Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
1 Gadet Francoise Pecheux Michel Lingua Inatingivel Discurso Historia Linguistica Campinas Pontes
Caricato da
ensinodelingamaternaCopyright
Formati disponibili
Condividi questo documento
Condividi o incorpora il documento
Hai trovato utile questo documento?
Questo contenuto è inappropriato?
Segnala questo documentoCopyright:
Formati disponibili
1 Gadet Francoise Pecheux Michel Lingua Inatingivel Discurso Historia Linguistica Campinas Pontes
Caricato da
ensinodelingamaternaCopyright:
Formati disponibili
The Morphosyntax of Transitions
OXFORD STUDIES IN THEORETICAL LINGUISTICS
GENERAL EDITORS
David Adger and Hagit Borer, Queen Mary, University of London
ADVISORY EDITORS
Stephen Anderson, Yale University; Daniel Büring, University of California, Los Angeles;
Nomi Erteschik-Shir, Ben-Gurion University; Donka Farkas, University of California, Santa
Cruz; Angelika Kratzer, University of Massachusetts, Amherst; Andrew Nevins, University
College London; Christopher Potts, Stanford University; Barry Schein, University of Southern
California; Peter Svenonius, University of Troms ; Moira Yip, University College London
RECENT TITLES
Diagnosing Syntax
edited by Lisa Lai-Shen Cheng and Norbert Corver
Pseudogapping and Ellipsis
by Kirsten Gengel
Syntax and its Limits
edited by Raffaella Folli, Christina Sevdali, and Robert Truswell
Phrase Structure and Argument Structure
A Case Study of the Syntax-Semantics Interface
by Terje Lohndal
Edges in Syntax
Scrambling and Cyclic Linearization
by Heejeong Ko
The Syntax of Roots and the Roots of Syntax
edited by Artemis Alexiadou, Hagit Borer, and Florian Schäfer
Causation in Grammatical Structures
edited by Bridget Copley and Fabienne Martin
Continuations and Natural Language
by Chris Barker and Chung-chieh Shan
The Semantics of Evaluativity
by Jessica Rett
External Arguments in Transitivity Alternations
by Artemis Alexiadou, Elena Anagnostopoulou, and Florian Schäfer
Control and Restructuring
by Thomas Grano
The Interaction of Focus, Givenness, and Prosody
A Study of Italian Clause Structure
by Vieri Samek-Lodovici
The Morphosyntax of Gender
by Ruth Kramer
The Morphosyntax of Imperatives
by Daniela Isac
Sentence and Discourse
edited by Jacqueline Guéron
Optimality-Theoretic Syntax, Semantics, and Pragmatics
From Uni- to Bidirectional Optimization
edited by Géraldine Legendre, Michael T. Putnam, Henriëtte de Swart, and Erin Zaroukian
The Morphosyntax of Transitions
A Case Study in Latin and Other Languages
by Víctor Acedo-Matellán
For a complete list of titles published and in preparation for the series, see pp .
The Morphosyntax
of Transitions
A Case Study in Latin and Other Languages
VÍCTOR ACEDO-MATELLÁN
1
3
Great Clarendon Street, Oxford, OX DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Víctor Acedo-Matellán
The moral rights of the author have been asserted
First Edition published in
Impression:
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
Madison Avenue, New York, NY , United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number:
ISBN –––– (Hbk)
–––– (Pbk)
Printed in Great Britain by
Clays Ltd, St Ives plc
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
Contents
General preface ix
Acknowledgements x
List of abbreviations xi
Introduction
. Aim and proposal
. Methodology
.. The advantages of a theoretical approach to the
grammar of unspoken languages
.. Data and corpus
. Structure
A neo-constructionist perspective on argument structure
. Endo-skeletal versus exo-skeletal approaches to the lexicon-
syntax interface
. Three neo-constructionist theories and a predecessor
.. Hale and Keyser’s theory of lexical syntax
.. Mateu ()
.. Borer (b)
.. Distributed Morphology
. The present framework
.. Argument structure is syntax
... No l-/s-syntax distinction
... Relational and non-relational elements
... Argument structure configurations
... Adjunction of roots to functional heads
.. The semantics of argument structure: a localist-aspectual
approach
... Structural and encyclopaedic semantics
... Interpretation of functional heads and arguments
... Against root ontologies
... Aspect and argument structure
... Locality domains for special meaning
. Summary
vi Contents
The syntax-morphology interface
. Words and structure
. Late insertion and the nature of roots
. Cyclic Spell-Out
. Operations at PF
.. Morphological Merger
.. Linearization and Vocabulary Insertion. Exponent-
defectiveness and PF crash
.. Interaction between Raising and Vocabulary Insertion.
Cross-linguistic variation
. Summary
Latin as a satellite-framed language
. Talmy’s () theory of change events and its adaptation to the
present framework
.. Talmy’s theory of (motion) events
.. Beyond events of motion
.. An asymmetric difference
.. Non-dynamic events and the s-/v-framed distinction
.. A syntactic interpretation of Talmy’s theory
... Syntactic structuring of events of change
... A morphological account of the s-/v-framed difference
. The surface shape of PathP in Latin
.. Verbal prefixes
.. Directional PPs
.. Verbal prefixes in combination with directional PPs
.. Verbal prefixes in combination with DPs
.. Directional DPs
.. APs
.. Case and directional PPs and DPs
. S-framed constructions in Latin
.. Complex Directed Motion Constructions
... CDMCs and situation aspect
... CDMCs and non-directed motion constructions in Latin
... The unaccusative nature of CDMCs
.. Figure Unselected Object Constructions
... The syntax and semantics of prefixed vs unprefixed verbs
... Conditions on the licensing of null objects: bibo ‘drink’
vs ebibo ‘drink up’
... Case alternations, situation aspect, and the merging
of roots
Contents vii
... Scopal relations between prefix and verb
.. Ground Unselected Object Constructions
... Case and situation aspect when the object is a Ground
... Transitive Ground UOCs in Latin
... Unaccusative Ground UOCs
.. The Locative Alternation
... Approaches to the LA
... The LA and the s-/v-framed distinction
... The LA and prefixation. The heterogeneity of the LA
.. Pseudoreversatives
. Summary
Weak satellite-framed languages
. The non-existence of complex adjectival resultatives in Latin and Slavic
.. Resultative constructions: initial clarifications
... Complex and simple resultative constructions
... Strong and weak resultative constructions
... Situation aspect in complex resultative constructions.
The AP as a result predicate
.. No complex AP resultatives in Latin
.. No complex AP resultatives in Slavic
.. No complex PP resultatives without a prefix. Internal
and external prefixes
. Latin and Slavic complex resultatives with an internal prefix
.. Latin complex resultative constructions
.. Slavic complex resultative constructions
.. The relation between resultativity and prefixation
... Resultativity without an internal prefix
... Atelic predicates and prefixation
... A contrast between Latin and Slavic. The role of
viewpoint aspect
. The role of morphology in the analysis of resultative constructions
.. The morphological properties of Path. The Split S-framedness
Hypothesis
.. The lack of complex AP resultatives in Latin and Slavic
.. Simple adjectival resultatives in Latin
. Summary
A revision of Talmy’s typology
. Another weak s-framed language: Ancient Greek
. Strong s-framed languages
viii Contents
.. German and Dutch
.. English
.. Icelandic
.. Finno-Ugric
. A typology of languages based on the morphology of Path.
Mandarin as a weak s-framed language
. Previous approaches and possible counterexamples
.. Snyder (, ), Beck and Snyder (a)
.. Horrocks and Stavrou (, ) and Horrocks ()
.. Kratzer ()
.. Svenonius ()
.. Son () and Son and Svenonius ()
... Korean: presence of complex adjectival resultatives,
absence of CDMCs
... Hebrew and Javanese: presence of CDMCs, absence
of complex adjectival resultatives
. Summary
Challenges and prospects
. The locus of cross-linguistic variation: Talmy or Snyder?
.. Complex Effected Object Constructions in English and Latin
.. Absence of CEOCs in v-framed languages. Theoretical
implications
. Reducing cross-linguistic variation to PF properties
Appendix: Latin telic predicates with prefixed manner-of-motion verbs
(section ..)
References
Index
General preface
The theoretical focus of this series is on the interfaces between subcomponents of the
human grammatical system and the closely related area of the interfaces between the
different subdisciplines of linguistics. The notion of ‘interface’ has become central in
grammatical theory (for instance, in Chomsky’s Minimalist Program) and in linguis-
tic practice: work on the interfaces between syntax and semantics, syntax and
morphology, phonology and phonetics, etc. has led to a deeper understanding of
particular linguistic phenomena and of the architecture of the linguistic component
of the mind/brain.
The series covers interfaces between core components of grammar, including
syntax/morphology, syntax/semantics, syntax/phonology, syntax/pragmatics, morph-
ology/phonology, phonology/phonetics, phonetics/speech processing, semantics/
pragmatics, and intonation/discourse structure, as well as questions of how the systems
of grammar involving these interface areas are acquired and deployed in use (including
language acquisition, language dysfunction, and language processing). It demonstrates,
we hope, that proper understandings of particular linguistic phenomena, languages,
language groups, or inter-language variations all require reference to interfaces.
The series is open to work by linguists of all theoretical persuasions and schools of
thought. A major requirement is that authors should write so as to be understood by
colleagues in related subfields of linguistics and by scholars in cognate disciplines.
In the current volume, Acedo-Matellán proposes a new theory of the difference
between verb-framed and satellite-framed languages. This distinction is usually taken
to be one that centres on how the semantics of lexical items is structured, but Acedo-
Matellán argues that the difference is fundamentally one of the morphological
exponence of syntactically present functional heads. He develops a theory of this,
constrained by adjacency and shows how cross-linguistic variation in argument
structure turns upon the different ways in which the systems of phonological and
phonetic form interpret the structures delivered by syntax to the interfaces. This
proposal leads to a new, more nuanced typology of the relationship between words,
broadly construed, and argument structures.
David Adger
Hagit Borer
Acknowledgements
This book is the result of a reorganization and modification of my PhD thesis
Argument structure and the syntax-morphology interface. A case study in Latin and
other languages, defended at the Universitat de Barcelona on December .
I acknowledge the support of my editor, Julia Steer, and of the series editors, David
Adger and Hagit Borer. Although many other fellow linguists deserve a heartfelt
acknowledgement for their help and support, I particularly thank Jaume Mateu and
Cristina Real-Puigdollers, who have greatly influenced the way I think about argu-
ment structure, for discussion of some important issues in this work. I am also very
grateful to two anonymous reviewers for their very useful comments. All errors are,
of course, attributable only to me.
I would also like to acknowledge the very friendly environment that I enjoy in my
current home, Queens’ College, where I have been writing the book, and also at the
University of Cambridge. Finally, I am deeply thankful for the unconditional love
and support of my friends, of my parents and my brother, and of my partner Thomas
Zannoni.
Víctor Acedo-Matellán
Cambridge, April
List of abbreviations
// first/second/third person
abl ablative
acc accusative
all allative
aor aorist
APC Absolute Participial Construction
AspQP Aspect Quantity Phrase
aux auxiliary
BTL second edition of the Bibliotheca Teubneriana Latina (Tombeur )
CDMC Complex Directed Motion Construction
COL Change of Location (alternant)
compar comparative
conj conjunction
COS Change of State (alternant)
dat dative
decl declarative
def definite
dim diminutive
DM Distributed Morphology
EA external argument
ECP Empty Category Principle
ela elative
EP Event Phrase
ess essive
ext external (prefix)
f feminine
S
FP shell functional projection
fut future
gen genitive
ger gerund
HMC Head Movement Constraint
ill illative
xii List of abbreviations
imp imperative
interr interrogative
inf infinitive
instr instrumental
int internal (prefix)
ipfv imperfective
LA Locative Alternation
loc locative
m masculine
mid middle voice
n neuter
neg negation
nom nominative
num number
opt optative
part particle
partve partitive
pass passive
pfv perfective
pl plural
pluprf pluperfect
prf perfect
prs present
pst past
ptcp participle
refl reflexive
sbjv subjunctive
s-framed satellite-framed
sg singular
SI Secondary Imperfective
superl superlative
sup supine (a nominal form of the Latin verb)
th thematic vowel
transl translative
UOC Unselected Object Construction
v-framed verb-framed
voc vocative
1
Introduction
. Aim and proposal
In this book I study the morphosyntax of events of transition cross-linguistically. My
point of departure is the morphosyntactic properties of this kind of predicate in Early
and Classical Latin. However, I compare this to other languages, mainly to Slavic,
Romance, and Germanic languages, in order to account for the cross-linguistic
differences in the expression of events of transition.
To achieve this aim I develop a theory of argument/event structure and a theory of the
syntax-morphology interface. Drawing on Mateu (), Borer (b) and Marantz
(), I argue for a view of argument/event structure where a basic distinction is drawn
between the elements carrying encyclopaedic content, the roots, and the syntactic con-
figuration built around functional heads. Argument structure properties exclusively
depend on the latter. Furthermore, the syntactic configuration provides the structural
semantics of the linguistic expression. I endorse a theory of the syntax-morphology
interface like the one proposed in the Distributed Morphology framework: morphology
is, by default, syntax, although some specific PF operations can disrupt the basic syntax-
morphology isomorphism—an isomorphism which, I argue, is inherent to the syntax-
semantics interface. Crucially, cross-linguistic variation is defended to depend exclusively
on that lack of isomorphism between syntax and morphology. In particular, it is
ultimately triggered by language-specific morphological properties of functional heads.
On the empirical side, I consider Talmy’s () theory of the cross-linguistic
expression of events of change, where a basic divide is drawn between the languages
in which the transition can be encoded by a non-verbal element—satellite-framed
languages—and the languages in which the transition must be encoded by the
verb—verb-framed languages. I couch Talmy’s theory of transition events within a
syntactic theory of argument structure, and I explore a wide range of constructions in
Latin—either presenting new data or giving a new perspective on data from the Latin
linguistics tradition—to show that Latin pertains to the class of satellite-framed
languages. I propose that the s-/v-framed distinction is explainable in morphological
terms. In particular, I make use of the theory of PF developed by Embick and Noyer
(, ) and Embick () within the Distributed Morphology framework.
The Morphosyntax of Transitions. First edition. Víctor Acedo-Matellán.
© Víctor Acedo-Matellán . First published by Oxford University Press.
Introduction
The distinction lies in the way that the head encoding transition within the vP—
Path—receives exponence during Vocabulary Insertion. In the case of v-framed
languages, Path can only receive an exponent (∅) when appearing as strictly left-
adjacent to the v head, as a prefix. This rules out any intervening material between
Path and v, accounting for the effect that these two heads are always lexicalized as a
portmanteau morph in these languages. On the other hand, in s-framed languages
this strict adjacency requirement does not exist for Path, so v and Path are free to be
phonologically realized independently from each other. Finally, I propose a refine-
ment of Talmy’s typology within the class of s-framed languages. First, there are
strong s-framed languages, like the Germanic languages, where v and Path are not
required to form one word, and, thus, allow constructions like complex adjectival
resultative constructions. Second, there are weak s-framed languages, like Latin,
where v and Path must form one word and may disallow, hence, constructions like
adjectival resultative constructions. This distinction is accounted for in terms of a
Path-to-v (PF) Raising operation for weak s-framed languages, which creates a
complex head. A three-way, gradual typology emerges encompassing strong s-
framed languages (no Path-to-v Raising required, no Path-v adjacency required),
weak s-framed languages (Path-to-v Raising required, no Path-v adjacency required),
and v-framed languages (Path-to-v Raising required, Path-v adjacency required).
. Methodology
.. The advantages of a theoretical approach to the grammar
of unspoken languages
This is, primarily, a study on theoretical linguistics, in particular, on how to handle
cross-linguistic variation in generative grammar. It is, secondarily, a study on Latin.
Since it has become a bit of a tradition in works like the present one to justify this
seemingly unnatural marriage, I shall also say a few words about it.
Needless to say, the main problem in doing generative grammar on an unspoken
language is the lack of native speakers. In particular, we do not have access to
competence, but only to performance, since we cannot elicit grammatical judge-
ments. Beyond the use of what ancient grammarians said about their language
(cf. Varro’s De lingua latina, On the Latin language) or any non-native competence,
built after years of exposition to the texts (Pfister , Miller :), we must rely
on closed corpora. But these data are, of course, natural, not experimental, and
deny us the precious gift of negative evidence, i.e. the starred sentence. Moreover,
we cannot be a hundred per cent sure that what has survived up to our times in the
manuscripts is undoubtedly positive evidence and we can only confide in the
expertise of the philologists to provide us with reliable editions.
I would like to assuage the dramatic scenario just depicted by pointing out how
generative grammar, or any well-articulated theory, for that matter, can shed light on
Methodology
the grammar of ancient languages. Interestingly, É. Kiss () notes that there have
been two major approaches to grammatical descriptions of unspoken languages. The
traditional approach is inductive, in that it builds a description from the data
available in the closed corpus. More recently, theoretical approaches, which are
deductive in nature, formulate hypotheses couched within a general theory of gram-
mar, and validate them against the data of the corpus. While the inductive approach
has proved useful in ‘listing and interpreting the morphemes of a language’ (É. Kiss
:) and in making generalizations concerning the different levels of grammar,
such an approach is, by necessity, considerably less heuristic than a deductive
approach. Specifically, it is only when equipped with a theory that we are in a
position to look for particular constructions—since we predict that they are possible
or not—and that we can thus ask ourselves why a particular construction is not
attested in the corpus. In this way, a deductive approach compensates for the lack of
negative evidence characteristic of corpora.
This work provides a perspicuous illustration of the advantage of a deductive
approach in addressing data from unspoken languages. As an example, I will show in
Chapter that Latin does not feature complex adjectival resultative constructions, i.e.
constructions like Sue hammered the metal flat, in which flat encodes the final state
attained by the metal and hammered encodes the way in which Sue brings the metal
to that state. As far as I know, this claim about how argument structure is expressed
in this language has never before been made in the Latin linguistics tradition or
elsewhere. Importantly, although the claim is empirical and arrived at through a
thorough corpus search that I shall describe in section .., I would never have made
it were it not for the fact that, from a particular theoretical perspective presented in
Chapters and , complex adjectival resultative constructions are expected to be
allowed in languages like Latin (s-framed languages). The theory leads us to the data.
In turn, the empirical finding in Latin leads me to non-trivial empirical and theor-
etical questions: do other s-framed languages disallow these constructions? Is
Talmy’s (, ) typology to be refined? Can I accomplish the refinement
through the theoretical tools that I assume?
.. Data and corpus
The Latin data in this study correspond to the periods of Early and Classical Latin,
spanning, respectively, from the third century BC to BC and from BC to the end
of the second century AD. Crucially, I follow Crocco Galèas and Iacobini () in
adopting a broad sense of the term Classical Latin, since this period is sometimes
taken to correspond to the first century BC, excluding the so-called Silver Latin period,
which encompasses the two first centuries of our era (see Haverling :). One of
the reasons to assume an extended version of the Classical Latin period and to add
also Early Latin to the relevant corpus has to do with what Crocco Galèas and
Introduction
Iacobini (:) call the ‘relative homogeneity in the control of the written norm’
(my translation), applicable to Classical Latin in the broad sense, and, hence,
encompassing Silver Latin. In particular, these authors point out that ‘in the literary
texts of the first two centuries of the Empire the prevalent norm is that of the Golden
Age [i.e. the Classical period in the narrow sense].’ (Crocco Galèas and Iacobini
:; my translation). A second reason is my suspicion that Late Latin (from the
third century to the sixth century AD) shows important differences as far as the
empirical domain of this work is concerned, i.e. argument structure and, secondarily,
Aktionsart. It will become clear in Chapters and that Latin makes use of verbal
prefixes in expressing argument structure changes and that there is a non-trivial
relation between prefixation and telicity. However, as Haverling (:) con-
cludes in a monumental work on the Aktionsart properties of unprefixed vs prefixed
sco-suffixed verbs, the event-structural function of the prefix is clearly lost by the end
of the second century AD—see also Barbelenet :– for an early observation in
the same vein. Taking into account these two reasons, I adopt the working hypothesis
that the periods of Early and Classical Latin (in the broad sense) constitute a
homogeneous language stage as far as the morphosyntactic expression of argument
structure and Aktionsart is concerned.
Unless otherwise stated, the data have been extracted from the CD-ROM corpus of
the second edition of the Bibliotheca Teubneriana Latina (BTL, Tombeur ), a
digitalized collection of the highly reliable Teubner’s Latin text editions. I restrict the
corpus searches to the Antiquitas subcorpus of the BTL, since this subcorpus
corresponds to the relevant period described above, from the first written texts
down to the end of the second century AD. The Antiquitas subcorpus contains a
total of , sentences. The procedure I have used to retrieve the data from the
BTL consists in searches for combinations of particular elements within the men-
tioned Antiquitas subcorpus. For instance, a search for telic instances of the prefixed
verb advolo ‘fly onto’ could involve the search of the combination of the sequence
‘advol*’, which yields all the registered forms of the verb without the inflectional
endings, and telicity-signalling expressions such as subito ‘suddenly’.
