit the meanings of
bill to a list of charges. However, right hemisphere interpretive processes will briefly generate beak and William also.
Psychoanalysis once again adds to this well-established architecture a dynamic unconscious component. Specifically, it adds
the idea that in some cases, a particular word (e.g., bill, rather
than check) may be produced due to unconscious activations
(see spreading activation) that are motivational, though
inaccessible to the working memory of the speaker.
Conclusion
The relation between psychoanalysis and language has been
explored by many writers using many approaches. In the preceding sections, I have tried to show that central functional
principles of psychoanalysis as they relate to language may be
685
Psycholinguistics
686
PSYCHOLINGUISTICS
Psycholinguistics studies the relationship between the mind
and language. Its concern is with the cognitive processes that
underlie the acquisition, storage, and use of language and their
correlates in observable neurological processes in the brain. The
field is thus heavily reliant upon principles and research methods adopted from cognitive psychology. However, it is essentially
multidisciplinary, drawing also upon linguistics, speech science,
phonetics, computer modeling, neurolinguistics, discourse
analysis (linguistic) and semantics.
Psycholinguistic inquiry is driven by the premise that language has developed in ways that reflect the structure of the
human mind. Certain shared cognitive routines are assumed
to underpin the production and reception of language, despite
variations between individuals in terms of level of vocabulary and powers of self-expression. Research thus tends to be
normative in direction, though due allowance is made for the
characteristics of the language being examined, the task being
undertaken, and the population under investigation. A favored
approach is to investigate small-scale components of a processing system in order to gradually build up a picture of the system
as a whole.
An interest in the psychology of adult language emerged
during the nineteenth century, with initiatives such as Paul
Pierre Brocas work (1863) on the location of language in the
brain and Francis Galtons (1883) on word association. A parallel interest in child language acquisition had developed earlier
from the Enlightenment debate between rationalist followers
of Descartes, who believed that much human knowledge was
innate, and empiricists such as Hume and Locke, who asserted
that it was entirely acquired.
Psycholinguistics
The term psycholinguistics probably dates from the 1930s,
but progress during the first half of the twentieth century was
discouraged by the dominant behaviorist view that the human
mind is unknowable. The field did not emerge as a discipline
in its own right until the mid-1950s. An impetus was given by
a series of essays in which George Miller mapped out possible
areas of inquiry. A major landmark was then Noam Chomskys
1959 rebuttal of the behaviorist assumptions expressed by B.
F. Skinner in his Verbal Behavior. Citing the speed with which
an infant masters a linguistic system and the poverty of the evidence available, Chomsky concluded that language is a genetically acquired faculty. His nativist stance (see innateness and
innatism) stimulated modern studies of child language and
began a controversy that continues to the present day.
Much early psycholinguistic inquiry was closely allied to linguistic theory. It explored aspects of Chomksys early generative grammar on the assumption that the rules represented
psychological reality: that is, that they provided a model of the
operations of the mind, as well as a linguistic account of grammatical structure. The findings were mainly negative or inconclusive, and the paths of linguistics and psycholinguistics began
to diverge, although a body of psychological linguists continues today to work within a framework of Chomskyan theory.
In its current form, psycholinguistics falls into a number of
distinct areas, but considerable overlaps between them give the
discipline coherence.
Language Processing
Processing research seeks to characterize the four language
skills of reading, writing, speaking (speech production)
and listening (auditory processing, speech perception)
in terms of the cognitive operations that underlie them. It also
investigates how vocabulary is retrieved from the mind (word
recognition) and how syntactic structures are assembled
or interpreted. Its findings have important applications for the
teaching of literacy (see teaching reading and teaching
writing), for language therapy, and for second language learning (see second language acquisition).
Drawing upon an information-processing precedent, accounts
often depict language users as taking linguistic material through
a series of levels of representation. For example, a listener
might be seen as building sounds into syllables, syllables into
words, words into clauses, and clauses into abstract meanings.
In fact, the process is not necessarily a sequential one, as the
language user is capable of operating at all these levels simultaneously. What is more, there is evidence that higher-level
knowledge influences processing at a lower level (knowledge of
the existence of a particular word might assist the recognition of
a string of sounds in a top-down way). Opinion continues to be
divided as to whether the relationship between the levels is fully
interactive or whether each level operates with a degree of independence (in a modular way). The benefit of the former is that
all sources of information become available at once; the benefit
of the latter is that rapid decision making is enabled.
The receptive skills of listening and reading entail two distinct operations. In the first, visual or auditory information is
decoded, a process that entails mapping from strings of sounds
or letters to known words. The user relies partly upon perceptual
Language Representation
A closely allied area of enquiry considers how linguistic knowledge is stored in the mind in a way that enables rapid matches to
be made. Words are said to be stored as lexical entries, containing
information about the items orthographical form and pronunciation, word class, inflections, and combinatorial possibilities.
There has been discussion of whether there are separate entries
for affixes such as un- in the word unhappy or whether words of
this type have their own entries. Also much discussed has been the
question of category membership: how a language user manages
to classify an item of crockery as a cup rather than a mug or bowl.
The entries in a users mental lexicon are massively interconnected,
enabling a process known as spreading activation, in which it
becomes easier to locate (e.g.) the words nurse, hospital, or patient
after recently hearing or seeing the word doctor. The connections
vary in strength, favoring words of high frequency and words with
a high probability of occurring together.
Traditional linguistic accounts assume that syntax is represented in the mind as a set of rules, but recent psycholinguistic
687
Psycholinguistics
thinking has swung toward the notion that speakers can only produce language as rapidly and accurately as they do if they make
extensive use of stored chunks of language already assembled
syntactically. The human mind appears better adapted for the
storage of enormous amounts of information than for rapid processing. exemplar theory suggests that linguistic knowledge
derives from multiple memory traces of individual language episodes. The attraction of the theory is that it accounts not only for
the speakers ability to retrieve chunks but also for the listeners
ability to deal with a range of voices and accents and to recognize the many forms that a single word can take in connected
speech. Instead of mapping back to one idealized template for
a phoneme or word, the listener draws upon vast numbers of
traces laid down by earlier encounters.
688
wild or confined to their rooms for long periods. There are also
studies of the course of language development in twins, blind
children, the children of deaf parents, and children acquiring a
pidgin language.
Psycholinguistics
Language Impairment
Psycholinguistics also studies the psychological factors that contribute to language disorders. One type of difficulty investigated
is developmental: It includes problems of speech production,
such as stammering, as well as disorders of reading and
writing, such as dyslexia and dysgraphia. Another type is
acquired as a result of accident, stroke, or surgery and includes
aphasia, an impairment of the ability to produce and/or understand speech. Also investigated are the effects of aging upon
language. In all of these areas, an important distinction has to be
made between loss of language from the mental store and disruption of the processes by which language is accessed.
Besides supporting the work of clinicians and therapists in
treating disorders, the study of language impairment feeds back
into other areas of psycholinguistics. Firstly, it sheds light on the
skills and cognitive processes that are essential to normal language operations but are absent or obstructed in the patients
studied. Secondly, it affords possible insights into the process of
language acquisition. Research into cases of specific language
impairment (SLI) has sought evidence as to whether language is
or is not genetically transmitted. Research into the linguistic competence of sufferers from syndromes such as Downs, Williams,
and autism has sought evidence as to whether language forms
part of general cognition or is independent of it.
Conclusion
Research is active in all the areas of psycholinguistics that have
been profiled. The methodologies employed are diverse and (as
noted) include experimentation, observation, elicited response,
verbal report, and grammaticality judgments. However, the
future direction of research will be greatly influenced by technological advances. The most important recent innovation has
been the use of brain imaging to validate psycholinguistic theories about the operations underlying language performance.
Researchers are able to identify different types of brain event,
distinguishing, for example, one associated with syntactic functions from one associated with semantic. A second development lies in the increased use of eye-tracking equipment not
only to study reading but also to record the way in which a language users attention is attracted to an object by semantic cues.
Computer science also makes a contribution. Psycholinguistics
employs computer modeling of linguistic performance, including neural networks that simulate spreading activation, connectionist word-recognition models based upon competition
between items, and connectionist learning models. It also draws
upon ongoing research into artificial intelligence for example,
studies of expertise, artificial speech recognition, or automatic
translation.
John Field
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Aitchison, Jean. 2008. The Articulate Mammal. 5th ed. London: Routledge.
Readable general introduction to many areas.
. 2004. Words in the Mind. 3d ed. Oxford: Blackwell. Accessible
overview of lexical storage and retrieval.
Clark, Eve V. 2003. First Language Acquisition. Cambridge: Cambridge
University Press. Extensive account of the field.
Deacon, Terrence. 1997. The Symbolic Species. New York: W. W. Norton.
In-depth account of neurolinguistics and language evolution.
Field, John. 2004. Psycholinguistics: The Key Concepts. London: Routledge.
Comprehensive reference guide for non-psycholinguists.
Gathercole, Susan E., and A. D. Baddeley. 1993. Working Memory and
Language. Hove, UK: Erlbaum. Key account of the part played by working memory.
Gleason, Jean Berko, and N. B. Ratner, eds. 1998. Psycholinguistics. Fort
Worth, TX: Harcourt Brace. Overviews of nine major areas.
689
Psychonarratology
Levelt, Willem J. M. 1989. Speaking. Cambridge, MA: MIT Press. Standard
comprehensive account.
Obler, Lorraine K., and K. Gjerlow. 1999. Language and the Brain.
Cambridge: Cambridge University Press. Short but wide-ranging introduction for non-specialists.
Perfetti, C. 1985. Reading Ability. New York: Oxford University Press.
Influential account of skilled reading.
PSYCHONARRATOLOGY
Definition
This term was first coined by P. Dixon and M. Bortolussi in 2001
in their chapter Prolegomena for a science of psychonarratology and developed further by Bortolussi and Dixon in their
book Psychonarratology: Foundations for the Empirical Study
of Literary Response (2003). It designates an interdisciplinary,
empirical approach to the study of literary response (see literature, empirical study of) and the processing of narrative. Generally, as the term suggests, it brings together two very
diverse fields: psychology and literary studies. However, while in
literary studies psychology typically refers to psychoanalysis,
psychonarratology draws on cognitive psychology and discourse
processing for its methodology. Within literary studies, it looks to
narratology for its conceptual and theoretical insights about
narrative prose but also draws on related fields, such as reception and reader response theories.
Theoretical Background
The conjoining of cognitive psychology and narratology has two
theoretical motivations. On the one hand, literary studies in general, and narratology in particular, seemed to have reached an
impasse from which the traditional stock of methods provided no
means of escape. Numerous scholars had understood that texts
are just collections of letters, words, sentences, and paragraphs,
that meanings emerge in the interaction between texts and readers, and that these interactions could be varied and complex. For
example, from within the phenomenological tradition, Roman
Ingarden conceived of literary texts as schematic structures that
needed to be concretized or completed by the reader, and W.
Iser, following in his footsteps, regarded literary texts as indeterminate objects that elicit gap-filling strategies on the part of
the reader. However, without appropriate methodologies for
assessing the cognitive processes of real readers, literary scholars were restricted to making intuitive and purely speculative
assumptions about the effects of particular aspects of literature
on its readers. This gave rise to notions of an ideal reader, a timeless, homogeneous entity possessing the ideal competence
required to process texts in some ideal fashion. Variants of this
ideal entity were expressed through a host of related terms: the
model reader, the super reader, and so on. Various branches of
literary studies, regardless of their specific goals or scope, inevitably came up against the same roadblock, namely, the need to
account for the effect(s) of texts on readers, without the required
tools for advancing beyond the purely hypothetical.
On the other hand, empirical research in discourse processing
generally neglected to address the processing of complex, realworld narratives and literary narratives in particular. Moreover,
to the extent that cognitive psychology has investigated such
690
Psychonarratology
undergraduate students in Alberta and California or all graduates of a community college who have taken no more than one
literature course. One of the main tenets of psychonarratology is
that scientific claims can only be made about a clearly delimited
population of readers. However, any given group of readers need
not be homogeneous and uniform; rather, each group may consist of any number of overlapping subgroups that may or may not
differ with respect to interesting aspects of narrative response.
Because populations are potentially heterogeneous, there
is no necessity that general statements about reader response
apply equally to all individuals. Nevertheless, this does not mean
that it is impossible to describe reader response. The solution
is to describe measurement distributions. With respect to any
given empirical measurement, one can describe the central tendency that is, what the measurements have in common and
the variability that is, how the measurements vary over individuals. There are standard statistical and analytical techniques
for developing these kinds of descriptions, not only for numerical measurements (such as reading time or rating scales) but also
for more open-ended responses. The scientific goal is, thus, to
generate theories that relate the description of measurement distributions to the observed characteristics of readers and texts.
THE TEXTUAL EXPERIMENT. The textual experiment provides a
methodology for understanding the relationship between textual features and reader constructions by systematically varying
those textual features. The essence of this approach is to measure reader constructions using, for example, questionnaires
and rating scales, and to assess how those constructions vary as
a function of the features of the text. When a particular reader
construction (e.g., the perceived justification of a characters
actions) covaries with a particular textual feature (e.g., the use
of free indirect speech style, a mode of representing some of
the form of a characters enunciated speech or thought without direct quotation), then one can conclude that the feature
causally contributes to the reader construction. An alternative
to textual experiments (in which the nature of the text is manipulated) is to assess reader constructions of different texts sampled from extant materials. However, any two sampled texts will
differ in a wide range of characteristics. As a consequence, it
is difficult to draw causal inferences about which features are
related to particular reader constructions. Such inferences are
much more sound in a textual experiment in which features are
manipulated. In order to apply this method properly, though, it
is essential that only a single feature be varied, and that other,
potentially causal, features are not inadvertently changed at the
same time.
THE CONVERSATIONAL NARRATOR. A theoretical hypothesis that
unifies many of the authors research results is the idea that readers represent the narrator as a conversational participant. This
follows from the theoretical advances of B. Bruce and others in
narratology in which the narrative is conceptualized as a communicative transaction between a narrator and a narratee. In the
psychonarratological version of this idea, readers are hypothesized to develop a mental representation of an individual that
could have produced the words of the narrative and, in many
circumstances, treat that representation much as they would
691
Psychophysics of Speech
Dixon, P, and Bortolussi, M. 2001. Prolegomena for a science of psychonarratology. In New Perspectives on Narrative Perspective, ed. Will van
Peer and Seymour Chatman, 27588. Albany: State University of New
York Press. A brief introduction to the psychonarratology framework.
Gerrig, R. J. 1992. Experiencing Narrative Worlds. New Haven, CT: Yale
University Press. An interdisciplinary attempt to examine what it means
to be transported mentally and emotionally by literary narratives.
Graesser, A. C, C. A. Bowers, B. Olde, K. White, and N. K. Person. 1999.
Who knows what? Propagation of knowledge among agents in a literary storyworld. Poetics 26: 14375. A good example of the application
of discourse-processing methodology to a problem in literary reading.
Holub, R. C. 1989. Reception Theory: A Critical Introduction.
London: Routledge. A very good, lucid, accessible introduction to theories of literary reception.
Iser, W. 1974. The Implied Reader: Patterns of Communication in Prose
Fiction from Bunyan to Beckett. Baltimore: John Hopkins University
Press. A seminal work describing the mechanics of readertext interactions; includes some key concepts such as the ideal reader, and
gaps.
Miall, D. S, and Kuiken, D. 1995. Aspects of literary response: A new
questionnaire. Research in the Teaching of English 29: 3758. A seminal effort to develop quantifiable indices of literary expertise.
Peer, Willie van, and Max Louwerse, eds. 2002. Interdisciplinary Studies
in Thematics. Amsterdam and Philadelphia: John Benjamins. Brings
together research on themes from a variety of fields in an attempt to
define the concept of theme.
Rimmon-Kenan, S. 1983. Narrative Fiction: Contemporary Poetics. New
York: Routledge. An accessible, solid introduction to narratological
basics.
Schmidt, Sigfried. 1981. Empirical studies in literature: Introductory
remarks. Poetics 10: 31736. A pioneering study that lays the theoretical groundwork for the empirical study of literature.
Singer, M. 1990. Psychology of Language: An Introduction to Sentence and
Discourse Processes. Hillsdale, NJ: Lawrence Erlbaum. An introduction
to discourse processing and cognitive psychology.
Vipond, D., and R. A. Hunt. 1984. Point-driven understanding: Pragmatic
and cognitive dimensions of literary reading. Poetics 13: 26177. A
pioneering investigation into how literary reading varies with reading
strategy.
PSYCHOPHYSICS OF SPEECH
The psychophysics of speech describes an interdisciplinary
approach to the understanding of speech perception. The
approach considers speech as a complex acoustic signal sharing
much in common with other complex perceptual events and posits that, as such, speech may be studied in the broader context
of general perceptual, cognitive, and sensorineural systems. This
approach is distinguished from those that consider speech to be
a special signal processed in a manner distinct from non-speech
sounds. The essence of a psychophysical approach is to determine
the extent to which speech perception makes use of general cognitive and perceptual processes before postulating mechanisms
specialized to the speech signal. Thus, understanding the psychophysics of speech may include the utilization of animal models of auditory behavior and physiology to examine how much of
speech perception may be accounted for by general rather than
specialized mechanisms, and the relation of speech perception to
neural coding at peripheral and central levels of processing.
Psychophysics often connotes bottom-up or peripheral processing, and, in fact, a great deal of research of the psychophysics
692
Qualia Roles
follow speech. In the case of human and non-human perception
of speech and non-speech contexts, speech perception appears
to be relative to and contrastive with the acoustics of context
sounds, whether speech or non-speech.
This portfolio of research findings is indicative of a psychophysical approach to speech perception in that it pays careful
attention to the spectrotemporal information available to listeners, it makes use of nonhuman animals as a means of examining
the generality of the mechanisms available to speech processing, and it examines the extent to which complex non-speech
signals may give rise to some of the same patterns of perception
as speech. Research relating the context-dependent coding of
acoustic signals to neural response (see phonetics and phonology, neurobiology of) adds to the understanding of how
phonetic context effects may arise from general characteristics of
the perceptual system. The constellation of available results suggests that general perceptual mechanisms play a role in phonetic
context effects.
In other domains, the psychophysical approach has contributed to the understanding of auditory representation, auditory
learning, and cross-modal processing as they relate to speech
processing. There remains much potential for understanding the
perceptual, cognitive, and neural underpinnings of speech communication from a general perceptual/cognitive perspective.
Lori L. Holt
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Diehl, R. L., and K. R. Kluender. 1989. On the objects of speech perception. Ecological Psychology 1: 12144. Discussion of a general
approach to speech perception.
Diehl, R. L., A. J. Lotto, and L. L. Holt. 2004. Speech perception. Annual
Review of Psychology 55: 14979. This review contrasts theoretical
approaches to speech perception.
Liberman, A. M. 1996. Speech: A special code. Cambridge, MA: MIT Press.
Discussion of the motor theory of speech perception, an alternative
account to a general approach to speech perception.
Lotto, A. J., and L. L. Holt. 2006. Putting phonetic context effects into context: A commentary on Fowler (2006). Perception and Psychophysics
68: 17883. Briefly reviews phonetic context effects in literature.
Schouten, M. E. H., ed. 2003. The nature of speech perception. Speech
Communication 41: 1271. A collection of 20 papers on the psychophysics of speech perception.
Q
QUALIA ROLES
Qualia structure is a system of relations that characterizes the
semantics of a lexical item or phrase. The notion of qualia structure is derived in part from the Aristotelian theory of explanation
(Moravcsik 1975). An important semantic concept within generative lexicon theory (GL), qualia roles are the major building
blocks for constructing word and phrasal meaning in a language
compositionally.
GL (Pustejovsky 1995) is a theory of linguistic semantics,
which focuses on the distributed nature of compositionality in
b.
b.
693
Quantification
They can both, however, be coerced to accomplishments
(transitions) by specifying a termination predicate, assigned to
the formal role (cf. Pustejovsky 1995), for example, John ran to
the store, Mary baked a cake.
Recently, researchers in computational linguistics and lexicography have adopted the notion of qualia roles as one organizing principle in the process of building resources for lexical
knowledge bases.
James Pustejovsky
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Copestake, A. and T. Briscoe. 1992. Lexical operations in a unificationbased framework. In Lexical Semantics and Knowledge Representation,
ed. J. Pustejovsky and S. Bergler. Berlin: Springer-Verlag.
Dowty, D. 1979. Word Meaning and Montague Grammar. Dordrecht, the
Netherlands: Kluwer.
Moravcsik, J. M. 1975. Aitia as generative factor in Aristotles philosophy. Dialogue 14: 62236.
Pustejovsky, J. 1995. The Generative Lexicon. Cambridge, MA: MIT Press.
Pustejovsky, J., and B. Boguraev. 1993. Lexical knowledge representation
and natural language processing. Artificial Intelligence 63:193223.
QUANTIFICATION
Quantification has been a central concern in Logic and
Language at least since Aristotle, who systematized all valid and
invalid syllogisms involving the forms All/ some/ no/ not all As are
Bs (see Kneale and Kneale 1962). In linguistics, quantificational
phenomena played a role in upsetting the architecture of standard
transformational grammar (Chomsky 1965) in which deep
structure determines semantic interpretation. Many transformations that were meaning-preserving on sentences involving referential terms were not so when applied to quantifiers:
(1)
(2)
694
(4)
Quantification
Quantitative Linguistics
Typology of Quantification
Jon Barwise and Robin Cooper (1981) hypothesized that all
languages use NPs interpreted as generalized quantifiers.
Research reported in Bach et al. (1995) identified several languages that falsify that hypothesis. Other ways of expressing
quantification include adverbs of quantification like usually,
mostly (Lewis 1975), adjectives (numerous, Russian mnogie
many), nouns (majority, lot, dozen), and verbal prefixes
(Evans 1995).
Cross-linguistic studies show that languages differ with
respect to syntactic positions and strategies for expressing different quantificational notions (Szabolcsi 1997), with respect to the
degree to which surface structure constrains semantic quantifier
scope, with respect to the range of meanings a bare NP like
horses can have, in the variety and interpretation of indefinites
(Haspelmath 1997; Chung and Ladusaw 2003), in interactions
between nominal quantification and verbal aspect, and in other
ways that are still being explored.
Barbara H. Partee
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bach, Emmon, et al., eds. 1995. Quantification in Natural Languages.
Dordrecht, the Netherlands: Kluwer Academic.
Barwise, Jon, and Robin Cooper. 1981. Generalized quantifiers and natural language. Linguistics and Philosophy 4: 159219.
Chomsky, Noam. 1965. Aspects of the Theory of Syntax. Cambridge,
MA: MIT Press.
Chung, Sandra, and William A. Ladusaw. 2003. Restriction and Saturation.
Cambridge, MA: MIT Press.
Clark, Robin. 2007. Games, quantifiers, and pronouns. In Game Theory
and Linguistic Meaning, ed. A.-V. Pietarinen, 20728. Amsterdam:
Elsevier.
Evans, Nick. 1995. A-quantifiers and scope in Mayali. In Quantification
in Natural Language, ed. E. Bach et al., 20770. Dordrecht, the
Netherlands: Kluwer.
Haspelmath, Martin. 1997. Indefinite Pronouns. Oxford: Oxford University
Press.
Heim, Irene. [1982] 1989. The Semantics of Definite and Indefinite Noun
Phrases. New York: Garland.
Hintikka, J., and G. Sandu. 1997. Game-theoretical semantics. In
Handbook of Logic and Language, ed. J. van Benthem and A. ter
Meulen, 361410. Cambridge, MA: MIT Press.
Kamp, Hans. [1981] 1984. A theory of truth and semantic representation. In Truth, Interpretation, Information, ed. J. Groenendijk, Th.
Janssen, and M. Stokhof, 141. Dordrecht, the Netherlands: Foris.
Keenan, Edward. 1996. The semantics of determiners. In The
Handbook of Contemporary Semantic Theory, ed. Shalom Lappin.
Oxford: Blackwell. A survey of the semantic properties of quantifier
expressions by a leader in the field.
Keenan, Edward, and Jonathan Stavi. 1986. A semantic characterization of natural language determiners. Linguistics and Philosophy
9: 253326.
Kneale, William, and Martha Kneale. 1962. The Development of Logic.
Oxford: Oxford University Press.
Lewis, David. 1975. Adverbs of quantification. In Formal Semantics
of Natural Language, ed. E. L. Keenan, 315. Cambridge: Cambridge
University Press.
Montague, Richard. 1973. The proper treatment of quantification in ordinary English. In Approaches to Natural Language, ed. K. J. J. Hintikka
et al., 22142. Dordrecht, the Netherlands: Reidel.
Newmeyer, Frederick. 1980. Linguistic Theory in America: The First
Quarter-Century of Transformational Generative Grammar. New
York: Academic Press.
Partee, Barbara H. 1986. Noun phrase interpretation and type-shifting
principles. In Studies in Discourse Representation Theory and the
Theory of Generalized Quantifiers, ed. J. Groenendijk et al., 11543.
Dordrecht, the Netherlands: Foris.
. 1995. Quantificational structures and compositionality. In
Quantification in Natural Language, ed. Emmon Bach et al., 541602.
Dordrecht, the Netherlands: Kluwer. A formal semanticists approach
to semantic typology of quantificational expressions.
Szabolcsi, Anna, ed. 1997. Ways of Scope Taking. Dordrecht, the
Netherlands: Kluwer.
QUANTITATIVE LINGUISTICS
While the formal branches of linguistics use the qualitative
mathematical means (algebra, set theory) and logics to model
structural properties of language, quantitative linguistics (QL)
studies the multitude of quantitative properties as the essential
basis for the description and understanding of the development
and functioning of linguistic systems and their components. The
objects of QL research do not differ from those of other linguistic
disciplines. The difference lies, rather, in the ontological points
of view (do we consider a language as a set of sentences with
their structures assigned to them, or do we see it as a system
that is subject to evolutionary processes in analogy to biological
organisms, etc.) and, consequently, in the concepts that form the
basis of the disciplines.
Differences of this kind enable researchers to perceive new
phenomena in their area of study. A linguist accustomed to
thinking in terms of set theoretical constructs is not likely to find
the study of such properties as length, frequency, age, degree
of polysemy, and so on interesting or even necessary. Zipfs
Law is the only quantitative relation that almost every linguist
has heard about, but for those who are not familiar with QL,
it appears to be a curiosity more than a central linguistic law,
which is connected with a large number of properties and processes in language. From a quantitative point of view, however,
it is quite natural to detect features and interrelations that can
be expressed only by numbers. There are, for example, dependences of length (or complexity) of syntactic constructions
on their frequency and on their ambiguity; of homonymy of
grammatical morphemes on their dispersion in their paradigm; and of the length of expressions on their age, the dynamics of the flow of information in a text on its size, the probability
of change of a sound on its articulatory difficulty, and so on.
In short, in every field and on each level of linguistic analysis,
phenomena of this kind are significant. They are observed in
every language in the world and at all times. Moreover, it can
be shown that these properties of linguistic elements and their
interrelations abide by universal laws of language, which
can be formulated in a strict mathematical way in analogy to
the laws of the natural sciences. Emphasis has to be put on the
fact that these laws are stochastic; they do not capture single
cases (this would be neither expected nor possible) but, rather,
695
Quantitative Linguistics
predict the probabilities of certain events or certain conditions
in a whole. It is easy to find counterexamples with respect to any
of the examples cited above. However, this does not mean that
they contradict the corresponding laws. Divergences from a statistical average are not only admissible but even lawful they
are themselves determined with quantitative exactness. This
situation is, in principle, not different from that in the natural
sciences, where the old deterministic ideas have been replaced
by modern statistical/probabilistic models.
The role of QL is to unveil corresponding phenomena, to systematically describe them, and to find and formulate the laws
that explain the observed and described facts. Quantitative interrelations have an enormous value for fundamental research and
can also be used and applied in many fields, such as computational linguistics and natural language processing, teaching language, optimization of texts, and so on.
Historical Background
The first scientific counts of units of language or text were published in the nineteenth century. The first theoretical insight, after
many years of merely descriptive counts of various kinds, was
offered by the Russian mathematician A. A. Markov, who created
the base of the theory of Markov chains in 1913. This mathematical model of the sequential (syntagmatic) dependence among
units in linear concatenation in the form of transition probabilities was despite its mathematical significance of only little
use for linguistics. In modern natural language processing, however, (hidden) Markov models are a central component of many
methods in language technology.
Later, quantitative studies of linguistic material were, in
the first place, a consequence of practical demands: Efforts to
improve second language training (see bilingual education)
and optimization of stenographic systems are examples. The
unveiled interrelations between frequency of words and the
ranks of the frequency class (alternatively: between frequency
and the number of words in the given frequency class) were systematically investigated by the aforementioned George Kingsley
Zipf. He was the first to set up a model in order to explain the
observations and to find a mathematical formula for the corresponding function the famous Zipfs Law. Among his publications, his books ([1935] 1968 and 1949) are considered the
most important. Zipf formulated (in different terms) innovative
thoughts on self-organization, the principle of language economy, and fundamental properties of linguistic laws long before
modern systems theory arose. His ideas, such as the principle
of least effort and the forces of unification and diversification,
are still important today (even if they suffer from certain shortcomings). Later, his model was conceptually and mathematically
improved by Benot Mandelbrot (cf. Rapoport 1982), the originator of fractal geometry. Zipfs body of thought inspired various
scientific disciplines and enjoys increasing exposure again.
C. E. Shannon and W. Weaver (1949) applied information theory to linguistics without much success. Physicist Wilhelm Fucks
(1955) was responsible for a turn toward theoretical considerations in German QL. He studied, among others, word-length
distributions and various phenomena of language, literature,
and music. In France, Charles Muller (1973, 1979) created a novel
approach for studying the vocabulary of a text. In Russia, Zipfian
696
Quantitative Linguistics
Radical Interpretation
The highest degree of order is achieved with the help of metrical concepts, which are needed if the difference between the
amounts of a given property possessed by two objects plays a
role. In this case, the values of the property are mapped to the elements of a set of numbers in which the relations between these
numbers correspond to the relations between the values of the
properties of the objects. In this way, specific operations, such
as subtraction, correspond to specific differences or distances in
the properties between the objects. This enables the researcher to
establish an arbitrarily fine conceptual grid within his or her field
of study. Concepts of this kind are called interval-scale concepts.
If a fixed point of reference is added (e.g., zero), ratio-scaled concepts are obtained that allow the operation of multiplication and
division. Only the latter scale enables a formulation of how many
times object A has more than B of a property.
R
RADICAL INTERPRETATION
Radical interpretation is one of the central concepts in the work
of the American philosopher Donald Davidson (19172003). For
Davidson, to interpret speakers means to understand their linguistic utterances (cf. [1975] 1984, 157), to assign meanings to
their words and, in a slightly extended usage, assign contents to
their propositional attitudes. Radical interpretation takes place
in a specific scenario in which a person encounters the speakers
of a completely unknown language L. The radical interpreter has
the task of devising a formal semantic theory (see semantics)
for L on the basis of data of a very specific kind: His or her evidence consists entirely of (all available) data about the linguistic
and nonlinguistic behavior of the speakers of L in its observable
circumstances. According to Davidsons method, radically interpreting a language automatically includes systematic ascriptions
of belief to its speakers.