Besides the BTL, I draw on the data and descriptions thereof found in the rich
tradition of works on Latin linguistics, from the nineteenth century onwards. Of
particular importance, also, are the Latin dictionaries: Gaffiot’s () Dictionnaire
Latin-Français and Lewis and Short’s () Latin Dictionary, available online at the
Perseus Digital Library Project (Tufts University; Crane ). I have also found data
in other online corpora, although I have always ascertained that the data were also
registered in the Antiquitas subcorpus of the BTL, and, accordingly, I have always
labelled them with the reference provided in the BTL. In particular, I have made use
of the Greek and Roman Materials database at the Perseus Digital Library Project, the
LacusCurtius database (University of Chicago; Thayer ) and the Itinera Elec-
tronica database (Université catholique de Louvain; Meurant ). For some of the
Structure
texts found in these corpora there is a translation available, which I have often taken
into account; however, I always provide a translation of my own for all Latin data, if
not otherwise stated.
. Structure
In Chapter I put forward a theory of argument structure and the syntax-
morphology interface. The theory to be presented pertains to the class of so-called
neo-constructionist theories, that is, theories where argument structure properties do
not emerge from lexical items, but are properties of the syntactic configurations built
by the computational system. I discuss three such previous theories that have
inspired my own, as well as Hale and Keyser’s (, , ) programme,
the first attempt (after Generative Semantics) to provide a syntactic explanation of
lexical facts.
In Chapter I deal with the syntax-morphology interface, adopting Embick and
Noyer’s (, ) and Embick’s () theories. I further adopt the idea that some
functional nodes may fail to be interpreted at PF, if the conditions for the insertion of
their exponents are not met, yielding a crashed derivation. Since those conditions are
stated as part of idiosyncratic, language-specific properties of the Vocabulary Items
of the nodes in question, the possibility of a natural explanation for cross-linguistic
variation emerges, based on the morphological properties of functional items.
Chapter attempts to show that Latin is an s-framed language, in the sense of
Talmy (): in predicates expressing a transition, the element conveying the
transition and the verb correspond to different phonological units. First I introduce
Talmy’s (, ) theory of transition events, and his distinction between
v-framed languages (like Romance, where the transition cannot be expressed as an
element different from the verb) and s-framed languages (like Latin). I make a
syntactic interpretation of Talmy’s theory and propose that the s-/v-framed distinc-
tion is to be accounted for in morphological terms: in v-framed languages Path has to
be linearly adjacent to v for Vocabulary Insertion to proceed. This makes it impos-
sible for v to associate with an independent root, giving rise to the effect that Path and
v form a portmanteau morph. In s-framed languages, on the other hand, Path does
not have to be strictly adjacent to v, which produces the effect that they are realized
through different exponents. After this theoretical introduction, I carry out an
investigation of the expression of events of change in Latin, and I show that this
language is indeed an s-framed one. I introduce data that, as far as I know, have not
been tackled before in the Latin linguistic tradition—e.g. Ground Unselected Object
Constructions or Pseudoreversatives. All the constructions receive a uniform analysis
based on the status of Latin as an s-framed language. Finally, within a scenario that
goes beyond Latin, I propose new hypotheses on the nature of phenomena like the
Locative Alternation.
Introduction
In Chapter I compare Latin with other languages as far as the morphosyntactic
expression of transition predicates is concerned. I provide empirical evidence that
Latin, although an s-framed language, does not feature complex adjectival resultative
constructions. I point out that a similar fact has been observed for another group of
languages, namely the Slavic languages. I furthermore note that neither of these
languages seems to allow complex PP resultative constructions if the verb is not
endowed with a prefix. I hypothesize that the disallowance of complex adjectival
resultative constructions and that of unprefixed PP resultative constructions are
related, and I propose that in these languages there is an affixal relation between
the v head and the Path head that blocks the generation of the mentioned construc-
tions. I call these languages weak s-framed languages, contrasting with strong
s-framed languages, which do not require any such affixal relation between v and
Path. In particular, I propose that the affixal relation between v and Path in weak
s-framed languages is accounted for through the properties of the Vocabulary Item of
Path in these languages, requiring, in its insertion frame, that Path be left-adjacent to
the verbal complex head, though not necessarily to v itself.
In Chapter I explore the empirical predictions of the proposal made in the
previous chapter in a number of languages, inside and outside Indo-European. Finally
I examine some of the works that have addressed the relationship of AP resultatives
with other resultative constructions, and I consider possible counterexamples.
In Chapter I describe the remaining challenges within the two main endeavours
to which the work contributes: the endeavour of understanding the nature of the
variation described by Talmy () and that of reducing cross-linguistic differences
to differences in the interpretation of syntactic structures.
The book is supplemented with an appendix including the exhaustive results of a
corpus research on telic motion constructions in Latin, as laid out in section ...
2
A neo-constructionist perspective
on argument structure
In this chapter I present the view of the lexicon-syntax interface that will be defended
throughout the book. I adopt a perspective often referred to as neo-constructionist
(Levin and Rappaport Hovav : ), where the computational system of the
language faculty creates structures independently of the semantic encyclopaedic
features of lexical items, and where the compositional semantics of those expressions
is directly read off the syntactic structure. The role of lexical items in the interpret-
ation of linguistic expressions is reduced to that of contributing their encyclopaedic
content. In section . I describe the two main types of theories of the lexicon-syntax
interface: the projectionist and the constructionist theories. In section . I examine
Hale and Keyser’s () theory of argument structure as a predecessor of three neo-
constructionist frameworks: the theory of relational syntax and semantics of argu-
ment structure put forward by Mateu (), the exo-skeletal model of event structure
by Borer (, b), and the Distributed Morphology model (Halle and Marantz
, Marantz , among others). In section . I put forward a model drawing on
the three models presented in section .. The central idea on which the theory is
built is the difference between elements conveying encyclopaedic content, roots, and
elements conveying grammatical content, functional heads.
. Endo-skeletal versus exo-skeletal approaches
to the lexicon-syntax interface
If an interface is a region where two cognitive systems meet, that is, where there is a
flux of information between both, the lexicon-syntax interface is the domain of the
linguistic knowledge where both lexical and syntactic knowledge are at stake.
A theory of the lexicon-syntax interface is, then, a theory of the relationship between
the meaning of lexical items (lexical knowledge) and the syntactic environments they
appear in (syntactic knowledge). The characterization of that relationship has been
approached in basically two different ways in the linguistic tradition: either from the
point of view of the semantics of the lexical item or from the point of view of the
The Morphosyntax of Transitions. First edition. Víctor Acedo-Matellán.
© Víctor Acedo-Matellán . First published by Oxford University Press.
A neo-constructionist perspective on argument structure
syntactic environment in which it appears. The first point of view is at the origin of
theories aiming at providing a necessary and sufficient characterization of the
semantic elements involved in a given lexical item that are relevant when determin-
ing its syntactic environment. Such theories are particularly concerned with the
design of appropriate lexical semantic representations that adequately register
those semantic elements crucial in determining the lexical item’s syntactic proper-
ties.1 To put it in Rappaport Hovav and Levin’s (: ) terms, ‘on this approach,
the lexical property of a verb that is taken to determine its syntactic behaviour is its
meaning (e.g., Levin ; Levin and Rappaport Hovav ; Pinker )’. Com-
plementarily, if the theory does not endow lexical items with a formal apparatus
marking the syntactic expression of their semantic information, it will design the
algorithms necessary for deriving the lexical item’s syntactic environment from the
mentioned syntactically relevant semantic elements. And, of course, it is possible that
both a representation of grammatically relevant properties of the lexical item and a
lexicon-(morpho)syntax mapping algorithm are provided.
Conversely, there are theories of the lexicon-syntax interface that try to uncover
which syntactic structures give rise to what semantic interpretations within a given
syntactic domain, taking in that way some of the weight of the semantic interpret-
ation from the lexical item itself and carrying it over to the syntax—in other words, to
functional categories and functional structure. In such theories, there is no need for
rich lexical semantic representations accounting for the lexical item’s syntactic
behaviour, or special algorithms relating the relevant aspects of meaning to morpho-
syntactic expression. There exists, however, a requirement of accurately describing—
often after enriching—the functional architecture of a sentence so as to account for
its syntax and its compositional semantics, abstracting from the conceptual content
of the lexical items it embeds. Borer (: ) calls the theories of the former kind
endo-skeletal theories, and those of the latter, exo-skeletal theories. This is not, of
course, the traditional nomenclature. Thus, Levin and Rappaport Hovav () call
the former theories projectionist, because the structure is projected from the lexical
item, while the latter are constructionist, because the compositional semantic and
syntactic properties are part of the construction, and not of the lexical item embed-
ded within. Borer’s () terms are based on the two basic types of skeletons we find
in the animal kingdom: the endoskeleton or internal skeleton, found in vertebrates,
and the exoskeleton or external skeleton, found typically in arthropods. Similarly, in
endo-skeletal theories, the structure is considered to be built from the inside, that is,
1
In fact, it was within this kind of theory that the difference between grammatically relevant and
grammatically irrelevant semantic aspects of a lexical item was first pointed out (see Pinker ;
Rappaport Hovav and Levin ). However, in some cases some lexical semantic aspect has been
considered grammatically relevant by one researcher and grammatically irrelevant by another one. See,
for instance, Mateu’s (: ff.) refutation of Levin and Rappaport Hovav’s () semantic notion of
internal/external causation as determining the realization of arguments.
Endo-skeletal vs exo-skeletal approaches
from the lexical items embedded in the structure, as it is through the properties of
these lexical items that linguistic expressions are built. In this sense, lexical items and
their properties constitute the structure’s skeleton (an endoskeleton). On the other
hand, in exo-skeletal theories functional structure is the skeleton—an exoskeleton—
of linguistic expressions, in that it is this structure that determines the (compos-
itional) semantic and syntactic features of the sentence. In turn, lexical items are
embedded within this exoskeleton. Here I will adhere to Borer’s terminology, and
I will reserve the term constructionist for the exo-skeletal models where the syntactic
structure corresponds, almost entirely, to lexically stored constructions. The exo-
skeletal models where structure is built by the computational system, that is, where
constructions are not primitive entities, will be called neo-constructionist—generative-
constructivist in Ramchand’s (: ff.) terms. Importantly, the discussion in the
present section is based almost entirely on the contrast between endo-skeletal and
neo-constructionist approaches, in spite of the use of the term exo-skeletal in
referring to the latter.2
In order to get a taste of how these general perspectives work out the relationship
between lexical semantics and syntax, let us have a look at the way they would
approach that relationship in the following sentence:
() The elephant broke the mirror.
In considering the relationship between the meaning of break and the syntactic
properties of the sentence it appears in, an endo-skeletal approach postulates a lexical
unit (stored among many others in some kind of lexicon), break, provided with a set
of idiosyncratic formal properties: a category V, a lexical semantic representation
and, perhaps, a subcategorization frame. The lexical semantic representation could
assume a variety of formats, for instance some kind of list of the theta-roles of the
participants in the event described by break. In the case of break two theta-roles
would be listed: the Agent or breaker and the Patient or thing broken. If a subcat-
egorization frame were also provided, it would contain information about the
insertion context of break, such as þ__NP, encoding the obligatoriness of an NP
in object position when break is inserted (all verbs have an (overt) subject in English,
so there would not be a need to state that for break).3 In most endo-skeletal models,
2
Examples of endo-skeletal theories are Williams (); Kaplan and Bresnan (); Pesetsky ();
Di Sciullo and Williams (); Wunderlich (); Grimshaw (); Van Valin (); Levin and
Rappaport Hovav (); Rappaport Hovav and Levin (); and Reinhart (, ), among others.
Examples of exo-skeletal theories are Ghomeshi and Massam (); Arad (, , , );
Kratzer (); van Hout (, ); Ritter and Rosen (); Ramchand (, , ); Travis
(); McIntyre (); Åfarli (); Starke (); Lohndal (); and De Belder and van Craenen-
broeck () among others.
3
Of course, break may appear in an intransitive context where the subject is the thing being broken, as
in The mirror broke. The endo-skeletal approach would probably posit mapping mechanisms (lexical or
not) to derive one alternant from the other.
A neo-constructionist perspective on argument structure
however, some general mapping mechanisms, either lexical or syntactic, convert the
list [Agent, Patient] into syntactic knowledge—both its phrasal category and its
syntactic function:
() Mapping from thematic roles to grammatical functions
a. Agent ! An NP subject (The elephant)
b. Patient ! an NP object (the mirror)
Such an approach predicts that, like break, other verbs with the same theta-grid
would resolve the mapping in the same way; crush, for instance, would incorporate
the same solution, at least as far as the sentence in () is concerned: [crush], [‘destroy
by exerting a hard pressure’], [Agent, Patient]. This is indeed the case: The elephant
crushed the mirror.
An exo-skeletal approach conceives of the structural properties of the expression
in () as responsible for some aspects of its semantic interpretation, such as the
notion of Agent or Theme, or its aspectual properties, and of its syntax, such as the
presence of an object or of a subject. Many of these structural properties are covert, of
course. In this approach, particular attention is paid to the presence of the same unit,
break, in other very different syntactic contexts, as in (), where the intended
meaning is ‘the elephant went in violently’:
() The elephant broke (*the mirror) in.
In contradiction of the prediction of the endo-skeletal approach, there seems to be no
possible projection of the Theme argument in (), an alleged idiosyncratic property
of break. The exo-skeletal approach would interpret the structure of the sentence in
() as disabling the appearance of the object, and would try to give an account of that
disallowance in terms of the syntactic structure. Probably, in the face of the avail-
ability of break in () and () the endo-skeletal account would propose two breaks, an
object-projecting break and a second lexical item break in, which would not count a
Theme within its theta-grid (hence, not projecting it in the syntax). The problem here
would be the failure to capture the generality that other verbs that, like break,
obligatorily project an object in certain structures (The elephant broke *(the mirror))
cannot project it when appearing with some particles. This is the case of smash,
another verb that cannot drop its object: The elephant smashed *(the mirror). Smash
is obligatorily intransitive when combined with through, as in () below, in the
interpretation in which the elephant is entering somewhere after going through
some entity (the sentence accepts the direct object in the interpretation in which
the elephant does not go through the mirror):
() The elephant smashed (*the mirror) through.
However, not only does the break case extend intra-linguistically, to other verbs
within the same languages, but also cross-linguistically. Thus, the break/break in
Endo-skeletal vs exo-skeletal approaches
alternation parallels the one found in Latin between rumpo ‘break’ and prefixed in-
rumpo “in-break” ‘break in’:
() Latin
a. Elephans *(speculum) rupit.
elephant.NOM mirror.ACC broke
‘The elephant broke the mirror.’
b. Elephans (*speculum) in-rupit.
elephant.NOM mirror.ACC in-broke
‘The elephant broke (*the mirror) in.’
If, as is probably assumed within the endo-skeletal approach, there are two lexically
listed (although related, as I said before) breaks, accounting for their different
argument structure properties, the question is why a similar listing obtains in a
different lexicon, namely that of Latin.4 Conversely, the exo-skeletal approach would
develop a theory of sentential architecture apt to host a position for the object in the
case of (), without resorting to any idiosyncratic properties of break. In doing this, it
might run the risk of either creating nonexistent structure or overgenerating.
Within such a scenario, a fundamental asymmetry arises between the articulations
of these two types of theory. While in the former type, the endo-skeletal, the interface
between the lexicon and the syntax is non-trivial, in the sense that it is the semantic
properties of lexical items that derives their syntactic properties, in the exo-skeletal
type the interface is considerably reduced, if it exists at all. In this sense, Borer (b:
) points out that ‘[c]ontrary to common assumptions, there is, in actuality, no direct
interface, as such, between the conceptual system and the grammar, in that proper-
ties of concepts do not feed directly into the determination of any grammatical
properties’. In attributing all non-purely conceptual semantic aspects of linguistics
expressions to the syntactic structure, paradoxically, exo-skeletal theories turn out
not to be theories of the lexicon-syntax interface any more, as they do not envision
any such interface. They attempt to explain problems of the relationship between
lexical semantics and syntax, dividing what has traditionally been packed together
as lexical semantics into compositional semantics and conceptual semantics, and
rethinking the former as an emergence of syntactic structure. Thus, in developing the
appropriate functional architecture, which is often phonologically covert, they seek to
explain the syntactic and compositional-semantic properties of the sentence.
The crucial difference just exposed is directly related to a difference in how each
type of theory conceives of the minimal units the syntax plays with. As exemplified
4
Within a classical constructionist approach (Goldberg , ), where constructions are primitive
lexically listed units, the cross-linguistic facts are difficult to accommodate. In general, lexical marking is a
problem when cross-linguistic parallels are found, since they remain, within frameworks that resort to
lexical marking, as mere coincidences.
A neo-constructionist perspective on argument structure
above, endo-skeletal theories typically work with units that, besides incorporating the
Saussurean relationship between the phonological information and the conceptual
information, also make explicit the semantic components (theta-roles, event struc-
ture, aspectual features, etc.) that are taken to be relevant for the construction of the
syntactic environment in which the lexical item appears. These theories must also
provide some formal code determining the syntactic behaviour of the lexical item,
which is either predictable from the grammatically relevant aspects of meaning or
not. By contrast, although there might be differences among various models, in exo-
skeletal theories lexical items are typically units endowed exclusively with encyclo-
paedic content, given that grammatically relevant aspects of meaning are claimed to
emerge from structural properties of the sentence.5
. Three neo-constructionist theories and a predecessor
The aim of this section is to provide a background for the theoretical framework that
I will assume throughout this work. I critically examine the three neo-constructionist
frameworks which I will draw on most heavily: that put forward by Mateu (),
Borer’s (, b) exo-skeletal model of event structure, and, finally, the imple-
mentation of the Minimalist Program represented by Distributed Morphology (Halle
and Marantz , Marantz , etc.). I start by introducing Hale and Keyser’s
() theory of lexical syntax as the first attempt to articulate a syntactic theory of
argument structure.
.. Hale and Keyser’s theory of lexical syntax
Hale and Keyser’s theory has undergone different phases during which it has
fluctuated between two theoretical poles, the one more semantic and the other
more syntactic. The syntactic stance has always been the foregrounded one, as the
following quotes from their seminal paper and their final work show:
[T]he proper representation of predicate argument structure is itself a syntax. That is to say, as
a matter of strictly lexical representation, each lexical head projects its category to a phrasal
level and determines within that projection an unambiguous system of structural relations
holding between the head, its categorial projections, and its arguments (specifier, if present,
and complements). Hale and Keyser (: )
We use the term argument structure to refer to the syntactic configuration projected by a
lexical item. It is the system of structural relations holding between heads (nuclei) and their
5
In this vein, Goldberg (: ) notes the need ‘to distinguish the semantics of argument structure
constructions from the verbs which instantiate them, and to allow the verbs to be associated with rich
frame-semantic meanings’. In Goldberg’s () constructionist framework, in fact, the semantics of verbs
are reduced to Frame Semantics (Fillmore , ).
Three neo-constructionist theories and a predecessor
arguments within the syntactic structures projected by nuclear items. While a lexical entry
is more than this, of course, argument structure in the sense intended here is nothing more
than this. Hale and Keyser (: )
The syntactic character of their theory is based on the idea that the same principles
that operate in syntax, accounting for both grammatical and ungrammatical syntac-
tic patterns, can also explain patterns in the lexicon, such as lexical gaps, argument
structure alternations, or the syntactic behaviour of verbal classes. In particular, Hale
and Keyser () propose that argument structure types reduce to four basic
syntactic configurations defined by the projecting properties of their lexical heads:
() Hale and Keyser (: )
a. [h h cmp] (V)
b. [h spc [h h cmp]] (P)
c. [h* spc [h* h* h]] (A)
d. h (N)
The configuration in ()a is headed by a category, h, that only takes complements. In
()b the heading category takes both a specifier and a complement. In ()c, h takes
only a specifier and must thus combine with an ancillary category (h*) of the type of
()a to project it. Finally, the configuration of ()d corresponds to a category with
zero valency, not taking any arguments. In the unmarked case, the configurations
in () are realized, respectively, as V(erb), P (adposition), A(djective) and N(oun),
in English. In () there is an example of an argument structure configuration,
namely, that corresponding to the predicate clear the screen, headed by the dead-
jectival verb clear:
() Hale and Keyser (: )
V’
V VP
NP V’
N V AP
screen A
clear
The A lexical head clear projects a specifier (screen) thanks to the fact that it is taken
as complement by a V head, characterized by the selection of a complement (see ()a
and ()c). In turn, the whole VP is taken as complement by another higher V which
transitivizes the predicate. In order to account for the fact that the verb clear is
pronounced as such, Hale and Keyser propose that this verb is formed by an instance
A neo-constructionist perspective on argument structure
of head movement that they call conflation, which takes clear up into the intermedi-
ate V and finally into the highest V:
() To clear the screen.
V’
V VP
cleari V NP V’
N V AP
screen ti V A
ti
In this theory thematic roles are not primitive, but interpretations of the positions
occupied by arguments in the configurations (and see Hale and Keyser : ff.
for considerations on the semantic interpretation of their argument structure
configurations).
The following is an example of how independently postulated syntactic principles
account for patterns of lexical well-formedness: while it is possible to derive a
predicate such as clear the screen as depicted in the next example, it is impossible
to derive such predicates as *to metal flat, meaning ‘to flattened (the) metal’, or *to
spear straight, meaning ‘to straightened (the) spear’ (Hale and Keyser : ). This
is due to the fact that, in order to derive these predicates, conflation should apply
from specifier position into the upper head:
() *To metal flat.