Background
The basic scenario of a field linguist trying to understand a
radically foreign language L on the basis of purely behavioral
data is introduced in W. V. O. Quines seminal work Word and
Object (1960). Here, the task is to construe a translation manual
for L, and Quine uses radical translation to consider how much
697
Radical Interpretation
of language can be made sense of in terms of its stimulus conditions, and what scope this leaves for empirically unconditioned
variation in ones conceptual scheme (Quine 1960, 26). Among
other things, he uses radical translation to argue for the indeterminacy of translation. This is the claim that on the
basis of the evidence available in radical translation, different
but equally correct translation manuals can be set up between
two languages L1 and L2, manuals diverging in a number of
places by translating a sentence of L1 into sentences of L2 which
stand to each other in no plausible sort of equivalence however
loose (p. 27). Davidson subscribes to indeterminacy for analogous reasons, even though in a more limited form (cf. Davidson
[1973] 1984, [1974] 1984).
The radical interpretation scenario derives its significance
for the theory of meaning, both in Quine and in Davidson, from
the foundational claim that meaning is entirely determined by
observable behavior (cf. Quine 1960, ix; Davidson 2005, 56).
According to this weak semantic behaviorism, the data available in radical interpretation are the data ultimately and entirely
determining the meanings of the expressions of the language
interpreted. In contrast to Quine, this determination remains
nonreductive in Davidson, but it is nevertheless both epistemic
and metaphysical in nature: The data available in radical interpretation are not only the ultimate evidence on the basis of
which meanings can be known; they are what (metaphysically)
determines or constitutes linguistic meaning. According to
Davidson, this is an individualistic affair not essentially involving social convention or a shared language; even though there
cannot be such a thing as a solitary speaker, one who never had
contact with other speakers, what a speaker means by his or her
words on an occasion of utterance is determined solely by his or
per own (dispositions to) behavior (cf. [1975] 1984, [1982] 1984,
[1992] 2001).
According to semantic behaviorism, meaning is determined
on a nonsemantic basis by data that can be described without
using semantic concepts, such as meaning or propositional
content. Davidson motivates his particular version of semantic
behaviorism partly by appeal to the essential publicness of language: The semantic features of language are public features.
What no one can, in the nature of the case, figure out from the
totality of the relevant evidence cannot be part of meaning
([1973] 1984, 135). According to Davidson, the relevant evidence
is evidence plainly available to an observer unaided by instruments (1994, 127). This restriction on the evidence stems from
the claim that terms like meaning and language are theoretical
terms deriving their significance from occasions of successful
linguistic communication (which do not, typically, involve the
use of any instruments).
Davidson thus transforms the basic question of the philosophical theory of meaning, the question What is it for words to
mean what they do? (1984, xiii), into two others: Given that we
can interpret the linguistic utterances of a speaker, what could
we know that would enable us to do this? How could we come
to know it? ([1973] 1984, 125). Radical interpretation addresses
the second of these questions and is supposed to show that there
is a method by which we can know meaning on the basis of the
nonsemantic evidence that, according to Davidson, determines
it (cf. 1994, 127).
698
T-theories can be constructed for significant parts of natural language (for a list of problems such as belief sentences or
counterfactuals, see Davidson [1973] 1984, 132). They are supposed to theoretically model semantic competence, but need not
be objects of knowledge for the speakers in any psychologically
realistic sense.
According to Davidson, the radical interpreter can devise a
T-theory for an unknown language L in two steps. On the basis
of their behavior, the interpreter can determine the sentences
of L that the speakers hold true in particular circumstances. This
amounts to detecting a propositional attitude: Holding a sentence true is having a belief, but so long as the sentence held
true remains uninterpreted, no meaning-theoretical question
is begged. In the second step, the radical interpreter uses data
about the circumstances under which speakers hold sentences
true to determine their truth conditions.
Holding a sentence true, however, is a product of two factors: what the sentence means and what the speaker believes to
be the case (cf. Davidson [1973] 1984, 134). Assigning a meaning to a sentence held true is ascribing a belief to the speaker.
Because of this interdependence of belief and meaning (p.
134), the ascription of belief needs to be restricted in relevant
ways if there is to be any evidence relation between holding true
and T-theory; so long as there is no such restriction, so long,
that is, as beliefs can be as absurd as the interpreter pleases, any
meaning can be assigned to any sentence. To establish such an
evidence relation is one of the main functions of the principle of
charity (see charity, principle of) (cf. Gler 2006, 340). It
tells the radical interpreter to assign truth conditions to the sentences of L such that the speakers of L have true and coherent
beliefs so far as that is plausibly possible. Since this can be done
only according to the interpreters own view of what is true,
application of charity amounts to agreement maximization or, as Davidson prefers, agreement optimization between
speaker and interpreter. The idea here is that some mistakes are
more destructive for understanding than others; an interpretation that avoids ascribing flagrant logical errors or very basic,
inexplicable perceptual mistakes is prima facie better than one
that does not. Ultimately, charity tells the interpreter to pick that
T-theory that stands in the relation of best fit (Davidson [1973]
1984, 136) to the totality of his or her data. In this way, the principle of charity fulfills its second main function: It allows for a
ranking of T-theories in terms of how well they fit the totality
Radical Interpretation
of the data, a ranking such that the best theory is the correct
one (cf. Gler 2006, 342). This amounts to a form of semantic
holism; the principle of charity determines all the meanings of
the expressions of L together, and on the basis of the totality of
the evidence (cf. Pagin 1997, 13, 18). indeterminacy, then, is
the claim that there can be more than one best T-theory for
any given natural language L.
Davidson provides what he calls a crude outline ([1973]
1984, 136) for the process of devising a T-theory on the basis of
data about holding true attitudes. It has three steps: First, the
logical form of the sentences of L is determined. Use of a
T-theory as a formal semantic theory requires paraphrasing the
expressions of L in the language of first-order quantified logic
(plus identity). The relevant evidence for this first step consists
of sentences that are held true (or false) under all circumstances
(candidates for logical truth or falsity) and of patterns of inference, that is, sentences held true on the basis of other sentences
held true.
The second step focuses mainly on sentences containing
indexicals, expressions whose interpretation depends on features of the context, such as I or here. Take the sentence It
is raining or its German equivalent Es regnet. Their truth value
varies with easily observable circumstances in the environment
of the speaker. The idea (according to Davidson [1973] 1984, 135)
is to take data of the form
(E) Kurt belongs to the German speech community and Kurt
holds true Es regnet on Saturday at noon and it is raining near
Kurt on Saturday at noon as evidence for a T-sentence of the
form
(R) Es regnet is true-in-German when spoken by x at time t if
and only if it is raining near x at t.
Reading
externalim currently dominating the theory of meaning and
content. However, foundational issues such as these remain
insufficiently explored; because of the role the shared environment plays in radical interpretation, Davidson thought of himself as a social and physical externalist, though clearly not of the
mainstream kind (cf. 2001). So long as a systematic comparison of these competing accounts of meaning determination
is lacking, it remains premature to simply write off semantic
behaviorism; prima facie, it is not even clear that Davidsonian
semantic behaviorism and mainstream (physical) externalism
are incompatible.
Kathrin Gler
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Davidson, Donald. [1967] 1984. Truth and meaning. In Davidson 1984,
1736.
. [1973] 1984. Radical interpretation. In Davidson 1984, 12539.
. [1974] 1984. Belief and the basis of meaning. In Davidson 1984,
14154.
. [1975] 1984. Thought and talk. In Davidson 1984, 15570.
. [1982] 1984. Communication and convention. In Davidson
1984, 26580.
. 1984. Inquiries into Truth and Interpretation. Oxford: Clarendon
Press.
. [1992] 2001. The second person. In Subjective, Intersubjective,
Objective, 10722. Oxford: Clarendon.
. 1994. Radical interpretation interpreted. Philosophical
Perspectives 8: 1218.
. 2001. Externalisms. In Interpreting Davidson, ed. P. Kotatko,
P. Pagin, and G. Segal, 116. Stanford, CA: CSLI.
. 2005. Truth and Predication. Cambridge, MA: Belknap.
Fodor, Jerry, and E. Lepore. 1994. Is radical interpretation possible?
Philosophical Perspectives 8: 10119.
Gler, Kathrin. 2006. The status of charity: I. Conceptual truth or aposteriori necessity? International Journal of Philosophical Studies
14: 33760.
Lepore, Ernest, and K. Ludwig. 2005. Donald Davidson: Meaning, Truth,
Language, and Reality. Oxford: Clarendon.
Lewis, David. 1974. Radical interpretation. Synthese 23: 33144.
Pagin, Peter. 1997. Is compositionality compatible with holism? Mind
and Language 12: 1133.
Quine, Willard V. O. 1960. Word and Object. Cambridge, MA: MIT Press.
Ramberg, Bjrn. 1989. Donald Davidsons Philosophy of Language,
Oxford: Blackwell.
Rawlings, Pierce. 2003. Radical interpretation. In Donald Davidson, ed.
K. Ludwig, 85112. Cambridge: Cambridge University Press.
READING
Reading is the process of decoding and comprehending written
language. Decoding, the conversion of written forms into linguistic messages, is central to this definition to the extent that
the comprehension of written language shares its features with
the comprehension of spoken language.
Reading connects printed information conveyed in a writing
system with the readers knowledge of the language encoded by
that system. Writing systems vary in their mapping principles, in
their implementation in a particular language (the orthography),
and in their visual appearance (the script). Alphabetic writing
systems map graphic units to phonemes. Syllabary systems,
699
Reading
represented by Japanese Kana, map graphic units to spoken
language syllables. Chinese is usually classified as logographic
because its graphic units (characters) correspond primarily to
morphemes. However, the fact that characters have components that provide syllable-level pronunciation information justifies an alternative designation of morpho-syllabic (DeFrancis
1989).
700
Reading Comprehension
Reading comprehension shares linguistic and cognitive processes with spoken language and correlates highly with it among
adults (Gernsbacher 1990). Because this correlation is based on
the use of equivalent texts across listening and reading, it may
miss the differences between ordinary spoken language and typical written texts that arise from their divergent syntactic structures, lexicons, and other aspect of their different registers. While
reading comprehension strongly depends on listening comprehension, reading and speech each place specific demands on
comprehension processes.
Reading comprehension processes begin with word identification and include context-relevant selection of word meanings and parsing (see parsing [human]), the basic process of
extracting grammatical relations among words in a sentence.
Beyond these word- and sentence-level basics, higher-level comprehension involves readers constructions of mental models
of text information. One mental model is based closely on the
language of the text, and another, the situation model, reflects
what the text is about (van Dijk and Kintsch 1983). The reader
builds a situation model from the linguistically based model (the
text base) by combining knowledge sources through additional
inference processes. A situation model may contain nonlinguistic representations, including spatial imagery (Glenberg, Kruley,
and Langston 1994) and the temporal organization of events
(Zwaan 1996), among others. Reading multiple texts that refer to
the same situation challenges the construction of a single situation model (Perfetti, Rouet, and Britt 1999) and requires additional comprehension skills in document use and evaluation
(Rouet 2006).
Because texts are never fully explicit, comprehension
research has had an enduring interest in inferences. Inferences
that link anaphora (e.g., pronouns) with their antecedents to
establish coreference are a routine part of comprehension. The
extent of elaborative and predictive inferences (Graesser, Singer,
and Trabasso 1994) is more in doubt (McKoon and Ratcliff 1992).
For example, the sentence The American tour group went to
London for its annual holiday may evoke an inference that the
group traveled by airplane, but whether a reader actually makes
this inference appears to be highly variable. Inferences about
causeeffect relations may be more likely than other kinds of
elaborative inferences (Trabasso and Suh 1993).
Comprehension skill is highly variable. Some children
appear to have a comprehension-specific problem (i.e., without
Reading
a decoding problem) that is general across reading and spoken
language (Nation and Snowling 1999; see also disorders of
reading and writing). The potential causes for comprehension problems include failures to make inferences during reading (Oakhill and Garnham 1988) and limitations in working
memory functions, among other factors (Nation 2005). Unstable
knowledge of word form and meaning (low lexical quality) also
contributes substantially to comprehension problems (Perfetti
and Hart 2001).
Learning to Read
In an alphabetic writing system, a child learns that letters and
strings of letters correspond to speech segments. For English,
this process is complicated by inconsistent orthography at the
letter-phoneme level, for example, the contrasts between choir
and chore and head and bead. Most European languages tend to
be coded by orthographies that more consistently map graphemes to phonemes, and learning to read reflects this fact; for
example, childrens errors reflect letter-to-phoneme decoding
procedures more than in English (Wimmer and Goswami 1994;
see also childrens grammatical errors).
Important for the alphabetic principle is phonemic awareness (see phonological awareness), the explicit understanding that the speech stream can be segmented into a set of
meaningless units (phonemes). Childrens phonemic awareness
correlates with early reading success, and phoneme segmentation instruction produces gains in reading. However, alphabetic
literacy experience itself affects awareness of phonemes, as
shown by studies of adults without exposure to alphabetic writing (Morais et al. 1979) and of Chinese who learned to read prior
to the introduction of the alphabetic Pinyin system (Read et al.
1986) as well as by longitudinal results that show a bidirectional
relation between phonological sensitivity and literacy (Perfetti
1992).
Theories of learning to read have usually referred to a series
of stages (Ehri 1991, 2005; Frith 1985; Gough and Hillinger 1980).
Alternative theoretical accounts emphasize the incremental
acquisition of decodable lexical representations and the role of
phonology to establish word-specific orthographic representations (Perfetti 1992; Share 1995; see also writing and reading, acquisition of).
Charles Perfetti
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Coltheart, Max, B. Curtis, P. Atkins, and M. Haller. 1993. Models
of reading aloud: Dual-route and parallel-distributed-processing
approaches. Psychological Review 100: 589608.
Coltheart, Max, K. Rastle, C. Perry, R. Langdon, and J. Ziegler. 2001. The
DRC model: A model of visual word recognition and reading aloud.
Psychological Review 108: 20425. Explains and defends the dual route
theory and its evidence.
DeFrancis, John. 1989. Visible Speech: The Diverse Oneness of Writing
Systems. Honolulu: University of Hawaii.
Ehri, Linn C. 1991. Learning to read and spell words. In Learning
to Read: Basic Research and Its Implications, ed. L. Rieben and
C. A. Perfetti, 5773. Hillsdale, NJ: Erlbaum.
. 2005. Development of sight word reading: Phases and findings.
In Snowling and Hulme 2005, 13555.
701
Realization Structure
Share, David L. 1995. Phonological recoding and self-teaching: Sine qua
non of reading acquisition. Cognition 55: 151218.
Siok, Wai T., C. A. Perfetti, Z. Jin, and L. H. Tan. 2004. Biological abnormality of impaired reading constrained by culture: Evidence from
Chinese. Nature 431 (September 1): 716.
Snowling, Margaret J., and C. Hulme, eds. 2005. The Science of Reading: A
Handbook. Oxford: Blackwell. A compendium of reviews across the
entire spectrum of reading-related research.
Stanovich, Keith E., and L. S. Siegel. 1994. Phenotypic performance profile of children with reading disabilities: A regression-based test of the
phonological-core variable-difference model. Journal of Educational
Psychology 86.1: 2453.
Trabasso, Tom, and S. Suh. 1993. Understanding text: Achieving explanatory coherence through online inferences and mental operations in
working memory. Discourse Processess 16: 334.
Van Dijk, Teun A., and W. Kintsch. 1983. Strategies of Discourse
Comprehension. New York: Academic Press.
Van Orden, Guy C., and S. D. Goldinger. 1994. The interdependence of
form and function in cognitive systems explains perception of printed
words. Journal of Experimental Psychology: Human Perception and
Performance 20: 126991.
Wimmer, Heinz, and U. Goswami. 1994. The influence of orthographic
consistency on reading development: Word recognition in English and
German children. Cognition 51: 91103.
Zwaan, Rolf. 1996. Processing narrative time shifts. Journal of
Experimental Psychology: Learning, Memory and Cognition
22.5: 11961207.
REALIZATION STRUCTURE
This term was coined by Keith Oatley (2002) to indicate how
one experiences a piece of literature. One does not just receive
or interpret it but realizes it, bringing it into being. The idea that
fiction involves such a realization or mental performance of the
piece by the reader or audience member was discussed in philosophy by Wolfgang Iser (1974) and in psychology by Richard J.
The relationship among the four aspects of a piece of literary prose or poetry can be illustrated by the diagram in Figure 1
(from Oatley 2002, 45). The implication of the layout of the diagram is that the event structure starts off a story, usually by means
of a setting, that the discourse structure and suggestion structure
occur simultaneously, and that the realization structure is a
resultant of the other processes.
Keith Oatley
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Brewer, William, and Ed Lichtenstein. 1981. Event schemas, story schemas and story grammars. In Attention and Performance. Vol. 9. Ed.
J. Long and A. Baddeley, 36379. Hillsdale, NJ: Erlbaum.
Event Structure
The events of the story world.
A creation of the author.
Discourse Structure
Suggestion Structure
ideas.
Realization Structure
The enactment in the mind of the reader or
watcher, which results from the structure
process and suggestion structure
being applied to the discourse structure.
Figure 1.
702
Such a relation to authority is often the sort of case that writers on zheng ming had in mind. In his famous treatment of the
rectification of names, Hsn Tzu (Xunzi) lamented the result of
verbal confusions wherein the distinction between the noble
and the humble is not clear and similarities and differences are
not discriminated. When this occurs, ideas will be misunderstood and work will encounter difficulty or be neglected (1963,
125). A similar practical concern is found in Mozi (Mo Tzu), but
with different political consequences. Thus, Mozi wrote against
those who distinguish names in the world in such a way as to
promote distinctions and discriminations. He favored those
who distinguish names in the world in such a way as to love
others and benefit others by advocating no discrimination
(quoted in Zhang 2002, 327).
The practical, political consequences of the rectification of
names are connected with the eight steps set out in the classic
Great Learning. The steps explain how to manifest ones virtue
by bringing order to a state. One brings order to ones state by
bringing order to ones family. One accomplishes this by developing oneself. That, in turn, results from rectifying ones mind/
heart, which derives from integrating ones thoughts. To do the
latter, one must extend ones knowledge, which is itself effected
by research. (On the eight steps, see Great Learning and Zhang
2002, 452. Note that the list of steps inverts their order in practice, with research, therefore, being the first step.) Knowledge is
at least in part a matter of knowing the right words and applying them properly. Thus, research is itself in part a matter of the
rectification of names. In research, one should seek the relevant
principles and connect them with names so that they will be
understood. Moreover, these principles, and our knowledge of
these principles, are not solely a matter of conceptualization.
As Zhang writes, names should match the reality, and reality
includes the actual use of the object (2002, 424). Zhang is referring to a Mohist idea. But the link between principle and practice
is much more widespread.
The eight steps also suggest the manner in which zheng
ming has consequences. In a given situation, our application of
a particular name affects our knowledge of the situation (Step
2). Moreover, it brings our understanding of the situation into
a complex of relations with other ideas and names (or words)
in such a way as to change the way our thoughts are integrated
(Step 3) and our heart or feeling is oriented (Step 4). We may
return here to the previous example. If I follow the lead of the
U.S. government and characterize the U.S. presence in Iraq as
liberation, then I understand its consequences in certain ways.
I also integrate this idea with other aspects of my thought
about the insurgents, about U.S. foreign policy, and so on. This,
in turn, has consequences for my emotional response to the
U.S. presence in Iraq, to the U.S. government, and so on ultimately with consequences for my support of that government.
The consequences are very different if I follow the lead of most
Iraqis and categorize the presence as occupation. Determining
which is correct, and using that term consistently in public discussion, is an instance of research leading to the rectification
of names, then to the integration of thought and orientation of
feeling.
Patrick Colm Hogan
703
704
being able to represent the relationship between the representation and what it refers to: to understand that a picture of
Niagara Falls stands for that visual scene, or that someones
belief that Santa Claus exists represents that potential state of
the world, or that rocks, the noun, refers to a set of stone objects.
Metarepresentation requires recursive embedding of representational relationships, but it is not identical to recursive embedding (Stone and Gerrans 2006). Metarepresentation may also be
uniquely human (Suddendorf 1999).
Marc Hauser, Noam Chomsky, and W. T. Fitch (2002) have
offered the hypothesis that recursion is the defining feature of
language, making it uniquely human. Other features of language,
however, do not follow directly from recursion and also seem to
be uniquely human, such as words, fine phonemic discriminations, and motor control of mouth, larynx, and so on (enumerated in Pinker and Jackendoff 2005; Parker 2006, Chapter 5).
Whether recursion is the single defining feature of language
or not, it might be uniquely human. Testing for recursive capacity directly is difficult. Instead, researchers rely on demonstrations of animals ability to do tasks dependent on explicit
recursion. Some claim that animals do implicit recursion in
certain tasks, for example, ants doing dead reckoning, but this is
difficult to substantiate. Although recursion is an efficient solution to many problems, unless one can test for the explicit content of the recursive steps in a computation, it is always possible
that animal brains solve problems using some other, nonrecursive computational technique. Thus, comparative research uses
tasks believed to depend on explicit recursion: mathematics,
theory of mind, problem solving involving interdependent steps,
mental time travel, or certain kinds of syntax (Corballis 2003;
Parker 2006). So far, no study has demonstrated that our closest
relatives, great apes, can do any of these tasks with the range and
flexibility of humans (Corballis 2003; Hauser 2005; Suddendorf
2006). For now at least, recursion can join a set of possibly unique
human cognitive capacities: other aspects of language, flexible
control of attention and inhibition, expanded working memory
capacity, and metarepresentation (Suddendorf 1999; Kawai
and Matsuzawa 2000; Hauser 2005; Pinker and Jackendoff 2005;
Stone and Gerrans 2006). Recursion may not be the key to unique
human cognition, but it is no less worthy of study for being one
of many keys.
Valerie Stone
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Anderson, A. K. 2007. Recursion in programming. Available online
at: http://www.usableresults.com. Retrieved January 15, 2007.
Corballis, M. C. 2003. Recursion as the key to the human mind. In
From Mating to Mentality, ed. K. Sterelny and J. Fitness, 15571. New
York: Psychology Press.
Hauser, M. 2005. Our chimpanzee mind. Nature 437.7055: 603.
Hauser, M., N. Chomsky, and W. T. Fitch. 2002. The faculty of language: What is it, who has it, and how did it evolve? Science
298.5598: 156979.
Kawai, N., and T. Matsuzawa. 2000. Numerical memory span in a chimpanzee. Nature 403.6765: 3940.
Parker, A.P. 2006. Evolution as a constraint on theories of syntax. Ph.D.
thesis, University of Edinburgh. Available online at: http://www.ling.
ed.ac.uk/~annarp. Retrieved February 4, 2007.
argument, the referent of F is the function that maps the referent of X onto the referent of F(X). Thus, the referent of a complex
expression is always the result of applying the referent of one of its
constituents, as a function, to the referents of its other constituents, taken as arguments. The referent of a sentence as a whole is
identified with its truth value. Thus, the referent of Chomsky is
Chomsky, the referent of is clever the function that maps each
object x onto truth if x is clever and onto falsehood otherwise, and
the referent (truth value) of Chomsky is clever is truth if and
only if Chomsky is clever.
It may seem surprising that the referent of a sentence is its
truth value, but it should be kept in mind that reference is used as
a technical concept within compositional semantics. Given the
use to which the concept is put, this is not an unnatural assumption: Frege was interested in a compositional semantics that
would tell us how the truth values of sentences are determined
by the referents of their parts, and all natural languages have
fragments in which, when a sentence has other sentences as parts,
the truth value of the whole depends only on the truth values of
the constituent sentences. Fragments of languages in which this
is the case, and in which the referent of a complex expression
in general depends only on the referents of its parts, are called
extensional. Thus, in an extensional fragment, expressions having the same referent can be substituted in any sentence without altering its truth value (contexts in which such substitutions
preserve truth value are also called extensional). Frege was primarily interested in constructing a semantics for the language
of mathematics, which is extensional, and so choosing truth
values as referents of sentences was natural. However, natural
languages as wholes are not extensional. In contexts involving
propositional attitudes, modality, and counterfactuals,
the substitution of clauses having the same truth value may alter
the truth value of the whole sentence. To account for such contexts, Frege held that each sentence or other expression has, in
addition to a referent, another kind of semantic value, which he
called the expressions sense (Sinn). The sense of a sentence is
what he called a thought, or, in contemporary terms, a proposition. In order to maintain a version of the principle of compositionality, he held that the truth values of nonextensional
sentences are determined in part by the senses of their constituents (see sense and reference).
For various reasons, Freges approach is now considered antiquated. Most recent work in formal semantics for natural languages is inspired by Alfred Tarskis work on the definability of
truth for formal languages. Richard Montague (1974) was the
first to apply Tarskis work productively to (fragments of) natural
languages. Here, extension is the preferred term. The extension of
a predicate is, again, the set of things to which it applies. Although
terminology varies, in this framework, too, one can speak of the
extension of almost any expression, including a sentence, so that
one identifies a sentences extension with its truth value. Applying
Tarskis approach, the aim is to recursively characterize not only
the truth conditions of sentences but also the entailments (logical consequences) for a language using the notion of extension: A
sentence S1 is said to entail a sentence S2 in language L if and only
if there is no assignment of extensions to the semantically simple
expressions of L (no model of L) under which S1 is true and S2
false. On this approach, the logical constants differ from other
705
Reference as Action
The third view we discuss maintains that reference depends
essentially on individual speakers (and possibly interpreters)
with variable interests: An appropriate slogan might be Words
dont refer; people do. One root of this view is found in the
work of the later Ludwig Wittgenstein, another in Descartes.
Those who defend it point out that it is difficult to find cases
of uniform wordworld relationships in the use of natural
languages. They grant that the practices of mathematicians
display uniformity, but these practices aside, reference varies
with time, context, speakers interests, and so on. They also
grant that some who offer theories of reference, such as Kripke
(1972), acknowledge a role for speaker intentions. But Kripke
and others incorrectly assume that ordinary speakers desire to
706
Reference Tracking
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Chomsky, Noam. 2000. New Horizons in the Study of Language and Mind.
Cambridge: Cambridge University Press.
Donnellan, Keith. 1966. Reference and definite descriptions.
Philosophical Review 75.3: 281304.
Frege, Gottlob. [1892] 1997. On Sinn and Bedeutung. In The Frege
Reader, ed. M. Beaney. Oxford: Blackwell.
Kripke, Saul. 1972. Naming and necessity. In Semantics of Natural
Language, ed. Donald Davidson and Gilbert Harman, 253355, 76369.
Dordrecht, the Netherlands: Reidel.
. 1976. Speakers reference and semantic reference. Midwest
Studies in the Philosophy of Language 2: 25576.
. 1980. Naming and Necessity. Cambridge: Harvard University
Press.
Montague, Richard. 1974. Formal Philosophy. New Haven and
London: Yale University Press.
Putnam, Hilary. 1975. The meaning of meaning. In Language, Mind,
and Knowledge, ed Keith Gunderson, 13193. Minneapolis: University
of Minnesota Press.
Strawson, P. F. 1950. On referring. Mind 59.235: 32044.
Register
grammatical markings, such as agreement morphology: subjectverb agreement, person, number, and gender (see gender
marking) agreement; the switch-reference system (Amele); or
topic/subject markers (Japanese). Discourse analysis (see discourse analysis [linguistic]) finds that reference usage
follows the constraints of information flow: The grammatical
subject of a transitive clause tends to be coded with a pronoun
in English, or zero anaphora in Chinese or Japanese, to present
given or accessible information (the light subject constraint), for
instance, He in the example, whereas the grammatical object
tends to be an NP carrying new information (e.g., one plant). This
allows easy processing of accessible information early in an utterance, while the rest of the utterance introduces the new referent,
thus facilitating reference tracking and discourse processing.
Experimental studies find that the discourse pattern of a
language engenders specific reference tracking strategies in its
native speakers. Therefore, speakers of different languages may
develop different cognitive strategies to track reference during
discourse processes.
Liang Tao
REFERENCE TRACKING
Reference tracking, or ANAPHORA resolution, concerns how
language users track who or what the speaker is referring to in
discourse. Because everyday language use generally concerns
who does what to whom, reference tracking is important in
studying human language and cognition.
Anaphora devices include noun phrases (NPs), pronouns,
and zero anaphora, whose identities depend on their antecedents in discourse. In the example Isabel went to China, and this
volunteer/she helped with midwifery training, This volunteer
is an NP that refers back to its antecedent Isabel. It could be
replaced by the pronoun she or by an empty slot (zero anaphora)
as in the sentence Isabel went to China and ____ helped with
midwifery training.
Pronouns and zero anaphora give less explicit information
than full NPs. Still, the reader/hearer benefits from the efficiency
of these devices in conveying information that has been introduced/given in the prior discourse or can be accessed/inferred
from the context. These devices are crucial for global cohesion
and local coherence in discourse. Experimental studies find that
without a specific need, replacing a pronoun with an NP for given
information may hinder understanding.
A discourse topic provides a basic means for tracking the
identity of a pronoun or zero anaphora because the topic tends
to recur as given information continuously. Cross-linguistic
studies find that people can track the identity of a pronoun or
zero anaphora even when its referent is not in the immediately
preceding clause but in the prior context. Therefore, although
language production may be linear due to human physical limitations, language processing and reference tracking are hierarchical cognitive processes.
Reference tracking requires the hearer to make inferences
from world knowledge about likely events, especially for languages that have no morphological markings (Chinese) yet
allow abundant zero anaphora, as in He grew only one plant,
but ___ blossomed well. Many languages (e.g., French, German,
and Turkish) make reference tracking easier with specific
REGISTER
Speakers of a language use different words and grammatical
structures in different communicative situations. For example,
we do not use the same words and structures to write an academic term paper that we would use when talking to a close
friend about weekend plans.
Researchers study the language used in a particular situation
under the rubric of register: a language variety defined by its situational characteristics, including the setting, interactiveness, the
channel (or mode) of communication, the production and processing circumstances, the purposes of communication, and the
topic.
Although registers are defined in situational terms, they can
also be described in terms of their typical linguistic characteristics; most linguistic features are functional and, therefore, they
tend to occur in registers with certain situational characteristics.
For example, first and second person pronouns (I and you) are
especially common in conversation. Speakers in conversation
talk a lot about themselves, and so they commonly use the pronoun I. These speakers also interact directly with another person,
often using the pronoun you.
There are many studies that describe the characteristics of
a particular register, such as sports announcer talk (Ferguson
1983), note taking (Janda 1985), classified advertising (Bruthiaux
1996), and scientific writing (Halliday 1988). Other researchers
707
Regularization
take a comparative approach, studying the patterns of register
variation, which seems to be inherent in human language.
corpus linguistics has been an especially productive
analytical approach for studying register variation. For example,
the Longman Grammar of Spoken and Written English (Biber et
al 1999) applies corpus-based analyses to show how any grammatical feature can be described for both its structural characteristics and its patterns of use across spoken and written registers.
In the multidimensional approach to register variation, corpusbased analysis is combined with sophisticated statistical analysis to analyze the patterns of linguistic variation that distinguish
among registers (see, e.g., Biber 1988, 2006; Conrad and Biber
2001).
Some studies have distinguished between registers and
genres. In these studies, the term register refers to a general kind
of language associated with a domain of use, such as a legal register, scientific register, or bureaucratic register. Register studies
have usually focused on lexico-grammatical features, showing
how the use of particular words and grammatical features vary
with the situation of use. In contrast, the term genre has been
used to refer to a culturally recognized message type with a
conventional internal structure, such as an affidavit, a biology
research article, or a business memo. Genre studies have usually
focused on the conventional discourse structure of texts or the
expected sociocultural actions of a discourse community.
Douglas Biber
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Biber, Douglas. 1988. Variation across Speech and Writing.
Cambridge: Cambridge University Press.
. 2006. University Language: A Corpus-Based Study of Spoken and
Written Registers. Amsterdam: John Benjamins.
Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad, and
Edward Finegan. 1999. Longman Grammar of Spoken and Written
English. London: Longman.
Bruthiaux, Paul. 1996. The Discourse of Classified Advertising. New
York: Oxford University Press.
Conrad, Susan, and Douglas Biber, eds. 2001. Variation in English: MultiDimensional Studies. London: Longman.
Ferguson, Charles A. 1983. Sports announcer talk: Syntactic aspects of
register variation. Language in Society 12: 15372.
Ghadessy, M., ed. 1988. Registers of Written English: Situational Factors
and Linguistic Features. London: Pinter.
Halliday, M. A. K. 1988. On the language of physical science. In
Ghadessy 1988, 16278.
Janda, R. 1985. Note-taking English as a simplified register. Discourse
Processes 8: 43754.
REGULARIZATION
Regularization is the process of becoming more regular or
rule governed. In a general sense, it implies that events in the
world (or in individuals) become more predictable and orderly.
Regularization in language is one of many kinds of language
change. It occurs when forms in the language are under pressure
to be consistent with a common pattern. The patterns that speakers use in the regularization of the existing language are typically
the same as those that are used with new forms that enter the language and are referred to as productive. Thus, the past tense -ed
708
Regularization
the official or preferred language, and it does not follow that
the chosen dialect will be any more regularized than less valued varieties. Standardized spelling, by contrast, is typically
a move to simplify and regularize spelling conventions (Wijk
1977). There are cases, however, where official bodies are delegated to choose or change words or expressions in order to
provide conformity with the rules of a language: For instance,
the Academy of the Hebrew Language in Israel is charged with
replacing words borrowed from other languages with words
based on Hebrew semantics and morphology, in this way
providing a kind of regularization of the lexicon of contemporary Hebrew. The French Academy (lAcadmie franaise)
compiles a dictionary of the French language and makes recommendations as to the admissibility of loan words, usually in
favor of words that are more French for instance, courier lectronique instead of e-mail. These attempts to control language
by decree are often not successful, especially in keeping out loan
words that are pervasive and international. In any language,
loan words themselves are subject to regularization that brings
them into conformity with the morphology and phonotactics of
the target language. For example, the English word baseball has
been incorporated into Japanese as beisuboru, and the umpires
call to play ball is rendered as purei boru.
A powerful regularizing force in a language is the communication pressure that results from becoming more cosmopolitan,
whereas an insular language is more likely to remain irregular
and complex. When a society becomes more multiethnic, there
is often language contact. In addition, a rapidly changing
society may have new technologies and terminologies, a mobile
population carrying language to new communities, and many
speakers who do not know the language well. All of these factors lead to linguistic economy, simplification, and regularization. When new words are coined, their inflections are regular.
Loan words from contact languages are regularized. And, like
first language learners, second language learners (see second
language acquisition) of whatever age tend to regularize
the language they are acquiring. It is very common for learners
of French, for instance, to regularize the irregular second person plural of dire (to say), producing vous disez instead of vous
dites. This regularized form is not permissible in standard French,
but it occurs in various nonstandard dialects and is the norm in
Louisiana Cajun French. When a language is spoken in a homogeneous, isolated community with few non-native speakers and
little outside contact, there is apparently a greater tendency for
complexity to remain in the inflectional system. A comparison
of modern Icelandic and Norwegian provides a good example of
this tendency: Both have Old Norse beginnings, but the relatively
isolated Icelandic has preserved complex morphology, including
many of the strong verbs of Old Norse, whereas comparatively
cosmopolitan Norwegian shows more regularization, with simpler inflectional morphology and many fewer Old Norse strong
verbs (Kusters 2003).
Psycholinguistic Studies
Some of the best evidence we have for the processes of paradigmatic regularization has come from psycholinguistic studies of
both children and adults. We know that adult speakers of English,
for instance, are able to produce appropriate inflectional endings
709
Relevance Theory
relies on exemplars or analogies: Speakers hear a variety of
words over time, some more than others, and their mental representation reflects the weight of these frequencies, the features
of the words, and the circumstances of their use. Frequently
encountered features are recognized and associated, and ultimately the learner produces language that matches the language
that has been heard.
Regardless of the theoretical model, it is clear that individuals
are sensitive to the characteristics of the language around them
and are able to generalize those characteristics to new instances.
By the time children are of preschool age, they have sufficient
knowledge of the most regular features of language to be able to
extend them to words they have never heard before. This kind of
knowledge underlies both productivity and regularization, and
it has implications for language change in general. Languages
everywhere are moved to become more regular in response to
communicative pressure. Speakers carry within themselves the
linguistic tools of regularization, which is a process that reveals
a fundamental characteristic of the way humans organize
information.
Jean Berko Gleason
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Baugh, Albert C. 1957. A History of the English Language. New
York: Appleton-Century-Crofts.
Berko, Jean. 1958. The childs learning of English morphology. Word
14: 15077.
Gleason, Jean Berko. 1966. Do children imitate? Proceedings of the
International Conference on Oral Education of the Deaf 2: 441448.
Washington, DC: Alexander Graham Bell Association for the Deaf.
Kusters, Wouter. 2003. Linguistic complexity: The influence of social
change on verbal inflection. Ph.D. diss., University of Leiden.
Marcus, Gary, S. Pinker, M. Ullman, M. Hollander, T. J. Rosen, and Fei Xu.
1992. Overregularization in language acquisition. Monographs of the
Society for Research in Child Development 57: 1182.
Pinker, Steven. 1991. Rules of language. Science 253: 5305.
Wijk, Axel. 1977. Regularized English: A Proposal for an Effective
Solution of the Reading Problem in the English-Speaking Countries.
Stockholm: Almqvist and Wiksell.
RELEVANCE THEORY
Relevance theory (RT) is best known for its account of verbal
communication and comprehension, which is employed by
people working in pragmatics. It also sets out a broad picture of the principles driving the human cognitive system as a
whole, and this plays a crucial role in underpinning the particular claims made about communication (Sperber and Wilson
[1986] 1995a; Wilson and Sperber 2004).
710
Relevance Theory
(3)
a. Ive eaten.
Expressing: Ive eaten dinner tonight.
b. Your knee will take time to heal.
Expressing: The knee will take a substantial amount of time to
heal.
c. The water is boiling.
Expressing: The water is very hot [not necessarily strictly at
boiling point].
as: Y might trip over, Y should take small steps, Y should keep an
eye on the path, and so on. The result is a concept, which we can
label uneven*, whose denotation is a proper subset of the denotation of the lexical concept uneven.
A distinctive RT claim in this context is that metaphorical
and hyperbolic uses of words involve a kind of concept broadening (loose use), and so fall within this single process of lexical
meaning adjustment. For instance, an utterance of the sentence
in (5) could be taken as an ordinary broadening (if, say, the run
referred to was a little less than 26 miles) or as hyperbolic (if it
was considerably less than the length of a marathon) or as metaphorical for a long, arduous, exhausting experience, whether
physical or mental:
(5)
It was a marathon.
711
712
Representations
REPRESENTATIONS
Definition
The core concept of a representation is made up of two distinguishable ideas.
(1) Any representation, r, must be about, or mean, or have as
its content, some distinct item s;
(2) r must be employable, in the absence of s, to guide the
behaviors of the consumers of r organisms or systems that
use r with respect to s.
To say that these ideas are distinguishable is not to say that they
are separable. On some accounts of representation, (1) is a function of (2). That is, a representation can have a distinct item as its
content only because of the way it is utilized by its consumers.
The relation between (1) and (2) is, however, disputed.
713
Representations
is neither necessary nor sufficient for content. It is not necessary because, typically, when I use the word dog, no such image
appears to me. It is not sufficient because the image is, itself, just
another symbol, and its content needs to be interpreted no less
than that of the word. An image of a dog can stand for many
things dog, furry creature, creature with four legs, and so on
and so the image can have content only if there is a further act
of interpretation on the part of its subject: an act that serves
to disambiguate the image. However, an act of interpretation is
itself an intentional state. Therefore, we have made no progress
in understanding how an intentional state has content; we have
simply replaced one intentional state with another.
These problems remain when we shift to more sophisticated
accounts of content. In one such account, content is understood
in terms of the following of rules: In using a sign, I consciously
or unconsciously follow a rule, a rule that determines how the
sign is to be applied in particular cases. Wittgenstein argued that
there is no fact about an individual that determines that he or
she is following one rule rather than another. Any behavior in
which I engage is compatible with an indefinite number of rules.
In continuing the mathematical sequence 2, 4, 6, 8, I could be following the n + 2 rule or the n + 2 if and only if n is less than 32,
if so n + 4 rule. And so on for an infinite number of ns. Similarly,
in applying the word dog, I could be following the Apply dog to
all and only dogs rule, or the Apply dog to all and only dogs
unless the dog is first seen after 2020 in which case apply cat,
and so on for an infinite number of permutations (see projectibility of predicates). Crucially, this point also extends
to mental rehearsals of a rule themselves just more subtle forms
of behavior. It would be implausible to suppose that whenever I
apply a rule I must mentally rehearse all the possible situations
in which the rule might apply for this would entail that whenever I followed a rule, I must be simultaneously thinking an infinite number of thoughts.
These examples are outlandish but are merely ways of making
graphic a simple point: Any rule is, logically, no different from
a word. The rule is just another symbol and, as such, stands in
need of interpretation if it is to have content. But interpreting is
a representational state. And so, in our attempt to understand
original representation, we have merely substituted the problem
of understanding the content of one representation with that of
understanding the content of another.
Wittgensteins response to this rule-following paradox
turns on the appeal to practice. This response is, however, deeply
problematic at least if understood as a constructive attempt to
explain the foundations of meaning. A practice is, as he put it,
what we do. But doing seems to be a form of acting. And actions
are essentially connected to, and presuppose, prior intentional
states of a subject. That is, both the status of an event as an action
and its identity as the particular action that it is depend on its
relation to a subjects intentional states. Therefore, the appeal to
action seems to presuppose representational states and so cannot explain what representations are (McDowell 1992; Hurley
1998; Rowlands 2006).
Naturalizing Representation
Wittgensteins legacy, therefore, is a convincing account of
what representation is not, coupled with a highly questionable
714
Representations
evolutionary products, and that we can therefore understand representation in terms of the concept of proper function. Suppose
we have a representational mechanism M capable of going into
a variety of states or configurations. The direct proper function of
mechanism M is, let us suppose, to enable the organism to track
various environmental contingencies for example, the presence
of predators. In the event that a predator is present, M goes into a
particular configuration, F. This state F of the mechanism has the
derived proper function deriving from the proper function of the
mechanism M of indicating the presence of a predator. That is,
it has the content, roughly, predator, there!
One strength of the teleological approach is the elegant manner in which it satisfies the normativity constraint. Another is
that it does not rely on the questionable assumption that all representation derives from mental representation. The historicalnormative account of representation can be applied directly to
linguistic forms independently of their connection to the intentional states of subjects. However, all attempts to naturalize
representation are controversial, and their success or failure is
a matter of continuing debate. The worries surrounding these
attempts can be divided into technical and foundational. With
respect to teleological approaches, for example, one technical
worry would be whether the approach can capture the fineness of
grain of certain content attributions, for example, whether it can
distinguish between the content of predator! and tiger! and
thing that will eat me! There is every reason to think that with
sufficient ingenuity, answers to these sorts of technical worries
will be (or, indeed, already have been) forthcoming.
Foundational worries are more serious. These concern
whether the sorts of natural relations invoked by these accounts
are the right sorts of things to explain the nature of representation.
Even if we were to identify items that satisfied ones preferred
naturalistic model of representation, it is argued, it would still be
an open question whether these items were, in fact, representations. Thus, some (e.g., McGinn 1991) argue that naturalistic
accounts provide only a criterion of representational content a
criterion that allows us to determine when one thing is about
another, and what thing it is about. But they fail to explain what
representation actually is. More positively, if we are to properly
understand representation, we also need to understand consciousness. And the prospects for a naturalistic interpretation
of the former are, therefore, tied to a naturalistic interpretation
of the latter.
Mark Rowlands
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Dretske, Fred. 1981. Knowledge and the Flow of Information.
Oxford: Blackwell.
. 1986. Misrepresentation. In Belief, ed. R. Bogdan, 1736
Oxford: Oxford University Press.
Fodor, Jerry. 1990. A Theory of Content and other Essays. Cambridge,
MA: MIT Press.
Grice, Paul. 1957. Meaning. Philosophical Review 66: 37788.
Hurley, Susan. 1998. Consciousness in Action. Cambridge: Harvard
University Press.
Lloyd, Dan. 1989. Simple Minds. Cambridge, MA: MIT Press.
McDowell, John. 1992. Meaning and intentionality in Wittgensteins
later philosophy. Midwest Studies in Philosophy 17: 3042.
715
716
717
Rhythm
RHYTHM
The notion of rhythm is widely present in language sciences and
an abundant literature ranging from acoustics to phonological
theory and neuropsychology is available, leading to several and
sometimes conflicting definitions. Nonetheless, most would
agree that rhythm involves the temporal organization of speech
and results from a threefold complex interaction among
the nature of the rhythmic atomic constituents;
the use of alternations between more and less prominent
constituents;
the pattern of regularity for the grouping of the constituents
into longer units.
According to this definition, rhythm is fundamental to languages
(it seems that no language may be defined as arrhythmic, even if
the last two proposed dimensions may be irrelevant in specific
languages).
In the twentieth century, phonetics searched mainly for the
acoustic correlates of rhythm units, while phonology with the
exception of metrical phonology usually considered rhythm as
a mere sequence of timing slots on which linguistic properties
are cast. In addition, cognitive science addressed distinct questions, namely, why languages are rhythmic and whether rhythm
plays a role in the cognitive processing of language.
718
719
720
SYNTACTIC REPRESENTATION
Linking
Algorithm
SEMANTIC REPRESENTATION
Discourse-Pragmatics
Coulson, Seana, and Wu Y. C. 2005. Right hemisphere activation of jokerelated information: An event-related brain potential study. J. Cog.
Neurosci. 17: 494506.
Federmeier, Kara D., and Marta Kutas. 1999. Right words and left
words: Electrophysiological evidence for hemispheric differences in
meaning processing. Cogn. Brain Res. 8: 37392.
Ferstl, Evelyn C., et al. 2005. Emotional and temporal aspects of situation
model processing during text comprehension: An event-related fMRI
study. J Cog. Neurosci. 17: 72439.
Gannon, Patrick J., et al. 1998. Asymmetry of chimpanzee planum temporale: Humanlike pattern of Wernickes brain language area homolog. Science 279: 2202.
Goel, Vinod, and Raymond J. Dolan. 2001. The functional anatomy of
humor: Segregating cognitive and affective components. Nature
Neurosci. 4: 2378.
Hutsler, Jeffrey, and R. A. W. Galuske. 2003. Hemispheric asymmetries
in cerebral cortical networks. Trends Neurosci. 26: 42936.
Joanette, Yves, Pierre Goulet, and Didier Hannequin. 1990. Right
Hemisphere and Verbal Communication. New York: Springer-Verlag.
Jung-Beeman, Mark. 2005. Bilateral brain processes for comprehending
natural language. Trends in Cognitive Sciences 9: 51218. This paper
proposes a neural microcircuitry explanation of computational asymmetries in language processing.
Leff, Alexander, et al. 2002. A physiological change in the homotopic
cortex following left posterior temporal lobe infarction. Ann. Neurol.
51: 443558.
Mason, Robert A., and Marcel Just. 2004. How the brain processes causal
inferences in text: A theoretical account of generation and integration
component processes utilizing both cerebral hemispheres. Psychol.
Sci. 15: 17
Mazoyer, B. M, et al. 1993. The cortical representation of speech. J. Cog.
Neurosci. 5: 46779.
Moffat, Scott D., Elizabeth Hampson, and Donald H. Lee. 1998.
Morphology of the planum temporale and corpus callosum in left
handers with evidence of left and right hemisphere speech representation. Brain 121: 236979.
Robertson, David A., et al. 2000. Functional neuroanatomy of the cognitive process of mapping during discourse comprehension. Psychol.
Sci. 11: 25560.
Schirmer, Annett, and Sonja A. Kotz. 2006. Beyond the right hemisphere: Brain mechanisms mediating vocal emotional processing.
Trends in Cognitive Sciences 10: 2430.
Seger, Carol A., et al. 2000. Functional magnetic resonance imaging
evidence for right hemisphere involvement in processing unusual
semantic relationships. Neuropsy. 14: 3619.
St. George, Marie, et al. 1999. Semantic integration in reading: Engagement of the right hemisphere during discourse processing. Brain 122: 131725.
Thompson-Schill, Sharon L., et al. 1997. Role of left inferior prefrontal
cortex in retrieval of semantic knowledge: A re-evaluation. Proc. Nat.
Acad. Sci. 94: 147927.
Vargha-Khadem, Faraneh, et al. 1997. Onset of speech after left hemispherectomy in a nine-year old-boy. Brain 120: 15982.
Virtue, S., J. Haberman, Z. Clancy, T. Parrish, and M. Jung-Beeman. 2006.
Neural activity of inferences during story comprehension. Brain
Research 1084: 10414.
Xu, J. et al. 2005. Language in context: Emergent features of word, sentence, and narrative comprehension. NeuroImage 25: 100215.
theoretical and descriptive constructs derive from the investigation of many non-Indo-European languages, yielding a theory
with rather different approaches to the analysis and explanation
of morphosyntactic phenomena. RRG posits only a single level
of syntactic representation, corresponding to the actual form of
the sentence; there are no abstract levels of representations or
derivations, as in Chomskyan generative grammar. Moreover,
no phonologically null elements are allowed in syntactic representations. The name of the theory comes from its view that
much of grammar is an interaction between semantics (role) and
discourse-pragmatics (reference). The organization of the theory
is in Figure 1. There is a mapping between the syntactic and
semantic representations, mediated by the linking algorithm.
Discourse-pragmatics plays a role in this mapping, but the role
varies across languages, leading to important typological differences among languages. It is represented in part by discourse
representation structures (DRSs) taken from discourse representation theory.
The syntactic representation is called the layered structure of
the clause and is composed of three interlocking projections: the
constituent projection (predicate, arguments, adjuncts), the
operator projection (grammatical categories like aspect, negation, tense), and the focus structure projection (information
structure of the sentence, that is, the potential focus domain
of the clause [where focus can potentially fall], the actual focus
domain, and the presupposed, non-focus part). The semantic
representation is an Aktionsart-based decompositional representation. thematic relations are defined in terms of positions in the decompositions. More important are the semantic
macroroles of actor and undergoer, which correspond to the
two primary arguments in a transitive predication, either one of
which can be the single argument of an intransitive predicate.
Based on extensive typological evidence, RRG rejects traditional
grammatical relations and instead posits a construction-specific
concept, the privileged syntactic argument (PSA) of a particular
construction.
The linking between syntax and semantics is bidirectional,
reflecting production and comprehension. In semantics-tosyntax linking, the elements in the semantic representation are
mapped into positions in a syntactic template, following both
universal and language-specific principles. An example from
English is given in Figure 2. In syntax-to-semantics linking,
information about the syntactic-semantic function of arguments
is derived from the overt morphosyntactic markings, and this is
linked to the argument positions in the lexical representation of
the predicate; a simple example is in Figure 3.
721
A ssertion
w, z
w, x, y, z
Sandy(w)
party(z)
w P at z
Sandy(w)
flowers(x)
Chris(y)
party(z)
Predicate focus
present x to y
w present x to y at z
SPEECH A CT
SENTENCE
SYNTA CTIC
INVENTORY
CLA USE
NP
NP
PP
PP
PRED
V
Sandy
at: A CC
presented
PSA :NOM
A CTIVE: 3sg
be-at(party
, [[do(Sandy
A CV
A CV
A CC
Non-Macrorole UNDERGOER
A CTOR
LEX ICON
to: A CC
Figure 2. Linking from semantics to syntax. Constituent and focus structure projections shown, along with discourse
representation structures.
SENTENCE
CLA USE
PrCS
CORE <--------NP
Parser
PP
PRED
NP
Voice? -- A ctive
PSA = A ctor
NUC
PERIPHERY
A DV
yesterday
A ctor
Lexicon
Not only has work been done using RRG on a wide variety of languages, but there has also been psycholinguistic and
neurolinguistic work done as well. RRG takes a cognitivist
approach to language acquisition (see Van Valin 1998, 2001),
as well as references on the RRG Web site). The application
of RRG to the study of sentence processing is investigated in
Bornkessel, Schlesewsky, and Van Valin (2004) and Van Valin
(2006).
Robert D. Van Valin, Jr.
722
Bornkessel, Ina, Matthias Schlesewsky, and Robert D. Van Valin, Jr. 2004.
Syntactic templates and linking mechanisms: A new approach to
grammatical function asymmetries. Available online at: http://linguistics.buffalo.edu/people/faculty/vanvalin/rrg/vanvalin_papers/
rrg_adm_CUNY04.pdf.
Van Valin, Robert D., Jr. 1998. The acquisition of wh-questions and
the mechanisms of language acquisition. In The New Psychology of
Language: Cognitive and Functional Approaches to Language Structure,
ed. Michael Tomasello, 22149. Mahwah, NJ: Lawrence Erlbaum.
Available online at: http://linguistics.buffalo.edu/people/faculty/vanvalin/rrg/vanvalin_papers/acqwh.pdf.
. 2001. The acquisition of complex sentences: A case study in the
role of theory in the study of language development. Chicago Linguistic
Society Parasession Papers 36: 51131. Available online at: http://linguistics.buffalo.edu/people/faculty/vanvalin/rrg/vanvalin_papers/
Acq_of_complex_sent.pdf.
.
2005.
Exploring
the
Syntax-Semantics
Interface.
Cambridge: Cambridge University Press.
. 2006. Semantic macroroles and language processing. In Semantic
Role Universals and Argument Linking: Theoretical, Typological and
Psycho-/Neurolinguistic Perspectives, ed. Ina Bornkessel, Matthias
Schlesewsky, Bernard Comrie, and Angela Friederici, 263302.
Berlin: Mouton de Gruyter. Available online at: http://linguistics.
buffalo.edu/people/faculty/vanvalin/rrg/vanvalin_papers/Sem_
Macroroles_&_Lang_Processing.pdf.
The best source for work and references relating to role and reference grammar is the RRG Web site. Available online at:
http://linguistics.buffalo.edu/research/rrg.html.
Rule-Following
RULE-FOLLOWING
In Ludwig Wittgensteins late writings, one finds numerous
interconnected remarks having to do with meaning, understanding, and rule-following remarks in which Wittgenstein
can seem to be continually circling around these topics without ever arriving at any definite conclusion. An exceptionally
clear and compelling way into this material was provided by
Saul Kripke in his influential 1982 book, Wittgenstein on Rules
and Private Language. By presenting Wittgensteins concerns
in the form of a single extended argument, Kripke brought
many readers to see that Wittgensteins discussions of rulefollowing bear significantly on central issues in (among other
things) contemporary philosophy of language. As Kripke
recasts Wittgensteins remarks, they amount to an argument
for a radical form of skepticism a skepticism according to
which there are no facts about what we mean by any of our
utterances or inscriptions. In what follows, a sketch is provided
of the skeptical argument that Kripke finds in Wittgensteins
remarks on rule-following. That done, an effort is made to
distinguish the actual, historical Wittgenstein from Kripkes
reconstruction of him.
Interpretationalism
Perhaps the most natural reply to the skeptic would make appeal
to the notion of a rule or interpretation in something like the
The problem with this sort of answer to the skeptics challenge is not that I may be misremembering the rule the interpretation of + that I learned as a child. The skeptic is happy
to grant that as a child I learned precisely this formulation and
that I remember it clearly to this day. But he questions how this
rule lets call it R ought to be understood. While the skeptic allows that R is the interpretation that Ive long assigned to
the plus sign, he claims that I am, at present, misapplying R.
He suggests that given what, in the past, I meant by the word
count, the result that a correct application of R would yield,
given 68 + 57, is 5.
At this point, I might reply to the skeptic by recalling an
interpretation of the word count that I also internalized as a
child. But he will just make the same sort of move again: Hell
grant me the remembered interpretation and suggest that I
am, at present, misapplying it. Thus, each new interpretation
that I adduce seems to require another one standing behind it,
and an infinite regress threatens. The moral the problem with
what might be called interpretationalism may be put as follows: However tempting it is to say that our words derive their
meanings from rules or interpretations, saying this just leaves
us with the question of where these rules or interpretations get
their meanings.
Dispositionalism
Another kind of reply to Kripkes skeptic would appeal to dispositions rather than to interpretations or rules. Thus, the skeptics challenge might be answered as follows: While I have never
before added 68 and 57, its nonetheless true that had I been
asked to add these numbers, I would have arrived at the answer
125. My having meant plus (rather than quus) by plus consists
in my having been so disposed disposed to answer questions
of the form What is x plus y? with the sum (and not the quum)
of x and y.
A problem with this sort of dispositionalism becomes apparent when one considers the fact that speakers are sometimes
disposed to make mistakes. Imagine that when Im sleepy, I
am disposed to answer 115 if someone asks me to add 68 and
57. We dont want an account of meaning according to which
it would follow from this that when Im sleepy, what I mean by
plus is not addition but, instead, some function that yields 115,
given 68 and 57 as arguments. Or consider a nonmathematical
example: It might be that I have a disposition to answer the question Is that a horse? affirmatively when Im shown an albino
723
Rule-Following
zebra or a donkey in dim light. Our account of meaning had better not commit us to claiming that the word horse in my idiolect therefore includes albino zebras and dimly lit donkeys in its
extension.
A further problem with the dispositionalist answer to Kripkes
skeptic (one that Kripke himself does not discuss) is this: The
dispositionalist at least as we (following Kripke) have imagined him or her takes it as a datum that for much of my life, I
have been disposed to answer 125 in response to being asked for
the sum of 68 and 57 (disposed, as Kripke puts it, when asked
for any sum x + y to give the sum of x and y as the answer [1982,
22]). But to say any such thing is to give a thoroughly intentional
characterization of my behavior to describe me not merely as
having a disposition to produce certain physically describable
movements and sounds (under some set of physically describable conditions), but as being disposed to answer a question in
a particular way, that is, to say something something with a
particular semantic content in response to being asked something. Thus dispositionalism, at least as weve imagined it, takes
for granted the phenomenon (semantic content, meaning) that
it pretends to vindicate and explain the very phenomenon that
Kripkes skeptic calls into doubt.
724
Schema
S
SCHEMA
A schema is a high-level conceptual structure or framework
that organizes prior experience and helps us to interpret new
situations. The key function of a schema is to provide a summary of our past experiences by abstracting out their important
and stable components. For example, we might have a schema
for a classroom that includes the fact that it typically contains a
chalkboard, bookshelves, and chairs. Schemas provide a framework for rapidly processing information in our environment. For
example, each time we enter a classroom, we do not have to consider each element in the room individually (e.g., chair, table,
chalkboard). Instead, our schemas fill in what we naturally
expect to be present, helping to reduce cognitive load. Similarly,
schemas also allow us to predict or infer unknown information in
completely new situations. If we read about a third grade classroom in a book, we can use our established classroom schema
to predict aspects of its appearance, including the presence of a
coatroom and the types of posters that might decorate the walls.
Schemas play an important role in language and linguistic
processing by helping to frame the semantic content of a situation. Even when linguistic input is sparse or vague, activation of
the appropriate schema can aid in the comprehension and retention of linguistically communicated material (see next section for
an example). In addition, schemas and scripts often help us to
define and interpret the discourse associated with particular contexts. In the classroom example, certain aspect of the communication between a student and teacher are captured by the schema,
including the facts that students should quietly raise their hand in
order to get the teachers attention and that the teacher will stand
facing the class and may call upon the student.
In a functional sense, schemas share much in common with
categories (see categorization) or concepts. However, a
distinguishing feature of schemas is that they are structured mental representations made up of multiple components. Schemas
typically contain various slots, which each take on any number
of values, and a set of relational structures that organize the slots
and represent their interconnections. The values of particular
slots are usually determined by the current context, percepts,
or situation. For example, the schema for a generic room might
include a slot for walls, doors, and windows, which could be filled
with specific values (i.e., wooden door, bay windows, etc.). Slots
that are left unspecified in the current situation are given default
values, which reflect expectations or inferences about unseen or
unknown information. Schemas derive their predictive power
through a process of filling in default values so that incomplete
knowledge about the current situation can be supplemented by
past experience. In addition, a slot may be filled by other schemas, allowing for the compositional construction of more
complex structures.
725
Schema
positive weights, while blackboards and beds might be linked
by a strongly negative weight. If a partial description or view of
a classroom activates the blackboard and pencil sharpener element, other classroom elements (such as desk or globe) will
also become active through their positive links to the observed
events, while classroom-irrelevant information (such as a bed)
would be inhibited. Structured attributes could form by subsets
of mutually inhibitory elements. For example, most classrooms
have either a blackboard or a whiteboard. Mutually inhibitory
weights between blackboard and whiteboard units can ensure
that only one value would fill this slot for at a time. Critically, the
parallel distributed processing (PDP) approach to schemas dispenses with the traditional structure of schema representations,
with slots, fillers, and relations favoring emergent and implicit
structures, such as partwhole relations and hierarchies.
726
symbol systems involve framelike structures that have slots, preserve fixed relationships between attributes, and may take on
default values. However, the objects upon which these framelike structures operate are direct, multimodal, sensory-motor
representations. Evidence in favor of this approach includes the
fact that brain areas representing particular concepts appear to
overlap with perceptual processing of that concept. For example,
damage to the visual cortex impairs conceptual processing of
categories that are primarily visual in nature. The fundamental
contribution of the perceptual symbols approach is to rethink
the relationship between conceptual and perceptual processing and to suggest how symbolic, schema-like mental structures
might emerge from perceptual experience.
Similarly, in cognitive linguistics, image schemas have
been proposed as a way to link bodily actions, perceptual experience, and semantic processing (Johnson 1987; Lakoff 1987).
Image schemas are an embodied prelinguistic structure of experience that can provide the basis of conceptual metaphors.
For example, one might have a generalized containment schema
that represents an object being inside another object (just like a
small ball might physically fit inside a cup). This representation
is assumed to be an abstracted, but perceptual, instantiation of
the concept and includes a number of structured spatial relationships. Generative meanings can be produced on the basis of this
schema through metaphorical mappings. For example, understanding a phrase such as a deep depression takes a long time
to get out of is accomplished through metaphor by the general
containment schema (so one might envision the depressed person rising out of the depression like an element out of its physical container; Johnson 1987). Like perceptual symbol systems,
image schemas emphasize perceptual, multisensory, embodied
content in schema-like representations.
Scripts
Bransford, J., and M. K. Johnson. 1972. Contextual prerequisites for
understanding: Some investigations of comprehension and recall.
Journal of Verbal Learning and Memory 11: 71726.
Brewer, W. F., and J. C. Treyens. 1981. Role of schemata in memory for
places. Cognitive Psychology 13: 20730.