V’
V VP
metali V NP V’
N V AP
ti A
flat
The derivation in () is precluded by a locality condition, the Empty Category
Principle, which states that empty categories must be governed and which was
meant to account for a variety of different syntactic phenomena. In particular, the
ECP would rule () out on the grounds that the trace ti is not properly governed, since
the VP counts as a barrier for government by metali (Hale and Keyser : ).
Three neo-constructionist theories and a predecessor
The ECP, a syntactic principle, can then explain a lexical fact: the non-existence, in
English, of verbs whose root designates an object submitted to a change of state and
which co-appear with an adjective expressing the resulting state. Crucially, this
explanation depends on the assumption that there is a level of representation of
the verb where its argument structure is syntactically displayed.
The scenario depicted seems to fit the characterization of a neo-constructionist
system, since argument structure properties and interpretation of arguments hang
on syntactic projections. However, two features of the theory militate against this
qualification. First, the status of the category A(djective) and, second, the l-syntax/-
s-syntax distinction. Since Mateu () accomplishes a successful theoretical
solution for the first problem, I leave the consideration thereof for section ..
and I concentrate now on the second problem. Consider first the following excerpts,
which form part of definitions of argument structure included in different works
by Hale and Keyser and which explicitly assume a lexical encoding of syntactic
properties:
() Hale and Keyser (: )
‘[A]s a matter of strictly lexical representation, each lexical head projects its
category to a phrasal level [ . . . ].’
() Hale and Keyser (: )
a. ‘[T]he syntactic configuration projected by a lexical item.’
b. ‘[T]he roster of syntactic properties listed for individual items in the lexicon
[ . . . ].’
c. ‘While a lexical entry is much more than this [ . . . ].’
() Hale and Keyser (a: )
a. ‘[ . . . ] the syntactic configuration projected by a lexical item.’
b. ‘[ . . . ] as part of their entries in the lexicon.’
c. ‘While a lexical entry is clearly more than this [ . . . ].’
() Hale and Keyser (: )
a. ‘[ . . . ] the syntactic configuration projected by a lexical item.’
b. ‘[ . . . ] the syntactic structures projected by nuclear items.’
c. ‘While a lexical entry is more than this [ . . . ].’
According to the above quotes, Hale and Keyser’s argument structure configurations,
much as being syntactic in a crucial sense, as I have argued before, are also part of a
lexical item. Thus, Hale and Keyser’s position with respect to the lexicon and its
relation to syntax can be considered innovative in that they seek to constrain the
possible range of argument structures available (and the number of theta-roles, verb
meanings, and lexical categories, for that matter) through independently established
principles of the syntax, but still heir to a classical conception of what a lexical item is
A neo-constructionist perspective on argument structure
in some respects: a projecting element.6 The co-existence of these two dimensions,
the lexical and the syntactic, is explained once one assumes the concept of l(exical)-
syntax, that is, the syntax applying at the lexical level, as opposed to s(entential)-syntax,
that applying to phrases. My purpose now is to show that l-syntax is a truly independ-
ent syntactic cycle. If this is true, to the extent that lexicon-syntax interface phenomena
are explained through an appeal to l-syntax, Hale and Keyser’s theory would depart
from a strict exo-skeletalism, where argument structure phenomena receive a plain
syntactic account.
Hale and Keyser have not commented much on the difference between l- and
s-syntax, much as the coexistence of these two seemingly independent syntaxes has
been considered a weak point of their theory by some syntacticians, as Hale and
Keyser themselves point out (: ). Besides some few references elsewhere (Hale
and Keyser : , , footnote ; Hale and Keyser : , footnote ), it is in
Hale and Keyser () that the difference between l- and s-syntax is most exten-
sively discussed. Here a contrast is made between l-syntactic representations and
d-structures (which are of course s-syntactic representations) and some operations
are proposed to derive the latter from the former. L-syntactic representations such as
the next one are configurations containing different elements:
() Hale and Keyser (: )
[V the sky [V V [A clear]]]
We can distinguish between (i) roots, such as clear, endowed with encyclopaedic and
non-defective phonological content; (ii) lexical heads, such as V; (iii) variable posi-
tions, such as the specifier position DP; and (iv) the different levels of projection of
the lexical head (here also marked as V). Roots provide phonological content to the
lexical heads by virtue of conflation. The most intuitive way of describing conflation
is in terms of movement, and in fact that is what Hale and Keyser have done most
times: in the case of (), clear raises up to the empty V head—and further up into
another empty V head in the case of the causative counterpart of the verb clear. This
movement, envisioned as an instance of head movement, crucially conforms to the
Head Movement Constraint (HMC), proposed by Travis ():
() Travis (: )
An X0 may only move into the Y0 which properly governs it.
6
See also Hale and Keyser (: ), where their research project is defined as stemming ‘from a
general program of study implied by the Projection Principle (Chomsky ) and the notion that syntax is
projected from the lexicon’. See also Hale and Keyser (b: footnote ) where they state that verbs must
be listed in the lexicon, much as their formation is syntactic. This is how they explain why not all
imaginable unergative birthing verbs are possible: The mare foaled, The shad roed, ?The kangaroo joeyed,
*The cat kittened, *The sow pigleted.
Three neo-constructionist theories and a predecessor
This is the fate of roots. What about the rest of the components of the argument
structure configuration? Hale and Keyser () propose that the argument nodes,
such as the one marked with DP in (), are variables where fully fledged phrases are
inserted at d-structure. The rest of the nodes are eliminated by some node-pruning
operation. Both the node-pruning mechanism and the fact that argumental positions
are refilled with DPs at d-structure clearly argue for the existence of some break
between l- and s-syntax. If, in addition to this, we take into account the fact, observed
by Hale and Keyser (: ), that there is no evidence that conflating elements
leave traces, in the s-syntactic sense of the term, we get quite a separate cycle of
syntactic computation.7 Besides the fact that l-syntax and s-syntax are different
because they constitute different cycles and l-syntax includes at least one
operation—conflation—that is not attested in s-syntax, Hale and Keyser resort to
an ontological difference between both based on the lexical, i.e., stored or static,
character of the former and the dynamic character of the latter, as can be gathered
from the following quote:8
The idea that the grammatical properties of a lexical item are syntactic in character, and that
they include dependencies of the type represented by the trace-antecedent relation, should not
be taken to imply that the use of a lexical item entails the actual application of movement rules
in processing or producing the sentence. Thus, the use of the verb saddle does not involve
performing a derivation, relating () and () [two representations of the verb saddle]. Rather,
the representation embodied in () and () is a static lexical representation of the relevant
grammatical properties of the verb saddle. It is, by hypothesis, present in the linguistic
knowledge of speakers of English who happen to know the verb. But it is not ‘accessed’ at
s-syntax. It is not visible there. Hale and Keyser (: )
I point out that the alleged staticness or ‘storedness’ of l-syntax is in prima facie
contradiction with its syntactic character. In particular, if knowledge of the syntactic
behaviour of a lexical item is really syntactic, then it cannot be in any case different
from the knowledge involved in the derivation of a sentence, as it should itself involve
a derivation. On the other hand, the knowledge involved in the derivation of a
sentence should be as static as Hale and Keyser claim l-syntax to be, if it is seriously
7
And observe that, before their revision of their concept of conflation in the third chapter of Hale and
Keyser (), the original sites of conflating elements could be occupied by overt material in s-syntax, as
in the account of cognate objects (like dance in She danced a silly dance). This insertion would add to the
counter-cyclicity of l-syntax with respect to s-syntax. See Haugen () for discussion.
8
Travis (: ), for instance, after accepting the Halekeyserian computational analysis of denom-
inal verbs like shelve (see section ..) states the following: ‘My conclusion will be that there is a principled
distinction which is not surprising—one [an l-syntactic process] appears to happen in the lexicon and is
therefore idiosyncratic, while the other [an s-syntactic process] arguably happens in the computational
system (i.e. syntax) and is therefore productive.’ The relevant point here is the opposition of ‘lexicon’ vs
‘computational system’.
A neo-constructionist perspective on argument structure
assumed that both l- and s-syntax pertain to the domain of competence and not to
that of performance. Similar remarks could be made of the following, later excerpt:
We will continue to use these diagrams, where convenient, with the understanding that they
are abstract informal representations of argument structure properties and not the represen-
tation of any actual point, initial, medial, or final, in the derivation of a verbal projection—they
could not be that, under the assumptions of a ‘bare phrase structure’ theory of lexical
and syntactic projection (Chomsky ) or under the assumption of ‘late insertion’.
Hale and Keyser (: )
Here they point out the assumptions of Bare Phrase Structure (Chomsky ) and
Late Insertion (Halle and Marantz ), as these refer typically to properties of
s-syntactic derivations (to be precise, Late Insertion refers to derivations in the
phonological branch of the derivation), and argument structure configurations do
not comply with them. They still oppose ‘abstract’ as a property of l-syntactic
representations and ‘actual’ as a property of s-syntactic derivations, and the same
fallacy obtains.
.. Mateu ()
Mateu () adopts Hale and Keyser’s configurational theory of argument structure
and endeavours to provide it with semantic content through his theory of relational
semantics. In particular, Mateu’s guiding principle in structuring his theory, and one
that makes him deviate from Hale and Keyser’s view in some nontrivial points, is the
following:
Meaning is a function of both (non-syntactically transparent) conceptual content and (syntac-
tically transparent) semantic construal. Mateu and Amadas (: )
This statement is the natural effect of the conceptual necessity that those aspects of
meaning that are compositional must be so in syntactic terms, while those aspects
of meaning that are not compositional cannot be stated in syntactic terms. That is,
semantic construal cannot be at the same time syntactically non-transparent, and
conceptual content cannot be at the same time syntactically transparent. In this way,
there is a strong (and natural) correlation between computation (syntax) and com-
positional meaning, on the one hand, and the non-computational elements of
linguistic expressions and non-compositional meaning, on the other. In conformity
with this statement, Mateu makes a crucial distinction between relational and non-
relational elements. Relational elements form a closed set, and constitute the articu-
lators of argument structure configurations, in that, besides being endowed with
certain highly abstract semantic content, they interrelate the building blocks of the
structure. Non-relational elements crucially do not have syntactic properties (not
even syntactic category), only conceptual ones: they cannot project a specifier or a
complement.
Three neo-constructionist theories and a predecessor
The relational heads proposed in Mateu () are basically two, although the
second comes in two varieties: one head, [r], is semantically interpreted as a non-
eventive relation, and projects both a complement and a specifier; the other is an
eventive head projecting a complement but only optionally projecting an external
argument (EA) as the specifier of some higher functional head (F). The EA-
projecting eventive head is [R], the source relation, while the one that does not
project it is [T], the transitional relation. These three heads are specified for a
value. If we leave this non-configurational property aside, the interpretation of [R],
[T] and [r] can be said to emerge purely from configuration. In particular, these
heads are to be found in the following configurations (F = functional head introdu-
cing the EA; X = a non-relational element):
() Argument structure configurations in Mateu ()
[F EA . . . F . . . [R X]]: unergative structure
[F EA . . . F . . . [R [X [r X]]]]: transitive structure
[T [X [r X]]]: unaccusative structure
Provided that each relational head is endowed with a non-configurational value,
the combinations in () to () obtain:
() Unergative predicates; based on Mateu (: )
a. John rolls (deliberately): [F John . . . F . . . [þR ROLL]]
b. John stank: [F John . . . F . . . [-R STINK]]
() Transitive predicates; based on Mateu (: )
a. John killed the horse: [F John . . . F . . . [þR [horse [þr KILL]]]]
b. John pushed the horse: [F John . . . F . . . [þR [horse [-r PUSH]]]]
c. John loved the horse: [F John . . . F . . . [-R [horse [-r LOVE]]]]
() Unaccusative predicates; based on Mateu (: )
a. John died: [þT [John [þr DIE]]]
b. The ball rolled: [þT [ball [-r ROLL]]]
c. John lived: [-T [John [-r LIVE]]]
As can be gathered from () to (), the þ value for [R] is associated with agentivity
(e.g., in John rolled deliberately vs John stank), the þ value for [T] is associated with
dynamicity (e.g., in The ball rolled vs John lived), and the þ value for [r] is related to
change and telicity (e.g., in John killed the horse or John died vs John pushed the horse
or The ball rolled). The combinations of () to () are not all the logical ones given
the number of relational heads and the number of values. As observed by Real
Puigdollers (: ), there are two surprisingly similar gaps in the paradigm of
transitives and in the paradigm of unaccusatives:
() *[F X . . . F . . . [-R [X [þr X]]]]
A neo-constructionist perspective on argument structure
() *[-T [X [þr X]]]
In semantic terms, a transitive non-agentive telic event (see ()) and an unaccusa-
tive stative telic event (see ()) do not seem to be possible. If we take into account the
fact that T and R are eventive, as opposed to r, which is non-eventive, we can collapse
() and () as the unavailability of the combination of a negatively valued eventive
head with a positively valued non-eventive head. As long as there is nothing in
Mateu’s system that prevents those combinations being formed, the question
emerges why they are not licit. In section ... I show that a more radically
configurational theory that does away with values for functional heads naturally
derives the facts in () and ().
One of the most salient advances of Mateu’s () theory with respect to Hale
and Keyser’s () is the reduction of the number of basic argument structure
configurations (see ()) based on the non-basic nature of the adjectival head
(h in ()c):
() Hale and Keyser (: )
a. [h h cmp] [realized as V in English]
b. [h spc [h h cmp]] [realized as P in English]
c. [h* spc [h* h* h]] [realized as A in English]
d. h [realized as N in English]
Mateu (: ff.) calls into question Hale and Keyser’s () distinction between
so-called locatum and location verbs like saddle and shelve, respectively, and
deadjectival verbs like clear. First he demonstrates the spurious character of the
locatum/location distinction. Hale and Keyser () argue that locatum and location
verbs differ in the nature of the abstract preposition they incorporate: locatum verbs
involve a preposition encoding a central coincidence relation, while location
verbs involve a preposition encoding a terminal coincidence relation. In a nutshell,
whereas locatum saddle may be paraphrased as ‘provide X with a saddle’, location
shelve may be paraphrased ‘as place X onto a shelf ’. Mateu argues, as does Harley
(), that this difference is not grammatically encoded, and that both types of
verbs correspond to the type [F EA . . . F . . . [þR [X [þr SADDLE/SHELVE]]]]. That
they encode a [þr] relation, inducing telicity, is argued for on the basis of the
following Catalan data:
() Catalan; Mateu (: –)
a. Ella ensellà el cavall {*durant/en} cinc segons.
she (in)saddled the horse {*for/in} five seconds
‘She saddled the horse in five seconds.’
b. En Joan encaixà cinc morts {*durant/en} dos minuts.
the Joan (in)boxed five dead (men) {*for/in} two minutes
‘Joan coffined five dead men in two minutes.’
Three neo-constructionist theories and a predecessor
Verbs incorporating an abstract preposition are, in Hale and Keyser’s () theory,
different from those incorporating an adjectival head, like clear. The difference is
related to the fact that only the latter are claimed to enter in the so-called Causative
Alternation:
() Hale and Keyser (: and )
a. The screen cleared.
b. *The book shelved.
c. *The horse saddled.
Configurationally, transitive clear has two V layers, a transitivizing one and an
unaccusative one. Thus, if the outer layer is taken off, the structure is still a verb,
and its specifier counts as the surface unaccusative subject (see ()a and ()a). The
presence of the internal V layer is due to the fact that the head A, which projects only
a specifier, needs the complement-projecting head V to project that specifier. On the
other hand, verbs involving a P projection have only one V layer, which is both the
verbalizing head and the transitivizing head (P, in projecting both a complement and
a specifier, does not need any other head to project) (see ()b, ()c and ()b):
() Hale and Keyser (: and )
a. [V the screen [V V A (= clear)]]
b. [V V [P the books/the horse [P P [N shelf/saddle]]]]
Mateu (), however, following Kiparsky (), argues that the facts in () are
due not to a grammatically encoded distinction, but to world knowledge. Thus, if the
action described by the predicate can be understood as non-agentive, locatum/
location verbs may license an unaccusative use (see (), where the helicopter is a
self-propelled object); the same applies to deadjectival verbs like clear, which may
(see ()d) or may not (see ()b) appear in unaccusative predicates on the grounds
of the same non-agentive/agentive reading:9
() Catalan; Mateu (: )
L’helicòpter aterrà tard.
the=helicopter (to)landed late
‘The helicopter landed late.’
9
In Acedo-Matellán (a) I provide more examples of uncontroversially locatum/location verbs
which, depending on the interpretation, may or may not enter into the Causative Alternation. Thus, for
instance, Cat. locatum em-perlar “in-pearl”, incorporating the prepositional prefix en- ‘in’, may be used to
mean ‘bead (a necklace)’ or ‘cover with pearl-like elements, like dew drops’. Thus, in the former use
emperlar invokes an agent-controlled scene, but not in the latter. Accordingly, emperlar may only appear as
intransitive in the latter use.
A neo-constructionist perspective on argument structure
() Levin and Rappaport Hovav (: –), in Mateu (: )
a. The waiter cleared the table.
b. *The table cleared.
c. The wind cleared the sky.
d. The sky cleared.
Once these facts have been acknowledged, there is no evidence that locatum/location
and deadjectival verbs differ grammatically. More generally, there remains no evi-
dence for a distinction between structures ()b and ()c. In particular, the h head
in ()c, which is defined as the head projecting a specifier but no complement, and
which is unmarkedly realized as A in English and many other languages, is non-basic.
Instead, it is amenable to a decomposition into an [r] relation (P in Hale and Keyser’s
terms) and a non-relational element (N in Hale and Keyser’s terms). I mention again,
lastly, what I pointed out in section ..: that this move has a welcome consequence
not sufficiently emphasized by Mateu (). Specifically, Mateu () eliminates
the undesirable situation of having an element (h in ()c) be relational and convey
conceptual content, simultaneously. In that sense, Mateu’s () theory can be
argued to fulfil the neo-constructionist desideratum of neatly separating roots
(non-relational elements) from the material able to create structure (relational
elements). See also Amritavalli and Jayaseelan (), Amritavalli (), and
Kayne () for the proposal that adjectives are to be analysed as non-primitive
categories, involving the combination of a non-relational element and an adposi-
tional element.
Finally, some lexicalist traces can be found in Mateu’s () theory that I wish to
refute. Turning back to the discussion on the telic nature of location/locatum verbs
(see () above), he points out some apparent counterexamples:
() Catalan; Mateu (: )
a. En Joan enfarinà les mandonguilles {durant/en} deu segons.
the Joan (in)floured the meatballs {for/in} ten seconds
‘Joan floured the meatballs {for/in} ten seconds.’
b. Ell engabià el seu ocell preferit {durant/en} un minut.
he (in)caged his bird favourite {for/in} one minute
‘He caged his favourite bird {for/in} one minute.’
These examples would jeopardize his proposal that both location and locatum verbs
incorporate a [þr] relation, inducing telicity. With respect to examples like ()b,
Mateu observes that the licensing of the durative adverbial is due to a measurement
of the resulting state: in this case, durant un minut ‘for a minute’ expresses the time
span spent by the bird in the cage after having been caged therein. With respect to
examples like ()a, Mateu points out that the non-relational element involved
refers to a mass entity, in this case flour (farina), and that this fact licenses an atelic
Three neo-constructionist theories and a predecessor
reading of the predicate. Thus, since the root does not refer to a bounded entity, the
action of putting that entity somewhere (the meatballs) cannot be measured out:
enfarinar “in-flour” ‘flour’ would turn out to be like ruixar ‘spray’, which can also
license an atelic reading for exactly the same reason in John sprayed the wall with
paint for five minutes (Mateu : ). Crucially, though, enfarinar cannot be said
to involve a [-r] relation—present in verbs like empènyer ‘push’—which would, on
the other hand, account for its atelic reading straightforwardly. The enfarinar/
empènyer ‘flour’/‘push’ dissociation and the enfarinar/ruixar ‘flour’/‘spray’ associ-
ation are based on diagnostics such as the following, involving licensing of adjectival
passives:
() Catalan; Mateu (: –)
a. Les mandonguilles estan enfarinades.
the meatballs PFV.be.PL (in)floured
‘The meatballs are floured.’
b. La paret està ruixada de pintura.
the wall PFV.be.PL sprayed of paint
‘The wall is sprayed with paint.’
c. *El carro està empès.
the cart PFV.be.PL pushed
According to this test, verbs like enfarinar pattern with verbs like ruixar in involving
a final state and licensing thereby the adjectival passive construction; on the other
hand, verbs like push, which do not involve a final state, disallow the adjectival
passive construction. Note, however, that the discussion is set, literally, in terms of
verbs, that is, lexical units, and in terms of what they involve as such. My claim here is
that neither does enfarinar ‘flour’ necessarily involve a [þr] head nor does empènyer
‘push’ necessarily involve a [-r] head. Accordingly, assuming Mateu’s primitives,
enfarinar ‘flour’ can be claimed to reflect either a [þR [X [þr X]]] configuration, in
which case a change of state is readily interpreted and telicity can thereby emerge, or
a [þR [X [-r X]]]] configuration, in which case no final state is entailed to be attained
and atelicity arises. I believe that what the diagnostics in () is really showing us is
that a very special context is needed for empènyer ‘push’ to be interpreted as
telic/change of state, unlike enfarinar ‘flour’ and ruixar ‘spray’. Thus, while it is
possible to conceive of a (bounded) quantity of flour or spray which would qualify as
standard in defining an end state for a flouring or spraying event, respectively, it is
considerably more difficult to evoke a standard ‘pushedness’. However, it is not
impossible, as the next example from Kratzer (), shows:
() German; Kratzer (: )
Dieser Kinderwagen ist schon geschoben.
this baby carriage is already pushed.