Johnson, M. 1987. The Body in the Mind: The Bodily Basis of Meaning,
Imagination, and Reason. Chicago: University of Chicago Press.
Lakoff, G. 1987. Women, Fire, and Dangerous Things. Chicago: University
of Chicago Press.
Lupyan, G. 2005. When naming means forgetting: Verbal classification
leads to worse memory. In Proceedings of the 27th Annual Meeting of
the Cognitive Science Society, ed. B. Bara, L. Barsalou, and M. Bucciarelli.
Hillsdale, NJ: Lawrence Erlbaum.
Mandler, J. 1984. Stories, Scripts, and Scenes: Aspects of Schema Theory.
Hillsdale, NJ: Lawrence Erlbaum.
Minsky, M. 1975. A framework for representing knowledge. In
The Psychology of Computer Vision, ed. P. Winston, 21177. New
York: McGraw-Hill.
Rojahn, K., and T. Pettigrew. 1992. Memory for schema-relevant
information: A meta-analytic resolution. British Journal of Social
Psychology 31: 2.
Rumelhart, D. 1980. Schemata: The building blocks of cognition. In
Theoretical Issues in Reading and Comprehension, ed. R. Spiro, B.
Bruce, and W. Brewer, 3358. Hillsdale, NJ: Lawrence Erlbaum. An
interesting and influential argument that schema may represent the
basic units of everyday knowledge in cognitive systems.
Rumelhart, D., P. Smolensky, J. McClelland, and G. Hinton. 1986.
Schemata and sequential thought processes in PDP models. In
Parallel Distributed Processing: Explorations in the Microstructure of
Cognition. Vol. 2. Ed. J. McClelland and D. Rumelhart, 757. Cambridge,
MA: MIT Press.
Sloutsky, V. M., and A. V. Fisher. 2004. When development and learning
decrease memory. Psychological Science 15: 5538.
Schank, R., and R. Abelson. 1977. Scripts, Plans, Goals and Understanding.
Hillsdale, NJ: Lawrence Erlbaum.
SCRIPTS
Let me tell you a simple story.
John went to a restaurant. He ordered lobster. He paid the check
and left.
727
728
Individual Differences
Whole theories have been built around social-psychological and
affective variables (for a review of research, see Dornyei 2005),
729
730
Self-Concept
. 1989. A nonstandard approach to standard English. TESOL
Quarterly 23.2: 25982.
. 1990. The Syntax of Conversation in Interlanguage Development.
Tubingen: Gunter Narr.
Schmidt, Richard W. 2001. Attention. In Cognition and Second Language
Instruction, ed. Peter Robinson, 332. Cambridge: Cambridge
University Press.
Schumann, John H. 1986. Research on the acculturation model for second language acquisition. Journal of Multilingual and Multiculutural
Development 7: 37992.
Segalowitz, Norman. 2003. Automaticity and second languages. In
Handbook of SLA, ed. Catherine J. Doughty and M. H. Long, 382408.
Oxford: Blackwell.
Selinker, L. 1972. Interlanguage. International Review of Applied
Linguistics 10.3: 20931.
Siegel, Jeff. 2003. Social context. In Handbook of SLA, ed. Catherine J.
Doughty and M. H. Long, 178223. Oxford: Blackwell.
Sorace, Antonella. 2003. Near-nativeness. In Handbook of SLA, ed.
Catherine J. Doughty and M. H. Long, 13051. Oxford: Blackwell.
Tomlin, Russell, and V. Villa. 1994. Attention in cognitive science and
SLA. Studies in Second Language Acquisition 15.2: 185204.
White, Lydia. 2003a. Second Language Acquisition and Universal
Grammar. Cambridge: Cambridge University Press.
. 2003b. Fossilization in steady state L2 grammars: Persistent
problems with inflectional morphology. Bilingualism: Language and
Cognition 6.2: 12941.
SELF-CONCEPT
The self and self-related perceptions are widely recognized for
their significance in explaining human behavior and in individuals mental well-being. Different disciplines study the self
from different perspectives. The present entry relies on findings
derived from cognitive-developmental psychological research.
The self is an explanatory and organizing framework that provides the person with a sense of individual identity and continuity. Early psychologists soon made the distinction between self
as a subject (I-self as active agent and as knower) and self as an
object of what is known (Me). The self as an object of knowledge
is reflected in the construct of self-concept. There is a lack of
a universally accepted definition of self-concept due mainly to
different theoretical perspectives. Self-concept is the persons
beliefs, attitudes, and perceived characteristics about the self. It
includes self-representations of abilities, past actions, goals,
intentions, interests, future self, ideal self, and so on. Perceived
competence is considered a key component of self-concept that
plays a significant role in motivation, learning, and achievement.
Therefore, self-concept represents descriptive as well as evaluative aspects of the self that are reflected in self-description of
ones strengths and weaknesses in given domains of functioning.
For instance, academic self-concept reflects individuals beliefs
and attitudes about their own competence in academic settings.
Before the 1990s, many studies often treated the various self
terms, such as self-concept, self-esteem, and self-efficacy, interchangeably. These constructs represent different aspects of the
self-system. Self-concept is essentially descriptive, and it includes
personal beliefs and attitudes regarding ones state of abilities.
Self-esteem is the overall evaluation of ones worth or value as a
person. It includes emotional appraisals of the self, ones likes or
dislikes as well as feelings of self-acceptance. Therefore, it reflects
731
Self-Concept
the affective aspect of the self-system. Self-efficacy is an individuals judgments and expectations for successfully performing
given tasks at designated levels of performance, and it should not
be regarded as synonymous with self-concept (Bandura 1997).
Self-concept is both a cognitive and a social construction.
Sources for self-concept are the individuals mastery experiences, internal and social comparison processes, and perceived appraisals by others (Harter 1999, 16694). For instance,
it has been shown that aspects of adolescents self-concept
regarding abilities in school language were influenced by their
performance in language and by perceived appraisals by
others (Dermitzaki and Efklides 2000, 62934). Self-concept is
considered a dynamic structure that changes with experience,
especially its more peripheral and situation-specific aspects. The
development of memory and language in toddlers allows for
the verbal expression and representation of the self as an object
and for gradually constructing an inner self by developing selfrepresentations.
Psychological research has documented that self-concept
affects cognition, emotions, motivation, and behavior, and it
is associated with social adjustment, leadership, and level of
anxiety. Self-concept has strong motivational power since it is
tied closely to how we choose to invest our time and effort and
whether we persist during our efforts. Therefore, positive selfconcept is associated with high achievements and active involvement in learning situations. In educational settings, academic
self-concept is considered an important mediator of academic
motivation and performance. Systematic relations have been
documented between the academic self-concept of students and
teachers and such factors as students intrinsic motivation, goal
setting, anxiety, course selection, effort spent, study skills, and
self-regulation, as well as teachers level of engagement and persistence in classroom activities.
Before the 1980s, self-concept was approached as a monolithic, unidimensional construct. Nowadays, it is believed that
self-concept is an entity with multiple facets and dimensions.
Various profiles of increasing multidimensionality that assess a
complex set of perceived competencies have been progressively
developed (see Byrne 1996). Although the multidimensionality of
self-concept is rarely disputed, researchers do not always agree
on the self-concept structure, that is, on the nature of relations
among the different aspects and dimensions of self-concept (see
Marsh and Hattie 1996). An influential theoretical model on the
self-concept structure is the hierarchical one, first proposed by R.
J. Shavelson and colleagues in 1976. The evidence is not conclusive. However, analysis using sophisticated statistical procedures,
such as confirmatory factor analysis, tends to support a potential
self-concept hierarchy. Within this model, there is a superordinate, general self-concept as a higher-order factor that subsumes
domain-specific self-concepts, which may be correlated, though
they can be interpreted as separate constructs. Academic, social,
emotional, and physical self-concepts are some of the main
domains identified within this model. Within a domain of action,
there are several levels of specific self-perceptions when moving from the top to the bottom of the hierarchy. For instance,
academic self-concept includes self-perceptions regarding various subdomains of knowledge, such as self-concept in English
language, self-concept in mathematics, self-concept in science,
732
Self-Organizing Systems
and so on. Self-concept in language reflects students beliefs and
attitudes about their competence in language domain. Selfconcept in language subsumes self-perceptions from more general aspects of the language domain, such as grammar, syntax,
reading, writing, and so on, to situation-specific self-perceptions, that is, with regard to the task at hand (see, for example,
Lau et al. 1999).
There are still many issues to investigate about self-concept
and its relations to human behavior and well-being. What we
should always bear in mind is that the fundamental goal of selfconcept research is to help persons function, grow, and adapt
better in various demanding settings.
Irini Dermitzaki
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bandura, Albert. 1997. Self-Efficacy: The Exercise of Control. New
York: Freeman.
Byrne, Barbara M. 1996. Measuring Self-Concept Across the Life
Span: Issues and Instrumentation. Washington, DC: American
Psychological Association.
Dermitzaki, Irini, and Anastasia Efklides. 2000. Aspects of self-concept
and their relationship to language performance and verbal reasoning
ability. American Journal of Psychology 113: 62138.
Harter, Susan. 1999. The Construction of the Self. A Developmental
Perspective. New York: Guilford.
Lau, Ivy Cheuk-yin, Alexander Seeshing Yeung, Putai Jin, and Renae Low.
1999. Toward a hierarchical, multidimensional English self-concept.
Journal of Educational Psychology 91: 74755.
Marsh, Herbert W., and John Hattie. 1996. Theoretical perspectives on
the structure of self-concept. In Handbook of Self-Concept, ed. Bruce
A. Bracken, 3890. New York: Wiley.
Shavelson, R. J., J. J. Hubner, and G. C. Stanton. 1976. Selfconcept: Validation of construct interpretations. Review of Educational
Research 46: 40741.
SELF-ORGANIZING SYSTEMS
A broad division is made between systems that organize themselves via their own internal constraints and systems that are
organized by others, that is, by agencies that are exterior to them.
Although no single definition is agreed upon as yet, the nature
of systems that make themselves up as they go along has been
intensively studied, and there is a general understanding of the
basic requirements.
A self-organizing (S-O) system is typically composed of many
differently sized parts operating at many time scales. The organized
states or patterns exhibited by an S-O system at more macroscopic
scales arise strictly from local interactions among its parts. The
interactions are multiple and nonlinear meaning that they produce disproportionate effects. Figure 1 shows that in a BelousovZhabotinsky (B-Z) chemical reaction involving dozens of molecule
types, microscopic interactions abiding few constraints result in
highly constrained collective behaviors (concentric circles or spirals
in a petri dish viewed from above) that then modulate the microscopic interactions in a circularly causal manner. These disproportionate effects are hidden in the nonlinearities and made manifest
by circumstances internal to the S-O system (specifically, fluctuations or noise) or external to the S-O system (e.g., energy flows
Self-Organizing Systems
733
Semantic Change
SEMANTIC CHANGE
The study of semantic change is a core area of historical semantics, that is, the diachronic study of the meaning of natural language expressions. Next to the study of meaning changes and the
principles governing such changes, historical semantics includes
the etymological reconstruction of the meaning of proto-forms,
and the semantic study of older states of languages regardless
of previous or later changes. This entry is concerned exclusively
with lexical change of meaning, given that the study of word
meaning has always taken up a central position in the field of
historical semantics. (See the entry on grammaticalization
for issues involving grammatical meaning.)
Historically speaking, the study of meaning change reigned
supreme within semantics in the era of prestructuralist semantics, from roughly 1830 to 1930, receded to the background in
the structuralist and generativist periods, and received
a new impetus with the emergence of cognitive approaches to
linguistics (see Nerlich 1992 and Ullmann 1962 for the older
traditions.) The contemporary field of diachronic lexical semantics may be charted according to the linguistic level on which
changes are envisaged: the level of the individual readings, that
of the word, and that of the lexicon at large.
734
Semantic Fields
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Blank, Andreas. 1997. Prinzipien des lexikalischen Bedeutungswandels
am Beispiel der romanischen Sprachen. Tbingen: Niemeyer.
Blank, Andreas, and Peter Koch, eds. 2003. Kognitive romanische
Onomasiologie und Semasiologie. Tbingen: Niemeyer.
Geeraerts,
Dirk.
1997.
Diachronic
Prototype
Semantics.
Oxford: Clarendon.
Nerlich, Brigitte. 1992. Semantic Theories in Europe, 18301930.
Amsterdam: John Benjamins.
Sweetser, Eve. 1990. From Etymology to Pragmatics. Cambridge:
Cambridge University Press.
Tournier, Jean. 1985. Introduction descriptive la lexicognetique de
langlais contemporain. Paris and Geneva: Champion, Slatkine.
Traugott, Elizabeth C., and Richard B. Dasher. 2005. Regularity in
Semantic Change. Cambridge: Cambridge University Press.
Ullmann, Stephen. 1962. Semantics. An Introduction to the Science of
Meaning. Oxford: Basil Blackwell.
SEMANTIC FIELDS
A semantic field consists of two parts: 1) a conceptual domain,
for example, color, containers, cooking (see concepts); and
2) lexical items that are mapped onto the conceptual domain.
Semantic field theory assumes that vocabularies of languages
are structured, not just a list of atomic units. A major part of field
theory provides an inventory of lexical relations, that is, how lexemes can be related to one another, such as by synonymy and
antonymy. The lexeme, the basic semantic unit, is generally a
word or an idiomatic phrase like wake up or kick the bucket.
Ferdinand de Saussure ([1916] 1966) is credited with the beginning of structuralism, where the elements are an interconnected system of signs. Jost Trier (1932) distinguished between
conceptual fields and lexical fields and showed how lexical items
divide up the conceptual field, focusing on paradigmatic relationships, while W. Porzig (1950) emphasized syntagmatic relationships. John Lyons (1963, 1968, 1977, Vol. 1. 231335) expanded
semantic field theory, stressing the role of context in semantic
analysis and embedding his work into a generative grammar. Adrienne Lehrer (1974), Peter R. Luzeier (1981), Richard
Grandy (1987), and Eva F. Kittay (1987), among others, have further developed, applied, or defended field theory.
A version of semantic fields has also been used by cognitive
semanticists, for example, Zoltn Kovesces (1986), George Lakoff
and Mark Johnson (1980), and many others. However, whereas
semantic field practitioners start with the lexemes and then analyze the structural relations and conceptual structure, cognitive
semanticists start with the conceptual structure.
animal
mammal
feline
canine
dog
wolf
fish
bird
tiger
lion
animal
domestic animal
wild animal
livestock
pet
cat
dog
bird
Siamese retriever
cow horse
pig
canary
mid
big
more
less
little
more
less
Usually a superordinate has several hyponyms. Other hyponyms of flower are lily, petunia, dahlia, geranium, and bachelor button. All the co-hyponyms are incompatible with one
another; that is,they form a contrast set (Grandy 1987). A is
a rose entails A is not a lily or petunia, or dahlia, and so on.
A taxonomy, especially a scientific taxonomy, is a subtype of
hyponymy. Figure 1 illustrates this structure. However, a different structure for animals, a functional one, is possible,
where animals are classified by their relationship to people, as
in Figure 2. In functional relationships, the kind of is weaker
than entailment. We think of a cat as a kind of pet, but a feral
cat is not a pet. Most birds are not pets, and few people consider snakes to be appropriate pets.
Antonymy is a special type of incompatibility that consists of
two opposing terms. Complementary antonyms divide the space
exhaustively, such as deadalive or singlemarried. In such cases,
A not B and B not A. If Al is married, Al is not single. If Bee
is single, Bee is not married. More interesting are the gradable
antonyms where the structure consists of a center point or center
region with bidirectionality. Figure 3 illustrates this relationship
with big and little. These antonyms are intimately connected to
comparison. A is bigger than B B is littler than A. There are various types of antonymy, which involve markedness and subtle
entailments (see Cruse 1986; Lehrer 1983; Murphy 2003).
Closely related to antonymy is converseness, which interacts
with syntax to provide sentences paraphrases. Examples are
735
Semantic Fields
Table 1.
hot
Actor-action:
Action-object:
Actioninstrument:
Cause-effect
surgeon-operate
chef-cook
dog-bark
warm
tepid
cool
burning
cold
freezing
frigid
more
less
more
less
play-game
drive-vehicle
He drives a large
vehicle.
see-eye
kochen1
kochen2 to boil
hear-ear
kill-die
sieden
simmer
dnsten
stew
to cook
braten fry, grill
rsten
roast, pan fry
backen bake
grillen
broil
gotowa1 cook
gotowa2 boil
sma y
fry
cook
dusi
boil
simmer
poach
fry
saut
broil
deep-fry =
French-fry
roast
736
grill
stew
stew
bake
Another part of the semantic analysis would contain the selection restrictions or collocations for each lexeme. Poach typically
occurs with foods like eggs or fruit. We bake hams but roast meat,
even though both are nowadays cooked in ovens. The phrase
roast meat is a relic of earlier periods when meat was cooked on
a spit over an open fire.
Temperature words illustrate a field with antonyms, as in
Figure 5. In the temperature field, there are two pairs of antonyms, an outer pair, hotcold, and an inner pair, warmcool.
The middle region has a lexeme, tepid, and its synonym, lukewarm, which are limited in distribution, used literally only for the
temperature of liquids. The outer words have various hyponyms,
only a couple of which are illustrated here. There are interesting
asymmetries in these structures (see Cruse 1986; Lehrer 1983).
Applications
Semantic field theory has at least four important applications: 1)
comparative lexicology, 2) pedagogy, 3) metaphor, and 4) philosophical semantics.
COMPARATIVE LEXICOLOGY. Lexemes in one language are rarely
equivalent in meaning to those of other languages. The space of
a semantic field can be structured in various ways, and field theory provides an excellent way of comparing the semantic structures of different languages. For example, German, Polish, and
Japanese show alternative structures to English cooking words
(see Figures 6, 7, and 8). In both German and Polish, the words
for cook have a general meaning for preparing food with heat
and also a specific meaning of cooking with water, for example,
boil. And whereas English has different lexemes for bake, broil,
fry, and roast, German braten covers fry, roast, and broil.
In Polish, pec includes bake and broil but has a separate word
Semantic Fields
nitaki cook
niru boil
taku boil <rice>
musU steam
itameru
aburu
stir fry cook with direct heat
for fry, smay, while in Japanese yaku refers to all these processes. Japanese also has a separate word for cooking rice.
That a language lacks a specific word for a concept in English is
of no consequence since the meaning can easily be specified by a
phrase. German im Ofen braten for roast and im Pfanne braten
for fry can be used whenever needed to clarify the sense.
PEDAGOGY. Closely related to comparative lexicology is pedagogy, especially in teaching foreign languages (see teaching
language and bilingual education). Although textbooks
frequently introduce related lexemes together, such as words for
parts of the body, spatial relations, furniture, and so on, a word
is typically translated by the nearest equivalent in the students
native language. However, most lexical items in different languages are rarely exact equivalents. A word in one language can
be translated by several in another language, depending on subtleties of meaning and on selection restrictions. A semantic field
analysis that compares and explicates partial synonyms can help
students greatly. For example, the two textbooks for advanced
English learners by B. Rudzka and colleagues (1981, 1985) offer
students sets of closely related words that are differentiated by
their definitions and collocations. One of their examples is a lexical set meaning roughly being insufficient, inadequate or thinly
distributed: meager, scant, sparse, frugal. Each of these words is
defined, semantically and pragmatically contrasted with others
in the set, and illustrated with relevant examples.
METAPHOR. A third application of semantic fields is to the study
of metaphor and extended word meanings. Often, when words
from one domain are extended metaphorically to another domain,
called bridge terms by Kittay (1987), other lexemes from the first are
also transferred to the second, retaining their relationship with one
another (Lehrer 1974; Lakoff and Johnson 1980; Kittay 1987). This
can be illustrated with the temperature terms hot, warm, cool, cold. In
addition to temperature (of weather, objects, sensations), they are all
used to refer to emotional states or friendship or passion or the lack
thereof. A sexually excited person can be described as hot and typically experiences a rise in body temperature. A cold or frigid person
does not experience a lowered body temperature, but the senses of
unfriendly or unresponsive derive from the meaning by virtue of the
antonymy of hot and warm. Finally, consider the sentence He traded
his hot car for a cold car. Hot car has at least two conventional metaphorical meanings: fast, for example, a Corvette, and stolen.
Because hot and cold are antonyms, one can figure out a plausible
meaning. He traded his fast/sports car for a conventional one (e.g., a
family sedan) or he traded his stolen car for a legally acquired one.
Extended metaphors are widely used in literature (see poetic
metaphor). In William Wordsworths poem On the Extinction
of the Venetian Republic, the history of Venice is described in
737
Semantic Memory
Saussure, Ferdinand de. [1916] 1966. A Course in General Linguistics.
Trans. W. Baskin. New York: Philosophical Library
Trier, Jost. 1932. Der Deutsche Wortschatz im Sinnbezirk des Verstandes.
Heidelberg: Winter.
SEMANTIC MEMORY
Semantic memory refers to a major division of long-term
memory that includes knowledge of facts, events, ideas, and
concepts. Thus, semantic memory covers a vast cognitive terrain, ranging from information about historical and scientific
facts, details of public events, and mathematical equations to the
information that allows us to identify objects and understand the
meaning of words.
The idea of semantic memory as a distinct form of memory has
a long history dating back at least to the late nineteenth century.
More recently, however, the idea of semantic memory as a separate memory system began in 1972 with Endel Tulvings distinction between semantic and episodic memory. While the notion
of episodic memory has undergone considerable evolution since
that original formulation (Tulving 2002), it remains helpful to
describe the properties of semantic memory in relation to episodic memory. In current formulations, episodic memory can
be thought of as synonymous with autobiographical memory.
Episodic memory is the system that allows us to remember (consciously recollect) past experiences and, perhaps, may also be
critical for imagining and/or simulating future events. Semantic
memories, in contrast, are devoid of information about personal
experience. Thus, unlike episodic memories, semantic memories
lack information about the context of learning, including situational properties like time and place, as well as personal dimensions like how we felt at the time the event was experienced. In
relation to episodic memory, semantic memory is considered to
be both a phylogenically and an ontologically older system. In
fact, rather than arising as an independent evolutionary development, it is commonly assumed that episodic memory emerged as
an add-on or embellishment to semantic memory.
738
Semantic Memory
are disproportionately associated with damage to the temporal
lobes. The most common etiology is herpes simplex encephalitis, a viral condition with a predilection for attacking antero-medial and inferior or ventral temporal cortices. Category-specific
knowledge disorders for animals also have been reported following focal, ischemic lesions to the more posterior regions of the
ventral temporal cortex associated with visual object perception.
In contrast, category-specific knowledge disorders for tools and
other manmade objects most commonly occur with focal damage to lateral frontal and parietal cortices of the left hemisphere associated with object manipulation. These anatomical
findings are broadly consistent with property-based views of
knowledge organization that maintain that category-specific
disorders occur when a lesion disrupts information about a
particular property or set of properties critical for defining, and
for distinguishing among, category members. Thus, damage
to regions that store information about object form, and formrelated properties like color and texture, will produce a disorder
for animals. This is because visual appearance is assumed to be a
critical property for defining animals and because the distinction
between different animals is assumed to be heavily dependent
on knowing about subtle differences in their visual forms. In a
similar fashion, damage to regions that store information about
how an object is used should produce a category-specific disorder for tools, and all other categories of objects defined by how
they are manipulated.
It is important to stress, however, that the lesions in patients
presenting with category-specific knowledge disorders are often
large and show considerable variability in their location from
one patient to another. As a result, these cases have been relatively uninformative for questions concerning the organization
of object memories in the cerebral cortex. In contrast, recent
functional neuroimaging studies of the intact human brain
have begun to shed some light on this issue.
739
740
Relational substantives:
KIND, PART
Determiners:
Quantifiers:
Evaluators:
GOOD, BAD
Descriptors:
BIG, SMALL
Mental predicates:
Speech:
Actions, events,
movement, contact:
DO, HAPPEN,
LIVE, DIE
Time:
MOVE, TOUCH
HAVE, BE (SOMEONE/SOMETHING)
Space:
Logical concepts:
Intensifier, augmentor:
VERY, MORE
Similarity:
LIKE/AS
Notes: Primes exist as the meanings of lexical units (not at the level of lexemes).
Exponents of primes may be words, bound morphemes, or phrasemes. They can be
formally complex. They can have different morphosyntactic properties, including word
class, in different languages. They can have combinatorial variants (allolexes). Each
prime has well-specified syntactic (combinatorial) properties.
inch (or become), not, and exist. Some took the form of binary
features, for example, +human, +contact. Often, componentialists employed plainly complex and/or arbitrary features, with no
plausible claim for primitive status. The componential tradition
did not yield a comprehensive set of semantic primitives that
could sustain any claim to be adequate for the semantic analysis
of the entire vocabulary. Most of the work was advanced on a provisional basis and in relation to small sets of data. Ray Jackendoff
(1990, 2001) is a partial exception. He has continued to elaborate
and diversify his account of abstract conceptual primitives over
some years, but without yet proposing a well-defined set.
741
Semantics
Arnauld, Antoine, and Pierre Nicole. [1662] 1996. Logic or The Art of
Thinking. Trans. Jill Vance Buroker. Cambridge: Cambridge University
Press.
Boguslawski, Andrzej. 1965. Semantyczne pojecie licebnika. Wroclaw:
Ossolineum.
. 1970. On semantic primitives and meaningfulness. In Sign,
Language and Culture, ed. R. J. A. Greimas, M. R. Mayenowa, and
S. Zolkiewski, 14352. The Hague: Mouton.
Goddard, Cliff, ed. 2006. Ethnopragmatics: Understanding Discourse in
Cultural Context. Berlin: Mouton de Gruyter.
. 2008. Cross-Linguistic Semantics. Amsterdam: John Benjamins.
Goddard, Cliff, and Anna Wierzbicka, eds. 2002. Meaning and
Universal Grammar Theory and Empirical Findings. Vols, 1 and 2.
Amsterdam: John Benjamins.
Jackendoff, Ray. 1990. Semantic Structures. Cambridge, MA: MIT Press.
. 2002. Foundations of Language. Oxford: Oxford University Press.
Peeters, Bert, ed. 2006. Semantic Primes and Universal Grammar:
Evidence from the Romance Languages. Amsterdam: Benjamins.
Schank, Roger C. 1972. Conceptual dependency: A theory of natural language understanding. Cognitive Psychology 3: 552631.
Srensen, Holger Steen. 1958. Word-Classes in Modern English.
Copenhagen: Gad.
Wierzbicka, Anna. 1972. Semantic Primitives. Frankfurt: Athenum.
. 1996. Semantics: Primes and Universals. Oxford: Oxford University
Press.
. 1999. Emotions Across Languages and Cultures. Cambridge:
Cambridge University Press.
. 2006. English: Meaning and Culture. New York: Oxford University
Press.
An extensive bibliography on semantic primes is available online
at: http://www.une.edu.au/bcss/linguistics/nsm.
SEMANTICS
Semantics, the study of linguistic meaning, has been traditionally
distinguished from syntax, the study of the grammatical structures of languages (see also grammaticality), and pragmatics, the study of the uses of language, though there are important
interconnections between them.
As natural languages like English, Chinese, and Arabic have an
infinite number of nonsynonymous sentences, some expressions
must be semantically primitive and others complex, given the finitude of speakers. An expression is a semantical primitive provided
the rules which give the meaning for the sentences in which it
does not appear do not suffice to determine the meaning of the
sentences in which it does appear (Davidson [1984] 2001, 9). A
language with such a division of terms is compositional and this
feature of it compositionality. In a compositional language,
complex expressions are understood on the basis of understanding primitive expressions and rules for their combination. This
is not a matter of explicit propositional knowledge but is, rather,
exhibited in a speakers competence in using words appropriately in response to his or her environment and what others say.
lexical semantics focuses on word meaning. Lexical
semantics is concerned with the semantical categories of words,
their structures, what individual words mean, meaning connections between words, and, more generally, what it is for words
to having meaning. Compositional semantics focuses on how we
understand complex expressions on the basis of their significant
parts. Lexical semantics contributes to compositional semantics
742
because a words semantical category determines how it contributes to the complex expressions in which it appears and also
because not all words are semantical primitives verb inflection
for tense, for example, involves a structural element.
Although we understand sentences on the basis of their contained words, there is a sense in which the sentence is the basic
unit of meaning in a language. Sentences are the main vehicle
for linguistic communication, and words have their point only
in giving sentences theirs. Sentences have their point, in turn, in
what they are used to do. It follows that speech-acts are the
basic unit of linguistic evaluation. A speech-act is an intentional
action with an illocutionary force and content. The illocutionary force of a speech-act is the dimension along which
we distinguish asserting, commanding, promising, and the like
(Searle 1979). Although speech-act theory is part of the subject
matter of pragmatics, broadly speaking, it must also be part of
the subject of semantics, since sentence meaning cannot be
understood apart from the central purposes it serves. The connection with language use is also seen in the need for rules to
handle context-sensitive elements in sentences, such as tense
and indexicals. For example, I am hungry expresses a different proposition depending on who uses it and at what time.
The rule attaching to I is that it refers to the person using it. The
rule for present tense is that it is about the time of the utterance of
the sentence. Speaker and time are two basic features of speech
contexts. These arguably suffice to determine other elements of
context relevant to interpretation. That refers to what the speaker
demonstrates in using it, there refers to the location the speaker
intends when using it, and so on.
The most basic question of semantics What is meaning? is best approached by way of asking other, more specific
questions. What different aspects can be distinguished in and
between the meanings of expressions? How do we understand
complex expressions on the basis of their constituents? What is
it to understand a word? What is it to speak a language? How is
word and sentence meaning determined in the mouth of a particular speaker? How is it determined across a linguistic community? The rest of this entry considers theories that address these
questions, beginning with formal compositional semantic theories (see formal semantics).
A compositional semantic theory gives the meanings of complex expressions on the basis of their primitive constituents
and mode of combination. A reasonable requirement is that it
produce for each declarative sentence of the studied language
the object language (L) a theorem of the form (M), where s is
replaced by a description of the object language sentence as
formed from its significant parts, and p is replaced by its translation into the metalanguage. (Something parallel is needed
for nondeclaratives if they are not reducible to declaratives see
Boisvert and Ludwig 2006.)
(M) s means in L that p.
Semantics
it for modern semantics is found in Gottlob Freges distinction
between sense and reference ([1892] 1997). An intensional
theory deals with the senses of expressions. Senses were traditionally held to determine, given how the world is, expressions
extensional properties. Frege introduced sense to account for the
cognitive significance of identity statements like The Morning
Star is The Evening Star. Since the referent of each term is Venus,
the significance has to be accounted for by appeal to something
else. Frege assigned each expression The Morning Star and
The Evening Star a distinct mode of presentation of Venus,
which he called their senses, and then generalized the notion to
expressions in other categories. New information is conveyed
about Venus, on this view, by way of the association of distinct
modes of presentation with one object. If, generally, extensions
are determined by terms intensions (a matter of controversy),
an intensional theory suffices for an extensional one. Intensional
theories, following Frege, typically reify expression senses and
assign them to expressions. These abstract intensional entities
are to be as finely individuated as meaning. They may be concepts (Fregean senses), properties, relations, propositions, or
similar objects (for example, functions from possible worlds to
extensions). Intensions of complex expressions are assigned on
the basis of the intensions of their parts. An extensional semantic
theory, in contrast, assigns referents to referring terms, extensions to predicates (sets of things or ordered n-tuples they are
true of, or something determinative of these), and truth values
to sentences (denoted with True or False). An extensional
semantic theory is not a full semantic theory, but a promising approach (discussed later) suggests that with appropriate
constraints, an extensional theory can achieve the aim of a full
semantic theory.