A neo-constructionist perspective on argument structure
In Kratzer’s (: ) words, ‘[a] natural setting for [()] would be a factory that
produces baby carriages and employs workers whose job it is to push new baby
carriages a few times to test their wheels’.
More generally, I think that diagnostic tests like the one in (), involving the
licensing of particular constructions, are not diagnostics about the membership of a
certain verb in a particular grammatically defined class: they could not be, once an
exo-skeletal perspective has been adopted, where category-free roots are freely
inserted in the structures generated by syntax, and hence, the only reason a root
does not fit into a structure is an incompatibility between the semantics emerging
from the structure and the conceptual content of the root. The adjectival passive
construction illustrated in () most probably involves some grammatical formative
like Mateu’s () [þr] relation, but enfarinar ‘flour’ or ruixar ‘spray’ or, more
specifically, the roots involved in them, do not.
.. Borer (b)
Borer (a, b, ) develops a theory of the lexicon-syntax interface char-
acterized by the idea that the conceptual system and the grammar do not interact.
Rather, the grammar yields structures where the units of conceptual content or
listemes, sound-meaning correspondences without any grammatical information
(notably, category and argument structure properties), act as mere modifiers. In
such a system, many instances of sequences commonly considered ungrammatical
are explained away as semantically aberrant, due to a clash between the interpret-
ation of the structure, which cannot be overridden, and the conceptual content of the
listemes.
Within the domain of functional categories there is a remarkably original advance
in Borer’s (a, b) theory. Borer (a: ) proposes that functional heads
are, in fact, open values, that is, variables which are in need of being assigned range by
an operator. The open values are labelled with a syntactic category and carry the
corresponding (functional) interpretation. Range assignment can be achieved basic-
ally in two ways: through direct or indirect range assignment. Direct range assignment
is accomplished when a grammatical formative is merged directly into the open
value, while indirect range assignment can be instantiated through adverbs or
discourse operators or, alternatively, through the specifier-head relation.
An example of indirect range assignment, and, in particular, of the specifier-
head subtype, is the induction of a telic reading in predicates by quantity DPs in
English. In Borer’s (b) system, a telic interpretation of a predicate depends
uniquely on the existence and licensing of a dedicated projection, AspQP (Aspect-
ual Quantity Phrase), headed by the open value Asp<e>#, which, in English, can be
assigned range if a DP with the right characteristics is merged as the specifier of
AspQP. Specifically, the DP must have a quantity interpretation. In turn, a quantity
Three neo-constructionist theories and a predecessor
interpretation is one that is neither divisive nor cumulative. For a predicate P to be
divisive it must describe a property ascribable to some entity and to any subdiv-
ision of that entity. For instance, the expression water can denote whatever amount
of water one can imagine and any subdivision of that amount, no matter the size.
The expression water, then, is divisive. It is also cumulative, since if the original
amount of water is increased in whatever degree, it will still fall under the
denotation of water. By contrast, the expression more than enough water is not
divisive, since for any amount of water counting as more than enough water, there
is always some portion that cannot be defined as more than enough water. On the
other hand, less than enough water is not cumulative, since, being applicable to
some amount of water, X, it cannot be applied to amounts bigger than X. The DPs
more than enough water and less than enough water count, thus, as quantity DPs.
Definite DPs like the water are also quantity (they are neither divisive nor
cumulative). Thus, the water specifies a definite amount of water (already intro-
duced in the discourse) and cannot, therefore, be applied to a smaller or a bigger
amount. In the next unaccusative example the quantity DP the flower assigns range
to Asp<e># through a specifier-head relation:
() Borer (b: )
EP
the flower
<e>E TP
the flower
wilt<pst><e>T AspQP
the flower
<e># VP
wilt
The same DP the flower moves to the specifier of EP, through that of TP, to provide
range to the eventive open value, <e>E. The open value for tense, <e>T, is assigned
range directly by the abstract past tense head feature <pst>, which triggers head
movement of the listeme wilt. The DPs assigning range to the relevant open values
receive an interpretation ‘as an entailment of the event structure’ (Borer b: ).
Thus, the specifier of AspQP is interpreted as Subject-of-quantity (in Tenny’s
terms, it measures out the event), since it is the subject of a quantity predicate, namely
AspQP. As a specifier of EP, the DP is interpreted as an Originator, as originating
the (wilting) event. These interpretations are, crucially, independent of the listeme
that ends up being the verb (wilt in ()). In this view, unaccusative predicates are
A neo-constructionist perspective on argument structure
characterized by involving a DP merged, successively, as Subject-of-quantity, speci-
fier of TP, and Originator, and, therefore, unaccusativity always involves telicity.10
Much as I am very sympathetic with the generalities of Borer’s (a, b) theory of
the lexicon-grammar interface, I would like to examine two problematic issues
involved in her conception of the relation between event structure and argument
structure: the status of subjects in telic predicates provided with a PP encoding a
bounded path and the nature of telicity itself.
As for the first issue, in cases where indirect range assignment to Asp<e># is
instantiated through a means different from the specifier-head relation, telicity
is predicted to arise in the absence of a Subject-of-quantity. Borer proposes that
this is the case with PPs expressing a bounded path in motion predicates such as the
following:
() Borer (b: )
a. John ran to the store.
b. Jane swam into the room.
c. Pat danced into the corridor.
In these cases the specifier of Asp<e># is indirectly provided range by the bounded
path PP, yielding the telic reading. The subject is, therefore, claimed to be directly
merged as the specifier of EP, in complete dissociation from the telicity of the
predicate. There are three problems with this account. First, the predicates in ()
are predicted not to be unaccusative. However, analogous predicates in Dutch select
the BE-auxiliary in the perfect tense (see ()b in comparison with ()a, without the
PP), strongly suggesting an unaccusative analysis for the predicate and, hence, an
analysis of the subject as originating lower than EP:
() Dutch; Borer (b: )
a. Jan heeft gesprogen.
Jan has jumped
b. Jan is in de sloot gesprongen.
Jan is in the ditch jumped
‘Jan has jumped into the ditch.’
Borer (b: , footnote ) does note that her analysis of () as (telic) un-
ergative predicates is in contradiction with the fact that similar predicates in Italian
allow ne-cliticization, a traditional unaccusativity diagnostic. However, she claims
that ne-cliticization does not necessarily signal unaccusativity, but, rather, a post-
verbal location of the subject. Even if her approach to () can escape the critique
10
The difference between unaccusative predicates and telic transitive predicates lies in the fact that the
latter involve two argumental DPs: the object, merged as the specifier of the aspectual head and the subject,
merged first as specifier of TP and then as specifier of EP.
Three neo-constructionist theories and a predecessor
based on ne-cliticization, it does not escape, I observe, that based on auxiliary
selection in Dutch, which she herself mentions as unaccusativity diagnostic
(b: ).
Second, the subjects in this kind of predicate turn out to be Subjects-of-quantity, as
evidenced by the fact that mass DP subjects (see marine life in ()) bleed telicity, in
spite of the presence of the bounded directional PP:
() Marine life swam into the cave (for hours/*in five minutes).
A third related problem is the fact that the subject is interpreted as a Figure with
respect to the DP embedded in the PP, which is understood as a Ground (see
section ...). Merging the subject directly as the Originator (at the specifier of
EP) does not account for this interpretation. On the other hand, the status of the
subject as an unaccusative subject seems to depend on the presence of the PP, as the
constrast in () tells us. We have reason to believe, therefore, that the subject and
the PP are structurally related at some level in motion predicates expressing a
transition, and that the properties of both the DP and the PP have a bearing on the
(a)telicity of the predicate. As it turns out, the syntactic relation and Figure–
Ground interpretation between the DP subject and the PP is straightforwardly
accounted for in theories proposing a Small-Clause projection where the PP acts as
the predicate and the surface subject is in fact the Small-Clause subject (see, among
others, Hoekstra : ; Hoekstra and Mulder : ; or Mateu and Rigau
: ). In section ... I develop a theory along these lines. In particular, I will
argue that in telic predicates endowed with argumental PPs, including the type
illustrated in (), both the specifier and the complement in the PP are eligible,
albeit under different syntactic circumstances, as responsible for the telic reading of
the event—see also section ... Crucially, I put forward a theory of transition vPs
as including a lower layer expressing a Figure–Ground configuration (PlaceP) and
an intermediate layer expressing a transition and licensing a telic reading of the
predicate (PathP).
The second observation has to do with the issue of whether telicity is always
grammatically represented. In particular, it is also unclear how Borer’s (b)
analysis can account for data such as the following:
() Italian; Mateu (a)
La giumenta {ha figliato/ *è figliata} in/??per due ore.
the mare(F) has foaled.M.SG is foaled.F.SG in/for two hours
‘The mare has foaled in two hours’
The above example is not unaccusative, since the HAVE-auxiliary is selected; but,
crucially, it is not atelic. However, there is no apparent licenser for AspQ either.
Rather, it seems, as argued by Mateu (a), that the telicity in the above example
is not grammatically represented and must depend solely on the conceptual
A neo-constructionist perspective on argument structure
properties of the root, here one referring to an entity unmarkedly interpreted as
bounded (figlio ‘son’).
.. Distributed Morphology
A glance at such works as Marantz (, ), Harley and Noyer (, ), or
Pylkkänen () reveals that Distributed Morphology (DM) is not simply a theory
of morphology, although maybe its motivations were, in the beginning, of a mor-
phological nature (see Halle , ; Halle and Marantz , ): it implies a
revision of the generative model of grammar, with particular attention to the syntax-
morphology interface, and basically assuming a minimalist design (Chomsky ).
The main tenet in the theory is that syntax is the only generative engine of the faculty
of language, and, hence, that whatever stores of idiosyncratic information must be
postulated are exclusively of a non-computational nature (but see below for a
qualification). In this way, it is denied that there could be any operations in the
lexicon, and, in fact, the traditional lexicon is split up into three different stores or
lists, as shown below (Marantz , ):
() Based on Marantz (: –)
a. List or Narrow Lexicon, containing bundles of purely morphosyntactic
features called morphemes.
b. List or Vocabulary, containing Vocabulary Items, which are rules of
correspondence between a phonological exponent and an underspecified
set of morphosyntactic features and other contextual instructions.
c. List or Encyclopaedia, containing Encyclopaedia Entries, which are rules
of correspondence between a phonological exponent and a set of world-
knowledge properties (for cat, for instance, ‘furry animal’, ‘domestic’, etc.).
Syntax exclusively operates with morphemes provided by the Narrow Lexicon to
yield hierarchic representations feeding both the phonological and semantic inter-
pretations of linguistic expressions. These morphemes, as mentioned above, are
bundles of abstract features taken from a common pool provided by UG. Marantz
(: ) contends that ‘[t]he sets of grammatical features are determined by
Universal Grammar and perhaps by language-particular (but language-wide) prin-
ciples. Since these sets are freely formed, subject to principles of formation, List is
“generative”.’ I note that, as long as one of the lists is generative, the goal of having a
single generative engine, expressed as the basic postulate of the theory, is not
achieved.11 On the other hand, Marantz (: ) characterizes the Vocabulary
and the Encyclopaedia as ‘non-generative but expandable’.
11
A critique based on Starke (). In Starke’s () nanosyntactic theory the nodes of the syntax are,
in fact, individual features, so there is no need for a pre-syntactic generative narrow lexicon.
Three neo-constructionist theories and a predecessor
No phonological or encyclopaedic information is present in syntactic computa-
tions: DM endorses the hypothesis of Late Insertion, by virtue of which phonological
information is retrieved once the syntactic representation is delivered at the PF
interface, after Spell-Out. At the moment of Vocabulary Insertion, the insertion of
Vocabulary Items into the nodes of the syntactic configuration, the distinction
between f-morphemes and l-morphemes becomes important (Harley and Noyer
, ). The former correspond to functional nodes like v or T, conveying
only morphosyntactic meaning like the values for number, tense, person, etc., and
triggering an almost automatic Vocabulary Insertion. For instance, the f-morphemes
of plural number in nouns and past tense may receive, in English, the phonological
exponents specified, respectively, by the following Vocabulary Items:
() Harley and Noyer (: )
/-s/ ⟷ {[Num], [pl]}
/did/ ⟷ [pst]
Vocabulary Insertion for f-morphemes is automatic, in the sense that there is not a
free choice of Vocabulary Items for a given f-morpheme. Rather, it is regulated
through a process of competition between different Vocabulary Items whose set
of contextual features must be a subset of those making up the f-morpheme (cf.
the Subset Principle, Halle ). On the other hand, Vocabulary Insertion into
l-morphemes (lexical morphemes) might seem arbitrary, non-deterministic: in prin-
ciple there is a choice as to inserting either cat, dog, table, or idea into an l-morpheme.
This aims at accounting for the fact that the phonological variation in roots is usually
significantly less dramatic than that in functional material. However, some authors
(Harley and Noyer , ) have emphasized the need to elaborate a theory of
licensing, where root Vocabulary Items are endowed with contextual specification
as to be insertable only in particular nodes. In that sense the difference between
f-morphemes and l-morphemes is significantly weakened. In Chapter I will follow
Harley (), among others, in assuming that roots are early inserted as abstract
indexes and that they receive phonological specification through Vocabulary Inser-
tion, as f-morphemes.
On the semantic side, the configuration generated by the syntax arrives at LF,
where it is automatically interpreted on the basis of both the featural content of
f-morphemes and their position in the configuration (which confers on them
different ‘flavours’, like ‘cause’ or ‘become’ for the v head—see Harley ;
Marantz ). Marantz (: ) emphasizes the fact that the semantic interpret-
ation of a linguistic expression partakes of both its LF representation and the
‘derivation as a whole’, in particular, ‘any and all unforced choices made’. Presum-
ably he is referring to the roots freely inserted during Vocabulary Insertion, for
which, as mentioned above, there is an unforced choice. Under a Late Insertion
approach to roots we must conclude that the only possible way for the conceptual
A neo-constructionist perspective on argument structure
system to access the non-compositional meaning encapsulated in roots like DOG or
CAT is by accessing Vocabulary Insertion, where the choice is made, and then
looking up the correspondent entry listed in the Encyclopaedia (for instance,
dog ⟷ [‘four legs’, ‘canine’, ‘pet’, ‘sometimes bites’, etc.]—see Harley and Noyer
: ). Of course, that architectural complication (graphically represented in
Harley and Noyer’s : diagram as the Encyclopaedia being linked by different
arrows) is not required if roots, as opposed to f-morphemes, are early inserted and,
hence, present before Spell-Out. See Marantz (, ); Embick (); Real
Puigdollers (); and Harley (), for discussion, and also section .. The
interpretation of roots turns out to be, to a certain extent, context-dependent.
Crucially, the context within which a special meaning of a root may be triggered is
locally defined. In Marantz (: ff.), for instance, the observation is made that
the little v (verbalizing) head defines one such domain, as vPs like take a leap are
interpreted as simple verbs like leap. On the contrary, the causative verb make can
only trigger idiomatic interpretation if the verb it embeds does not itself project an
external argument. For instance, make ends meet receives an idiomatic interpret-
ation ‘earn and spend equal amounts of money’ due to the special meanings
retrieved for the roots involved (END, MEET) within a local domain (vP). That the
meanings can be retrieved is possible because unaccusative meet does not involve
the projection of a head selecting an external argument, which would count as a
boundary between make and ends meet. That boundary is present in constructions
like make (someone) swim/fly a kite/etc., which, accordingly, may only receive an
interpretation in which make is a causative verb and the embedded verb retains its
usual meaning—see also Harley (). Crucially, much as special meaning might
be triggered for roots within well-defined contexts, the whole structure is not
assigned a special meaning. That could never be the case, since the LF-semantics
inherent to the configuration generated by syntax is compositional and cannot be
overridden. Marantz (: ff.) makes the claim, for instance, that in the idiom
kick the bucket a special interpretation is retrieved for kick and bucket (specifically,
for KICK and BUCKET). However, the meaning associated with a transitive structure
with a definite DP as object, that is, the LF of that expression, is computed, and,
thus, kick the bucket is not the same as die (cf. He was dying for days/*He was
kicking the bucket for days). Finally, the local domain in which a particular
interpretation of a root is triggered has eventually come to be identified with the
phase (Chomsky ). Accordingly, there has been theorizing, within the DM
tradition, about what categories define phases, based on the evidence of particular
interpretations arising within well-defined contexts (cf. Arad , ; Marantz
, , ; Embick ; or Real Puigdollers , among others.).
In the next chapter I will expand on the way DM has dealt with syntax-
morphophonology mismatches, since this will be one of the main issues of the
present study.
The present framework
. The present framework
In this section I present the framework within which I approach the argument
structure phenomena dealt with in the book. Although I have been primarily inspired
by the configurational theory of thematic interpretation to be found in Hale and
Keyser () and Mateu (), I also draw on insights from Borer (b) and
DM. Thus, on the one hand, assuming as desirable a theory of grammar with only
one generative engine (cf. Marantz ), I endeavour to do away with the l-/s-syntax
distinction. On the other hand, I emphasize Borer’s (b) view of roots as
grammatically opaque elements and I also try to incorporate her insights on the
syntactic representation of event structure. First, I will show how argument structure
is syntactically built. Then I will discuss how the syntactic configuration is inter-
preted semantically, leaving its morphological interpretation for Chapter .
.. Argument structure is syntax
... No l-/s-syntax distinction In section .. I argued that l-syntax, as por-
trayed in the works of Hale and Keyser, constitutes an independent cycle of syntactic
computation. I will assume, along with the DM framework, that there is only one
generative engine responsible for the generation of every (morpho)syntactic object.
In particular, roots and DPs will be shown to be merged as arguments (that is, as
complements or specifiers of argument structure configurations), and, hence, to be
interspersed in the configuration. I am, of course, not arguing for a cycle-less syntax.
Rather, the cyclic effects described at the level of the word and those described at the
level of bigger units should be accounted for through the same mechanism (Marantz
, , , ). In particular, the phase as a cycle (Chomsky ) should
account for any phonological and semantic opacity effects traditionally attributed to
the word/non-word, lexicon/syntax or l-syntax/s-syntax distinction (cf. Marantz
). Phases are mostly important, within this work, as locality domains for
semantic and phonological interpretation (see sections ... and .).
... Relational and non-relational elements I adopt Mateu’s () important
distinction between relational and non-relational elements as the basic building
blocks of argument structure. Relational elements are functional heads, universally
provided by UG, and are able to project structure. There are two kinds of relational
elements within the vP: an eventive v head and the adpositional heads Path and
Place. The v head may acquire ‘flavours’, that is, different semantic interpretations
depending on configurational properties. In particular, if it takes a PlaceP as com-
plement, it is interpreted as stative, and if it takes a PathP, it is interpreted as a
transition. Following a seminal idea in Kratzer () (see also Borer b;
Pylkkänen ; and Harley , among many others), I take the external argument
to be introduced by an independent head, Voice, merged on top of vP.
A neo-constructionist perspective on argument structure
A single adpositional projection, Place, is interpreted as a predicative relation
between two entities. A Path head taking PlaceP as complement is interpreted as a
transition and may induce a telic reading of the resulting predicate (see section
...). In this way, Place corresponds, semantically, to Hale and Keyser’s central
coincidence relation, while Path corresponds to their terminal coincidence relation
(Hale ; Hale and Keyser ). Roughly, while a central coincidence relation like
the one in Sue is in the room involves stasis, a terminal coincidence relation like the
one in Sue goes into the room involves change (Hale and Keyser ). I use the terms
Place and Path to parallel (not entirely, though) a distinction made within studies of
the PP, from the seminal work of Jackendoff () through works such as Koopman
(), Svenonius (), or Gehrke (), among others.12
Non-relational elements are unable to project structure, and are of two kinds:
roots (represented in small capitals in this work) and DPs. Roots are deprived of
category and cannot project, unlike in other exo-skeletal frameworks, such as
Harley (, ). They are grammatically opaque, as are Borer’s (a,
b) listemes. Since roots cannot project, there is no syntactic object of the
form RootP.13 DPs, on the other hand, may be expanded by adjuncts, but no new
structure is created thereby. That non-relational elements should be of these two
kinds is a natural consequence of eliminating the l-/s-syntax distinction: once a
single computation is assumed, the merger of roots and DPs is expected to be
interspersed in the structure. Non-relational elements appear either at complement
or specifier position, although roots are precluded from the specifier position, since,
not having functional structure, they cannot be proper subjects (see, for instance,
Espinal and Mateu ).14
... Argument structure configurations Application of the operation Merge
to relational and non-relational elements yields the different types of vP which
correspond to the different argument structure configurations, as illustrated in
() to ():
12
Ultimately, the difference between Path and Place could be argued to be purely configurational, a
single category p being interpreted as Place and an ulterior one being interpreted as Path. In turn, the
categories v and p could themselves be conflated into one relational head, the distinction derived also from
configurational properties. See Boeckx () for related discussion.
13
See De Belder () and De Belder and van Craenenbroeck () for an interesting proposal in
which roots are not present during the syntactic derivation, their properties (including inability to project)
being derived from a particular theory of the operation Merge and a reformulation of the Subset Principle.
For other proposals doing away with the relational/non-relational dichotomy in syntax, see Real Puigdol-
lers (); Acedo-Matellán (); and Acedo-Matellán and Real-Puigdollers ().
14
Marantz (, ) has recently rejected the idea that roots may occupy argumental positions. See
Acedo-Matellán () for an exposition of Marantz’s () arguments that roots can only be modifiers,
and a refutation of this proposal on empirical and theoretical grounds.
The present framework
() Unergative/Transitive creation/consumption event
a. Sue danced.
VoiceP
Sue Voice’
Voice vP
v dance
b. Sue did a dance.
VoiceP
Sue Voice’
Voice vP
v a dance
c. Sue pushed the car.
VoiceP
Sue Voice’
Voice vP
pP vP
p the car v push
() Atelic (stative) unaccusative event
a. Dinosaurs existed (for a long time).
vP
v PlaceP
Dinosaurs Place’
Place exist
b. Sue is in Barcelona.
vP
v PlaceP
Sue Place’
Place Barcelona
Place in
A neo-constructionist perspective on argument structure
() Unaccusative event of change of state/location
a. The sky cleared.
vP
v PathP
The sky Path’
Path PlaceP
The sky Place’
Place clear
b. Sue went to Barcelona.
vP
v = went PathP
Sue Path’
Path = to PlaceP
Sue Place’
Place Barcelona
() Stative transitive event
a. Sue loves peaches.
VoiceP
Sue Voice’
Voice vP
v PlaceP
peaches Place’
Place love
b. Sue kept the car in the garage.
VoiceP
Sue Voice’
Voice vP
v = kept PlaceP
the car Place’
Place the garage
Place in
The present framework
() Transitive event of change of state/location
a. The strong winds cleared the sky.
VoiceP
The strong winds Voice’
Voice vP
v PathP
the sky Path’
Path PlaceP
the sky Place’
Place clear
b. Sue shelved the books.
VoiceP
Sue Voice’
Voice vP
v PathP
the books Path’
Path PlaceP
the books Place’
Place shelf
c. Sue put the books onto the shelf.
VoiceP
Sue Voice’
Voice vP
v = put PathP
the books Path’
Path = -to PlaceP
the books Place’
Place the shelf
Place on
A neo-constructionist perspective on argument structure
Some remarks must be made about how these configurations relate to syntactic facts.
First, I follow Hale and Keyser’s () or Mateu’s () proposal that unergative
predicates (see ()a) are underlyingly transitive predicates. Specifically, within the
present proposal, unergative verbs like dance correspond to a vP where Compl-v is a
root, and not a DP/NP. The structure of unergative verbs as transitives is forced by
the properties of the system: it is not possible for a functional head to project a
specifier without projecting any complement, since the first DP/root merged with a
functional head must be its complement (and roots are independently ruled out as
specifiers, as I pointed out in ...). Hence, unergatives must be transitives (that is,
they must feature a complement—a root).15
In Hale and Keyser (: ff.) transitive activity verbs like kick are provided
with the following lexical-syntactic configuration (where F stands for a functional
category introducing the external argument), identical to that provided for stative
verbs like love—see Mateu () and Acedo-Matellán and Mateu () for an
analysis along similar lines:
() Sue kicked the door.
[FP Sue [F’ F [vP v [PP the door [P’ P KICK]]]]]
The P head in the above configuration is to be understood as involving a relation
of central coincidence. Hale and Keyser paraphrase predicates like kick someone as
‘give someone a kick’. It is hard to accept, though, that stative predicates like love
and dynamic, activity predicates like kick should involve the same syntactic
configuration, much as both types of predicates are atelic—a fact captured, within
this system, through the absence of the Path projection. A basic difference, for
instance, is that stative predicates can hardly drop their objects, while activity
predicates do so with relative ease. In fact, stative verbs seem to be forced to adopt
an activity interpretation when they are used intransitively, as in the second
example with love:
() Object drop in kick vs love
a. Sue kicked for several minutes.
b. ??Sue loved for years.
Following an idea in Marantz (), I consider transitive activity predicates such
as kick the door as involving an unergative configuration, the verbal root being
merged as the sister of v. The overt object is in fact an adjunct merged with the
vP through a null preposition interpreted as a central coincidence relation (see
also ()c). The adjunct-hood of this DP is what explains the fact that it can be
omitted:
15
The same rationale underlies the treatment of particles as ‘unergative’ prepositions. See section ....
The present framework
() Sue kicked the door.
[VoiceP Sue [Voice’ Voice [vP [pP p the door] [vP v KICK]]]]
The structure is paraphrasable as ‘do kicking on/at the door’. This paraphrase
reveals the presence of the preposition, which can actually be overt, as pointed
out by Marantz (), and with no apparent modification of the meaning (see
also Levin for cross-linguistic evidence of this type of preposition in non-
prototypical objects):
() Anderson (: ) in Levin and Rappaport-Hovav (: )
a. The farmer plowed the field.
b. The farmer plowed in the field.
Unaccusative predicates (see () and ()) are those not projecting a VoiceP.
Unaccusatives may be causativized (transitivized) if Voice is added and a DP
merges as Spec-Voice, as shown through the contrast between ()a and ()a.
The difference between an unaccusative structure with PlaceP as Compl-v ()
and one with PathP as Compl-v () has to do with the interpretational differ-
ence between a stative predicative relation and a transition (see sections ...
and ...). However, a unifying syntactic phenomenon for all unaccusatives,
hence for both () and (), is the fact that these predicates select or admit
selection of a BE-auxiliary for the perfect tenses in languages like Italian, as shown
below:
() Italian; Sorace (: )
a. I dinosauri {sono esistiti/ ??hanno esistito}
the dinosaurs are exist.PTCP.PST.M.PL have.PL exist.PTCP.PST.M.SG
milioni di anni fa.
millions of years ago.
‘Dinosaurs existed million years ago.’
b. Maria {è venuta/ *ha venuto} alla festa.
Maria is come.PTCP.PST.F.SG has come.PTCP.PST.M.SG to.the party
‘Maria came to the party.’
Finally, observe that the DP at Spec-Place rises to Spec-Path when it is available (for
instance, the sky in ()a). This movement, and the semantic interpretation of the
structures above will be discussed in section ....16
16
In the representations I have also abstracted away from other movements, for instance movement of
the internal argument for case-reasons (to the second Spec-Voice or to Spec-T).
A neo-constructionist perspective on argument structure
... Adjunction of roots to functional heads Alongside the complement position,
roots may appear as adjuncts to the functional heads. This is the case in ()c of
section ... above: the root ON is adjoined to the functional Place head. That the
preposition on should involve a root, that is, a non-relational element, might seem
striking at first, but once a strict delimitation between conceptual and grammatical
content is accepted, it must be acknowledged that the difference between, say, in the
box and on the box cannot be grammatical in nature, and that the choice between
both is of the same status as that between The cat is on the mat and The dog is on the
mat. My proposal that the spatial value of adpositions is encoded as a root adjoined
to a functional element is in accordance with Baker’s (: ) or Svenonius’s
() observation that ‘P is essentially a functional category, despite its association
with encyclopedic information’ (Svenonius: : ). Actually, Baker himself
suggests that
English might have a relatively large number of prepositions on the surface because it permits
relational nouns to conflate into an abstract P head prior to lexical insertion. This proposal
would capture nicely the fact that preposition seems to be a hybrid category in English, neither
clearly functional nor clearly lexical. Baker (: )
Similarly, within the domain of Latin prepositional prefixes, the separation of the
conceptual content of the prefix and the structure in which it is embedded finds an
echo in Carvalho’s (: ) distinction between signifié lexical and signifié cate-
goriel of the prefix.17
The dissociation of prepositions into a functional and a non-functional element
straightforwardly implements the well-established idea that particles are intransitive
prepositions (see Cappelle : ff. and references cited therein). In particular,
while PPs like on the shelf correspond to PlaceP structures in which the root of the
preposition is adjoined to Place and Compl-Place is a DP (the shelf), particles like on
correspond to PlacePs where the root of the preposition sits directly at Compl-Place.
The difference is illustrated below:
() The books (are) on the shelf.
[PlaceP The books [Place’ [Place Place ON] the shelf]]
() The lights (are) on.
[PlaceP The lights [Place’ Place ON]]
17
I conjecture that the fact that inventories of adpositions contain much fewer elements than those of
nouns is due to the fact that the set of possible or cognitively salient spatial relations, conveyed by
adpositions, is much smaller than the set of entities, conveyed by nouns. For more discussion on the
functional or lexical status of P, see Koopman () or Den Dikken (), among others. See also
Roßdeutscher () for an analysis of contentful prepositions as involving a root and a functional head p.
The present framework
Thus, particles, whether affixal (as in Latin) or not, turn out to be, specifically,
unergative prepositions, as illustrated in () (see Kayne ).
A root can also adjoin to v. Thus, the roots DANCE and HAMMER are adjuncts to v in
()a and ()b, respectively:
() Root-adjunction to v
a. Sue danced into the room.
vP
v PathP
v dance Sue Path’
Path = to PlaceP
Sue Place’
Place the room
Place in
b. Sue hammered the metal flat.
VoiceP
Sue Voice’
Voice vP
v PathP
v hammer the metal Path’
Path PlaceP
the metal Place’
Place flat
Root-adjunction to v, which will be crucial for understanding the data dealt with in
this work, is designed to capture so-called lexical subordination constructions (Levin
and Rapoport ), that is, constructions involving a complex event where the main
event is identified with an accompanying co-event. Thus, for instance, in ()a the
unaccusative event whereby Sue enters the room is accompanied by a subordinate
event of dancing (although the dancing, it should be noted, is not linguistically
represented as a separate event, that is, through a separate v head). For similar
treatments of lexical subordination, see Embick (); McIntyre (); Zubizarreta
and Oh (); and Mateu (b, ), among others. For a reformulation in
terms of Bare Phrase Structure, see Acedo-Matellán ().
A neo-constructionist perspective on argument structure
.. The semantics of argument structure: a localist-aspectual approach
... Structural and encyclopaedic semantics An important distinction must be
drawn between the semantic interpretation of the configurations delivered by the
syntax, as shown in section .., and the conceptual semantics encapsulated within
roots. Let us call the former structural semantics, following Harley and Noyer (),
and the latter, encyclopaedic semantics, since it must be listed, for every root, in a
storage called Encyclopaedia (Marantz : ). It is the integration of the encyclo-
paedic content of the roots with the structural semantics read off the syntactic
configuration that corresponds to the semantic interpretation of the whole derivation
(Marantz : ). These two dimensions of meaning correspond to compositional
and non-compositional meaning, respectively. In particular, I follow Marantz ()
in the idea that syntax alone is responsible for the derivation of compositional
meaning (that is, compositional meaning is built up or derived), while the Encyclo-
paedia alone is responsible for the storage of non-compositional meaning (that is,
non-compositional meaning is stored and underived). Thus, any object created by the
syntax must bear compositional meaning, although, of course, it embeds minimal
pieces endowed with non-compositional meaning.
... Interpretation of functional heads and arguments As was briefly introduced
in section ..., v is an eventive head, introducing an event in the structural
semantics. This event might be interpreted as externally originated (brought
about), if a DP—the external argument—is merged as Spec-Voice (see ()a), and
as non-externally originated, if Voice is not projected (see ()b):
() Externally vs non-externally originated events
a. The strong winds cleared the sky.
b. The sky cleared.
In turn, the head v receives a variety of interpretations according to the nature of its
complement. It is interpreted as a creation/consumption event when its complement
is a root or a DP:
() Creation/consumption event
a. Sue danced.
b. Sue did a dance.
v is interpreted as a transitive atelic event if its complement is a PlaceP, embedding a
root or a DP as Compl-Place:
() Transitive atelic event
a. Sue loves peaches.
b. Sue kept the car in the garage.
The present framework
Finally, v is interpreted as an externally originated change-of-state/-location
event if its complement is a PathP, again, embedding either a root or a DP as
Compl-Place:
() Externally originated change of state/location
a. The strong winds cleared the sky, Sue shelved the books.
b. Sue put the books on the shelf.
When Voice is not projected, v is interpreted, if its complement is a PlaceP, as an
intransitive stative event, either involving a root at Compl-Place or a DP:
() Stative or atelic unaccusative event
a. Dinosaurs existed (for a long time).
b. Sue is in Barcelona.
In turn, if the complement of unaccusative v is a PathP, it is interpreted as a non-
externally originated change of state/location, embedding either a root or a DP as
Compl-Place (I assume that the preposition to in ()b is a direct phonological
realization of Path in English):
() Unaccusative event of change of state/location
a. The sky cleared (in five minutes).
b. Sue went to Barcelona.
A single adpositional projection is a PlaceP, which establishes a predicative relation
between two entities. Thus, in The sky cleared (in five minutes) and The sky is clear
there is a predicative relation between The sky and the root CLEAR. Similarly, in Sue
went to Barcelona and Sue is in Barcelona there is a predicative relation between Sue
and (in) Barcelona. The Path head takes a PlaceP as complement, introducing a
transition and inducing telicity in the predicate if a quantity DP is internally merged
as its specifier. See section ... for more details on situation aspect and argument
structure.
Arguments, be they DPs or roots, are semantically interpreted as a result of the
position they occupy in the structure. This interpretation does not correspond to
traditional theta roles, but it is more abstract in nature. Next I list these interpret-
ations, each one of them linked to a precise position in the configuration:
() Interpretation of DPs and roots
a. Originator: a DP at Spec-Voice
Sue danced, Sue did a dance, Sue pushed the car, Sue kept the car in the
garage, The strong winds cleared the sky, Sue shelved the books, Sue put the
books on the shelf
b. Effected Object: a DP or root at Compl-v
Sue did a dance, Sue danced
A neo-constructionist perspective on argument structure
c. Figure: a DP at Spec-Place
Dinosaurs existed, The sky cleared, Sue went to Barcelona, Sue is in
Barcelona, Sue kept the car in the garage, The strong winds cleared the
sky, Sue shelved the books, Sue put the books on the shelf
d. Central Ground: a DP or root at Compl-Place when no PathP is projected
Dinosaurs existed, Sue loves peaches, Sue is in Barcelona, Sue kept the car
in the garage
e. Terminal Ground: a DP or root at Compl-Place when PathP is projected
Sue went to Barcelona, The sky cleared (in five minutes), The strong winds
cleared the sky, Sue shelved the books, Sue put the books on the shelf
f. Measurer: a DP raised from Spec-Place to Spec-Path
Sue went to Barcelona, The sky cleared (in five minutes), The strong winds
cleared the sky, Sue shelved (the) books, Sue put (the) books on the shelf
g. Co-event/Conformation: a root adjoined to a functional category (v or Place)
Sue danced/tiptoed into the room, Sue hammered/trod the metal flat, The
book was in/on the box
These interpretations are in part localistic and in part aspectual, that is, Aktionsart-
related. The notions Figure and Central or Terminal Ground are localistic. The
Figure, in Talmy’s () terms, is the entity that is located or moving with respect
to some other entity, which is the Ground. For instance, Sue is a Figure and Barcelona
is a Ground both in Sue went to Barcelona and Sue is in Barcelona. The relation
between Figure and Ground can also be metaphorical, in terms of the predication of
some property: the Figure is an entity to which some property, encoded by the
Ground, is ascribed. Thus, the sky and clear are, respectively, a Figure and a Ground
in The sky cleared in five minutes and in The sky is clear.
The Ground, in turn, can be either a Central Ground or a Terminal Ground, a
localistic-aspectual distinction. A Central Ground corresponds to a location/state
that corresponds to a static description, as in The sky is clear. A Terminal Ground
corresponds to a final or resulting location/state. For instance, in Sue went to Barcelona
and The sky cleared in five minutes it is entailed that Sue ends up in Barcelona and that
the sky ends up in a pragmatically defined state of clearness after five minutes.
The Originator, the Effected Object, and the Measurer are event-structural
notions. An Originator is the entity that originates the event, as, for instance, is
The strong winds in The strong winds cleared the sky. An Effected Object is an entity
that comes into existence or disappears as the event evolves. For instance, in Sue
danced, the root DANCE, an Effected Object, refers to the activity of dancing, which
unfolds along with the event introduced by v. In Sue did a dance, the DP a dance
is the Effected Object, with the same interpretation. Lastly, a Measurer, a DP at
Spec-Path, is an entity that induces a measure for the transition introduced by
The present framework
PathP. Thus, for instance, in Sue shelved the books in five minutes or The sky cleared
in five minutes, the books and The sky are Measurers (they move to Spec-Path from
their original Spec-Place position, where they are interpreted as Figures) in that they
establish a measure for the events of shelving and clearing. Thus, these events will be
completed as soon as the entities denoted by the Measurers attain the location/state
denoted by PlaceP, that is, when all the books denoted by the books are shelved and
when the whole entity of the sky denoted by The sky is clear. However, note that I also
call Measurer a non-quantity DP like books in Sue put books on the shelf or Marine
life in Marine life swam into the cave for hours. In these predicates there is also a
transition encoded by PathP, but since the quantity conveyed by the object is not
definite, telicity cannot arise. See section ... for more details on the relation
between Path and (a)telicity and the interpretation and syntax of the Measurer.
As I pointed out in section ..., the roots adjoined to functional categories, like v or
Place, provide a conceptual specification of their abstract meaning. A root adjoined to v
is thus interpreted as a (Manner) Co-event in that it specifies the way in which the event
introduced by v is carried out. Thus, in Sue hammered the metal flat, the externally
originated event of change of state (of a metal which becomes flat) is identified with a
hammering activity, since v forms an adjunct structure with root HAMMER. A root
adjoined to Place identifies the type of spatial relation that a Figure holds with respect
to a Ground, its Conformation, in Talmy’s () terms (see section ...).
I point out, finally, a crucial difference between Mateu’s () theory and the present
theory, which concerns the interpretation of functional heads (relational heads in
Mateu’s terminology). Recall from section .. that relational heads are endowed
with either a þ or a - value, characterizing agentivity/non-agentivity (for R), transi-
tion/non-transition (for T), and telicity/atelicity (for r). Recall, also, that among struc-
tures featuring the r relation, two structures were missing in Mateu’s () model:
() *[F X . . . F . . . [-R [X [þr X]]]] (a transitive non-agentive telic event)
() *[-T [X [þr X]]] (an unaccusative stative telic event)
I claim that to the extent that the present account eliminates (non-configurational)
features in the interpretation of relational heads, the non-existence of the above
combinations is explained away. With respect to (), since I have not taken agentivity
to be linguistically represented I do not make a difference between þR (Sue sings: [F
Sue . . . F . . . [þR SING]]) and -R (Sue stinks: [F Sue . . . F . . . [-R STINK]]). Thus, I have
no non-existing combination to account for. As regards (), the þT/-T difference
relates to a dynamic/stative difference. However, I do not encode this difference on the
eventive head. Rather, an unaccusative predicate, if expressing a transition, is endowed
with a double p-projection (PathP) (cf. clear); if not expressing a transition, it is
endowed with a single p-projection (PlaceP) (cf. exist). In this scenario a configuration
equivalent to that in () could never be generated.
A neo-constructionist perspective on argument structure
The table below summarizes again the correspondences between the syntactic
position of arguments (DPs or roots) and their interpretation:
() Syntactic positions and semantic interpretation
Syntactic position Semantic interpretation
Spec-Voice Originator
Compl-v Effected Object
Spec-Place Figure
Central (no Path projected)
Compl-Place Ground
Terminal (Path projected)
Spec-Path Measurer
Adjunct to v Co-event
Adjunct to Place Conformation
... Against root ontologies I argue that roots must be treated on a par with
DP arguments. That means that roots, as DPs, receive a particular interpretation
depending on their position in the structure. For instance, a root like HAMMER may be
interpreted as an Effected Object (see ()), Terminal Ground (see ()) or Co-event
(see ()), depending on the configuration where it is merged:
() Sue hammered (the metal) for hours.
[VoiceP Sue [Voice’ Voice [vP [pP p the metal] [vP v HAMMER]]]]
() Sue hammered the metal in five minutes.
[VoiceP Sue [Voice’ Voice [vP [PathP the metal [Path’ Path [PlaceP the metal [Place’
Place HAMMER]]]]]]]
() Sue hammered the metal flat.
[VoiceP Sue [Voice’ Voice [vP [v v HAMMER] [PathP the metal [Path’ Path [PlaceP the
metal [Place’ Place FLAT]]]]]]]
In (), the root HAMMER is understood as an Effected Object, since it is the comple-
ment of v; as such, it describes the result of an activity and, accordingly, is compatible
with an atelic reading of the predicate. In (), the root is understood as a Terminal
Ground, since it is embedded in a PathP. Therefore, it depicts a final state (‘the state
of being hammered’), which, accordingly, habilitates a telic interpretation. Finally, in
() the root is interpreted as a Co-event by virtue of its being merged as an adjunct
to v: it specifies the way in which the event, here an externally originated change of
state, takes place.