An influential formal framework in linguistics is Montague
semantics (Montague 1974). Montague semantics employs formal logic and the ontology of possible worlds (introduced to
provide a semantics for statements about necessity and possibility [Kripke 1963]; see also possible worlds semantics)
to develop formal intensional compositional theories of natural
languages. The approach originates in Frege ([1892] 1997) and
was developed in important work by Alonzo Church (1951) and
Rudolf Carnap (1947). The basic idea is that semantic compositionality is functional application: When significant expressions
are concatenated, one is a functional term and the other an argument term. Each expression has an intension and an extension.
The intension is a function from possible worlds to extensions.
This formally models expression sense as a determinant of extension relative to how the world is. The extension of an expression
is an entity of a sort appropriate for its semantic type.
To provide a rough idea of the approach, take as an example
the sentence Every ball rolled. The extension (referent, essentially) of each word, in Montagues approach, is a function. The
sentence is formed by first concatenating Every with ball,
where ball is an argument for the function denoted by Every,
which yields as its value another function denoted by Every ball
(thus, it is of the form Every(ball) where Every() represents
the function and ball supplies its argument). Every ball is then
concatenated with rolled, which is treated as supplying the
argument for the function denoted by Every ball. Thus, Every
ball rolled is treated as having the form (Every(ball))(rolled).
The extensions of ball and rolled are functions from individuals (or for Montague, strictly speaking, individual concepts) to
truth values (the function determines a set of individuals the word
is true of, and so a predicate extension). Their intensions, in turn,
are functions from possible worlds to such functions. Every is
treated as referring to a second-order function. It takes the intensions of common nouns like ball as arguments to yield a function
that takes intensions of verbs like rolled as their arguments and
yields truth values. The intension of Every, in turn, is a function from possible worlds to functions of the sort it denotes. The
extension of Every(ball), in particular, is a function that takes
the intension of a predicate F (e.g., rolled) to True if, and only
if, F (e.g., rolled) is true of anything that ball is true of (for our
example, if, and only if, every ball rolled).
The point of taking argument terms generally to supply their
intensions is that it provides a uniform treatment that handles
so-called intensional contexts, in which one cannot intersubstitute on the basis of sameness of extension. Consider an intensional transitive verb such as want. One can want a basketball
without wanting an orange ball, even if all and only basketballs
are orange balls (which means that a basketball and an orange
ball have the same extension). Want then must not take extensions as arguments, for a function yields one value per argument.
However, the intensions of a basketball and an orange ball
differ, since there are possible worlds in which some basketballs
are not orange. So if want takes the intensions of the terms
that follow it as arguments, we can explain why we cant substitute in the context and be guaranteed to preserve truth value.
On this approach, compositionality is understood as a matter
of the meanings of complex expressions being a function of the
meanings of their syntactic constituents; that is, it requires only
that there be a mapping from the intensions of the constituents
of complex expressions (in a corresponding order) to the intensions of those complex expressions.
Two points are worth noting about this framework. First, treating intensions as functions from possible worlds to extensions
does not allow us to distinguish between the meanings of necessarily coextensive expressions, such as triangle and trilateral. This might be overcome by some refinement of the notion
of intension (e.g., Thomason 1980). Second, the approach does
its work by assigning a referent to every significant expression,
whether or not it appears to be a referring term. Two questions
arise about this feature of the approach (and others that assign
intensional entities to ostensibly nonreferring expressions, e.g.,
situation semantics). The first is whether it is necessary. The
second is whether it is sufficient. We consider both questions
by looking at a more ontologically parsimonious approach proposed by Donald Davidson ([1984] 2001, Chap. 2).
Davidsons approach also exploits formal logic. The central
idea is that an axiomatic truth theory for a language that is
known to satisfy certain constraints can be used to provide interpretations of its sentences in a way that reveals how we understand complex expressions on the basis of their parts. Alfred
Tarski ([1934] 1983) showed how to define a truth predicate for a
formal language a predicate that had in its extension all and only
the true sentences of the language. The desideratum he placed
on a successful definition (Tarskis Convention T) required
that it entail all instances of the schema (T), where is T is the
743
Semantics
defined predicate, s is replaced by a structural description of an
object language sentence and p by a metalanguage sentence
that translates it. (S), for example, is an instance of (T), where
is the symbol for concatenation (AB =df the concatenation of A with B).
(T) s is T if and only if p
(S) Laneigeest blanche is T if and only if snow is white.
Davidson noted that if we had a theory that met Tarskis constraints and could identify formally the sentences of the form (T)
satisfying them, then we would have in effect a way of providing
an interpretation for every sentence of the language. For if p
translates s, means that can replace is T if and only if without loss of truth, yielding a sentence of the form (M) for every
object language sentence. Moreover, if the axioms meet a similar
constraint, the proof of the theorem for a sentence would reveal
its compositional structure, that is, how its meaning giving truth
conditions were determined by meaning giving axioms and its
structure.
Tarskis approach can be modified to apply to natural languages containing context-sensitive expressions (Larson and
Segal 1995; Lepore and Ludwig 2005, 2007). It can also be
extended to nondeclarative sentences (Boisvert and Ludwig
2006). Since the approach gives a central place to a truth theory for a language, it is often called truth-theoretic semantics or
truth conditional semantics, though this label is also
applied to Montague semantics. Notably, it requires no more
ontology than is needed for an extensional semantic theory.
If this approach can be elaborated successfully, it shows that
assigning intensional entities to all expressions is not essential to
compositional semantics.
There are reasons to think that assigning intensional entities to
expressions is also not itself sufficient. A compositional meaning
theory must enable understanding of the object language, but the
utility of a theory that assigns intensional entities to expressions in
enabling understanding depends on the use of terms we already
understand to denote their meanings. This shows that reference to
the meaning entities per se is not sufficient. A simple example illustrates the difficulty. Let an italicized expression denote its meaning. Let f(x, y) be a function that takes the meaning of an adjective
in the place of x and of a noun in the place of y to the meaning
of the expression gotten from adjectival modification of the noun.
Suppose as axioms: i) The adjective rouge means red, ii) the noun
boule means ball, and iii) for any adjective A and noun N, for any
meanings x and y, if x is the meaning of A and y is the meaning of
N, then NA means f(x, y). It follows that boule rouge means
f(red, ball). We see that f(red, ball) denotes the meaning of red
ball, which we understand, and so come to understand boule
rouge. However, if instead the meanings were denoted using the
meaning of rouge and the meaning of boule instead of red
and ball in our axioms, the theorem would still assign the right
meaning, but knowing it would not suffice to understand boule
rouge. Thus, understanding is not conveyed (where it is) by the
reference to entities but by appropriate matching of understood
terms with terms of the object language understood to be the same
in meaning (see Lepore and Ludwig 2006).
What semantical primitives mean in a natural language is ultimately a matter of their use in a linguistic community. Further
744
Semantics, Acquisition of
Carnap, Rudolf. 1947. Meaning and Necessity. Chicago: University of
Chicago Press.
Church, Alonzo. 1951. A formulation of the logic of sense and denotation. In Structure, Method and Meaning: Essays in Honour of H. M.
Sheffer, ed. P. Henle, H. Kallen, and S. Langer, 324. New York: Liberal
Arts.
Davidson, Donald. [1984] 2001. Inquiries into Truth and Interpretation.
2d ed. New York: Clarendon.
Frege, Gottlob. [1892] 1997. On Sinn and Bedeutung. In The Frege
Reader, ed. M. Beaney. 15171. Oxford: Blackwell.
Grice, Paul. 1957. Meaning. Philosophical Review 66: 37788.
. 1969. Utterers meaning and intentions. Philosophical Review
78: 14777.
. 1989. Studies in the Way of Words. Cambridge: Harvard University
Press.
Kripke, Saul. 1963. Semantical considerations on modal logics. Acta
Philosophica Fennica 16: 8394.
Larson, Richard, and Gabriel Segal. 1995. Knowledge of Meaning.
Cambridge, MA: MIT Press.
Lepore, Ernest, and Kirk Ludwig. 2005. Donald Davidson: Truth, Meaning,
Language and Reality. New York: Oxford University Press.
. 2006. Ontology in the theory of meaning. International Journal
of Philosophical Studies 14.3: 32131.
. 2007. Donald Davidsons Truth-Theoretic Semantics. New
York: Oxford University Press.
Montague, Richard. 1974. Formal Philosophy: Selected Papers of Richard
Montague. Ed. R. Thomason. London: Yale University Press.
Quine, Willard Van Orman. 1960. Word and Object. Cambridge, MA: MIT
Press.
Searle, John. 1979. A taxonomy of illocutionary acts. In Expression and
Meaning, 129. Cambridge: Cambridge University Press.
. 1983. Intentionality: An Essay in the Philosophy of Mind.
Cambridge: Cambridge University Press.
Tarski, Alfred. [1934] 1983. The concept of truth in formalized
languages. In Logic, Semantics, Metamathematics, 152278.
Indianapolis: Hackett.
Thomason, R. 1980. A model theory for propositional attitudes.
Linguistics and Philosophy 4: 4770.
SEMANTICS, ACQUISITION OF
Without special training or carefully sequenced input, by the
age of four or five children are effectively adults in their abilities
to produce and understand novel sentences and to judge the
truth or falsity of endlessly many statements. There are two main
accounts of this remarkable acquisition scenario, one emphasizing the contribution of innate knowledge (see innateness
and innatism)and the other emphasizing childrens abilities
to extract generalizations from the input. The alternatives can
be traced back to the nature versus nurture debate about how
knowledge is acquired in any cognitive domain.
One approach views language acquisition on a par with the
acquisition of social skills, learning to count, learning to read,
and so on. This nurture approach views language development
as recruiting domain-general learning mechanisms and highlights the availability of relevant cues in the input to children.
These cues serve as the basis for the linguistic generalizations
that children form. The alternative nativist approach attributes
domain-specific knowledge to children and highlights the contributions of human nature to the emergence of linguistic knowledge. On this view, there is a considerable gap between childrens
Disjunction
In classical logic, disjunction has the truth conditions associated with inclusive-or, such that a statement of the form A or
B is true if only A is true, if only B is true, and if both A and B
are true. However, the findings from several studies have been
interpreted as showing that even children 1012 years old assign
the truth conditions associated with exclusive-or to natural language or (e.g., Braine and Rumain 1983). Although Martin Braine
and Barbara Rumain acknowledge the view that equates or with
standard logic, they ultimately reject this view on the grounds
that coherent judgments of the truth of or-statements emerge
relatively late and are not universal in adults (1983, 291). The
conclusion that children assign the exclusive-or (not both)
interpretation to disjunctive statements is unwarranted, however, because it rests largely on tasks that involve requests to perform actions, which encourage a choose one interpretation.
But an act-out methodology has limited value in determining the
full range of truth conditions that are associated with a particular linguistic expression (Crain and Thornton 1998). At most, the
act-out task provides evidence that the subjects grammar allows
one interpretation (the one that is acted out), but findings from
this task cannot be used to infer that children do not have other
interpretations available to them. Children may simply favor one
reading over others in a certain experimental context.
A research methodology has been devised to distinguish
between a preference for one interpretation and the availability of alternative interpretations. It is called the truth value
judgment task (Crain and Thornton 1998). In this task, children
judge whether or not statements made by a puppet, played by
one experimenter, are accurate descriptions of what happened
in stories acted out by a second experimenter. For example, suppose that Bunny Rabbit ate a carrot and a pepper. The puppet
745
Semantics, Acquisition of
could then produce a variety of different statements: Bunny
Rabbit ate a carrot and a pepper versus Bunny Rabbit ate a
carrot or a pepper. Notice that the statement with disjunction is
true from a logical point of view since the truth conditions associated with disjunction in logic are those of inclusive-or. In fact,
children younger than six or seven consistently judge or-statements to be true in circumstances where and-statements are also
true (e.g., Chierchia et al. 2004). By contrast, adults consistently
judge or-statements to be false in circumstances where andstatements are true. Adult judgments are influenced by a pragmatic principle of cooperation. This principle dictates that
a speakers intended meaning should be conveyed as directly
as possible. In circumstances in which both or-statements and
and-statements are true, and-statements convey the intended
meaning more perspicuously since or-statements would be true
in other circumstances as well. Since hearers infer that a cooperative speaker would use an and-statement if he or she wanted
to convey the both interpretation, the use of or is taken to imply
exclusivity (not both).
Negation
One of the laws of classical logic governs negative statements
with disjunction: (A B) A B. According to this law, a
negative disjunctive statement, (A B), generates a conjunctive
entailment, A B. This entailment only holds if disjunction is
interpreted as inclusive-or. In human languages, too, we can determine whether or not disjunction (e.g., English or, Japanese ka)
corresponds to inclusive-or by asking whether negative statements with disjunction license conjunctive entailments.
Consider (1).
(1)
English-speakers judge (1) to be true in circumstances (as indicated by ) in which Suzi did not see Max eat sushi and she
did not see Max eat pasta. So, in English, negative statements
with disjunction generate a conjunctive entailment, just as a
negative disjunctive statement generates a conjunctive entailment in classical logic. In (1), the clause that contains disjunction, Max eat sushi or pasta, is embedded in the clause with
negation, Suzi didnt see. In the syntax, the clause with
negation is higher than the clause that contains disjunction.
Semantically, the critical observation is that (1) generates a conjunctive entailment.
Cross-linguistic research has found that conjunctive entailments are also licensed in other languages in statements in
which negation appears in a higher clause than the clause
that contains disjunction; that is, in statements of the form not
S[A or B]. Statements of this form are interpreted as excluding both the possibility of A and of B. This is clearly the case
in English, as (1) illustrates. Moreover, when such statements
are translated into Japanese, German, Russian, Hungarian,
and so forth, the corresponding statements also carry the same
conjunctive entailments. The cross-linguistic findings invite us
to conclude that disjunction has the truth conditions associated with inclusive-or in all human languages (see Crain and
Thornton 2006).
746
Semantics, Acquisition of
the same conjunctive entailment that everyone who ordered
pasta became ill and everyone who ordered sushi became ill
(contrary to fact). This shows us that disjunction licenses a conjunctive entailment in the subject phrase of the universal quantifier everyone who ordered pasta or sushi.
(3)
We now turn to the literature on child language. Several studies have investigated the truth conditions that children associate
with disjunction in the subject phrase and in the predicate phrase
of the universal quantifier. Using a truth value judgment task, one
study asked children four and five years old to evaluate sentences
like those in (5) and (6) (Gualmini, Meroni, and Crain 2003).
(5)
(6)
Focus Expressions
The final topic is the acquisition of focus operators: English only,
Japanese dake, Chinese zhiyou. The meaning of focus operators
is quite complex. Consider the statement: Only Bunny Rabbit
ate a carrot or a pepper. This statement expresses two propositions. One is called the presupposition, and the other is
called the assertion (e.g., Horn 1969, 1996). The presupposition is
derived simply by deleting the focus expression from the original
747
Semantics, Acquisition of
the test sentences over 90% of the time. This finding suggests
that children assigned disjunctive (not both) truth conditions
to or in the presupposition in sentences like (7). In contrast,
the same children rejected the test sentences in Condition
II over 90% of the time. The high rejection rate by children
in Condition II suggests that they know that the disjunction
operator or creates conjunctive entailments in the assertion of
sentences with only/dake. The findings invite the inference that
children have adultlike knowledge about the semantics of only/
dake and are able to compute its complex semantic interaction
with disjunction.
To conclude, the contribution of logic to human languages
is far from transparent due to the intrusion of pragmatic factors
and due to cross-linguistic variation. Children appear to be more
logical than adults, however, because children are less sensitive to these factors, at least initially. Therefore, studies of child
language can sometimes be more revealing about the interplay
of logic and language than studies of adults. Yet there are also
examples of universal linguistic principles that reveal properties
in common between logic and human languages for both children and adults.
Stephen Crain and Rosalind Thornton
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Braine, Martin, and Barbara Rumain. 1983. Logical reasoning. In
Handbook of Child Psychology. Vol. 3: Cognitive Development. Ed.
J. Flavell and E. Markman, 261340. New York: Academic Press.
Chierchia, Gennaro, Maria Teresa Guasti, Andrea Gualmini, Luisa
Meroni, and Stephen Crain. 2004. Semantic and pragmatic competence in children and adults interpretation of or. In Experimental
Pragmatics, ed. I. Noveck and S. Wilson, 283300. London: Palgrave.
Crain, Stephen, and Rosalind Thornton. 1998. Investigations in Universal
Grammar: A Guide to Experiments in the Acquisition of Syntax and
Semantics. Cambridge, MA.: MIT Press.
. 2006. Acquisition of syntax and semantics. In Handbook of
Psycholinguistics (2d ed.), ed. M. Traxler and M. Gernsbacher, 1073
110. Oxford: Elsevier.
Crain, Stephen, Takuya Goro, and Rosalind Thornton. 2006. Language
acquisition is language change. Journal of Psycholinguistic Research
35: 3149.
Goro, Takuya. 2004. The emergence of universal grammar in the emergence of language: The acquisition of Japanese logical connectives and
positive polarity. Manuscript, University of Maryland at College Park.
Goro, Takuya, and Sachie Akiba. 2004. The acquisition of disjunction and positive polarity in Japanese. In West Coast Conference on
Formal Linguistics 23, ed. V. Chand, A. Kelleher, A. Rodrguez, and B.
Schmeiser, 25164. Somerville, MA: Cascadilla.
Goro, Takuya, Utako Minai, and Stephen Crain. 2006. Bringing out the
logic in child language. In Proceedings of North Eastern Linguistic
Society 35, ed. L. Bateman and C. Ussery, I: 24556. Amherst, MA: GLSA
Publications.
Gualmini, Andrea, Luisa Meroni, and Stephen Crain. 2003. An asymmetric universal in child language. In Proceedings of Sinn und Bedeutung
7, ed. Matthias Weisgerber, 13648. Constance, Germany: Konstanz
Linguistics Working Papers.
Horn, Laurence. 1969. A presuppositional approach to only and even.
Proceedings of the Chicago Linguistic Society 5: 98107.
. 1996. Presupposition and implicature. In Handbook
of Contemporary Semantic Theory, ed. S. Lappin, 299319.
Oxford: Blackwell.
748
749
Semantics, Neurobiology of
750
SEMANTICS, NEUROBIOLOGY OF
semantics is the study of meaning, and unsurprisingly, for
different disciplines the term itself has come to mean different
things. The relation between neurobiology and semantics can
also be construed in any number of ways. Here, we refer to it
as the study of the ways in which knowledge is established and
organized in the brain.
Currently, most data in the field are derived from studies
of language. These studies can be summarized in the context
of two main goals: a) understanding the principles by which
semantic knowledge that is, information about objects and
words is organized, and b) understanding how the brain
enables compositional operations over this knowledge. With
respect to the first goal, researchers have addressed the organization of conceptual knowledge in the brain and its relation
to taxonomic ontology, as well as the neural mechanisms for
operating on it. With respect to the second goal, neurobiological investigations of compositional processes address issues
such as comprehension of nominal combinations, metaphor,
and sentences. At the level of discourse, empirical research
has examined the comprehension of sentences in context, and
the relation between sentence comprehension and the evaluation of sentence meaning. Brain research has also addressed
the comprehension of meaningful nonlinguistic semiotic entities, such as graspable objects, hand gestures, environmental
sounds, pantomime, and music. Furthermore, the neural processes underlying compositional processes in language may
be, at least partially, subserved by networks that perform more
basic functions, such as the monitoring of predictability in an
input stream. Thus, investigating issues both within and outside the purely linguistic realm could lead to a better understanding of semantic processes in language and interpretive
processes in general.
In what follows, we review the research in a) the organization of conceptual knowledge, b) sentence-level and discourselevel processes, and c) research in nonlinguistic domains. Our
goal is to outline a general framework for the neurophysiology
of semantics.
Semantics, Neurobiology of
Later research revealed that Wernickes area was not a concept center (i.e., a repository of semantic information) but
had a more complex role. Indeed, current research suggests that
no single cortical area serves as a centralized store of semantic
knowledge. This distributed nature of semantic memory has
been revealed by two scientific methods: lesion studies and
neuroimaging. Until the advent of neuroimaging, the neural basis of semantic knowledge was evaluated by establishing
associations between lesions to specific brain regions and the
concomitant behavioral deficits. This lesion deficit model, developed in the late nineteenth and early twentieth centuries, was
advanced significantly by a series of pivotal studies by Elizabeth
Warrington and her colleagues showing that multifocal lesions
(e.g., as caused by herpes simplex encephalitis) and focal lesions
(e.g., stroke, tumors) could result in larger deficits for some
semantic categories than others (for reviews, see Forde and
Humphreys 2002). The literature reports (relative) category-specific deficits along various axes: for example, abstract versus concrete, living versus nonliving, or nouns versus verbs (cf. Gainotti
2000, for meta-analysis). The existence of such categorical deficits from focal lesions, albeit of a graded nature, has led to multiple interpretations and is the focus of ongoing imaging studies.
These neuroimaging studies have shown that accessing semantic
information for such categories as tools, faces, or body parts is
associated with peaks of activity in particular regions that differ
across categories (e.g., Yovel and Kanwisher 2004). Attesting to
the conceptual essence of this knowledge, certain regions are
active whether the items for judgment are presented as words or
pictures (cf. Martin 2001). However, it is important to note that
many of these studies were designed to evaluate whether there
are neural regions that are more relatively sensitive for one category than another; the extent of common activity is typically of
less interest, though it is of major theoretical importance. Indeed,
a number of studies have shown that the distributed population
codes in less active regions also demonstrate sensitivity to categories (e.g., Haxby et al. 2001).
Not only is knowledge about different categories distributed,
but accessing different features of the same item can also lead
to activity in different cortical areas. For instance, naming a referent is associated with neural activity patterns in cortical areas
different from those implicated in access to other types of conceptual knowledge. Although deficits in naming are more likely
to be associated with lesions to the left hemisphere than to the
right, deficits in recognition, for example, the ability to say meaningful things about an object, are associated with more evenly
distributed lesions (H. Damasio et al. 2004; Gainotti 2000).
Thus, in the brain, there is a distinction between knowing object
names a very unique feature of a referent and more general
sorts of knowledge. Feature knowledge may also be distributed
among different regions, in particular, those associated with
perception and motor function. In support of this possibility, it
has been shown that verifying whether a certain color or a certain action holds for an object activates different brain regions
(Martin 2007).
Are these perceptual- or motor-related units of information
independent, or are they linked via a more abstract, modalityindependent form of semantic knowledge? A recent case study
(Coccia et al. 2004) argues for the latter. It reported a gradual
751
Semantics, Neurobiology of
While frontal and temporal regions are essential to semantic
processing during sentence comprehension, recent data suggest that modality-specific brain regions also play an important
role. M. Tettamanti and colleagues (2005) found that sentences
describing actions performed by the human body evoke greater
activity in posterior left IFG than do abstract sentences, suggesting that this region codes for action on a level that is sufficiently
abstract to be accessible to language. This finding is notable
because the homologue of this region in the monkey contains
mirror neurons, which fire during both observation and execution of certain goal-directed actions (see mirror systems). The
researchers also show that sentences referring to mouth, hand,
and leg actions evoked neural activity in premotor regions associated with the performance of the described actions (cf. AzizZadeh et al. 2006 for similar results; cf. Buccino et al. 2005 and
Hauk, Johnsrude, and Pulvermller 2004 for related findings).
Language comprehension, at least in the action domain, may
therefore be supported by motor systems used to perform the
actions referenced in those sentences. These findings are consistent with theoretical approaches in the embodied cognition
framework (see cognitive linguistics; embodiment).
The interpretation of sentences relies on an integration of
world knowledge with the proposition conveyed by the sentence.
A central issue is whether at any stage of comprehension there is
a level of semantic interpretation that corresponds to the state of
affairs denoted by a sentence, but that is relatively encapsulated
from world knowledge or belief. P. Hagoort and colleagues (2004)
showed that false statements (in virtue of world knowledge) and
nonsensical statements (in virtue of lexical knowledge) are recognized as anomalous equally fast. Specifically, an electrophysiological signature associated with the ease of integrating a word
into context (the N400) was almost identical for the word white
in the false statement Dutch trains are white as it was for the
word sour in Dutch trains are sour. This suggests that lexical
and world knowledge are integrated during the same time frame.
Functional magnetic resonance imaging (fMRI) showed that both
violations were associated with increased activity in anterior left
IFG. Related research (Berkum, Hagoort, and Brown 1999) has
shown that the recent discourse context similarly constrains the
integration of words during sentence comprehension. Thus, the
construction of sentence representation seems to immediately
incorporate world knowledge and prior discourse context.
The integration of meta-logical information with proposition
content has not been extensively researched. One exception is a
study that investigated the neural mechanisms underlying successful encoding of propositions said to be either true or false
(Mitchell, Dodson, and Schacter 2005). This study began from
behavioral findings showing that false sentences are more likely to
be (mis)recalled as true than vice versa, which suggests that their
encoding requires additional cognitive effort (cf. Gilbert 1991).
In an fMRI study, J. P. Mitchell, C. S. Dodson, and D. L. Schacter
found that accurate recall of falsity (i.e., correctly remembering
that a statement was false) was associated with increased activity
in IFG (more on the left) and left medial temporal regions during
encoding (see meaning and belief).
Several neuroimaging studies have examined processes of
meaning integration at the discourse level, but this domain of
research is in its initial stages. Discourse comprehension at the
752
multiple-sentence level demands integration of ongoing discourse with prior context and world knowledge. An early study
(St. George et al. 1999) implicated the right hemisphere in such
processes, demonstrating that middle temporal regions in this
hemisphere showed reduced activity when a confusing text was
clarified by an informative title (interestingly, the left hemisphere
homologue showed increased activity). However, another study
employing a conceptually similar manipulation where texts
were clarified by pictures did not replicate this result, finding
increased activity in medial brain regions (Maguire, Frith, and
Morris 1999). Other studies show that the integration of consecutive discourse ideas is associated with activity in the anterior
temporal pole (bilaterally), IFG (bilaterally), and perhaps also
the dorso-medial prefrontal cortex. These regions differentiate
between narratives and unlinked sentences and also differentiate between sentences that differ in their consistency with prior
context (Ferstl, Rinck, and von Cramon 2005; Xu et al. 2005;
see coherence, discourse). Transitions between narrative
events have also been associated with activity in the right temporal regions (Speer, Zacks, and Reynolds 2007).
Two studies (Kuperberg et al. 2006; Mason and Just 2004)
have attempted to identify cortical regions mediating the integration of world knowledge by examining neural activity during
comprehension of causal statements. Both reported that comprehension of statements related by a causal relation of intermediate strength was associated with greater neural activity than
comprehension of statements in which the causal relation was
very weak or very strong. These findings are consistent with the
notion that scenarios with intermediate causal strength demand
particular enrichment with world knowledge. However, the
regions identified by these studies did not overlap. The picture
emerging from the study of discourse-level relations is that these
rely additionally on brain regions not typically involved in the
comprehension of single, context-independent sentences (JungBeeman 2005).
Semantics, Neurobiology of
pictures. Both deficits were associated with damage to posterior regions in the left superior temporal gyrus (STG) and the
inferior parietal lobule. A. Cummings and colleagues (2006)
used a method with very high temporal resolution (EEG/ERP)
to assess whether matching pictures to words or sounds is
associated with similar time lines of neural processing. When
either words or sounds mismatched a picture, they found a
typical index of mismatch processing (an N400 ERP component). Attesting to the high similarity between meaningful
words and meaningful sounds, there was no difference in the
onset time of this component, and if anything, it peaked earlier
for mismatching sounds than for mismatching pictures. These
results are consistent with other findings (Saygin et al. 2003)
showing that environmental sounds are recognized faster than
the corresponding word referents, which suggests that they are
not recoded as lexical items. As a whole, these studies demonstrate that similar networks process meaningful auditory
information, whether expressed verbally or not. Knowledge
of environmental sounds may also be organized in a distributed manner. Certain regions are particularly sensitive to such
categories as animal sounds or sounds associated with actions
performed by body parts (Gazzola, Aziz-Zadeh, and Keysers
2006; Kraut et al. 2006).
Compositional processes underlying language comprehension may also be instantiations of more basic processes that are
not specific to language. C. Humphries and colleagues (2001)
demonstrated that both meaningful sentences and meaningful
environmental sound sequences were associated with neural
activity in bilateral temporal regions associated with semantic
processing, including posterior aspects of the superior temporal gyrus and sulcus. Sentence comprehension was associated
with increased activity in several temporal regions. However,
as discussed, the increased activity for sentences as compared
to environmental sounds could be due to easier semantic access
afforded by environmental sounds. Studies of the neurobiology of music processing similarly suggest some overlap with a
semantic system subserving language comprehension. Listening
to structured musical pieces as opposed to scrambled versions
results in increased activity in a region of IFG known to play
an important role in the integration of information over time
(Levitin and Menon 2003). And music, like spoken language, can
activate broad semantic domains (Koelsch 2005). More basic
functions can also potentially explain activity in certain regions
during more and less surprising linguistic information: It has
been shown that decreased predictability of events in a nonverbal input stream is associated with greater activity in Wernickes
area and its right homologue, as well as posterior IFG bilaterally
(Bischoff-Grethe et al. 2000). Thus, the neural regions subserving
semantic processes in language may have a more general role in
interpreting meaningful communication streams.
Uri Hasson and Steven L. Small
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Aziz-Zadeh, L., S. M. Wilson, G. Rizzolatti, and M. Iacoboni. 2006.
Congruent embodied representations for visually presented
actions and linguistic phrases describing actions. Current Biology
16.18: 181823.
753
Semantics, Universals of
Humphries, C., K. Willard, B. Buchsbaum, and G. Hickok. 2001. Role
of anterior temporal cortex in auditory sentence comprehension: An
fMRI study. NeuroReport 12.8: 174952.
Jung-Beeman, M. 2005. Bilateral brain processes for comprehending
natural language. Trends in Cognitive Sciences 9: 51218.
Koelsch, S. 2005. Neural substrates of processing syntax and semantics
in music. Current Opinion in Neurobiology 15: 20712.
Kraut, M. A., J. A. Pitcock, V. Calhoun, J. Li, T. Freeman, and J. Hart, Jr.
2006. Neuroanatomic organization of sound memory in humans.
Journal of Cognitive Neuroscience 18.11: 187788.
Kuperberg, G. R., B. M. Lakshmanan, D. N. Caplan, and P. J. Holcomb.
2006. Making sense of discourse: An fMRI study of causal inferencing
across sentences. NeuroImage 33: 34361.
Levitin, D. J., and V. Menon. 2003. Musical structure is processed in
language areas of the brain: A possible role for Brodmann area 47 in
temporal coherence. NeuroImage 20.4: 214252.
Maguire, E. A., C. D. Frith, and R. G. M. Morris. 1999. The functional neuroanatomy of comprehension and memory: The importance of prior
knowledge. Brain 122: 183950.
Martin, A. 2001. Functional neuroimaging of semantic memory. In
Handbook of Functional Neuroimaging of Cognition, ed. R. Cabeza and
A. Kingstone, 15386. Cambridge, MA: MIT Press.
. 2007. The representation of object concepts in the brain. Annual
Review of Psychology 58: 2545.