The present framework
Assuming that roots are freely merged as arguments—again, with the proviso
that they are excluded from specifier position—root ontologies, that is, classifica-
tions of roots according to the possibilities they display of being inserted in the
structure as based on their semantic properties, turn out to be just a descriptive
artefact. Root ontologies are assumed in works such as Harley (), Levinson
(, ), and Rappaport Hovav and Levin (). Harley (), for instance,
proposes that instrument-naming verbs, such as hammer or rake, involve a root
that names an instrument (a hammer, a rake), and so the root is not merged in an
argumental position, for instance, in a position where it is interpreted as a final
state. Following Mateu and Acedo-Matellán (), and Acedo-Matellán and Mateu
(), I claim instead that, if HAMMER or RAKE name an instrument, that fact clearly
belongs to the encyclopaedic semantics and, hence, cannot determine where in the
structure the root is merged. Thus, for instance, Acedo-Matellán and Mateu ()
argue that instrument-naming roots can identify a resulting state in a change-of
state predicate. Verbs such as brush or rake can readily be used with depictive
secondary predication, which, according to Rapoport () and Mateu (), is
only compatible with change-of-state predicates, and not with activity predicates like
those headed by push:
() Acedo-Matellán and Mateu (: –)
a. Don’t brush the coat wet or you’ll ruin it.
b. He raked the field dry.
c. ??He pushed the mare pregnant.
In addition, roots like BREAK, which very intuitively name a result and not a manner
(Rappaport Hovav and Levin ; Rappaport Hovav and Levin ), can none-
theless be used as encoding manner co-events. For instance, in the following example
there is no entailment that the hammer head actually broke, so the verb break cannot
be said to encode a result state predicated of any overt participant in the event:
() McIntyre (: )
The hammer head broke off.
I assume that in the above example the root BREAK has been merged as an adjunct to v,
whereby it is interpreted as a co-event of the main eventuality of the hammer head
separating from the hammer (see section . for more examples with break).
Facts such as these suggest that the interpretation of a root as encoding Co-event,
(Terminal) Ground, or Effected Object depends on where it is merged in the
structure, rather than on a deterministic marking.18 See sections ... and ..
for more related discussion.
18
For further theoretical and empirical arguments against root ontologies and, in fact, against any
diacritic marking on roots, see Acquaviva (). See also Marantz (), who shows that the impression
A neo-constructionist perspective on argument structure
... Aspect and argument structure I assume a two-component theory of aspect
in the sense adopted by Smith (), MacDonald (), and Borik (), among
others: a theory that distinguishes between situation or inner aspect and viewpoint or
outer aspect. Situation aspect has to do with properties internal to the event and, hence,
can be related to what has traditionally been called the type of situation or Aktionsart.
Situation aspect is what distinguishes between states (The sky is clear), activities (Sue
danced), achievements (Sue spotted Jane in the crowd), and accomplishments (The
strong winds cleared the sky) (cf. Vendler ). In this work I concentrate on the
Aktionsart property of (a)telicity, the property distinguishing events with an explicit
endpoint—telic events—and those without an explicit endpoint—atelic events.
On the other hand, viewpoint aspect encodes properties external to the eventuality:
it is related to how the eventuality is presented (Comrie ). Specifically, if the
viewpoint aspect is imperfective, only an internal part of the event is asserted; if it is
perfective, the whole event is asserted, with initial and final bounds. This can be
exemplified from Latin, where the contrast is marked morphologically:
() Latin; Plaut. Merc. and Caes. Gall. , , , in Pinkster (: , )
a. Quo nunc ibas?
To_where now go.IPFV.SG
‘Where were you going to?’
b. Orgetorix [ . . . ] suam familiam [ . . . ] co-egit.
Orgetorix his.F.ACC.SG household(F)ACC.SG together-lead.PRF.SG
‘Orgetorix gathered his household.’
The imperfective ibas in ()a licenses an interpretation in which the going event is
visualized from within, and is not asserted to have been carried out. By contrast, in
()b the perfect form coegit yields an interpretation in which the gathering event is
seen as completed.
Situation aspect and viewpoint aspect are independent from each other. Specific-
ally, telic events can be either imperfective (see ()a) or perfective (see ()b), while
atelic events can be also imperfective (see ()a) or perfective (see ()b), as shown by
the next Catalan examples, which incorporate the traditional test of temporal in- and
for-adverbials:
() Catalan; imperfective and perfective telic predicates
a. En Pol pintava un quadre en dues hores.
the Pol paint.IPFV.SG a picture in two hours
‘Pol was painting/used to paint a picture in two hours.’
of the existence of different classes of roots boils down to (contextually determined) allosemy. Finally, see
Schäfer (: ) for the position that the suitability of a given root for the alternants of the Causative
Alternation depends on its conceptual properties, rather than on any grammatical marking.
The present framework
b. En Pol va pintar un quadre en dues hores.
the Pol PFR.SG paint.INF a picture in two hours
‘Pol painted a picture in two hours.’
() Catalan; imperfective and perfective atelic predicates
a. En Pol ballava durant hores (cada dia).
the Pol dance.IPFV.SG during hours every day
‘Pol used to dance for hours every day.’
b. En Pol va ballar durant hores.
the Pol PFR.SG dance.INF during hours
‘Pol danced for hours.’
Finally, situation aspect is linked to properties traditionally called lexical (i.e.,
related to particular verbs or verb classes), while viewpoint aspect is usually highly
grammaticalized, and expressed through inflectional morphology (that is, morph-
ology that enters into paradigms). In this work, where the term lexical could only
refer to idiosyncratic, non-grammatical properties of roots, the distinction between
situation aspect and viewpoint aspect is structural: situation aspect is encoded
within the vP, while viewpoint aspect is encoded above the vP, in an aspectual
head, Asp, situated between Voice and T (see, for instance, Demirdache and Uribe-
Etxebarria ).
As for situation aspect, I argue that it is calculated partially from argument
structure properties. Drawing partly on Borer’s (b) account, I take telicity to
emerge from a certain configuration involving the projection of a PathP—but see
below for a qualification of the case of telic unergative predicates. This projection
yields the interpretation of a transition, with a resulting location/state, the Terminal
Ground, which is taken as the endpoint for the eventuality. However, a PathP,
though forcing the interpretation of a transition, is not enough to yield a telic
interpretation: a DP with the relevant quantificational properties, a quantity DP, in
Borer’s (b) terms, is what licenses that interpretation (Verkuyl , ). The
DP must have a quantity interpretation (see section ..) in order for the event to be
measured out (Tenny ; Borer b) and, hence, to be telic. Consider the
following example:19
19
Path is of course not completely equivalent to Borer’s (b) AspQ: on the one hand, AspQP, though
entailing a measured change, does not entail the interpretation of a final location/state. On the other hand,
Borer contends that although in some languages the only way to license AspQ is by merging a DP
conveying a definite quantity as its specifier, in some other languages/constructions AspQ is argued to be
licensed independently, through particles, for example (see Borer b, chapters and ).
A neo-constructionist perspective on argument structure
() Sue put {the books/books/paper} onto the shelf.
VoiceP
Sue Voice’
Voice vP
v = put PathP
(the) books/paper Path’
Path = -to PlaceP
(the) books/paper Place’
Place the shelf
Place on
The Path head, when c-commanded by v, triggers movement of the nearest DP in
its c-command domain, usually the Figure DP at Spec-Place. However, as will be
argued in section ... on the basis of Latin data, the Ground may move to Spec-
Path when the Figure is not present. It is at this position that the Figure or Ground
DP is interpreted as a Measurer for the event. Hence, the Measurer interpretation is
dissociated from the Figure or Ground interpretation, as shown by the following
examples:
() Figure vs Ground DP as the Measurer
a. Pour the water out of the bucket in three minutes.
b. Pour the bucket out in three minutes.
This dissociation motivates providing different structural positions for the Meas-
urer, the Figure, and the Ground, and positing movement to Spec-Path to explain
why a single DP can be simultaneously interpreted as Figure and Measurer or
as Ground and Measurer. In turn, observe that the projection of PathP and,
hence, the possibility that the predicate is telic depends, crucially, on the projec-
tion of a PlaceP. This state of affairs, together with the assumption that
telicity arises when a quantity DP from PlaceP is merged as Spec-Path, naturally
accounts for the fact that in intransitive telic predicates endowed with an argu-
mental PP (PlaceP) the subject is a Subject-of-quantity, pace Borer (b) (see
section ..).
Three possibilities arise as to the type of DP internally merged as Measurer and the
type of inner-aspectual interpretation yielded in conjunction with PathP: that the DP
is a quantity description (the books, some books, three books, etc.), a bare plural
(books), or a mass DP (paper):
The present framework
() Different kinds of Measurers (Spec-Path)
a. Sue put {the/some/three books} onto the shelf in ten minutes.
b. Sue put books onto the shelf {for ten minutes/in five seconds}.
c. Sue put paper onto the shelf for/*in ten minutes.
When a quantity DP is merged as Measurer, it licenses a telic interpretation of
the event. For instance, in ()a a quantity of books that qualifies as quantity
(the books or three books is neither cumulative nor divisive; some books, on the
other hand, is cumulative but is not divisive) is asserted to have been put on the
shelf, and the event is over (in ten minutes) when all the books are on the shelf.
When a bare plural is merged as Measurer, two interpretations may emerge:
an atelic one, which depends on the fact that there is no definite number of
elements (books, in ()b), and a telic one, called by MacDonald (: ) the
Sequence of Similar Events interpretation, which hangs on the fact that the
transition codified by PathP may be measured out by each book. Thus, in
()b the telic interpretation involves an indefinite number of telic events of
putting each book onto the shelf in five seconds. Finally, when a mass DP is
merged as Measurer, although the transition codified by PathP is entailed to
take place, the whole event cannot be measured out, since the Measurer conveys
an indefinite quantity. For example in ()c some paper is entailed to end up on
the shelf. In other words, and quite crucially, ()c cannot entail that the
amount of paper is moved towards the shelf by Sue for ten minutes without
ever reaching the shelf. However, since the amount of paper is not quantity, the
event cannot be measured out and atelicity arises.
Telicity is licensed also in predicates involving a DP or root Effected Object, at
Compl-v:
() Telicity with Effected Objects
a. Sue did a dance in an hour.
b. Sue dined (‘had her dinner’) in an hour.
However, the same predicates seem to accept an atelic reading:
() (A)telicity with Effected Objects
a. Sue did a dance for an hour.
b. Sue dined (‘ate at dinner’) for an hour.
Following Mateu (a), I propose that the telicity emerging in predicates without a
PathP projection is not grammatically represented. Rather, in the absence of a Path
head forcing a resultative interpretation of the predicate, and thus, possibly, telicity,
the object is free to be conceived as bounded or not, yielding telicity and atelicity,
respectively. Mateu (a) provides relevant evidence from auxiliary selection in
Italian. In particular, he observes that intransitive telic birthing verbs like figliare
A neo-constructionist perspective on argument structure
‘foal’ select the HAVE auxiliary in the perfect tense, like atelic activity verbs like
cantare, and unlike run-of-the-mill intransitive telic verbs like arrivare ‘arrive’,
which select BE. As was pointed out already at the end of section .., this kind of
example ((), repeated below) is problematic for a theory like Borer’s (b), in
which telicity is always grammatically represented:
() Italian; Mateu (a)
La giumenta {ha figliato/ *è figliata} in/??per due ore.
the mare(F) has foaled.M.SG is foaled.F.SG in/for two hours
‘The mare has foaled in two hours’
Telicity here depends solely on the fact that the Effected Object refers to a bounded
entity (figlio ‘son’). See Harley () for much relevant discussion.
Atelicity can be claimed to emerge from a greater variety of situations in com-
parison with telicity. First, Effected Object predicates license an atelic interpretation,
when they are roots, quantity DPs, bare plurals, or mass DPs:
() Atelicity with Effected Objects
a. Sue danced for an hour.
b. Sue did a dance for an hour.
c. Sue did dances for an hour.
d. Sue did work for an hour.
Predicates with a single p-projection, PlaceP, and, hence, a Central Ground, are
atelic, since they cannot present the location/state as final or resulting. This atelicity
obtains independently of the quantificational properties of the DP merged as
Spec-Place:
() Atelicity with Central Grounds
a. {These people/People} have been in Barcelona for a day.
b. Sue has loved {Jane/tomatoes} for years.
A predicate involving a PathP may be compatible with durative adverbials in three
circumstances. The first has already been illustrated in ()b and ()c: a non-
quantity Measurer (books, paper) yields an atelic interpretation in which the
transition encoded by PathP is entailed to have been partly carried out but, since
the quantity denoted by the DP is not definite, the transition corresponding to the
whole event cannot be calculated and, hence, the event—or, rather, the description
thereof—cannot be telic. On the other hand, predicates with a PathP structure
may license a durative adverbial by virtue of their embedding a PlaceP, which, as
has been pointed out, establishes a predicative relation. In particular, a PlaceP
embedded within a PathP may license an interpretation in which the resultant
location/state is measured by the for-temporal adverbial (Binnick , cited by
Dowty : ):
The present framework
() Atelicity emerging from the resulting location/state
a. George shelved the book for an hour. (MacDonald : )
b. Sue sat down on the couch for a moment.
In ()a the book is entailed to remain on the shelf for an hour after it has been put
there and in ()b Sue sits on the couch for a moment after she has sat down.
Finally, durative adverbials are also licensed in predicates involving PathP struc-
tures, as in the following case:
() MacDonald (: )
George shelved the book for an hour.
The relevant interpretation here is called by MacDonald (: ) Sequence of
Identical Events interpretation: for an hour long a succession of identical events of
shelving the same book is entailed to have been carried out by George.
To conclude the section, I would like to return now to the mechanism via which Path
raises the nearest DP in its c-command domain to Spec-Path. Note that the condition
for Path to behave in such a way, that is, as a probe in search of a goal, is that PathP is a
sister to v. The probing powers of Path in search of a Measurer DP are claimed to
depend, therefore, on the presence of v. This parallels Chomsky’s () proposal on the
primordial role of C in relation to T: C is the phase head, and the (real) probe, and T is a
repository of the phi-features contained in C, through which C triggers movement of a
DP to Spec-T. The intuition behind the proposal for v and Path is quite transparent:
Path introduces a transition because the phrase it heads is c-commanded by v. There are
clear empirical reasons for this, as presented in Chapters and : a preposition that is
external to vP does not trigger telicity and, morphologically, it does not trigger prefix-
ation to v in Latin and Slavic. By contrast, a PathP, being a sister to v, triggers telicity (if a
quantity DP is merged at Spec-Path, as described above) and prefixation in Latin and
Slavic. I shall not pursue the technical implementation of such a proposal, in terms of
feature inheritance or otherwise, but I note that the vP-internal/external nature of the
PP will become particularly relevant in the discussion on directed motion constructions,
dealt with in section .., and that on the relation between atelicity and prefixation in
Latin and Slavic in section .... See Real Puigdollers () for further discussion of
the syntactic parallelism between the pairs C/T and v/Path.
... Locality domains for special meaning With Marantz (, ), I claim
that the special meaning ascribed to either word-sized units or bigger units must boil
down to contextually determined special meaning for roots, and that those special
meanings, like any non-compositional meaning, are listed in the Encyclopaedia.
Indeed, on the one hand, the Encyclopaedia cannot store chunks of structure,
since, from a strictly derivational point of view, structure cannot be stored (see
section .. for a critique of the l-/s-syntax difference in the same spirit); on the
other hand, structure cannot carry special meaning, since it depends uniquely on
A neo-constructionist perspective on argument structure
functional heads, whose semantic interpretation is determined by features provided
by UG. In particular, the Encyclopaedic entry of a given root may list a special
meaning of that root providing the context within which that meaning is triggered.
Crucially, though, the context is a local domain: the phase.
Latin prefixed verbs provide an example of how the phase delimits a domain where
special meaning of roots can be triggered. In particular, prefixed verbs in Latin show
idiosyncratic meanings presumably not derived from the sum of the parts (the prefix
and the verb).20 Thus, for instance, the verb ob-curro ‘against-run’, is found with the
fairly transparent meaning of ‘run to meet, meet after a run’, derived from curro ‘run’
and ob ‘against, in front of, facing’ (see ()a); however, it also licenses the special
meaning ‘present itself, occur’ (see ()b):
() Latin; Caes. Gall. , , and Cic. Orat.
a. Ut [...] calones [...] etiam inermes
that soldier’s_servant.NOM.PL even unarmed.NOM.PL
armatis oc-currerent.
armed.DAT.PL against-run.IPFV.SBJV.PL
‘That the soldiers’ servants, although unarmed, ran against the armed men’.
b. Haec tenenda sunt oratori —saepe
this.ACC.N.PL hold.PTCP.FUT.PASS.ACC.N.PL be.PL orator.DAT often
enim oc-currunt.
since against-run.PL
‘These things shall have to be regarded by the orator, since they often
present themselves.’
In Chapter I will argue that predicates headed by verbs like oc-curro ‘against-run’
correspond to a non-externally originated change of location/state. For instance,
()a is analysed as follows:
() Analysis of ()a
vP
v PathP
v curr calones Path’
Path PlaceP
calones Place’
Place ob
20
This is a claim made also for prefixes in the Slavic languages, particularly for so-called internal
prefixes, merged, by assumption, within the vP. See section .. for relevant examples and references.
The present framework
The semantic transparency of ()a is reflected in the analysis of (). The root OB
‘against’ is interpreted as a Terminal Ground, since it is embedded in a PlaceP in turn
embedded within a PathP structure: it depicts the final location of the Figure calones
‘soldiers’ servants’. I assume that the dative armatis ‘armed men’ is a possessive dative
understood as inalienably possessing the region identified by PlaceP—see section
.. for details on the so-called directional dative. The Figure calones raises to Spec-
Path, where it is interpreted as a Measurer for the event: the event is over when all the
calones ‘soldiers’ servants’ end up in front of the armed men (armatis). The predicate
is unaccusative, since Voice is not projected. To v is adjoined the root CURR, which
specifies the way in which the change of location takes place (running). On the other
hand, ()b is not less transparent than ()a, and it receives a similar analysis:
() Analysis of ()b
vP
v PathP
v curr haec Path’
Path PlaceP
haec Place’
Place ob
I claim that the structural semantics of the verb in ()b is the very same as that in
()a. It could not be otherwise, since the meaning inherent to syntactic configur-
ations simply cannot be overridden. Both describe a telic change of state/location.
However, since the roots CURR ‘run’ and OB ‘against’ find themselves within the same
local domain for interpretation, they can trigger special meanings for each other. In
particular, the Encyclopaedic entries of both CURR and OB specify that a special
metaphorical meaning may be triggered in the presence of each other. Possibly
CURR ‘run’ is bleached out into conveying something like suddenness, while OB
‘against’ is reduced to a deictic marker. The Encyclopaedia need not specify the
extension of the domain within which that special meaning may be triggered: that is
provided by the syntax. Specifically, both roots are ‘visible’ to each other if and only if
they fall within the same Spell-Out domain.21
21
Interestingly, as pointed out by García Hernández (: –, : ), among others, the
conceptual meaning of verbal prefixes in Latin is different—more conservative—from that of their
corresponding prepositions. Thus, while the prefix de- conveys a downward orientation (García
Hernández , ) the preposition de can be paraphrased as ‘from’ or ‘away from’. This fact would
be easily accommodated in a theory in which the prefix and the verb—and not the verb and the
preposition—find themselves in an environment that is local enough for them to trigger a special meaning
in each other.
A neo-constructionist perspective on argument structure
. Summary
In this chapter I have made explicit my assumptions on the nature of the lexicon-
syntax interface. I have begun by introducing a fundamental distinction between
endo-skeletal and exo-skeletal theories. The former propose that the syntactic and
semantic properties of linguistic expressions are but a projection of lexical items,
while for the latter they emerge, largely, from the structure itself, lexical items being
reduced to conveyors of grammatically opaque, encyclopaedic content. After intro-
ducing the seminal work of Hale and Keyser I have revised three neo-constructionist
models: Mateu’s () theory of the relational syntax and semantics of argument
structure, Borer’s (b) syntactic theory of event structure, and the DM version of
the Minimalist Program for the architecture of grammar. I have then presented a
neo-constructionist model in which argument/event structure configurations are
created in the syntax through the application of free Merge. Structure is created on
the functional heads v, Place, and Path. Roots and DPs are merged in argumental
positions, a circumstance derived from an abandonment of the l-/s-syntax distinction
of the Halekeyserian model. Roots and DPs receive an argumental interpretation
according to the position that they occupy in the structure. Crucially, roots cannot
project structure, unlike some implementations of the DM model. As in any other
Minimalist account, the structures generated by the syntax are interpreted at the
interfaces. As far as semantic interpretation is concerned, I have emphasized the
distinction between structural semantics, emerging from the structure, and encyclo-
paedic semantics, encapsulated in the roots. I have also paid attention to the
aspectual interpretation of configurations, establishing that a Path projection is
responsible for a telic interpretation of the event if a quantity DP is merged at its
specifier.
3
The syntax-morphology interface
In this chapter I outline a theory of the syntax-morphology interface based on
the architecture proposed in DM (Embick and Noyer ; Embick ;
Marantz ). I adopt the DM view that the morphological dimension of
linguistic expressions is construed on the basis of a previously built syntactic
representation, and that these two representations are, by default, isomorphic
(Embick and Noyer ). However, it is, of course, well-known, and corres-
pondingly observed within the DM tradition (Halle and Marantz ), that
syntax/morphology mismatches do occur and that, hence, the interface can be
non-isomorphic or non-transparent. With respect to such cases of mismatch,
Embick and Noyer (: ) point out that
one of the primary tasks of morphological theory is to identify the set of PF operations
that are responsible for these deviations from the default case. Although this option
weakens the theory by allowing PF to alter syntactic structures, it does so in a way that
maintains the most direct possible correspondence between syntactic and morphological
(i.e. PF) structures.
The PF branch of grammar consists of a series of operations that may generate
the mentioned lack of isomorphism between the morphological representation
and the syntactic representation delivered at LF. Finally, in a departure from
what is standardly assumed in DM, and coming closer to theories like Borer’s
(a, b, ) or Nanosyntax (Fábregas ; Ramchand ; Ramc-
hand and Svenonius ), I assume that although all syntactic representations
are available universally, particular languages may not have the means to inter-
pret some of these representations at PF, which gives rise to cross-linguistic
variation.
. Words and structure
It is often taken for granted that words, as units that can be pronounced in isolation,
are the atoms of syntactic computation. But the most superficial look at the relation
between so defined phonological words and the units assumed as syntactic atoms
The Morphosyntax of Transitions. First edition. Víctor Acedo-Matellán.
© Víctor Acedo-Matellán . First published by Oxford University Press.
The syntax-morphology interface
tells us otherwise. For instance, as shown in (), the Latin conjunction -que ‘and’
encliticizes to the word on its left and triggers stress shift, revealing that the whole
string is behaving like a phonological word:
() Latin; Nespor and Vogel (: ), in Julien (: )
virum [ˈwi:ɾum] / virum=que [wi:ˈɾumkwe]
man.ACC.SG man.ACC.SG=and
Thus, virumque behaves prosodically in exactly the same way as any other word of
more than two syllables where the penultimate syllable is heavy. However, on no
sound syntactic account could -que and the host be analysed as one and the same
syntactic atom. Out of the domain of clitics, situations exist where arguably the same
components can be found within a phonological word or distributed in different
phonological words, depending on the context, as those italicized in the following
pairs of sentences:
() Marantz (), in Newell (: )
a. John cried.
b. Did John cry?
() Marantz (), in Newell (: )
a. John is bigger.
b. John is more intelligent.
() Marantz (), in Newell (: )
a. John took a leap.
b. John leapt.
These are some of the very numerous cases of the indirect relation between
prosodic words and syntactic atoms. In this vein, I defend the view that phrases
interact syntactically and semantically with sub-word units, in consonance with
one of the main tenets of DM: Syntactic Hierarchical Structure All the Way
Down. In Harley and Noyer’s (: ) words, this tenet ‘entails that elements
within syntax and within morphology enter into the same types of constituent
structures (such as can be diagrammed through binary branching trees)’. As
was claimed in Chapter , DPs and roots may both occupy argumental posi-
tions in the structure. Observe the predicates in (). It is argued that they
correspond to the same configuration and, hence, yield the same structural
semantics:
() Latin and English
a. Marcus ex-iit.
Marcus out-go.PRF.SG
‘Marcus went out.’
Late insertion and the nature of roots
b. Marcus went out.
vP
v (= i/GO) PathP
Marcus Path’
Path = to PlaceP
Marcus Place’
Place ex/out
Specifically, the same predicative relation is claimed to hold between the unaccusative
subject Marcus and the locative pieces ex- and out. However, the morphophono-
logical packaging of the material is different in ()a and ()b: while the sequence
ex- ends up prefixed to the verb in Latin, its English counterpart out remains an
independent word in English. These facts support a view in which words are the
result of a variety of packaging mechanisms at PF operating on the representation
yielded by the syntax. Since the application of these mechanisms meets morpho-
phonological requirements of the nodes, cross-linguistic variation can be reduced to
how those nodes are phonologically specified.
. Late insertion and the nature of roots
One of the tenets of DM is that the phonological information is not present during
syntactic computation: this information is lately inserted after Spell-Out, the moment
when the representation is shipped off to the interfaces. That this is desirable for
functional material is supported by the fact that the phonological shape of functional
heads is highly sensitive to syntactic properties (see the seminal work of Bonet
for Catalan pronominal clitics) and the fact that the formal variation is sometimes
too dramatic to be handled with readjustment rules operating on early inserted
material—as is the case with suppletive allomorphy.
It has been debated whether roots are also subject to Late Insertion. Embick
() provides evidence from Latin that roots should be early inserted, that is,
that the choice of root should be made during the syntactic computation (see also
Halle ). In particular, Embick () shows that the choice of root for Latin
verbs determines aspects of their morphosyntax in the perfect tenses: while the
majority of verbs present synthetic forms for the perfect (see ()a), so-called
deponent verbs, that is, verbs that are morphologically passive notwithstanding
their active interpretation, present analytic forms for that tense, composed of a
past participle (agreeing in phi-features with the subject) and a form of the verb
sum ‘be’ (see ()b):
The syntax-morphology interface
() Latin
a. amo ‘I love’ / amavi ‘I have loved, I loved’
b. hortor ‘I order’ / hortatus sum ‘I (masc.) have ordered, I ordered’
Embick () further demonstrates that deponency is orthogonal to argument
structure and lexical semantics. Thus, for instance, hortor, in spite of its exclusively
passive morphology, appears in both transitive (see ()a) and passive sentences
(see ()b):
() Latin; Caes. Civ. , , and Varro in Prisc. GL. II, , (in Embick : )
a. Regemque hortatus est, ut [ . . . ] legatos.
king.ACC.SG=and order.PRF.SG.M that ambassador.ACC.PL
ad Achillam mitteret.
to Achilla.ACC send.IPFV.SBJV.SG
‘And he ordered the king to send ambassadors to Achilla.’
b. Ab amicis hortare-tur.
by friend.ABL.PL urge-IPFV.SBJV-PASS.SG
‘He was urged by friends.’
Embick concludes that deponency is an idiosyncratic property, and that, therefore, it
must be ‘arbitrarily associated’ with certain roots. Since the synthetic/analytic dis-
tinction within the perfect tense can be argued to respond to a distinction in syntactic
configuration related to movement of the Asp(ect) head to T, deponency has to be
present in the computation, and the root is necessarily early inserted.
I believe that there are still other reasons to assume that roots are early inserted,
that is, that the choice of root is made before the derivation is shipped off to
the interfaces. Importantly, it is the only way to preserve an inverted Y model of
grammar. Indeed, if roots are inserted into blank l-morpheme nodes after syntax, at
PF, how could the semantic interpretation access it, since it constitutes an independ-
ent branch? In order for the non-compositional meaning of roots to be integrated
within the structural semantics emerging from the syntactic configuration, the
choice of particular roots must have been made before. In this debate it is
crucial to distinguish roots as elements merged in the syntax from their actual
phonological and semantic interpretation. Harley () provides evidence that
the phonological and semantic interpretation of roots depends on their syntactic
context. On the phonological side, they even show suppletion, against the prediction
originally made by Marantz (). Thus, it seems that, while roots must be in the
derivation from early on if the architectural problem pointed out above is to be
solved, they must also be lately inserted at the interfaces. Harley, following work by
Pfau () and Acquaviva (), proposes to meet these requirements by positing
that roots are early merged as abstract indexes, devoid of both phonological and
semantic features. It is at the interfaces where a particular index, say , triggers a
Cyclic Spell-Out
particular phonological and semantic interpretation according to the syntactic con-
text in which it is found. I adopt this position, and will represent roots with small
capital letters (CAT) to distinguish them from their exponents, which I represent as
the correspondent written form (cat).1
. Cyclic Spell-Out
One of the most important issues dealt with from the inception of DM—and, of
course, in previous theories of morphology (Siegel ; Allen ; Kiparsky ,
among others)—is the investigation of the domains in which specific morphophono-
logical effects are triggered (Marantz , , , ; Newell ; Embick
, among others). Examination of patterns of allomorphy and other phonological
phenomena (and also of allosemy—Arad ; Marantz ) has shown that
linguistic expressions are phonologically computed in chunks of a particular size.
Specifically, an element contained within one domain D cannot trigger allomorphy
on an element contained in a different domain D. For instance, it can be argued on
interpretive grounds (see Borer : for extensive discussion), that gerund-
forming -ing attaches to a vP, which counts as a cyclic domain. This is the reason why
-ing cannot trigger allomorphy in the root inside the vP or vice versa. The suffix -ing
contrasts with other nominalizing suffixes attaching directly to the root. In these
cases allomorphy of the root and of the affix is indeed attested:
() Outer and inner nominalizers in English (Embick : )
a. [nP [vP laugh] ing], [nP [vP marry] ing], [nP [vP destroy] ing]
b. [nP laugh-ter], [nP marri-age], [nP destruct-ion]
Determining the size and the nature of the domains in which special form and
meaning are triggered is a matter of debate (see, for instance, Borer and
Marantz ). Since DM is based on the idea that morphophonology is an inter-
pretation of the syntactic output, it is only natural that the determination of the
domains for special phonological effects is attributed to syntax itself. Since the advent
of phase theory in the Minimalist Program (Chomsky , ) DM theoreticians
have endeavoured to use phases as the Spell-Out domains they needed. Phases are, by
definition units of Spell-Out, that is, stretches of structure shipped off to the
interfaces. Following standard assumptions (Embick , Marantz ), I will
take the heads v, n, and a to define phases, that is, to be cyclic heads in Embick’s
() terminology. I will also consider DP as a cyclic domain (Embick ).
1
See Borer () for arguments against the indexical, non-phonological notation of roots. See De
Belder (); Acedo-Matellán and Real-Puigdollers (); and De Belder and van Craenenbroeck ()
for late-insertion theories of roots.
The syntax-morphology interface
Following Embick (: ff.), I assume that a cyclic domain headed by cyclic
node x is composed of x, the complement of x, and the set of all non-cyclic heads
(W, Z . . . ) between x and the immediately higher cyclic head y (I represent cyclic
domains within braces):
() y {W Z x COMPL}
According to the same author, Spell-Out proceeds in the following fashion: when
cyclic head y is merged, the lower cyclic domain headed by x is spelled out. For
instance, any DP (a cyclic domain) within the complement of v (a cyclic head) is
spelled out upon merger of v. This is the reason why DPs are phonologically
processed independently of the rest of the material in the vP.
. Operations at PF
.. Morphological Merger
Within DM a range of operations have been proposed to account for syntax-
morphology mismatches. Here I will concentrate on Morphological Merger
(Marantz , ). Marantz (: ) defines Morphological Merger as follows:
‘at any level of syntactic analysis (D-Structure, S-Structure, phonological structure), a
relation between X and Y may be replaced by (expressed by) the affixation of the
lexical head of X to the lexical head of Y’. This operation aimed at capturing cases
where morphemes appear dislocated from the position in which they are interpreted:
clitics, inflectional morphemes like causative affixes, or elements in words showing
so-called bracketing paradoxes (Pesetsky ). Embick and Noyer (, )
propose that Morphological Merger translates into somehow different ‘varieties’
depending on whether it applies before or after Vocabulary Insertion, that is, before
or after the insertion of the exponents of the linguistic expression. Assuming a Late
Linearization Hypothesis (Embick and Noyer : ), that is, assuming that the
set of terminal heads are provided with linear order and adjacency at Vocabulary
Insertion, it turns out that Morphological Merger is relativized with respect to
linearization itself: before Vocabulary Insertion it affects terminals that bear a purely
structural relation and after Vocabulary Insertion it affects terminals that bear a
linear and adjacency relation. In this work I will refer to the ‘variety’ of Morpho-
logical Merger that takes place before Vocabulary Insertion. An illustration thereof is
Embick and Noyer’s treatment of the presence of inflection (T) in the verb in English.
Independent evidence shows that the verb does not move to T in this language, in
overt syntax. To account for the surface adjacency of T and v, Embick and Noyer
(: ) propose that Morphological Merger lowers T to v, which is the head of
T’s sister, vP:
() Mary [TP t [vP loudly play-ed the trumpet.]]
Operations at PF
This example shows that Morphological Merger, when applying before Vocabulary
Insertion, is only sensitive to structural contiguity and not to linear adjacency:
although the vP adjunct loudly linearly intervenes between T and v, it does not
block Morphological Merger. However, when a different head appears between the
two relevant heads, an intervention effect does block Merger (Embick and Noyer
: , footnote ). Thus, if NegP projects between T and vP, T cannot move to
v, and do-insertion must take place in order to license the T node:
() *Mary [TP t [NegP not [vP play-ed the trumpet.]]]
() Mary [TP did [NegP ’nt [vP play the trumpet.]]]
Both in Marantz () and in Embick and Noyer (, ), Morphological
Merger is illustrated with cases in which, given a head X and its phrasal sister YP, X is
affixed onto Y, the head of YP, as illustrated above. This is called Lowering in Embick
and Noyer (, ):
() Lowering of X0 to Y0 (Embick and Noyer : )
[XP X0 . . . [YP . . . Y0 . . . ]] v [XP . . . [YP . . . [Y Y0þX0] . . . ]]
However, in Marantz’s original formulation of Merger there is no restriction as to the
‘upward’ or ‘downward’ sense in which Merger should operate. By virtue of this
formulation the relationship between XP and YP could very well be traded by the
raising of Y to X, that is, by ‘PF head movement’:
() PF Raising of Y0 to X0
[XP X0 . . . [YP . . . Y0 . . . ]] ! [XP [Y Y0þX0] . . . [YP . . . ]]
In this work I will adopt Raising as a variety of Morphological Merger before
Vocabulary Insertion.2 Raising can be illustrated with the derivation of a verb of
change of state like Catalan aparèixer ‘appear’. After Spell-Out, there is a successive
movement of the terminal nodes, beginning with the root, up to v (I am ignoring here
the functional heads above v):
2
Assuming Raising at PF evidently brings us to the still unsettled debate of whether there is also head
movement in the syntax proper. Chomsky () has argued that there is not, mostly on theoretical
grounds: head movement does not respect the Extension Condition (Chomsky ), that is, it does not
target the root of the derivation. Moreover, it does not seem to trigger semantic effects, which I find is the
most compelling empirical argument against the view that head movement takes place in the syntax. See
Matushansky (); Roberts (); and Bauke () for arguments in favour of head movement having
to take place in the syntax.
The syntax-morphology interface
() Catalan; PF-derivation of aparèixer ‘appear’
a. vP after Spell-Out
vP
v PathP
Path’
Path Place
Place
Place apareix
b. Successive Raising to v
v
Path v
Place Path
apareix Place
c. Linearization
APAREIX-Place-Path-v
d. Vocabulary Insertion
apareix-∅-∅-∅
See section .. for details on linearization and Vocabulary Insertion.
.. Linearization and Vocabulary Insertion. Exponent-defectiveness
and PF crash
The structure yielded by syntax and by any Morphological Merger operation must be
flattened out at some point in order to be pronounced (Marantz ; Embick and
Noyer ; Embick ). Following Embick () I assume that linearization of
the terminal nodes takes place before the insertion of their exponents (see Arregi and
Nevins , for relevant argumentation). I represent the relation ‘linearly precedes’
through the symbol ‘>’, although I will use ‘-’ to make clear that the linear precedence
is between elements within one complex word:
() [a [b [c d]]] ! a > b > c > d
The assumption that Vocabulary Insertion follows linearization predicts that the
instructions governing the former mechanism, that is, Vocabulary Items, cannot
make reference to structure, but to linear adjacency. Thus, if a Vocabulary Item
corresponding to a functional node F, with exponent e, has an insertion frame, it can
specify at most what is to precede or follow F:
Operations at PF
() Possible shapes of a Vocabulary Item with insertion frame
a. F ⟷ e / C > _
b. F ⟷ e / _ > C
However, the extent of C, that is, the extent of the stretch to which a Vocabulary Item
can make reference in the insertion frame is currently debated. While Embick ()
argues for a condition of strict adjacency between terminal nodes, such that C in the
above schemas can only correspond to a terminal node, Merchant (), following
Bye and Svenonius (), provides data from Greek and English that a non-adjacent
node can also condition the exponence of F, if and only if it finds itself within a span
adjacent to F (Bye and Svenonius ), that is, an adjacent stretch of contiguous
terminal nodes (a > b > c) within the same extended projection:
() F⟷e/_>a>b>c
In this work I provide data suggesting that a Vocabulary Item can specify adjacency
of the node either to a span or to a node, that is, that the condition of adjacency can
be more or less strict according to the Vocabulary Item in different languages.
As standardly assumed, Vocabulary Insertion is regulated by the Subset Principle
(Halle , Halle and Marantz ), which establishes that, given a node F to which
more than one Vocabulary Item refers, only the Item specified with the biggest subset of
the features of F will insert its exponent in F. DM authors usually devise, for a given node
F, an elsewhere Vocabulary Item, with no insertion frame, that executes insertion in a
context where no other more specific exponent can be inserted. This aims at accounting
for the unmarked realization of a given morpheme. For instance, as illustrated in Embick
(: ), a past Tense node in English is by default realized as -d. This is captured by the
following elsewhere Vocabulary Item, with no insertion frame:
() T[past] ⟷ -d [cf. wait-ed, stay-ed, etc.]
The above exponent will be inserted in all contexts in which the following ones, by
virtue of their insertion frames, cannot be inserted:
() T[past] ⟷ -t / _ {LEAVE, BEND, . . . } [cf. lef-t, ben-t, etc.]
() T[past] ⟷ -∅ / _ {HIT, SING, . . . } [cf. hit-∅, sang-∅, etc.]
However, in this study I consider and illustrate the possibility that some functional
node F may not have an elsewhere Vocabulary Item, that is, an exponent insertable
where no other exponent is. Such a node F could be argued to be exponent-defective, in
that its exponence is systematically dependent on the association of F with another
functional node, as encoded in the insertion frame of the Vocabulary Item(s) of F. As
one might expect, it might be the case that particular structures do not provide the
contextual conditions for an exponent-defective node to be realized, which could lead
to a ‘PF crash’, that is, to a failure to interpret morphologically the linguistic expression
The syntax-morphology interface
at hand. In later chapters I provide case studies of such failure in the abovementioned
languages. I will argue that the functional node Path in languages like Romance
languages, Latin, Slavic, and Ancient Greek is exponent-defective in this sense. For
instance, in Romance languages the Path node has a limited set of Vocabulary Items all
of which establish, as a condition for the insertion of the exponent, that Path be strictly
adjacent to the node v (e represents here different exponents):
() Path ⟷ e / _-v
In these languages Path does not have an elsewhere exponent, i.e., an exponent
insertable independently of any insertion frame, as it does in English (i.e., to). Hence,
if Path is not adjacent to the node v when Vocabulary Insertion takes place, it will not
receive any exponent, leading to a failure in the interpretation of the structure. The
theory developed here is, therefore, a crashing theory. Importantly, contemplating
crashes at PF, that is, the possibility that some structures built by the syntax cannot be
lexicalized, is not a hallmark of DM, in which, as pointed out above, for a given node
F there is always a default Vocabulary Item that can insert an exponent (frequently a
null one) in F where no other can. However, something like a PF crash is assumed in
theories like Borer’s (a, b, ), in which it is argued that cross-linguistic
variation is also reduced to the (un)availability of particular phonological represen-
tations of functional nodes or combinations thereof (see, for instance, Borer b:
). Similarly, in Nanosyntax, it is assumed that languages only differ in what lexical
representations they have available for a particular freely built syntactic representa-
tion (Fábregas ; Ramchand and Svenonius ). The notion of PF crash based
on an uninterpretable exponent-defective node can be modelled through Fábregas’s
(: ) Exhaustive Lexicalisation Principle, which states that ‘Every syntactic
feature must be lexicalised.’
.. Interaction between Raising and Vocabulary Insertion.
Cross-linguistic variation
Embick and Noyer (, ) propose that the operations of the PF-branch of the
derivation respect a sequence based mainly on whether they apply before or after
Vocabulary Insertion. The rationale behind this hypothesis is that there are oper-
ations not sensitive to phonological material and that must therefore apply before
Vocabulary Insertion and there are operations sensitive to phonological material that
must apply after Vocabulary Insertion. I assume the cartography shown in (),
partly illustrated in the last section by ():
() A (partial) cartography for PF
a. Raising
b. Linearization
c. Vocabulary Insertion
Operations at PF
The operations before Vocabulary Insertion are in fact sensitive to hierarchical
configuration, rather than to linear precedence or phonological properties. Raising
brings a head to the immediately higher head, forming a complex head therewith.
Afterwards, Linearization flattens out the structure, establishing a precedence rela-
tion among the set of terminal nodes. Finally, Vocabulary Insertion introduces
exponents into the nodes according to the featural and contextual specifications of
their Vocabulary Items.
Operations before Vocabulary Insertion (Raising, here) do not operate necessarily
in every language or for every derivational output. Rather, I take them to operate
freely, their outputs being filtered out during Vocabulary Insertion according to
specific properties of Vocabulary Items in specific languages (or classes of languages).