Mason, R. A., and M. A. Just. 2004. How the brain processes causal inferences in text: A theoretical account of generation and integration component processes utilizing both cerebral hemispheres. Psychological
Science 15.1: 17.
Mitchell, J. P., C. S. Dodson, and D. L. Schacter. 2005. fMRI evidence for
the role of recollection in suppressing misattribution errors: The illusory truth effect. Journal of Cognitive Neuroscience 17.5: 80010.
Rothi, L. J., L. Mack, and K. M. Heilman. 1986. Pantomime agnosia.
Journal of Neurology, Neurosurgery and Psychiatry 49.4: 4514.
Saygin, A. P., F. Dick, S. M. Wilson, N. F. Dronkers, and E. Bates. 2003.
Neural resources for processing language and environmental
sounds: Evidence from aphasia. Brain 126.4: 92845.
Speer, N. K., J. M. Zacks, and J. R. Reynolds. 2007. Human brain activity time-locked to narrative event boundaries. Psychological Science
18.5: 44955.
St. George, M., M. Kutas, A. Martinez, and M. I. Sereno. 1999. Semantic
integration in reading: Engagement of the right hemisphere during
discourse processing. Brain 122: 131725.
Tettamanti, M., G. Buccino, M. C. Saccuman, V. Gallese, M. Danna, P.
Scifo, et al. 2005. Listening to action-related sentences activates
fronto-parietal motor circuits. Journal of Cognitive Neuroscience
17.2: 27381.
Tyler, L. K., E. A. Stamatakis, P. Bright, K. Acres, S. Abdallah, J. M. Rodd, et
al. 2004. Processing objects at different levels of specificity. Journal of
Cognitive Neuroscience 16.3: 35162.
Van Petten, C., and B. J. Luka. 2006. Neural localization of semantic context effects in electromagnetic and hemodynamic studies. Brain and
Language 97.3: 27993.
Xu, J., S. Kemeny, G. Park, C. Frattali, and A. Braun. 2005. Language in
context: Emergent features of word, sentence, and narrative comprehension. NeuroImage 25.3: 100215.
Yovel, G., and N. Kanwisher. 2004. Face perception: Domain specific,
not process specific. Neuron 44.5: 88998.
SEMANTICS, UNIVERSALS OF
A semantic universal is any aspect of meaning that is somehow represented in all languages. To first characterize the two terms of this
subject separately, semantics pertains to conceptual material as
754
it is organized by language. It ranges from single linguistically represented concepts to principles of conceptual organization.
Universality has properties varying along three parameters.
The level of a universal is absolute (Joseph Greenbergs
term, e.g., 1963) if it is realized in every individual language (see
absolute and statistical universals). But its level is here
termed abstractive if the universal is part of the language faculty
but not of every language. This second abstractive level can in
turn be realized in the three following ways, each exemplified
later in this entry. An abstractive universal is inventorial if each
language draws its own selection of elements of a certain type
from a relatively closed inventory of such elements. An abstractive universal is typological if languages fall into a typology
on the basis of manifesting one or another out of a relatively
small set of (combinations of) certain elements in an otherwise
common system. An abstractive universal is implicational
(Greenbergs term) if one linguistic feature can appear in a language only if another feature itself abstractive also appears
there (see implicational universals).
Another parameter of universality is its weighting. At the
extremes of weighting, a universal is either positive or negative.
At such extremes, an absolute universal feature is either present
in or absent from all languages, while an abstractive universal
feature is either part or not part of the language faculty. But while
some features pertaining to an abstractive universal may appear
in all or no languages, other features can range from appearing in most down to few languages behavior that must also be
regarded as part of the language faculty. Traditionally, such nonextremes have been termed tendencies. But to promote the idea
of a gradient hierarchy from positive down to negative, and to
highlight the fact that such tendencies are themselves properties
of the language faculty, such nonextremes will here be termed
positively tending abstractive universals or negatively tending
abstractive universals with whatever further modifiers it might
take to indicate the degree.
A final parameter of universality is its subject the particular
linguistic phenomenon that manifests the universal property.
For a semantic universal, this can range from a specific concept,
through a conceptual category or set of such categories, to a conceptual system and can extend as well to principles of conceptual
organization.
All the forms of universality in this framework presumably
have a cognitive basis, and this is addressed where feasible in the
following.
The framework can be illustrated by much within the present authors work. A positive absolute universal is that the morphemes of every language are divided into two subsystems,
the open class, or lexical, and the closed class, or grammatical
(see Talmy (2000a, Chap. 1). Open classes have many members
and can readily add more. They commonly include (the roots
of) nouns, verbs, and adjectives. Closed classes have relatively
few members and are difficult to augment. They include bound
forms inflections, derivations, and clitics and such free forms
as prepositions, conjunctions, and determiners. Closed classes
can also be implicit, as with word order patterns, lexical categories, and grammatical relations.
A semantic difference correlates with this formal difference.
The meanings that open-class forms can express are almost
Semantics, Universals of
unrestricted, whereas those of closed-class forms are highly constrained. This constraint first applies to the conceptual categories
to which they can refer. For example, many languages around
the world have closed-class forms in construction with a noun
that indicate the number of the nouns referent, but no language
has closed-class forms indicating its color. A positive abstractive
universal, accordingly, is that the grammatical morphemes of a
language can represent an approximately closed set of conceptual categories, such as those for number, gender, tense, aspect,
causality, and status. Excluded are color, and indefinitely many
more such as food and religion the corresponding negative
absolute universal.
A further semantic constraint on grammatical forms is that,
even within represented conceptual categories, only certain
member concepts can be grammatically expressed and not
others. Thus, a positive abstractive universal is that, within the
category of number, a bound closed-class form can represent
such concepts as singular, dual, plural, and paucal, while a free
closed-class form can also represent such concepts as no, some,
many, and all. But the corresponding negative absolute universal is that no languages closed-class forms represent concepts
such as even, odd, dozen, countable, or any other concept pertaining to number.
Open-class morphemes are not subject to these same semantic constraints on categories or member concepts. This is shown
by the existence of such morphemes as food, even and odd.
Even open-class morphemes exhibit a few semantic constraints, however. Thus, on the one hand, the meaning of a morphemic verb can incorporate particulars of aspect that can, in
turn, interact with external closed-class aspectual forms. On the
other hand, it is a negative absolute universal that the meaning of
no morphemic verb incorporates a tense that can, in turn, interact with external closed-class tense forms. If not for this exclusion, we might expect to find a verb like (to) went that could be
used in a construction like I am wenting to mean I was going,
or in a construction like I will went to mean I will have gone.
Comparably, proper nouns like Manhattan or Shakespeare
exist that refer to a specific bounded portion of the space-time
continuum. But a negative absolute universal is that there are no
proper verbs or proper adjectives with the same property.
That is, both can be type specific but they are token neutral. Thus,
there is never a verb like (to) Deluge referring uniquely to the soconceived spatio-temporally bounded event of the biblical flood,
as in some sentence like After it Deluged, Noah landed the ark.
Due to the semantic constraints on the closed-class subsystem, the total set of conceptual categories and their respective sets of member concepts that can ever be represented by
closed-class forms constitutes an approximately closed inventory. This inventory is universally available. No language has
closed-class forms representing all of the conceptual categories
and member concepts in the inventory. Rather, each language
draws in a unique pattern from the inventory for its particular set
of grammatically expressed meanings. Accordingly, this inventory is a positive abstractive universal of the language faculty,
not an absolute universal overtly manifested in all languages.
In turn, the pattern in which individual languages select their
particular sets of grammatically expressed conceptual categories and member concepts from the inventory is governed by a
755
Semantics, Universals of
earlier (and causal) event in the subordinate clause, leaving
the later (and caused) event in the main clause, and never the
other way around. The exception is that in addition to an aftertype conjunction, which obeys the constraint languages can
also have a before-type conjunction. Thus, beside We left after
we ate, English has We ate before we left. But all other cases follow a negative absolute universal. Thus, beside a because-type
conjunction that obeys the constraint, as in English We stayed
home because they arrived, English has no inverse conjunction
lexicalized to express the hyphenated phrase in *They arrived
to-the-occasioning-of-the-event-that we stayed home and seemingly no other language does either. Comparably, beside an
although-type conjunction, as in We went out even though they
arrived, there is never a conjunction lexicalized to represent the
hyphenated phrase in *They arrived in-ineffective-counteractingof-the-event-that we went out. Talmy (2000a, Chap. 5) bases a
cognitive account for this unidirectional lexicalization on an
earlier events natural function as ground for a later event as figure. And it details a similar unidirectionality in conjunctions for
the temporal inclusion of one event in another, the contingency
of one event on another, and the substitution of one event for
another.
Based on their formal and semantic differences, treated so
far, a further major finding is that the two types of form classes exhibit a functional difference. In the conceptual complex
evoked by any portion of discourse, the open-class forms contribute most of the content, while the closed-class forms determine most of the structure. This division of labor in cognitive
function amounts to a positive absolute universal: The openclass subsystem as a whole represents conceptual content and
the closed-class subsystem as a whole represents conceptual
structure (for illustration, see Talmy 2000a, Chap. 1). Although
individual open- and closed-class forms in general or in a particular sentence may perform the opposite functions, the subsystems overall are universally dedicated to their respective content
and structure functions. Further, the function of the closed-class
subsystem to structure conceptual content presumably accounts
for the negative absolute universals that constrain its semantics.
The crucial conclusion again a positive absolute universal is
that the closed-class subsystem is perhaps the most fundamental conceptual structuring system of language.
The main focus so far has been on semantic abstractive universality that involves an inventory, but we now switch to a type
that involves a typology. Talmy (2000b, Chaps. 1, 2, 3) examines
such typologies for an extended Motion event (the capitalized
form covers both motion and location). This larger event consists of a Motion event proper and a co-event that usually represents the manner or the cause of the Motion. In turn, the event
of Motion consists of four components the moving or stationary figure, its state of Motion (moving or being located), its path
(path or site), and the ground that serves as its reference point.
What may be a positive absolute universal is that every language
has coordinated lexicalization patterns and syntactic constructions to represent all the components of an extended Motion
event directly and colloquially over at most two clauses. But the
particular lexical and syntactic categories in which the semantic
components are characteristically represented is not an absolute
universal. Rather, each language has a characteristic pattern of
756
Semantics, Universals of
Semantics-Phonology Interface
SEMANTICS-PHONOLOGY INTERFACE
Systems theories in widespread use within the cognitive and
neurosciences (see hippocampus) face two major challenges: determining when systems differ and determining
where and how different systems interact. Both problems have
been solved for the phonological and sentential-semantic
systems. The semantics-phonology interface has been established via three complementary criteria independent activity,
connectivity, and error frequence that may help resolve other
system-boundary disputes. Under these theoretical and empirical criteria, the sentential-semantic versus phonological systems
are language-memory and comprehension-production systems.
That is, the sentential-semantic system contains units for comprehending, storing, retrieving, and producing morphemes,
words, phrases, and propositions, and the phonological
system contains units for comprehending, storing, retrieving,
and producing syllables, phonological compounds, and segments. These are the three criteria for viewing the semantic-sentential versus phonological systems in this way.
757
Semantics-Phonology Interface
Semantics-Pragmatics Interaction
758
SEMANTICS-PRAGMATICS INTERACTION
It seems unlikely that there will ever be consensus about the
extent to which we can reliably distinguish semantic phenomena from pragmatic phenomena. But there is now broad agreement that sentence meaning can be given in full only when
a sentence is studied in its natural habitat: as part of an utterance by an agent who intends it to communicate a message.
Here, we document some of the interactions that such study
has uncovered. In every case, in order to achieve even a basic
description, it is necessary to pool semantic information, contextual information, speaker intentions, and general pragmatic
pressures.
Space limitations preclude discussion of presuppositions
and speech-acts, two important classes of phenomena for
which semantics and pragmatics are so thoroughly intertwined that analyses of them invariably draw information from
both domains.
In a broad range of cases, pragmatic information is required
just to obtain complete and accurate meanings for the words and
phrases involved. indexical expressions are clear examples. In
order to determine the proposition that is expressed by an utterance of (1), we must look to the context to fix the speaker:
(1)
I am here.
We must also appeal to the context to obtain the intended meaning of here (in this room, in this city, etc.). Which meaning we
select will be shaped by considerations of informativity and
relevance. For example, (1) is likely to be trivially true if here is
construed as picking out planet Earth, and speakers will therefore avoid that interpretation until interplanetary travel becomes
routine.
Similar factors influence anaphora resolution. If a speaker
utters (2), his or her addressees will be guided to a referent for
she by their assumptions about intentions as well as discourse
coherence:
(2)
The interpretive steps resemble those for indexicals, but pronouns tend to be more ambiguous in practice, and, perhaps as a
Semantics-Pragmatics Interaction
result, languages depend on complex systems of conventions to
narrow their referents.
This kind of context dependency is so widespread in language
that one might feel hard-pressed to find an example that lacked
it. We do not know what is expressed by Its cloudy until we fix a
time and place (and a standard for cloudiness). We cant be sure
of the truth or falsity of the counterfactual If kangaroos had no
tails, they would fall over unless we know which possibilities to
consider. (Suppose they had crutches or jet packs.) If a politician says Many people support my proposal, we need to know
whether there are additional implicit restrictions (people in the
city, people who own cats), and we need to have a sense for the
current numerical standards for many. The sentence Wood is
strong enough can be evaluated only if we can flesh out its meaning so that it specifies what wood is claimed to be strong enough
for (Bach 1994). Comparable examples can be constructed for
just about any area of natural language.
Pragmatic information can enrich a speakers message in
ways that extend far beyond determining its central descriptive
content. The primary meaning classification here is the conversational implicature. The dialogue in (3), based on one
from Grice (1975), illustrates:
(3)
Semiotics
responses, see Levinson 2000, Chierchia 2004, Russell 2006, and
Bach 2004.)
The facts discussed here are important for the principle of
compositionality, which says that the meaning of a complex
expression is a function of the meanings of its parts and their
mode of composition. In order to fix the meaning of here in an
utterance of (1), for instance, we had to look far beyond the item
itself. We gathered information from the context, and we drew
inferences about the speakers intentions. Compositionality thus
demands that the context itself be one of the parts involved in
the construction of complex meanings. The example is not especially remarkable; we find intricate context dependency of this
sort throughout the compositional semantic system.
Christopher Potts
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bach, Kent. 1994. Conversational impliciture. Mind and Language
9: 12462.
. 1999. The semanticspragmatics distinction: What it is and why it
matters. In The SemanticsPragmatics Interface from Different Points
of View, ed. Ken Turner, 6584. Oxford: Elsevier.
. 2004. Context ex machina. In Semantics vs. Pragmatics, ed.
Zoltn Szab, 1544. New York: Oxford University Press.
Chierchia, Gennaro. 2004. Scalar implicatures, polarity phenomena,
and the syntax/pragmatics interface. In Structures and Beyond: The
Cartography of Syntactic Structures. Vol. 3. Ed. Adriana Belletti, 39103.
New York: Oxford University Press.
Grice, H. Paul. 1975. Logic and conversation. In Syntax and Semantics.
Vol. 3: Speech Acts. Ed. Peter Cole and Jerry Morgan, 4358. New
York: Academic Press.
Horn, Laurence R. 2005. The border wars. In Where Semantics Meets
Pragmatics, ed. Klaus von Heusinger and Ken P. Turner, 2148.
Oxford: Elsevier.
Kadmon, Nirit. 2001. Formal Pragmatics. Oxford: Blackwell.
Levinson, Stephen C. 2000. Presumptive Meanings: The Theory of
Generalized Conversational Implicature. Cambridge, MA: MIT Press.
Russell, Benjamin. 2006. Against grammatical computation of scalar
implicatures. Journal of Semantics 23: 36182.
Sadock, Jerrold M. 1978. On testing for conversational implicature.
In Syntax and Semantics. Vol. 9: Pragmatics. Ed. Peter Cole, 28197.
New York: Academic Press.
SEMIOTICS
This term is remarkable in the history of intellectual life because
it was coined, almost simultaneously, by a French-speaking
Swiss linguist, Ferdinand de Saussure, and an English-speaking
American logician, Charles Sanders Peirce, around the turn of
the twentieth century. In this, the advent of semiotics as a field of
study rivals the almost simultaneous creation of the calculus by
Isaac Newton and Gottfried Leibniz in the late seventeenth century with the difference that it does not possess the precision
of Newtons and Leibnizs mathematical method but remains an
area of inquiry that is subject to great contestation. Moreover,
while Newtons and Leibnizs calculus was essentially the same,
Saussure and Peirce emphasize different aspects of semiotics. Still, they both aim at a science of meaning, and one aim of
this entry is to weave together their work and legacies. I begin,
then, with Peirces definition of semiotics as the study of signs.
759
Semiotics
Peirce noted that semiotic was the formal doctrine of signs
that is closely related to logic: a sign, Peirce writes, is something which stands to somebody for something in some respect
or capacity (193158, 2: 227, 2: 228). Moreover, signs, like logic,
exist within a system (or systems) that can be discerned and analyzed: Semiotics seeks to discover and describe the ways that
signs give rise to and communicate meaning.
760
Finally, Peirce describes the interaction among the representamen, the interpretant, and the object as semiosis (193158,
5: 484). Since the interpretant, as Peirce says, is an equivalent
or more developed sign, this model of the sign lends itself to
what Umberto Eco describes as unlimited semiosis (1976, 69).
Both Peirce and Saussure were working in the context of the
emergence of philosophical phenomenology, and it is clear from
these early remarks of Peirce that his starting point (as it was
Saussures as well) was the phenomenology the experience of
meaning in human affairs. On first view, the scientific abstractions and formal systematizations seem at odds with the tracing
of phenomenal experience, but semiotics attempts to discover
a systematic accounting for the everyday experiences of meanings in the world. Just as it is virtually impossible for a literate
person not to apprehend meanings in the alphabetic markings
on a page, it is virtually impossible for people not to experience
their lives and the world in which they live in terms of meaning.
Peirces pure sensuousness of qualities, and even the pure whatness of unequivocal objects or unequivocal attention, almost
always (perhaps always) present themselves in relation to the
future-oriented signaling of signs that will be meaningful for
someone else (including ones momentary future self).
In his posthumous Course in General Linguistics ([1916] 1959),
Saussure coins the term semiology as part of his examination of
the systematic (i.e., structural) comprehension of language
(see structuralism). It is possible, he writes, to conceive
of a science which studies the role of signs as a part of social life.
It would form part of social psychology and, hence, of general
psychology. We shall call it semiology (from the Greek semeon,
sign). He goes on to say that linguistics is only one branch of
this general science. The laws which semiology will discover will
be laws applicable in linguistics ([1916] 1983, 1516).
Already here in Saussures exposition, one can see areas of
contest for the science of semiotics: While Saussure claims that
linguistics is a subset of the more general study of the sign, others Roland Barthes, in particular claim that linguistics is the
defining case of semiotics insofar as linguistics is the most systematic organization of signs and, thus, most precisely isolates
the defining features of any sign system and creates a vocabulary
for its understanding. Moreover, in this exposition, the contest of
situating semiotics in relation to language, as Saussure does, or
in relation to logic, as Peirce does, also arises. Peirce, in contrast
to Saussure, conceived of semiotics as the science of representation of which logic was a branch, though almost coextensive
with it; the aim of semiotics for Peirce was to distinguish the possible kinds of sign functions (or processes of semiosis), particularly in mathematics and the sciences. Because of his interest in
empirical sciences he focused not only on physics and chemistry but even spent years as a surveyor and cartographer he was
far more interested in sign functions in relation to reference than
was Saussure: No version of his category of index exists within
Saussures linguistics.
Semiotics
Any scientific organization isolates the defining features
of any system of understanding and creates a vocabulary for its
comprehension (in phonology, Roman Jakobson and N. S.
Trubetzkoy called them distinctive features); in this way, systematic science seems the antithesis to phenomenology. More
specifically, the issue of the functioning of a vocabulary or (in
mathematics) a notational system, rather than simple experience, is closely tied to semiotics insofar as the sign system of a
particular science helps delineate its object(s) of study. Thus,
Bertrand Russell has argued, speaking of the periodic table, that
it has been found that, when the order derived from the periodic
law differs from that derived from the atomic weight, the order
derived from the periodic law is much more important (1923,
23; see also Schleifer 2000, 167). That is, Russell like Peirce, a
philosophical mathematician and logician asserts that the sign
system of the periodic table reveals more truth than empirical
measurement.
In A Theory of Semiotics (1976), Eco offers a related description
of semiotics in relation to lying, rather than truth. Semiotics, he
says,
is concerned with everything that can be taken as a sign. A sign
is everything which can be taken as significantly substituting for
something else. This something else does not necessarily have to
exist or to actually be somewhere at the moment in which a sign
stands in for it. Thus semiotics is in principle the discipline studying everything which can be used in order to lie. If something cannot be used to tell a lie, conversely it cannot be used to tell the
truth: it cannot be used to tell at all. (1976, 7)
With this statement, we can see why semiotics is so useful for literary studies and the humanities in general since literature (and
perhaps philosophy and history as well) describes possible
worlds. Yet semiotics aspires to be a science, which is to say,
to systematically account for phenomena in a manner that is, at
once, accurate, simple, and generalizable. Eco seeks such a complete system when he claims that a theory of the lie should be
taken as a pretty comprehensive program for a general semiotics (1976, 7).
761
Semiotics
is in this significance that the difference between linguistics and
semiotics can be discerned. In semiotics, unlike language, the
signifiers are not necessarily systematically developed but found
ready-to-hand in distinctions that can be appropriated to meaning. As Barthes has argued, in opposition to human language, in
which the phonic substance [the phoneme] is immediately significant, and only significant, most semiological systems probably involve a matter which has another function beside that of
being significant (bread is used to nourish, garments to protect)
(1967, 68). Barthes here is pursuing the argument that food and
fashion are semiotic systems; in fact, he suggests that a system of
meaning can take up most anything as its signs. And he is also
demonstrating (in his analogy between phonology and semiotics), as already mentioned, that linguistics has systematically
developed the scientific distinctions and discourse that must
be applied to semiotics (1985, xi), while Saussure imagined that
semiology would be an encompassing science lending its methods to linguistics.
In any case, Saussurean linguistics assumes that the basic elements of language and of signification more generally can be
studied only in relation to their function, rather than their cause.
Thus, the signifier implies a signified and vice versa, and such
reciprocal presupposition (see A. J. Greimas 1983) means that
the functioning of the sign, rather than any prior cause, governs
its understanding. Such a relational definition allows Saussure to
explain the problem of the identity of units of language and signification: The reason we can recognize different pronunciations
of the word stop as the same word (even when some aspirate
the /t/) is that the word is not defined by inherent qualities, but
rather it is defined in relation to other elements in a system the
structural whole of the English language.
This further suggests another governing assumption in
Saussurean linguistics and semiotics, what Saussure calls the
arbitrary nature of the sign. By this, he means that the relationship between the signifier and signified in language is never
necessary (or motivated): One could just as easily find the
sound signifier arbre as the signifier tree to unite with the concept of tree. But more than this, it means that the signified is
arbitrary as well: One could as easily define the concept tree
by its woody quality (which would exclude palm trees) as by its
size (which excludes the low woody plants we call shrubs). This
relationship is not necessary because it is not based upon inherent qualities of signifier or signified: The sign itself is governed
by systematic relationships. Moreover, these assumptions about
the nature of language also suggest why reference, the obvious
sense we all have that words refer to things in the world, is simply
excluded from Saussurean semiotics, which remains indeed in
many ways initiates a constructivist understanding of meaning. Unlike that of Peirce, in the Saussurean tradition of linguistics and semiotics which includes the linguistics of Jakobson,
the structuralism of Lvi-Strauss, even the post-structuralism
of Jacques Derrida (see deconstruction) reference remains
a problem.
762
763
764
Sentence
particulars of the determination of reference by the constituent
senses (i.e., their modes of presentation) to advance to truth will
have greater cognitive value than thoughts for which this is not
required, and any thoughts that so differ in cognitive value will
consequently enter into distinct sorts of judgments, differing in
whether they can extend our knowledge or not. Thus, since to
judge its truth we do not have to attend to the semantic connection of sense and reference, the thought expressed by Cicero
is Cicero lacks the modicum of cognitive value possessed by
Cicero is Tully.
The point is completely general; 2 + 3 = 5 and 5 = 5 also
differ in cognitive value in just this way. And so Frege now has the
result he needs: 2 + 3 = 5 and 5 = 5 have different mathematical content; the difference in their cognitive values belies that
one has significant mathematical content, while the other is no
more than an expression of the law of self-identity. Consequently,
assuming that arithmetic equalities are a species of identities is
no bar to reduction of arithmetic to logic, given the distinction,
founded in Freges understanding of content and our epistemic
access to it, of sense and reference.
Robert May
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Dummett, Michael. 1973. Frege: Philosophy of Languge.
London: Duckworth.
. [1975] 1978. Freges distinction between sense and reference.
In Truth and Other Enigmas, 11645. Cambridge: Harvard University
Press.
Frege, Gottlob. [1879] 1952. Begriffsschrift, a Formalized Language of
Pure Thought Modelled upon the Language of Arithmetic. Trans. Peter
Geach. In Translations from the Philosophical Writings of Gottlob Frege,
ed. Peter Geach and Max Black, 120. Oxford: Basil Blackwell.
. [1884] 1950. The Foundations of Arithmetic. Trans. J. L. Austin.
Oxford: Basil Blackwell.
. [1891] 1952. Function and concept. Trans. Peter Geach. In
Translations from the Philosophical Writings of Gottlob Frege, ed. Peter
Geach and Max Black, 2141. Oxford: Basil Blackwell.
. [1893] 1952. On sense and reference. Trans. Max Black. In
Translations from the Philosophical Writings of Gottlob Frege, ed. Peter
Geach and Max Black, 5678. Oxford: Basil Blackwell.
. [1893] 1964. The Basic Laws of Arithmetic. Trans. Montgomery
Furth. Berkeley and Los Angeles: University of California Press.
. [1897] 1979. Logic. Trans. Peter Long and Roger White. In
Posthumous Writings, ed. Hans Hermes, Friedrich Kambartel, and
Friedrich Kaulbach, 12651. Chicago: University of Chicago Press.
. [1918] 1977. Thoughts. Trans. P. T. Geach and R. H. Stoothof. In
Logical Investigations, ed. P. T. Geach, 130. Oxford: Basil Blackwell.
Heck, Richard, and Robert May. 2006. Freges contribution to philosophy of language. In The Oxford Handbook of Philosophy of Language,
ed. Ernest Lepore and Barry Smith, 339. Oxford: Oxford University
Press.
Lotze, Hermann. 1888. Logic. Oxford: Oxford University Press.
Russell, Bertrand. 1903. Principles of Mathematics. London: George Allen
and Unwin.
. 1905. On Denoting. Mind 14: 47993.
SENTENCE
This entry starts with analytical issues associated with the concept of sentence and then turns to more general issues.
Sentence
Traditionally, the sentence is the domain within which purely
syntactic relations, constituent structure, and grammaticality are defined. It is the one independent syntactic entity.
It expresses a predication and is often regarded as the linguistic
vehicle for the expression of a propositonal attitude and
thus the performance of a speech-act.
That much would generally be held to be true cross-linguistically, though how it is realized in syntax and morphology
differs from language to language. We focus here on English.
Traditionally, English sentences consist of one expression functioning as subject and another as predicate, the latter centered
on a verb which must be finite (i.e., inflected for tense/subject
agreement). What further expressions must figure in the predicate depends on the type (subcategory) of the verb: none with
intransitive verbs (Tom [laughed]), one with transitives (Tom
[ate the pies]), two with ditransitives (Tom [gave them the
pies]), for example. Sentences are declarative (all of those examples), interrogative (What did Tom eat?), exclamative (What a
long meeting that was!), or imperative (Eat those pies!), where
the subject is not overt because understood as referring to the
addressee.
Sentences may contain sentence-like constituents, traditional subordinate clauses. In these, the subject may be nonovert
because understood (compare Tom wants [Anna to go] with
Tom wants [to go], where Tom is the subject of to go), and the
verb need not be finite (to is the infinitive particle). Subordinate
clauses contribute to the speech-act performable at sentence
level but dont in themselves allow for the performance of a
speech-act. A subordinate interrogative clause, for example, is
not a question (Tom asked [who ate the pies]). Simple sentences
consist of one clause (sentence equals clause here). Complex
sentences consist of a main clause and any number of subordinate clauses. (See Burton-Roberts 2010, a basic introduction to
English sentence analysis, and Huddleston and Pullum 2002, a
comprehensive grammar, in which sentence is abandoned in
favor of clause.)
In what follows, we trace how sentence has fared in Chomskian
generative grammar (CGG, henceforth). Noam Chomsky
(1957) was traditional in taking S as the symbol to be defined
by any grammar the initial symbol and thus the root node in
any phrase structure tree. In defining S for a language L,
a grammar was said to generate the sentences of L and thereby
generate L. The definition consisted of successive rewrite rules,
the first being [S NP-VP], where NP (noun phrase) functions
as subject and VP (verb phrase) as predicate, though predicate
is seldom used in this context (see predicate and argument ). Chomsky (1965) adopted an alternative initial rule,
which informed subsequent developments: [S NP-AUX-VP].
Here, AUX is the locus of tense/agreement, as in S[NP[Tom]
AUX[has] VP[eaten the pies]]. Subordinate clauses were treated
as embedded sentences. To accommodate clause-introducing
complementiser expressions (e.g. that/whether/when in Anna
knows that/whether/when Tom laughed), an extension of S was
introduced: S (S-bar). This was defined/expanded by the rule
[S Comp-S], designed to capture the fact that [that/whether/
when [Tom laughed]] is in some sense sentential (clausal), while
distinguishing it from the basic clause itself (Tom laughed). In
terminology adopted later, S is a projection of S. This projection
765
Sentence
Sentence Meaning
766
SENTENCE MEANING
This expression occurs in two different contexts, in contrast with
word meaning or in contrast with speakers meaning.
767
768
769
770
Parkinsons Disease
Studies of deaf signers with Parkinsons disease provide evidence
for the importance of extrapyramidal motor systems and basal
ganglia in the mediation of signed languages (Poizner and Kegl
1992; Brentari, Poizner, and Kegl 1995). Motor deficits observed
in subjects with Parkinsons disease are not specific to language
but are evidenced across the general domain of motor behaviors.
Signers with Parkinsons disease have been described as signing
in a monotonous fashion with a severely restricted range of temporal rates and tension in signing. Accompanying the restrictions
in limb movements are deficits in the motility of facial musculature, which further reduces expressivity in signing.
Neuroimaging Studies
Neuroimaging techniques like positron emission tomography
(PET) and functional magnetic resonance imaging (fMRI) provide unique contributions to our current understanding of the
neurological processing of signs. These studies reaffirm the
importance of left hemisphere anterior and posterior brain
regions for sign language use, and emphasize that some neural
areas appear to participate in language perception and production, regardless of the modality of the language.
LEFT HEMISPHERE. Sign production tasks are especially likely
to recruit the left hemisphere. For example, when signers name
objects (Emmorey et al. 2003), generate verbs to go with nouns
(e.g., chair sit) (McGuire et al. 1997; Petitto et al. 2000; Corina
et al. 2003), or sign whole sentences (Braun et al. 2001), their left
hemispheres show significant increases in blood flow, relative
to control tasks. This heightened blood flow reflects, in part, the
Conclusion
Aphasic and neuroimaging data converge on a relatively consistent picture of sign language processing: Left hemisphere
regions, such as Brocas and Wernickes areas, are undisputedly
recruited in a similar fashion for both sign and speech production and comprehension. However, evidence further suggests
that sign comprehension may rely upon neural regions that
extend into the left parietal cortex and may engage the posterior
right hemisphere to an extent not observed for spoken language.
These differences may be related to demands of visual-spatial
processing in signing.
David P. Corina and Heather Patterson
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bavelier, Daphne, David Corina, Peter Jezzard, Vince Clark, Avi Karni,
Anil Lalwani, Josef P. Rauschecker, Allen Braun, Robert Turner, and
Helen J. Neville. 1998. Hemispheric specialization for English and
ASL: Left invarianceright variability. NeuroReport 9: 153742.
Blumstein, Shelia E. 1973. A Phonological Investigation of Aphasic Speech.
The Hague: Mouton.
Braun, A. R., A. Guillemin, L. Hosey, and M. Varga. 2001. The neural
organization of discourse: An H215O-PET study of narrative production
in English and American sign language. Brain 124: 202844.
Brentari, Diane, Howard Poizner, and Judy Kegl. 1995. Aphasic and
Parkinsonian signing: Differences in phonological disruption. Brain
and Language 48: 69105.
Capek, Cheryl M., Daphne Bavelier, David Corina, Aaron J. Newman,
Peter Jezzard, and Helen J. Neville. 2004. The cortical organization
of audio-visual sentence comprehension: An fMRI study at 4 Tesla.
Cognitive Brain Research 20: 11119.
771
772
Child-Directed Signing
Child-directed speech the language that hearing mothers use
when talking to young hearing children has certain characteristic features, including simplicity, brevity, and accuracy (Snow
1977). It is also closely tied to ongoing events and activities in
which a child and a communicative partner are engaged (Harris
1992). Researchers have asked whether child-directed signing
has analogous characteristics. In order to answer this question,
they have focused on the language used by native-signing parents to their deaf children. There are now published studies of
child-directed signing in the United States, the UK, Australia,
and Japan, and their results have led to a remarkably consistent
picture of child-directed signing (Spencer and Harris 2006).
Child-directed signing, like child-directed speech, relates to
ongoing activity and events. There are, however, some important
differences between the two that arise from the fact that signing is perceived primarily through the function of sight, whereas
speech is heard. This has important implications for the way in
signs being produced before spoken words, and they argued for
a sign advantage over spoken language. Other investigators
(Petitto 1988; Volterra, Beronesi, and Massoni 1994) have argued
that the apparent sign advantage arises because parents overinterpret motor behaviors that are produced by both deaf and
hearing infants. For example, many infants open and close their
hand or clap their hands together in the first months of life. Since
these movements are very similar to signs, it is easy to assume
that they are signs. This presents a methodological challenge for
sign language researchers that has been resolved by applying
strict criteria for identifying sign production that mirror those
used for identifying early-acquired words
A follow-up analysis of the original Orlansky and Bonvillian
data (Bonvillian, Orlansky, and Folven [1990] 1994), which
adopted just such criteria, led the authors to conclude that there
was no sign advantage. In a longitudinal study of infants from 9
to 18 months of age, Patricia Spencer and her colleagues directly
compared the age of first signs and spoken words (MeadowOrlans, Spencer, and Koester 2004). A sign was defined as a
hand movement that was not imitated, occurred in a clearly
communicative context, and included at least two of the three
primary elements of a sign (location, handshape, movement;
see sign languages ), while making allowances for the fact
that, like spoken words, childrens early signs may differ from
the adult form (see Morgan, Barrett-Jones, and Stoneham 2007).
The rate of development for signed and spoken language proved
to be remarkably similar, with the majority of both deaf and
hearing children producing single-unit linguistic communications at 18 months and only a minority beginning to combine
units.
The content of early vocabulary in signed and spoken languages is also similar. Like the early vocabulary of young hearing
children acquiring English, that of deaf children of deaf parents
acquiring American Sign Language (ASL) refers to objects and
events in the everyday environment of infants and their families,
such as familiar people, food, clothing, and toys, and is biased
toward object names (nouns) rather than action words (verbs).
As the children reach their second birthday and vocabulary
size increases, a difference does begin to emerge, with the deaf
infants beginning to show a greater proportion of verbs than the
hearing children (Anderson and Reilly 2002). In the case of spoken languages, children are more likely to acquire verbs in their
early vocabulary when these occur relatively frequently in parental speech. Both Korean-speaking (Gopnik and Choi 1995) and
Mandarin Chinesespeaking mothers (Tardif 1996) use verbs
more frequently than do English-speaking mothers, and this is
reflected in the relatively high proportion of verbs that appear
in the early vocabularies of their children. It may well be the
case that the higher proportion of verbs in early ASL vocabulary
is related to a greater use of verbs in child-directed ASL than in
child-directed English.
Margaret Harris
773
774
Sign Languages
Waxman, Robyn, and P. Spencer. 1997. What mothers do to support
infant visual attention: Sensitivities to age and hearing status. Journal
of Deaf Studies and Deaf Education 2.2: 10414.
SIGN LANGUAGES
When deaf people get together in a community, a sign language
emerges. The sign languages of linguistic investigation are these
naturally developing languages. They stand in contrast to the
invented sign systems developed by educators to teach language. The sign systems may be designed to represent, for example, English on the hands, using some vocabulary borrowed from
the natural sign language but following English grammar. The
natural sign languages such as American Sign Language (ASL)
in the United States, or British Sign Language (BSL) in the UK
have distinct grammatical systems. In this light, the fact that sign
languages do not necessarily correspond with spoken languages
(e.g., ASL and BSL are distinct, despite the common spoken language) should not be mysterious.
Some sign languages are related historically, just as spoken
languages have historical ties (see historical linguistics).
One fairly well-known family tree involves French Sign Language
(Langue des Signes Franaise, LSF), which has descendants in
much of Europe and the United States due to the fact that in the
nineteenth century, graduates of the French National Institution
for the Deaf, who all used a common sign language, were dispersed to a number of countries and helped to establish schools
there. In many cases, there was no common sign language across
groups of deaf people until the schools were formed and attracted
a community. Whatever signs were used previously combined
with LSF, and the national sign language grew out of this connection (Lane 1984).
In part because of these historical relations, a signer of ASL
may have an easier time communicating with a signer of, say,
Swedish SL, than hearing speakers of English and Swedish.
However, it is important to note that the sign languages are distinct, each having its own vocabulary and grammar. There is no
universal sign language, and couldnt be, for the same reasons
that there is no universal spoken language.
Sign Languages
HC
Figure 2.
MOTHER (A SL)
IDEA (A SL)
Figure 1.
working with a team of researchers at the Salk Institute, studied
ASL from both linguistic and psycholinguistic viewpoints (see,
e.g., Klima and Bellugi 1979). Other linguistic and psycholinguistic studies of various sign languages soon followed. Conferences
on sign language research were held starting in the late 1970s,
and in 1986, the main sign language research conference series,
Theoretical Issues in Sign Language Research (TISLR), began.
This meeting is the major international gathering of sign language
researchers, held in a different location every two to four years.
The growth of this meeting, the establishment of several journals
and many edited books devoted to sign language research, and
the formation of the Sign Language Linguistics Society are indications of the recent surge in growth in the field.
about location and handshape specified for each M and H segment. Wendy Sandler (1989) advanced the theory of sequentiality by changing the segments to movements and locations (L),
with the typical sign consisting of a sequence of LML (movement from one location to another). In the sign IDEA (see Figure
1), then, the first segment is the temple location, and the sign
moves to a few inches away from the temple. The extended pinky
handshape (HC) is used throughout the sign. On this model, the
handshape is specified in a separate hierarchically structured
unit that connects and spreads across the LML units, as shown
in Figure 2, following the principles of autosegmental phonology. This permits phenomena such as handshape assimilation in
compounding to be efficiently accounted for by the delinking of
one handshape and spreading of another.
Diane Brentaris (1998) model starts with different assumptions. On her account, the primary division in a signs characteristics is between those elements that move (the prosodic
elements) and those that do not (the inherent features). While a
sign typically moves from one location to another, it is also possible for the handshape or orientation to change, either with path
movement or without. Some such prosodic element (path movement and/or handshape/orientation change) is required for a
(monomorphemic) sign to be licit. Brentari captures this requirement for path, handshape, or orientation to change by grouping
these prosodic elements together in the representation.
The models summarized here are based largely on data from
American Sign Language and Israeli Sign Language (ISL), but
they are intended as models of signs more generally. It is clear
that different sign languages have different inventories of phonological primitives, particularly handshapes, but strikingly different patterns of organization or phonological processes have
not been identified.
One major theoretical issue that has been addressed in sign
language phonology is the syllable. The syllable is an important component of spoken language phonology as it is an organizing unit that has both universal and language-particular
aspects. There is also an intuitive component to the syllable, and
it is often taken advantage of in poetry and other areas of language use. So it might be natural to ask whether there is an analog to this important unit in sign languages.
The answer seems to be yes and no. Researchers have identified units that serve a similar function to spoken language syllables with respect to timing, generally consisting of one LML
unit. There are constraints on these units that point out the need
for positing their existence apart from the word or morpheme.
However, sign language syllables are not bound by sonority in
the same way that spoken language syllables are, and there is no
clear equivalent to a sonority cycle. That is, in spoken languages,
the nucleus of the syllable (generally the vowel) is the most sonorous element, while elements further away (in either direction)
775
Sign Languages
Figure 3.
I-ASK-HER
SHE-ASKS-HIM
776
Sign Languages
One issue that has come up in connection with this debate is
the role of the nonmanual marking that accompanies wh-questions and other structures. These nonmanual markings are often
described as grammatical, in contrast to emotional facial expressions. Some researchers consider them a part of prosody, akin to
intonation making them clearly related to syntactic structure
but also not strictly determined by syntax alone (see Sandler and
Lillo-Martin 2006 for review and references).
Aside from the generative approach, there has been a growing
body of research on sign linguistics from a cognitive grammar
perspective. This research attempts to account for the construction of meaning in language using reference to mental spaces,
taking into consideration gradience and optionality as well as systematicity. Since the use of spatial locations is integral to signing,
this approach has been very appealing to some researchers.
Sinewave Synthesis
still have parents who were themselves exposed from birth. Most
deaf children are born to hearing parents who do not know sign
language or the deaf community. The parents may be advised to
expose their children to sign language, but even if they chose to
do so, there may not be programs available for their children to be
exposed to fluent signers for more than a few hours per week. Deaf
signers begin to learn sign language at a wide variety of ages, and
their input providers themselves have a range of skills. To make
matters more complicated, deaf signers frequently must communicate with others who are not fluent in sign language, and they
have varying degrees of multilingualism, including knowledge of
the dominant spoken language(s) in their community.
This situation leads to a number of questions concerning
how to define a native speaker and which dialects/registers of
the language to use as a model. Disagreements among linguists
about basic facts are not uncommon, and signers frequently discuss varying judgments on particular examples at sign linguistics
conferences.
One response may be to attempt to narrowly define native
signers and contexts of data collection (e.g., having only naturalistic data or having only native signers interview consultants);
another response may be to attempt to explain the range of judgments observed in linguistic terms. In any case, sensitivity to the
difficulty of collecting reliable data is of great importance.
Diane Lillo-Martin
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Aronoff, Mark, Irit Meir, and Wendy Sandler. 2005. The paradox of sign
language morphology. Language 81.2: 30134.
Brentari, Diane. 1998. A Prosodic Model of Sign Language Phonology.
Cambridge, MA: MIT Press.
Emmorey, Karen. 2002. Language, Cognition, and the Brain: Insights from
Sign Language Research. Mahwah, NJ: Lawrence Erlbaum.
Klima, Edward S., and Ursula Bellugi. 1979. The Signs of Language.
Cambridge: Harvard University Press.
Lane, Harlan. 1984. When the Mind Hears: A History of the Deaf. New
York: Random House.
Liddell, Scott K., and Robert E. Johnson. 1986. American Sign Language
compound formation processes, lexicalization, and phonological remnants. Natural Language and Linguistic Theory 8: 445513.
Meier, Richard, Kearsy Cormier, and David Quinto-Pozos, eds. 2002.
Modality and Structure in Signed Language and Spoken Language.
Cambridge: Cambridge University Press.
Neidle, Carol, Judy Kegl, Dawn MacLaughlin, Benjamin Bahan, and
Robert G. Lee. 2000. The Syntax of American Sign Language: Functional
Categories and Hierarchical Structure. Cambridge, MA: MIT Press.
Sandler, Wendy. 1989. Phonological Representation of the Sign: Linearity
and Non-linearity in American Sign Language. Dordrecht, the
Netherlands: Foris.
Sandler, Wendy, and Diane Lillo-Martin. 2006. Sign Language and
Linguistic Universals. Cambridge: Cambridge University Press.
Stokoe, William C., Dorethy Casterline, and Carl Croneberg. 1965.
A Dictionary of American Sign Language on Linguistic Principles.
Washington, DC: Gallaudet College Press.
SINEWAVE SYNTHESIS
Sinewave synthesis is a technique for creating digital acoustic
signals by computationally combining simulated pure tones of
777
Sinewave Synthesis
Situation Semantics
778
SITUATION SEMANTICS
Situation semantics was originally developed by Jon Barwise and
John Perry in Situations and Attitudes (1983). semantics is conceived as providing a systematic theory of connections between
structures of a particular language and entities in the world the
language is used to talk about.
A situation is a limited part of reality; situations are the sorts of
things we perceive, reason about, and live in, for example, what is
going on now in this office, what has happened in America since
World War II, or the games of the Chicago Cubs in 1919.
Such situations are classified by the issues provided by a
scheme of individuation and classification: spacetime locations,
properties and (other) relations including argument roles
(which here we assume to have an order) and individuals. The
general form of an issue is as follows:
Whether, at space-time location l, the relation R holds among
individuals a1,an.
For example,
Whether, on July 4 at noon in Las Vegas, Elwood kisses Gertrude.
Situation Semantics
are constraints.
Meaning is a matter of constraints. Examples:
The fact that the stump has 100 rings means that the tree that
grew from it was 100 years old when cut.
The fact that the bell is ringing means that dinner is ready.
Barwise, Jon, and John Etchemendy. 1987. The Liar. New York: Oxford
University Press.
Barwise, Jon, and John Perry. 1983. Situations and Attitudes. Cambridge,
MA: Bradford/MIT Press. Reprint with new introduction by CSLI
Publications, Stanford, CA, in 1999.
Devlin, Keith J. 1991. Logic and Information. Cambridge: Cambridge
University Press.
Gawron, Jean Mark, and Stanley Peters. 1990. Anaphora and Quantification
in Situation Semantics. Stanford, CA: CSLI Publications.
Israel, David, and John Perry. 1990. What is information? In
Information, Language and Cognition, ed. Philip Hanson, 119.
Vancouver: University of British Columbia Press.
Ginzburg, Jonathan, and Ivan A. Sag. 2001. Interrogative Investigations.
Stanford, CA: CSLI Publications.
Perry, John. 1986. From worlds to situations. Journal of Philosophical
Logic 15: 83107.
779
780
Sociolinguistics
SOCIOLINGUISTICS
Sociolinguistics is an interdisciplinary research area concerned
with covariance between language and social factors. Since its
beginnings, it has sought to explain how variably speakers use
language and what the patterns of the variation are. It is characterized both by the diversity of its subject matters and by the
heterogeneity of its approaches to them.
Despite their diversity, the different brands of sociolinguistics subscribe to the position that a language is inherently variable, and they reject the assumption that a grammatical system
applies uniformly to all speakers and all settings. The social factors usually invoked to account for the variation include gender
(see gender and language), age (see age groups), level of
formal education, and profession. The settings are distinguished
into categories, such as family gathering, in which vernacular
(i.e., colloquial) speech is the norm; medical appointments, at
which speakers do not hold the same status and do not speak
exactly the same variety; and university lectures, at which the
control of the floor is not equally distributed between instructor
and students, and the language is typically not vernacular. Much
of this boils down to a common distinction between formal and
informal settings, correlated roughly with the opposition standard versus colloquial, casual, and nonstandard speech, among
a host of other distinctions. Another common characteristic
of all the different brands of sociolinguistics is their emphasis
on empirical observation, rather than on elicitation; research
781
Sociolinguistics
is usually focused on naturally occurring speech in everyday
situations.
In North America, a view of sociolinguistics has emerged that
tends to limit it primarily to variationist sociolinguistics, which
has focused on structural factors that bear on the speakers
choice from a range of competing variants, for instance, when the
copula is likely to be omitted in a vernacular utterance where it is
required in standard English. This tradition has led to a reclassification of approaches to the covariance of language and setting,
also known as the ethnography of communication, into linguistic anthropology. This may also have been encouraged by more
invocations of pragmatics to account for linguistic behavior. In
Europe, in contrast, the sociolinguistic landscape is far more contrasted. For example, though influenced by the variationist paradigm, James and Lesley Milroy have developed in the UK their
own brand of sociolinguistics, known as social network analysis
(SNA). They explain language change and maintenance through
the study of individuals social ties within a given network. The
latter varies according to size (dense and multiplex) and intensity (weak vs. close-knit). Contrary to the received doctrine, SNA
argues that close-knit structure functions as a conservative force
that prevents linguistic change, whereas speakers with weak ties
within a network are most exposed to external pressure and,
therefore, will favor innovations. In France, on the other hand,
sociolinguistic theory has hardly been influenced by the variationist paradigm, although William Labovs books of the early
1970s were translated into French toward the end of the same
decade. A brainchild of the sociolinguistics school is praxematics
(hardly known outside France) developed in the 1970s by Pierre
Lafont. It is primarily interested in the production of identity and
meaning, focusing not only [on] the specific meaning produced
and its implications, but also [on] the conditions and dynamics of
the process that enables a speaker to construct meaning (Lafont
2003 [1978], 86). Within this framework, analyses are speaker or
author oriented and combine techniques of discourse analysis
(see discourse analysis [human]) with interactional sociolinguistics. Indeed, they study both oral and written practice.
Another distinction is often made nowadays between quantitative sociolinguistics, another name for variation analysis, and
qualitative sociolinguistics, which focuses almost exclusively on
the correlation among form, function, and sociological factors.
A quick examination of recent textbooks on or dictionaries of
sociolinguistics also reveals that there is no general agreement
on the disciplines boundaries. Classifications of what counts as
sociolinguistics vary according to author and school of thought.
For some, creole studies, discourse analysis, conversation
analysis, and language contact, for instance, belong in
sociolinguistics, whereas for others, they are separate fields of
investigation that happen to discuss topics that overlap with the
concerns of sociolinguistics. Discipline boundaries are often ideological, serving interests that are not always strictly academic.
Part of the fuzziness in the boundaries of sociolinguistics
stems from the history of its development since the 1960s and
the diversity of disciplines from which it has evolved.
Historical Landmarks
As in the history of many disciplines, it is a convergence of several
factors, both intellectual and social that led to the development
782
Sociolinguistics
macro- versus micro-sociolinguistics, with the sociology of language listed in the first category and todays sociolinguistics in
the second. One may wonder if this distinction is really tenable
since it is based mainly on a matter of scale. Indeed, many studies overlap both approaches.
Primarily associated with the work of Fishman, the sociology
of language is concerned with the correlation between language
as an entity and social factors at a macrolevel, such as a nation
(see nationalism and language) or ethnic group (see ethnolinguistic identity). Whereas sociolinguistics looks at
language structure and use in order to get a better understanding of society, the sociology of language looks through the lens of
social organization to understand language dynamics. The latter,
therefore, applies sociological research techniques and models
to the study of language. Among the subject matters associated
(nonexclusively) with the sociology of language are language
planning, bilingualism, language maintenance and shift, and
language conflicts.
While sociologists have reproached sociolinguists for their
naive conception of society and their uncritical adoption of disputable social categories (e.g., social class and gender) to explain
language dynamics, sociolinguists have in turn deplored the lack
of linguistic training among sociologists. Both allegations happen to be well founded and are prompting scholars interested
in understanding the correlation of language and social life to
develop a better integrated approach to both disciplines so as
not to reduce language to a caricature of society and vice versa.
Variationist Sociolinguistics
The distinction between quantitative/variationist and qualitative sociolinguistics is now well established, although these
approaches have increasingly been combined in recent work,
paving the way for derived research paradigms. A prominent
concern in variationism has been language variation and
change, also the title of its main journal, focusing on the extent
to which patterns of variation are indicative of the direction
of change in a language community and on the social factors
that bring about the relevant changes. Overall, statistical methods are used to uncover the distribution patterns of linguistic variants across groups of speakers. Extensive descriptions
have shown that linguistic variation is not only conditioned by
some linguistic environments but also correlated with social
parameters such as age, occupation/profession, and gender,
notwithstanding the relevant speakers particular interactional
networks.
Variationist studies also show that social stratification operates within a population through linguistic differentiation (e.g.,
the pronunciation of [r] in fourth floor). Labovs (1972) influential study on linguistic patterns and social stratification in New
York City showed that the pronunciation of postvocalic /r/ is a
marker of the highest social group. Some linguistic features not
only reflect membership in particular social classes and/or ethnic groups but are also associated with distinct social values (see
prestige), including stigmatization. The largest body of work
produced within this research paradigm is phonological,
although significant studies were produced on syntax, chiefly
on African American English, Caribbean English creoles, and
white American nonstandard English varieties.
Variationist analyses are based on vernacular speech, characterized by Labov as a less-controlled way of speaking in informal
settings, such as conversations at home with family or friends.
Preference has been given to nonstandard language varieties,
a tradition that is at variance with the original meaning of the
French word vernaculaire as the primary communication style
of a speaker, regardless of its level of formality. The emphasis on
spontaneous vernacular speech radically contrasts with the introspective and/or elicited data on which the generative school has
relied. Elaborate fieldwork techniques (e.g., informal interviews,
telephone surveys) have been developed to increase the size of
the corpus and maximize the reliability of data, while making
sure that the production of utterances is not influenced by the
presence of the interviewer/fieldworker. The problem addressed
by this latter precaution is termed the observers paradox: How
can the researcher participate in or observe the production of
data without influencing the process itself? For instance, how
can he or she make sure that the subjects can continue to communicate naturally in their vernacular without making adjustments to the fact that they are being observed (by an outsider)?
The quantitative paradigm has also come under criticism
especially for assuming a simplistic, unrefined model of society, as well as for positing uncritically such categories as social
class and race/ethnicity, which are considered problematic in
sociology. Language is approached as a mere reflection of social
structures, leaving little space for speakers personal distinctiveness. While providing extensive and fine-grained descriptions of
language variation, the variationist paradigm has not given an
accurate sociological account of how changes occur in a community, failing especially to distinguish between the initiators
and spreaders of change. These need not be the same individuals, and the factors that introduce change in a population are
not identical with those that spread them. Researchers working
on language variation and change have also been criticized for
not discriminating between normal variation, inherent to any
population of speakers, and changes in patterns of variation due
to the introduction or loss of some variants or caused by demographic changes. Only the latter kind of variation can be correlated with change, not the former.
The quantitative paradigm employed in variationist sociolinguistics focuses primarily on the structural factors that bear on
the distribution of variants and, secondarily, on social factors
that frame the linguistically relevant phenomena. In contrast,
qualitative sociolinguistics (including interactional sociolinguistics and related approaches, such as ethnography of speaking)
focuses mainly on the ways that various social parameters regulate how speakers use language and/or what it means to them.
Statistics play no role in this research paradigm.
Interactional Sociolinguistics
Associated primarily with the work of John Gumperz since the
early 1960s, interactional sociolinguistics seeks an understanding of the interpretive processes at work in face-to-face or group
interactions. One of the questions that it addresses is how conversationalists (i.e., interlocutors) use linguistic resources and
sociolinguistic knowledge in order to produce and interpret
discourse in context. Using direct empirical observations and
naturally occurring speech as their primary data, interactional
783
Sociolinguistics
sociolinguists also draw on ethnographic and anthropological
research methods.
As argued by Gumperz (1982), the activity of speaking is inherent in that of interpreting. This implies that interpretation
and meaning are approached as collaboratively achieved by the
conversationalists in the course of their exchange. Information
is conveyed through multiple channels of verbal and nonverbal
coding, and the interlocutors resort to contextualization cues
(namely, lexical, grammatical, and prosodic features; code and
style shift; laughter, gaze, and so on; see also gesture), which
signal how a chunk of discourse should be interpreted.
Interactions are analyzed in relation to a multilayered context: a) the immediate context of production (e.g., the interactants, the setting); b) the intraconversational context as rendered
visible by the interactants using contextualization cues; and c) the
broader sociopolitical context in which the interaction is embedded. For instance, the interaction of a Middle Eastern traveler
with an American immigration officer in 2010 is undoubtedly
different from how it was before 2001. Attention to the broader
sociopolitical context is one of the salient trademarks of interactional sociolinguistics, which distinguish it from conversation
analysis.
Social relations rely heavily on verbal or written interactions,
which depend on shared linguistic and cultural knowledge.
Failure to recognize these interdependencies leads to misunderstandings or breakdowns of the relevant interactions. As amply
documented by Gumperz and his associates, misunderstandings can be socially consequential for a speaker, who may be
denied access to some rights and benefits in society. This is particularly evident in social settings involving asymmetric relationships, for instance, a defendant versus a judge, a doctor versus a
patient, a job seeker versus an employer. An asylum seeker may
be denied asylum simply because his or her story does not meet
the expected narrative schema (see story schemas, scripts,
and prototypes). Although this shared linguistic and social
knowledge is often not explicitly verbalized during the interaction, it plays a crucial role in the way speakers display the social
identities they assume and ascribe new ones to their co-interactants, for example, competent versus noncompetent speakers.
Interactional sociolinguistics provides a useful framework of
analysis that helps interpret situated interactions by connecting them to broader social dynamics. Many scholars engaged in
this field seek to unveil the mechanisms of social inequality
that are partially maintained and, in some cases, even produced
through language.
784
785
Speech-Acts
786
Rice, Mabel, K. Wexler, and S. Hershberger. 1998. Tense over time: The
longitudinal course of tense acquisition in children with specific
language impairment. Journal of Speech, Language, and Hearing
Research 41: 141231.
Tager-Flusberg, Helen. 2004. Do autism and specific language impairment represent overlapping disorders? In Developmental Language
Disorders, ed. Mabel Rice and S. Warren, 3152. Mahwah, NJ: Lawrence
Erlbaum.
Tomblin, J. Bruce, N. Records, P. Buckwalter, X. Zhang, E. Smith, and
M. OBrien. 1997. The prevalence of specific language impairment
in kindergarten children. Journal of Speech, Language, and Hearing
Research 40: 124560.
SPEECH-ACTS
Speech-acts are the basic building blocks of language activity. In
using language publicly to communicate one performs speech-acts.
To think privately, (arguably) one performs (internal) speech-acts.
The minimal moves of conversational exchange are self-standing
speech-acts performed with sentences known as ILLOCUTION
ARY acts (Austin 1962; Searle 1969). The basic illocutionary acts
are assertions (describing, stating, concluding), directives (orders,
requests, suggestions, questions), commissives (threats, promises),
expressives (apologies, thankings, congratulatings), and declarations (baptisms, marriage pronouncements). There is also what
some theorists call the locutionary or propositional act: the act of
conveying some propositional content with a sentence. An illocutionary act is typically a part of a more encompassing act, which is
the production of some effect in the speaker in virtue of the illocutionary act: convincing or threatening. This is called a perlocutionary act. In addition to these speech-acts, we can also include
subsentential speech-acts: say, acts of referring, using property
names, or predicates, if we think predicates refer.
In what follows, I consider the distinction between sense and
force, assertion, conditional illocutionary acts, locutionary acts,
the sentential speech-act of conventional implicature distinguished from converstional implicature performatives,
and the role of speech-act theory in understanding meaning.
SENSE AND FORCE Fundamental to the orthodox understanding
of speech-acts is the distinction between sense and force (Stenius
1967; Searle 1969; Searle and Vanderveken 1985). On this view,
a sentence uttered in an illocutionary act, say, an assertion,
has two components of content the proposition and its force.
Propositions are meant to be the following:
(a) The objects of propositional attitudes, as in what
subjects believe, fear, hope, and so on.
(b) The contents of sentences in embedded contexts, as in
either P or Q, where P and Q are not performed in illocutionary acts yet still have content.
(c) The common contents of mood-modified sentences. The
following sentences all have a common content, the thought
that Fred jumps: Fred jumps. Is Fred jumping? Fred, Jump!
(d) The ultimate objects of which ascriptions of truth or falsity are made.
In contrast, force is that aspect of meaning to do with the
deployment of sentences that possess propositional content.
Speech-Acts
Thus, assertoric force is one use of a sentence, say, in which we
intend to express a true proposition. Imperative force is that
where we intend that a proposition be made true by an audience. And so on. The force behind uttered sentences does not
necessarily correspond to their grammatical mood: declarative,
imperative, or interrogative.
What exactly are forces? We can focus that question in relation to assertion. Assertion is one kind of illocutionary act. For
Paul Grice (1958), illocutionary acts are essentially acts of communication: they involve communicative intentions, that
is, intentions directed toward an audience. These intentions are
reflexive in form: the speaker U aims to get an audience H to gain
a certain state r say, a belief, desire, intention by attempting to
get the audience to recognize this very intention. (See Grice 1969,
Strawson 1964, and Schiffer 1972 for examination of the structure
of such intentions.) It is not clear how the analysis should go, or
whether we should involve intentions in this way at all.
ASSERTION The task, then, for an analysis of assertions according to Grice is to determine which state r U intends H to gain
when U asserts something. Grice (1971) toys with two proposals:
Brandoms analysis, it seems, deals with speakers who are indifferent to the communicative effect of their utterance and who
may be insincere. But there are worries. One is circularity: Both
(i) and (ii) specify conditions that really make reference to the
assertion of S, and it is not entirely obvious how this reference
can be removed. Brandom seems to suggest some kind of mutual
delineation between inference and assertion, but it is not obvious
that this works. Another concern is what a speaker does in order
to obligate him- or herself. Merely uttering a sentence with a certain tone of voice does not do that. A theory of what the speaker
does might amount, in itself, to a theory of assertion.
Evidently, assertoric force has yet to be convincingly clarified.
(See Pagin 2004, 2005 for more discussion of the challenges.)
a belief that P;
a belief that U believes that P.
The answer to (ii) is that there is no sound of rain. So, does utterance of (i) comprise a question embedded in a conditional? That
cannot be. To successfully perform a question with Why cant
we hear the rain outside? one must believe that we cannot hear
the rain outside, which, in this context, requires that one believe
there is rain outside. Rather, in (i) the if-clause provides a proposition from which the presupposition of the question why cant
we hear the rain? can be derived. The interrogative sentence is
in the scope of if, but no question is performed with the (consequent) sentence.
What speech-act are we performing? The natural thought is
that we are performing a conditional question, that is, a conditional illocutionary act. One concept of conditional illocutionary
acts is that an illocutionary act, A condition on P, is an act that
amounts to an illocutionary act, A if P, and otherwise is no illocutionary act at all (see Belnap 1973). An alternative idea is that in
conditionals like (i), the speaker performs a proto-illocutionary act
in the scope of a supposition. Proto-illocutionary acts are kinds of
precursors to illocutionary acts that can embed, but are no locutionary acts in the sense described here (see Barker 2004).
787
Speech-Acts
truth-apt they can be judged true or false even though they
are not asserted. The standard view is that truth-apt sentences
are sentences encoding propositions. By propositions we mean
entities whose intrinsic natures are separate from pragmatic factors, and encoding is that semantic relation binding propositions
to sentences. The problem with this idea is that orders, arguably,
also encode propositions (see Price 1988). So why are orders not
truth-apt?
One idea is that a sentence is truth-apt just in case it encodes a
proposition, is declarative in mood, and embeddable in an unrestricted way in logical compounds. Sentences produced as orders
cannot be antecedents of conditionals or subject to negation, and
so arent truth-apt (Wright 1992). But many truth-apt sentences,
say, rhetorical questions, are truth-apt but do not embed and are
not declarative. And some sentences, such as epistemic modals
like There may be life on Mars, do not happily embed as the antecedents of conditionals or in the scope of negation.
Another idea, suggested by W. J. Alston (2000), is that declarative sentences represent, in a constituent-isomorphic way,
the structure of the proposition they encode, whereas imperatives do not. But any such theory would seem to be misguided.
Perfectly meaningful declarative sentences can fail to be truthapt because they are used to perform orders or performatives
(see later section). Interrogative sentences can be used to make
assertions and be truth-apt. This suggests that truth-aptness is
more a matter of pragmatic features than syntactic features and
the supposed internal composition of propositions. Thus, the
exact speech-act form of utterances of sentences that we call
truth-apt is far from clear.
Conventional Implicature
Conventional implicature introduces the prospect of another
kind of sentential speech-act. In asserting a declarative sentence,
two contents may be produced: what is said, and what is not said
but merely indicated. Moreover, the distinction between what is
said and what is indicated can occur with respect to the conventionally determined content of the sentence. Thus, in uttering He
is an Englishmen. He is therefore brave, the speaker says that he is
an Englishman and that he is brave, and indicates, through therefore, that his being brave follows from his being an Englishmen.
The non-said content in such sentences is conventional implicature (see Grice 1971). Expressions that introduce content of this
second kind include particles like even, but, therefore, and so on.
Implicated content makes no contribution to truth conditions.
Thus, consider the following:
(iii)
(iv)
788
Bach points out that we hesitate in our truth evaluations of sentences like (vi), in which the subsidiary proposition is judged
false:
(vi)
According to Bach, this is, rather, like our hesitation in the face
of sentences like (iii) and (iv). So, implicature-bearing sentences
are just a species of mutlipropositional sentences. That suggests
that we do not need any special speech-act for conventional
implicature.
The argument does not work (Barker 2003). Consider sentences like:
(vii) If the Dalai Lama is a monk who, unlike Mother Teresa, is
pious, then the Dalai Lama is unlike Mother Teresa in
being pious.
If the consequent of (vii) is false because Mother Teresa is pious,
we must conclude that the antecedent is false, contrary to Bachs
claim. The same kind of argument cannot run for conventional
implicatures because they do not unpack in this way in conditionals. Witness the oddness of
(viii) If even Granny got drunk, then Granny is a surprising
instance of drunkenness.
Its oddness can be traced to the fact that what is implicated in the
antecedent is transferred into said-content in the consequent.
If Bachs claims are false, we are left with the prospect that
conventional implicature is a different kind of sentential speechact, one whose nature is still not clarified (though see Barker in
press).
Performatives
One of the original topics of speech-act theory was performative utterances. These were described by J. Austin (1962) as
speech-acts in which speakers did things, rather than merely
said things. The following sentences can be used to issue
performatives:
(ix) I order you to leap.
(xi)
Speech-Acts
epistemic statuses. There are detailed felicity conditions on performatives relating to these statuses (see Austin 1962 and Searle
and Vanderveken 1985).
Performatives are puzzling to some because they comprise
sentences that describe the very activity undertaken through
utterance of the sentence. In uttering (xi), I declare the games
open and, indeed, describe what I am doing in uttering (xi). That
suggests to some that performatives are truth-evaluable (see
Searle 1989). But one has to say that they do not appear intuitively to be truth-evaluable. Declarative sentences can be used in
nonassertoric ways. For example, one can order someone onto a
boat by uttering You shall be on that boat in two minutes. I am not
performing an assertion with this sentence. My utterance is not
truth-evaluable, even if it describes a future event of your getting
on the boat. The mood of a sentence is not always a guide to the
illocutionary act I perform with the sentence (see Strawson 1964;
Davidson 1984), and mere descriptive adequacy is not sufficient
for truth.
789
Speech-Language Pathology
790
Bo, L.-J., S. Maeda, and J.-L. Heim. 2001. Neanderthal man was not morphologically handicapped for speech. Evolution of Communication
3: 4977.
Buhr, Robert D. 1980. The emergence of vowels in an infant. Journal of
Speech and Hearing Research 23: 7394.
Hewitt, Gwen, A. MacLarnon, and K. E. Jones. 2002. The functions of
laryngeal air sacs in primates: A new hypothesis. Folia Primatologica
73: 7094.
Holloway, Ralph L. 1983. Human paleontological evidence relevant to
language behavior. Human Neurobiology 2: 10514.
Laitman, Jeff, R. C. Heimbuch, and E. S. Crelin. 1979. The basicranium
of fossil hominids as an indicator of their upper respiratory systems.
American Journal of Physical Anthropology 51: 1534.
Lieberman, Daniel, R. C. McCarthy, K. M. Hiiemae, and J. B. Palmer.
2001. The ontogeny of postnatal hyoid and larynx descent in humans.
Archives of Oral Biology 46: 11728.
Lieberman, Philip. 2006. Toward an Evolutionary Biology of Language.
Cambridge, MA: Belknap.
Lieberman, Philip, and E. S. Crelin. 1971. On the speech of Neanderthal
man. Linguistic Inquiry 2: 20322.
McCarthy, Robert C., D. S. Strait, F. Yates, and P. Lieberman. A recent
origin for fully modern human speech capabilities. Proceedings of the
National Academy of Sciences USA. In press.
Negus, Victor E. 1949. The Comparative Anatomy and Physiology of the
Larynx. New York: Hafner.
Tobias, Phillip V. 1971. The Brain in Hominid Evolution. New
York: Columbia University Press.
SPEECH-LANGUAGE PATHOLOGY
Speech-language pathology, as a profession, engages in the
research, diagnosis, and treatment of communication and
swallowing disorders. Speech-language pathologists (SLPs)
assess, diagnose, and treat disorders of articulation, voice, fluency, language, and swallowing. Additionally, they have a role
in the prevention of communication and swallowing disorders,
for example, via education, early identification and intervention,
and family counseling. SLPs collaborate with other health-care
professionals, including clinical psychologists, neuropsychologists, audiologists, physical therapists, occupational therapists,
nurses, neurologists, radiologists, and nutritionists.
Communication disorders may result from impairment in
any modality spoken, written, and manual and may occur
both receptively and expressively (e.g., Owns, Metz, and Haas
2007). Among the speech and language disorders studied and
treated by SLPs are language impairment in children (ranging, for example, from mild developmental delays to language
impairment secondary to autism spectrum disorders, mental
retardation, and hearing loss); articulation and phonological
problems (such as the inability to produce certain sounds or
the systematic substitutions of sounds); voice disorders (such
as vocal fold nodules and vocal fold paralysis); language and
speech impairment (aphasia, dysarthria) resulting from neurogenic causes (for example, degenerative disease, stroke, traumatic brain injury); disorders of fluency (i.e., stuttering); and
deficits resulting from head and neck cancers (such as those following laryngectomy). Swallowing disorders typically co-occur
along with neurological deficits like stroke or degenerative diseases, such as amyotrophic lateral sclerosis (ASL) or Parkinsons
disease. They also occur with head and neck cancers and with
Speech-Language Pathology
developmental delays. Whereas hearing loss is diagnosed and
treated by audiologists, SLPs play a vital role in assessing and
treating the concomitant language, voice, and articulation problems. Cognitive (for example, mental retardation) and social
aspects of communication (for example, behavioral problems,
autism), as well as multilingualism (see bilingualism and
multilingualism) and a diverse cultural background, have
an impact on the individual and must be considered in terms of
treatment, diagnosis, and research.
Thus, the professions scope is broad and covers the entire life
span from infancy to old age. SLPs are employed in a variety of
settings, including day-care centers (adult and child), rehabilitation centers, public and private schools, community and hospital
clinics, health-care agencies, universities, research laboratories,
and private practice.
In the United States, the American Speech-LanguageHearing Association (ASHA) is the professional, scientific,
and credentialing association for more than 123,000 members
nationally and affiliates internationally. Established in 1925,
ASHA maintains close affiliations with similar organizations in
the United Kingdom, Europe, and Asia. ASHA has set standards
for licensing and certification and for accreditation of academic
programs. A masters degree from an ASHA-accredited university program is required for ASHA certification and for a license
to practice in most states.
1993; Kay, Lesser, and Coltheart 1996), areas of deficit are identified using linguistic theories of language processing, regardless
of anatomy (for example, disturbances of word meaning, disorders of sentence production). Finally, the functional approach
(e.g., Chapey 2002) emphasizes the effectiveness of communication, the importance of the setting and the interlocutors, and
the levels of intelligibility and functional language use. In the
evaluation process, SLPs consider aspects relevant to the specific population being evaluated (for example, academic skills
for school age children; functional communication for adults
with aphasia).
Once a problem is detected, SLPs need to define the goals of
intervention. If treatment is warranted, it is initiated and may
include both individual and/or group therapy. The duration of
treatment may vary from a relatively short term to a period of
months or even years, as with aphasia.
In providing speech-language treatment, SLPs strive to
improve the clients communication skills and facilitate the
interaction between speaker and listener. The clinician provides
verbal and visual stimuli (such as a question or a picture) to elicit
language production, and then prompts the desired language
behavior, modeling it if needed, and responding to the communication attempted by the client. SLPs provide clients with
continuous opportunities to communicate, to produce language
forms affected by the communication disorder, to increase the
use of desired language skills, and to decrease undesired language behaviors. Varying approaches to treatment have been
taken, targeting the aforementioned disorders. For example,
when treating articulation disorders, the clinician may adopt
the motor approach (e.g., Van Riper 1972) or the linguistic, phonological approach (e.g., Hodson 1992). Under the motor
approach, the treatment would aim to train the client to perceive
and produce the impaired sounds, whereas under the linguistic
approach, the treatment would address classes of sounds and
phonological rules, with the assumption that the impairment is
at the linguistic conceptual level.
791
Speech-Language Pathology
The second area of research, focused on the evaluation of
treatment efficacy, examines the extent to which improvement
in communication skills following intervention can be attributed to the treatment received by the individuals. Efficacy and
outcome measures have been defined as an increase in scores
on standardized tests and/or on the experimental measures
generated for the study, or as a measurable change in the clients and interlocutors evaluation of the clients communication abilities (e.g., Ross and Wertz 1999). Moreover, in addition
to assessing treatment efficacy, the goal of treatment studies is
to generalize the results from the studied sample to the larger
population. This generalizability of the results is a crucial aspect
of the treatment evaluation. Whereas research studies have provided evidence associating SLP treatments with improvement
in communication, some of the studies present methodological
concerns that may render the findings inconclusive (Thompson
2006). Variables that contribute to inconsistent results of treatment efficacy include the specific measures used to demonstrate
the change following treatment, the comparability of the control
group(s) employed, and the influence of extraneous factors (for
example, education, intelligence, motivation, and amount of
language use).
792
An alternative approach to the study of individuals with communication disorders is the single-subject experimental design
(e.g., Kearns 2000). Here, studies might include a single participant or several single cases. The participants serve as the subjects
and their own control, as phases of alternative treatment methods are compared and contrasted. Treatment can be tailored to
the individual in question, and the individuals gains under each
treatment program are evaluated. Following several experimental cases, the treatment outcomes for the individual participants
across the treatment phases are integrated and examined. This
approach has proved efficient in assessing treatment methods
by considering data from a number of individuals without compromising the results by collapsing differing individuals into
participant groups. If generalization is found in repeated administration of the treatment with the same individual and across
individuals from a relatively similar population, predictions can
be extended to the larger population.
EVIDENCE-BASED PRACTICE. In the twenty-first century, leaders
in the field of SLP and communication disorders have begun to
emphasize the importance of incorporating research evidence
into the process of making clinical decisions and treatment
choices (ASHA 2005). The goal of evidence-based practice (EBP)
is to integrate evidence from up-to-date, well-executed research
studies with clinical experience and intuitions when making
decisions about the most desirable treatment. EBP, accepted
worldwide, has been put forward with the intention of improving the intervention provided by SLPs. EBP essentially allows
the clinician to discard inefficient or unacceptable treatment
methods, even if endorsed by others, and to select appropriate
and useful intervention techniques that are well supported by
research. Whereas in earlier years experts opinions and clinical traditions had been highly valued, the new EBP approach
has brought to the forefront the need for systematic research
evidence. Therefore, clinicians who embrace this approach
evaluate and interpret the best-available research data while
considering the clients preference, environment, culture, and
values.
To maximize the efficiency of EBP, SLPs consider converging
evidence from a number of studies and from a number of cases.
They pay attention to the design of the research studies and to
their experimental control. When assessing the results of treatment studies, SLPs weigh the differing treatment approaches
employed in the study and assess the results by considering variables, such as the type of treatment provided, its length, its frequency, and the degree of feasibility and relevance of the research
evidence to the practice in question. Also useful are meta-analysis techniques, such as best-evidence analysis, which assess all
available evidence but attribute greater weight to the results of
better-designed and better-executed studies. In addition to the
quality of the research published, it is important to consider the
bias in published studies toward significant results (for example,
the finding of a significant difference between two groups is more
likely to be published than the finding of a lack of a difference),
as well as experimental, rather than observational, data. Thus,
important observations and findings might be missing from the
published literature. An increasing number of research studies
that provide data for evidence-based practice have characterized
Speech-Language Pathology
Speech Perception
SPEECH PERCEPTION
Speech perception researchers study how language users identify spoken language forms. Perceivers identify them on the basis
of acoustic and, sometimes, visual and lexical information.
Language forms enable speakers to make their linguistic
communications public. They include, among others, consonants and vowels (phones) that compose word forms and word
forms themselves.
Experimental psychologists interest in speech perception
stemmed from research that began in the 1940s at Haskins
Laboratories, where Alvin Liberman and Franklin Cooper developed a reading machine for the blind (Liberman 1996, 28). It
793
Speech Perception
was not feasible then for a device to read text aloud. Instead,
the machine scanned text and generated a unique non-speech
sound for each orthographic symbol: an acoustic alphabet.
Liberman and Cooper tried many symbol-to-sound mappings,
but the machine consistently failed to provide learnable acoustic
sequences.
One problem was rate. Unless sounds were sequenced so
slowly that the output lacked practical utility, they merged perceptually into a blur. Liberman asked: Why is speech perceivable
at faster rates?
Using tools for displaying speech visually (the sound spectrograph) and for transforming visual patterns to sound (the pattern
playback), Haskins researchers discovered that speech is not an
acoustic alphabet. Moreover, phones, apparently, lack invariant
acoustic signatures. The culprit is co-articulation: Talkers temporally overlap vocal tract (articulatory) gestures for serially nearby
phones.
This exposes how speech differs from acoustic alphabets but
does not explain why speech is easier to perceive. To discover
how speech is perceived, Haskins researchers initiated a research
program to explore listeners perceptions of simplified acoustic
speech patterns.
A provocative finding was categorical perception (Liberman
et al. 1957). An acoustic cue was varied in even steps from that
for one consonant-vowel syllable (in the first experiment, /be/,
bay) to those for others (/de/ and /ge/). Listeners heard continuum members and identified the consonants, and they heard
three syllables in succession and decided which was different
(discrimination). Although syllables changed in equal acoustic
steps along the continuum, identification responses changed
abruptly. Additionally, discrimination was near chance when
the oddball syllable was identified as sharing its consonant
with the two identical syllables (e.g., all were identified as bay).
Discrimination improved markedly when the oddball was from
a different category (e.g., bay vs. day).
For Liberman, these findings supported his motor theory
of speech perception, which he had developed on the basis of
earlier, quite different, findings. According to the theory, listeners hear articulatory gestures, not the acoustic cues they
cause, and they achieve that percept because their speech motor
system is active in perception. Categorical perception supported
the theory, because although the acoustic cue varies continuously, the specified articulatory gestures change abruptly, like
the identification responses. A lip gesture produces /b/, a tongue
tip gesture /d/, and a tongue body gesture /g/. Within-category
discrimination is difficult because a given gesture type (e.g., lip)
has to be discriminated from another token of the same type.
Categorical perception intrigued other researchers but not
because they judged the motor theory plausible. In the theory,
consonant production is strictly categorical; accordingly, so
should perception be, but within-category discrimination is typically above chance. Thus, listeners do have perceptual access to
acoustic differences, contrary to the motor theory. For example,
Patricia Kuhl (1991) and Joanne Miller (e.g., Allen and Miller
2001) showed that listeners give distinct goodness ratings to
acoustically different members of the same category.
Another troublesome outcome for motor theorists derived
from research with animals. Kuhl and James Miller (1978) found
794
Speech Perception
affect what listeners experience hearing. Researchers had already
shown that, in noise, speech is more intelligible if the listener can
see the speakers face. However, the McGurk effect shows an
influence of facial speech gestures on perception of discrepant
noise-free acoustic speech signals. For example, acoustic /ba/
dubbed onto visual /va/ is heard as /va/.
This finding has fostered countless follow-ups. In a novel
study, H. Yehia, T. Kuratate, and E. Vatikiotis-Bateson (e.g.,
2002) tracked movements on the face during speech. Their network models learned to link these articulatory motions to the
associated acoustic signal. After learning, the network generated predicted acoustic signals from facial articulations that are
highly correlated with the real acoustic patterns. Therefore, there
is substantial phonetic information in facial speech gestures.
Unknown yet is how much of it is used by perceivers.
It is known that viewers are especially sensitive to dynamic
information on the face. Lawrence Rosenblum and his colleagues
(e.g., Rosenblum and Saldana 1996) found audiovisual integration when they presented faces filmed in the dark with pointlight patches on them dubbed onto acoustic signals. Motionless
point-light faces were not identifiable as faces. In motion, they
provided phonetic information to listeners. (The acoustic realm
provides an analogue. Robert Remez and his colleagues [e.g.,
Remez et al. 1981] showed that caricatures of acoustic signals
for sentences, with formant trajectories replaced by sinewaves
at the formants center frequencies, are identifiable. Sinewave
speech lacks most traditional acoustic cues, but, like point light
faces, preserves dynamic phonetic information.)
Another source of information for speech perceivers is lexical. William F. Ganong (1980) published the seminal finding
here. His listeners heard, for example, an acoustic continuum
from giss to kiss. Listeners also heard members of a gift-to-kift
continuum. They identified more of the consonants as k in the
first continuum, that is, when the /k/ end of the continuum was
a word.
This experiment has spawned many follow-ups, including
a finding by Lawrence Brancazio (2004) that lexicality fosters
McGurk integrations. When acoustic bench is dubbed onto
video dench, the expected integration is dench, a non-word. If
the acoustic signal and video are besk and desk, respectively, the
expected integration is desk, a word. More McGurk integrations
occurred in the latter condition.
Debate has addressed the architecture of the speech perceptual system such that lexical effects on phonetic identification
occur. Most theorists assume that phones are identified, and
these identifications then support lexical access. The question
is whether there is also a top-down influence of lexical knowledge on phone perception. Given an initial phone that is ambiguous between /g/ and /k/ followed by iss, there is a word kiss in
the lexicon, but no giss. If there is top-down information flow,
then kiss can augment the evidence favoring /k/. Alternatively,
there need be no top-down flow. Rather, a decision that the consonant is g, but a poor token of it, can be revised to k when the
lexicon is later accessed. Whether there is a top-down flow of
information in speech perception remains subject to intensive
investigation (e.g., Norris, McQueen, and Cutler 2000; Samuel
and Pitt 2003).
Carol A. Fowler
795
Speech Production
Maye, J., J. F. Werker, and L. Gerken. 2002. Infant sensitivity to distributional information can affect phonetic discrimination. Cognition
82: B10111.
Miyawaki, K., W. Strange, R. Verbrugge, A. M. Liberman, and J. J. Jenkins.
1975. An effect of linguistic experience: The discrimination of [r]
and [l] by native speakers of Japanese and English. Perception and
Psychophysics 18: 33140.
Nazzi, T., J. Bertoncini, and J. Mehler. 1998. Language discrimination by
newborns: Towards an understanding of the role of rhythm. Journal
of Experimental Psychology: Human Perception and Performance
24: 75666.
Rivera-Gaxiola, M., J. Silva-Pereyra, and P. K. Kuhl. 2005. Brain potentials to native and non-native speech contrasts in 7- and 11-month-old
American infants. Developmental Science 8: 16272.
Thiessen, E. D., and J. R. Saffran. 2003. When cues collide: Use of stress
and statistical cues to word boundaries by 7- to 9-month-old infants.
Developmental Psychology 39: 70616.
. 2007. Learning how to learn: Infants discovery of stressbased strategies for word segmentation. Language Learning and
Development 3: 75102.
Vouloumanos, A., and J. F. Werker. 2004. Tuned to the signal: The
privileged status of speech for young infants. Developmental Science
7: 2706.
SPEECH PRODUCTION
Speech production is the most skillful action most of us perform.
It involves the coordinated actions of articulators of the vocal
tract, larynx, and respiratory system. Research issues focus on
the language units that are produced, the aims (acoustic or
articulatory) of making those units publicly available, and
the extent to which those aims are compromised by the tendency
that speakers have to co-articulate speech gestures (that is, to
overlap articulatory gestures temporally).
Language Forms
Language forms, that is, consonants and vowels and the larger
units they form, are the means that languages provide for making
linguistic utterances publicly available to other language users.
A major theoretical issue is to determine what they are, in two
senses. What kinds of things are they (cognitive, articulatory,
acoustic), and what are the units?
By most accounts, consonants and vowels are categories in
the minds of language users that are defined by their featural
attributes. For example, /b/ is a voiced, bilabial, stop consonant.
/i/ (as in the word keep) is a high, front, unrounded vowel. When
speakers talk, they implement those featural attributes as actions
of the vocal tract. The two lips close for /b/. This achieves the
bilabial place of articulation, and the complete closure at the lips
makes the consonant a stop. The vocal folds of the larynx open
and close to implement voicing. Except that the vocal folds are
abducted (rather than adducted), /p/ is made in the same way.
A different, albeit controversial, idea is that consonants and
vowels are not categories in the mind; they are the public actions
that talkers engage in when they speak (e.g., Goldstein and
Fowler 2003).
As for the issue of what the units are, an important source of
information has been spontaneous errors of speech production
(e.g., Garrett 1980; Shattuck-Hufnagel 1979). These are errors
produced by people who are capable of errorless speech. They
appear to reveal many of the units proposed to be units by linguists. Accordingly, errors can involve whole words (for example,
from Fromkin 1973, We have a laboratory in our own computer
when the intended utterance was We have a computer in our
own laboratory), but they also can involve individual consonants and vowels (morage in the fountains for forage in the
mountains). Rarely, they involve syllables and single features
(glear plue sky for clear blue sky; Fromkin 1973). However,
these units reveal themselves in other ways. Syllables appear to
serve as frames in speech production planning. When consonants
or vowels move in errors (as in morage in the fountains), they
preserve their position in syllables; that is, syllable-initial consonants move to syllable-initial positions, and final consonants
move to final positions. When one consonant or vowel substitutes for another, it tends to be featurally similar to the consonant or vowel it replaces.
For the most part, speech errors have been collected on the
fly by individuals with a pen and paper who transcribe errors that
they hear. The speech-errors waters have recently been muddied, however, by evidence (Mowrey and MacKay 1990; Pouplier,
2003) that, when collected in the laboratory by means that allow
articulation to be observed, speech errors look somewhat different, and far less tidy, than the errors collected by transcription. Whereas errors collected by transcription involve shifts in
position or substitutions of whole consonants, vowels, or words,
errors observed in the laboratory can be of components of segments (e.g., muscular activations for components of a consonant;
Mowrey and MacKay 1990) and of intrusions of lesser or greater
articulatory movements suggestive of a segment (Pouplier 2003),
and they do not necessarily preserve the phonotactic constraints
of the language. For example, in research by M. Pouplier (2003),
talkers repeatedly produced such word pairs as top cop, and their
errors often involved intrusion of a tongue body gesture (for the
/k/ in cop) during the tongue tip gesture for the /t/ in top. The
intrusive gesture varied in magnitude; in any case, it created a
consonant cluster (/tk/ or /kt/) that is disallowed in the language
of the participant speakers. Those findings, though important,
may not challenge the idea, however, that consonants, vowels,
and larger units are psychologically real components of planning
for speech production.
In any case, researchers now focus on methods that do not
involve collecting speech errors but may provide converging
evidence for the units that error collections have suggested (e.g.,
Levelt, Roelofs, and Meyer 1999).
797
Speech Production
similar acoustic signals. Finally, talkers sometimes compensate
for articulatory perturbations by using wholly new ways of preserving the acoustic characteristics of segments. This implies
that what talkers control is the acoustic signal. In this account,
then, acoustic-perceptual targets underlie the control of articulatory movements.
In the DIVA model of speech production, targets of speaking are normalized acoustic signals reflecting resonances of the
vocal tract (formants). The normalization transformations create
formant values that are the same for men, women, and children
despite formant differences in natural speech that reflect differences in vocal tract size. Formants characterize vowels and
sonorant consonants but not, for example, stop or fricative consonants; therefore, the model is restricted to an explanation of
just those classes of phones.
To learn a mapping from articulation to acoustics, the young
DIVA model, like young infants, babbles that is, produces consonant-vowel-like sequences (see babbling).Through learning,
the articulation-to-acoustics mapping is inverted so that acoustic-perceptual targets can underlie the control of articulatory
movements.
An alternative theoretical account is that articulatory gestures are the primary targets of speech production. This kind
of account is supported, for example, by a finding that talkers
compensate for articulatory perturbations that have no audible
acoustic consequences (Tremblay, Shiller, and Ostry 2003). In
that research, compensation also occurred in a silent speech
condition, which, of course, had no acoustic consequences at all.
These results appear inconsistent with a hypothesis that speech
targets are acoustic.
There is also a more natural speech example of preservation
of inaudible articulations. In an investigation of an X-ray microbeam database, C. Browman and L. Goldstein (1991) found
examples of utterances such as perfect memory in which transcription suggested deletion of the final /t/ of perfect. However,
examination of the tongue tip gesture for the /t/ revealed its
presence. Due to overlap from the bilabial gesture of /m/, however, acoustic consequences of the /t/ constriction gesture were
absent or inaudible.
In the task-dynamic model of speech production (e.g.,
Saltzman and Kelso 1987; Saltzman, Lofqvist, and Mitra 2000;
Goldstein and Fowler 2003), the aims of speakers are to establish
transient dynamical systems in the vocal tract that achieve the
constriction locations (e.g., at the lips for /b/) and degrees (e.g.,
complete closure for /b/) of intended consonants and vowels.
798
of neighboring vowels that co-articulate with it, the lips invariably close when /b/ is produced. Thus, for theoretical accounts
in which speakers goals are articulatory, it appears that gestures
for consonants and vowels are realized by synergies or coordinative structures (e.g., Easton 1972) that achieve the gestures in
flexible, equifinal ways. Accordingly, when the jaw is perturbed
(prevented from raising) during production of /b/, within 2030
ms of the perturbation extra activation of muscles of the lips occur
so that the lips can do some of the work of lip closure that the jaw
otherwise would have performed (Kelso et al. 1984).
Co-articulation is limited in another way that prevents it from
destroying the achievement of essential gestural concomitants
of consonants and vowels. Different consonants and vowels
exhibit different degrees of co-articulation resistance (Bladon
and Al-Bamerni 1976; Fowler 2005; Recasens 1984, 1985) that
is, resistance to co-articulatory overlap by neighbors. The degree
of resistance appears to reflect the extent to which the neighbors might prevent the segments articulatory goals from being
achieved. For example, D. Recasens (1984) found that resistance
by consonants to co-articulatory overlap by vowels was correlated with the extent to which the consonantal constriction
required action of the tongue body, the primary articulator of the
neighboring vowels.
This research on co-articulation suggests that, viewed at an
appropriately coarse-grained level of description, gestures for
consonants and vowels are invariantly achieved despite coarticulation. Speech production is an elegant interweaving of
gestures for successive consonants and vowels that leaves their
essential properties intact.
Carol A. Fowler
WORKS CITED AND SUGGESTIONS FOR FURTHER READING
Bladon, R. A. W., and A. Al-Bamerni. 1976. Coarticulation resistance in
English /l/. Journal of Phonetics 4: 13750.
Browman, C., and L. Goldstein. 1986. Towards an articulatory phonology. Phonology Yearbook 3: 21952.
. 1991. Gestural structures: Distinctiveness, phonological processes, and historical change. In Modularity and the Motor Theory
of Speech Perception, ed. I. G. Mattingly and M. Studdert-Kennedy,
31338. Hillsdale, NJ: Lawrence Erlbaum.
Easton, T. 1972. On the normal use of reflexes. American Scientist
60: 5919.
Fowler, C. A. 2005. Parsing coarticulated speech in perception: Effects of
coarticulation resistance. Journal of Phonetics 33: 195213.
Fromkin, V. 1973. Speech Errors as Linguistic Evidence. The
Hague: Mouton.
Garrett, M. 1980. Levels of processing in speech production. In
Language Production. Vol 1: Speech and Talk. Ed. B. Butterworth, 177
220. London: Academic Press.
Goldstein, L., and C. A. Fowler. 2003. Articulatory phonology: A phonology for public language use. In Phonetics and Phonology in Language
Comprehension and Production: Differences and Similarities, ed. N. O.
Schiller and A. Meyer, 159207. Berlin: Mouton de Gruyter. This chapter summarizes the controversial idea that language forms are public
actions of the vocal tract.
Guenther, F., M. Hampson, and D. Johnson. 1998. A theoretical investigation of reference frames for the planning of speech. Psychological
Review 105: 61133. This paper defends the idea that intended language forms are acoustic.
Spelling
Hockett, C. 1955. A Manual of Phonetics. Bloomington: Indiana University
Press.
Kelso, J. A. S., B. Tuller, E. Vatikiotis-Bateson, and C. A. Fowler. 1984.
Functionally-specific articulatory cooperation following jaw perturbation during speech: Evidence for coordinative structures. Journal
of Experimental Psychology: Human Perception and Performance
10: 81232.
Levelt, W., A. Roelofs, and A. Meyer. 1999. A theory of lexical access in
speech production. Behavioral and Brain Sciences 22: 175.
Mowrey, R., and I. MacKay. 1990. Phonological primitives: Electromyographic speech error evidence. Journal of the
Acoustical Society of America 88: 12991312.
Ohala, J. 1981. The listener as a source of sound change. In Papers from
the Parasession on Language and Behavior, ed. C. Masek, R. Hendrick,
R. Miller, and M. Miller, 178203. Chicago: Chicago Linguistics
Society.
Pouplier, M. 2003. Units of Phonological Encoding: Empirical Evidence.
Ph.D. diss., Yale University.
Recasens, D. 1984. V-to-C coarticulation in C