Thus, the specification of Vocabulary Items derives cross-linguistic variation within
the domain of argument structure from a strictly morphological point of view. For
instance, as claimed above, in Romance the Path head only has Vocabulary Items
establishing a strict linear adjacency between Path and v:
() Path ⟷ e / _-v
According to the insertion frame of this Vocabulary Item, at the time of Vocabulary
Insertion, Path must be strictly left-adjacent to v, that is, a prefix. This can be brought
about through an instance of Raising bringing Path onto v before Vocabulary
Insertion:
() Path-to-v Raising
vP
v PathP
Path v PlaceP
Linearization takes place after Raising, effecting the sequence Path-v. Importantly, if
Raising does not take place, or if some overt node linearly intervenes between Path
and v, Vocabulary Insertion into Path cannot be effected, since the contextual
requirement is not met and Path is exponent-defective, that is, it does not have an
elsewhere Vocabulary Item to rescue its phonological interpretation. The realization
of Path in these languages (and also in Basque, Japanese, Hebrew, etc.) seems to be
intimately dependent on that of v. I argue that this is the reason behind the well-
known awkwardness of predicates like the following one in Romance (Talmy ,
; Mateu and Rigau , among many others):
() Catalan
*Ella ballà a l’habitació. (In the directional sense.)
she danced at the=room
The syntax-morphology interface
() Catalan; PF derivation of ballar a (l’habitació) ‘dance into the room’
a. vP after Spell-Out
[vP [v v BALL] [PathP [Path’ Path [PlaceP [Place’ Place {}]]]]]
b. Successive Raising to v
[vP [v Path [v BALL v]] [PathP [Path’ [PlaceP [Place’ Place {}]]]]]
c. Linearization
Path-BALL-v > Place
d. Vocabulary Insertion
?-ball-∅ > a
In the above predicate, the root BALL ‘dance’ is sister to v in a complex v head,
providing the specification of a co-event (in this case, a dancing event accom-
panies the main event of getting into the room; see section ...). I assume, as
does Embick (), that roots raise to the immediately upper functional category
(see section . for more extended discussion). This implies that after Path has
raised to v and the whole structure has been linearized, the root linearly inter-
venes between Path and v. Importantly, as argued by Embick () and Marantz
(), an element does not linearly intervene when phonologically null. It is not
the case here, however, since the root BALL is phonologically overt. As a result,
the contextual specification of Path is not met, and PF fails to interpret the
structure.
A similar, though not identical effect can be found in languages like Latin, Slavic,
or Ancient Greek. In these languages, too, the Path head is exponent-defective,
although in a different way. I assume that one of the Vocabulary Items for Path in
these languages reads as follows (see section . for a revision of this Vocabulary
Item):
() Path ⟷ ∅ / Place-_-[ . . . v . . . ]
The above Vocabulary Item establishes that Path has to be right-adjacent to
Place and left-adjacent to a span containing v, that is, a (trivial or not) stretch of
contiguous nodes containing v. Provided there is no other Vocabulary Item for
Path in these languages, the specification of the above Vocabulary Item has the
effect of filtering out configurations in which the Path head is not prefixed onto
v or a span containing v and where the Place head is also not prefixed onto Path.
However, the strict adjacency requirement imposed on Path and v in Romance
is not active in these languages. Thus, for instance, these languages admit
directed motion constructions with a Manner Co-event like those banned in
Romance, as long as the specification of the final location appears prefixed onto
the verb:
Operations at PF
() Latin; PF derivation of ad-equito “at-ride” ‘ride up to’
a. vP after Spell-Out
vP
v PathP
v equit Path’
Path PlaceP
Place’
Place ad
b. Raising
v
Path v
Plac Path equit v
ad Place
c. Linearization
AD-Place-Path-EQUIT-v
d. Vocabulary Insertion3
ad-∅-∅-equit-∅
The raising of the terminal nodes onto v, to form a complex head, provokes the
prefixation effect typical of languages like Latin or Slavic, and this will be amply
discussed in later chapters. Importantly, as will be shown in Chapter , what is not
found in these languages is equivalent constructions in which the final location is not
prefixed onto the verb:
() Latin; made-up example
#Caesar ad portas equitavit.
Caesar.NOM at gates.ACC ride.PRF.SG
‘Caesar rides towards the gates.’ / *‘Caesar rides up to the gates.’
The above example is not ungrammatical in Latin, but it does not convey the resultative
reading in which Caesar actually arrives at the gates (on horseback)—see Van der Heyde
() for an early observation along these lines. At most, the PP ad portas ‘towards the
3
Here and elsewhere I ignore thematic vowels (here a) in Latin and Romance. Following Oltra-Massuet
() and Oltra-Massuet and Arregi (), I assume that thematic vowels are introduced postsyntacti-
cally (see Embick for the notion of dissociated morpheme). The issue is tangential to the current
discussion.
The syntax-morphology interface
gates’ is interpreted as a directional adjunct, not as entailing the final location. Thus, this
example cannot involve the structure in ()a. The reason for the failure of the
resultative interpretation, I argue, is that the prefixation requirement on Path, described
above, cannot be met. The only derivation compatible with the sentence is one in which
ad portas ‘towards the gates’ is an adjunct to an activity vP involving the root EQUIT ‘ride’:
() Ad portas equitare ‘ride towards the doors’.
vP
pP vP
p portas v equit
p ad
Since Raising is defined in structural terms, and applying from one head to the
immediately higher c-commanding head, it cannot target anything within the
adjunct PP. This in turn explains why in motion constructions involving a prefixed
verb, such as that in (), the adjunct reading of the prefix (in this case a ‘towards’
reading) is out, and the only interpretation possible is that of a change of location.
In Chapters , , and we will have more opportunities to see how the morpho-
logical interpretation of syntactic structures gives rise to observable systematic cross-
linguistic differences within the realm of argument structure.
. Summary
In this chapter I have presented a (partial) theory of the PF-interpretation of linguistic
structure. It is this interpretation, I argue, that cross-linguistic differences are restricted
to. I have assumed that, unlike the syntax-semantics interface, the syntax-morphology
interface can be non-isomorphic. PF computation operates in cycles defined by the
categorizers like v and a. Importantly for the main discussion in Chapter , the material
computed in a given cycle is not available for another cycle. I have discussed a series of
PF operations. A Raising operation, taking place before Vocabulary Insertion, brings
the nodes together, yielding complex words. This operation applies freely, its results
being filtered by the exponents available at Vocabulary Insertion, which takes place
after linearization. Crucially, I have argued that some functional heads in certain
languages are exponent-defective, that is, they do not count with an elsewhere
Vocabulary Item when the rest of Vocabulary Items do not meet the contextual
conditions of insertion. In these cases, if Vocabulary Insertion cannot take place, the
derivation crashes at PF. Differences in the specification of Vocabulary Items for
functional heads are then at the basis of the explanation of patterns of cross-linguistic
variation in argument structure, as will be shown in subsequent chapters.
4
Latin as a satellite-framed language
In this chapter I use the theoretical tools introduced in Chapters and to analyse a
wide range of argument structure phenomena in Latin. A quick glance at the
Dictionnaire Latin-Français by Gaffiot () shows that many composite verbal
lexical entries in Latin receive a periphrastic definition in French. Importantly, the
correspondence between the morphological components of the Latin verb and the
syntactic components in the Romance periphrasis appears to be systematic. The
following entries involving the prefix ex- ‘out’ illustrate the fact:
() Latin; Gaffiot ()
a. ex-cutio
out-shake.SG
‘Faire sortir ou tomber en secouant’ (‘make go out or fall shaking’)
b. ex-cudo
out-beat.SG
‘faire sortir en frappant’ (‘make go out beating’)
c. e-repo
out-crawl.SG
‘sortir en rampant, en se traînant’ (‘go out crawling’)
In the above examples, the prefix ex- ‘out’ (with the form e- in erepo), seems to
correspond, in the French translation, to a whole verb, namely (faire) sortir ‘(make)
go out’, while the semantic content of the simple verb in each case is translated as a
manner adverbial (en secouant ‘shaking’; en frappant ‘beating’; en rampant, en se
traînant ‘crawling’). For motion events in general, while Latin expresses the trajec-
tory and final location within one morpheme and the ‘kind’ of motion—shaking,
beating, and crawling, respectively, in ()a to ()c—within a different morpheme
(namely, the verb: quatio ‘shake’, cudo ‘beat’, repo ‘crawl’), French lexicalizes the
trajectory and final location in the form of an independent and monomorphemic
verb—as sortir ‘go out’, entrer ‘go in’, etc.—and the kind of motion is conveyed by an
optional adjunct. This difference in the expression of the components of a (motion)
event shown by Latin and French actually corresponds to a typological difference
The Morphosyntax of Transitions. First edition. Víctor Acedo-Matellán.
© Víctor Acedo-Matellán . First published by Oxford University Press.
Latin as a satellite-framed language
claimed by Talmy () to divide many of the world’s languages into two blocks:
satellite-framed languages (Latin-like languages) and verb-framed languages (French-
like languages). After introducing Talmy’s insightful observations on the cross-
linguistic expression of events of change, I model his theory in terms of the one
put forward in Chapter . It will be argued that cross-linguistic differences are purely
morphological and, as such, derive from the language-specific morphological speci-
fication in the Vocabulary Items of functional nodes. In section . I describe the
possible morphosyntactic manifestations of PathP in Latin. The bulk of the chapter is
devoted to showing the validity of Talmy’s (: ) observation that Latin is a
satellite-framed language. I explore and analyse, to that end, a set of constructions
involving change or transition (in my terms, a PathP).
. Talmy’s () theory of change events and its adaptation
to the present framework
.. Talmy’s theory of (motion) events
Talmy (: ff.), in a revision and expansion of his earlier, highly influential
work (Talmy , ) on the relation between meaning and surface form in the
expression of events, proposes that any motion event has a semantic structure
integrating a set of distinct components. I will illustrate this view with the following
sentences:
() The cat walks into the basket.
() There stood a cat in the basket.
In each of these sentences there is something that moves or is stationary: the cat. This is
the Figure component. The object that is taken as a reference for the movement
or stationariness of the Figure is the Ground, here the hat, in both sentences. The
Figure and the Ground are spatially related to each other by the Path component,
which in () is expressed by (in)to and in () is expressed by in. Lastly, the Motion
component is encoded, in the above sentences, in walks and stood, respectively. In the
former case it is typed as movement proper, MOVE, while in the latter it conveys
stationariness and is typed as BEAT. Importantly, Talmy considers that the core part of
the motion event (the one that distinguishes different events) lies in either the Path
alone or the Path together with the Ground. This is what he calls the Core Schema.
Talmy (, ) further breaks the Path component down into a Vector
subcomponent, a Conformation subcomponent, and a Deictic subcomponent.
The Vector expresses the sense in which the relation between Figure and Ground
is established. The types of Vector are given the names of certain abstract pre-
positions: such as AT, which specifies a contact relation between the Figure and the
Ground, TO, which specifies that the sense is towards the Ground, FROM, which
Talmy’s (2000) theory of change events
specifies that the Ground is the starting point of movement, VIA, which signifies that
the Ground is something located in the Path, but which is neither the starting point
nor the end point, etc. In () the Vector is TO, and is codified in the -to morph of
into, while in () the Vector is AT, and lies in the preposition in.
The Conformation creates a geometrical shaping of the Ground, which comes then
to be conceptualized as a volume, an enclosure, a plane, etc. The conformation in
both () and () is the one corresponding to an enclosure, and could be paraphrased
as INSIDE. Note that, in both cases, it is expressed by the preposition in, which in ()
encodes, in addition, the Vector AT, and in () is morphologically attached to the TO
Vector encoder -to. A volume conformation, which we could dub SURFACE, applied
to the same motion event, could yield The cat walks onto the hat and There stood a cat
on the hat, respectively. I will argue that the different interpretation of the Conform-
ation is not grammatically represented, and that it is instead encoded as a choice of
different roots.
The Deictic component conveys whether the sense of the Path is towards the
speaker or away from the speaker. The verbs to come and to go exemplify, respect-
ively, a þSPEAKER (towards the speaker) and a -SPEAKER Deictic.1
A last important element must be mentioned, which, although not itself a com-
ponent of the motion event, is very often associated with it. It is what Talmy calls the
Co-event, that is, an event that is related in some way to the Motion event, which is
considered, in turn, the Framing event. That relation can be of different types:
causation, manner, etc. In the case of () and (), the Co-event expresses manner,
more specifically, the way in which the movement or the stationariness takes place,
a walking event in () and a standing event in (). Note that in both sentences this
Co-event is expressed via the verb (the root of the verb), together with the Motion
component, MOVE and BEAT, respectively.
Having put forward the main elements involved in the expression of motion,
I now want to introduce the major cross-linguistic difference referred to in the
introduction to this chapter. Talmy (, : –) proposes that languages can
be ascribed to groups in which there is a systematic encodement, in a single
morpheme, of the same components of a motion event. Specifically, he focuses on
the Core Schema, and describes two possibilities as to its surface (syntactic) expres-
sion: the Core Schema can be expressed within the verb, conflated—that is, fused
into the same piece—with the motion component, or it can be expressed through
an independent element of the predicate that he calls satellite, ‘the grammatical
category of any constituent other than a noun phrase or prepositional-phrase
complement that is in a sister relation to the verb root’ (Talmy : –).
Languages that primarily opt for the first way of encoding the Core Schema are
1
The technical names INSIDE, SURFACE, þSPEAKER, and -SPEAKER are creations of my own
(Talmy : refers to þSPEAKER as hither and to -SPEAKER as hence).
Latin as a satellite-framed language
called verb-framed languages, while languages that choose the second way are called
satellite-framed languages.2 What is of relevance to the present discussion is that
there is a kind of complementary distribution between the expression of the Core
Schema and the expression of the Co-event, such that in v-framed languages the Co-
event is not conflated in the verb, and usually appears in an adjunct phrase, while in
s-framed languages the Co-event can be readily expressed within the verb, as is the
case with the manner Co-event in () and () above. Although we have already seen
how an s-framed language distributes the Core Schema and Co-event components,
in () and (), let us now introduce a minimal pair involving Catalan (a v-framed
language) and English (as pointed out already, an s-framed language) expressing a
motion event with a manner (of motion) Co-event:
() Catalan and English
a. La pilota va [entrar]verb: Motion+Core Schema
the ball PST.SG go_in.INF
[rodolant.]adjunct: Co-event (manner)
rolling
b. The ball [rolled]verb: MotionþCo-event (manner) [in.]satellite: Core Schema
As glossed in the examples, the Catalan sentence expresses the trajectory of the ball
(the Core Schema, here equivalent to a trajectory ending up in some enclosure)
within the verb, while the manner in which it moves along that trajectory is encoded
in an independent and optional gerund phrase. In English, those same components
of the motion event are expressed in a different way: the Core Schema is separated
from the verb and is expressed as a satellite, while the manner Co-event is fused
together with the Motion within the verb. This different morphosyntactic structuring
of the motion event is correlated, as Talmy (: ff.) observes, with certain facts
about the lexicon of each type of language. For instance, Catalan (and, in general,
v-framed languages) has a great variety of roots expressing directed motion at its
disposal, each corresponding to a particular Core Schema component, while English
lacks many of those specialized verbs: Cat. entrar, ‘go in’; sortir, ‘go out’; treure, ‘take
out, off ’; ficar, ‘put in’; etc.
.. Beyond events of motion
As pointed out by Talmy (: ) himself, the s-/v-framed distinction does not
apply exclusively to motion events. In particular, it can be extended to events
expressing change, in general. From this perspective, the Figure is the entity under-
going change, the Core Schema is the actual change of state, with the Ground being
2
See Talmy (: –, –) for another major typological group of languages, namely languages
in which it is the Figure component that gets lexicalized into the verb.
Talmy’s (2000) theory of change events
the final, resultant state, the Motion component is to be identified with the event
itself, and the Co-event is the way in which the change of state takes place. The
examples from German and Spanish following illustrate, respectively, how s-framed
and v-framed languages express events of change of state:
() German and Spanish; Talmy (: )
a. Der Hund hat [den Schuh]Figure [kaputt]Core schema -[gebissen.]Event+Co-event
the dog has the shoe in_pieces bite.PTCP.PST
‘The dog bit the shoe to pieces.’
b. El perro [destrozó]Event+Core schema [el zapato]Figure [a mordiscos.]Co-event
the dog destroy.PRF.SG the shoe to bites
.. An asymmetric difference
As can be shown through a comparison of s-framed English and v-framed Catalan,
the s-/v-framed distinction happens not to be symmetric, that is, it does not yield two
groups of opposing languages. The asymmetry appears to consist in a wider avail-
ability of the v-framed strategy, which is allowed in typically s-framed languages
like English (Mateu ). The s-framed pattern, on the other hand, is precluded in
v-framed languages like Romance. Thus, English does have directional verbs, which,
not surprisingly, are mostly Latinate: to enter, to exit, to remove, etc. It can also
express events of change of state within a verb, as in The wind cleared the sky, The sun
melted the snow, etc. The opposite, however, is not found in v-framed languages:
while they can make use of the v-framed strategy, they cannot make use of the
s-framed strategy. Hence, typically s-framed constructions involving the expression
of a Co-event within the verb are ungrammatical in these languages:
() Catalan
*El vent bufà el cel clar.
the wind blew the sky clear
‘The wind blew the sky clear.’
() Catalan
*En Joan martellejà el metall pla.
the Joan hammered the metal flat
‘Joan hammered the metal flat.’
In section ... I provide a morphological analysis of this asymmetry.
.. Non-dynamic events and the s-/v-framed distinction
Up to now I have restricted my attention to events of transition, which seem to be the
locus of the s-/v-framed cross-linguistic variation. In fact, there is evidence that for
stative events v-framed languages like Romance allow a single morph corresponding
to a BEAT Motion component together with a Co-event. I am referring to predicates
Latin as a satellite-framed language
like the following (see Hoekstra and Mulder for French and Torrego for
Spanish):
() Catalan; Mateu (: )
En aquesta coral n’hi canten molts, de nens.
in this choir PARTVE=LOC sing..PL many.PL of child.PL
‘There are many children who sing in this choir.’
According to Rigau (), in predicates such as (), the verb canten ‘they sing’ bears
an existential stative meaning close to that found in there-existential sentences.
Hence, a good paraphrase for () is the English translation provided underneath.
On the other hand, and according to Mateu (: ff.), there is evidence that
the construction is of unaccusative nature, as hallmarked by the possibility of
en-extraction (see () itself), and the licensing of postverbal bare plural subjects, as
shown below:
() Catalan
En aquesta coral hi canten nens.
in this choir LOC sing.PL children
‘Children sing in this choir.’
It is also telling, in this respect, that in Italian these constructions resist HAVE-
selection when put in the perfect:
() Italian; Centineo (: –), in Mateu (: )
??Ce ne ha nuotato molta, di gente, in quella piscina.
LOC PARTVE has swum many of people in that swimming pool
Importantly, Mateu (: ) highlights Centineo’s (: , footnote ) obser-
vation that, upon elicitation, some native informants attempted to use essere, the BE
auxiliary, in examples like (). I will assume with Mateu () that this type of
construction is unaccusative. I will analyse them as such, and I will explain why they
are fine in v-framed languages in section ....
.. A syntactic interpretation of Talmy’s theory
... Syntactic structuring of events of change When one tries to cast Talmy’s
ideas into a theory such as the one proposed in Chapter , one of the first challenges
is that of selecting as functional elements only those components proposed by
Talmy that seem to be grammatically relevant, and of assigning to others the
status of roots—that is, elements whose content is invisible and irrelevant to gram-
mar. In so doing, the range of the ontology of the components of events is greatly
reduced, deriving many of the nuances from configurational properties. The
Talmy’s (2000) theory of change events
correspondence between the components involved in both theorizations is laid out in
the table below:
() A comparison between Talmy’s proposal and the present one
Components in Talmy’s proposal Interpretation within the present model
MOVE v taking as complement a PathP
Motion
BEAT v taking as complement a PlaceP
Figure Spec-Place
Ground Compl-Place
Dynamic Path Path
Vector —
Subcomponents
Conformation Root adjoined to Place
of Path
Deictic Compl-Place
Non-dynamic Path Place
Co-event Root adjoined to v
Talmy’s MOVE/BEAT distinction is derived from the configuration: while v introduces
the event (motion or otherwise), it is understood as dynamic or stative, respectively,
if v takes a PathP or a PlaceP as complement. Note that Place does not have a
correspondence in Talmy’s theory. The Figure and Ground are, respectively, the
specifier and the complement of the same head, Place, accounting for their predica-
tional relation. As for the Deictic component, I take it to be a certain kind of Ground.
For instance, a verb such as arrive is analysable as a predicate of change of state/
location where the Ground, that is, Compl-Place, is itself a Deictic whose reference
coincides with that of an element already mentioned or with the speaker, by default.3
As to the Conformation and Co-event components, they are treated as roots adjoin-
ing to Place and v, respectively. We saw in section ... that roots can appear as
adjuncts to the functional heads, specifying the kind of event or of locational
predication (if the predication is in fact locative). All these components are repre-
sented in the analysis of the following sentence (I am neglecting movement from
Spec-Place to Spec-Path and the morphological operations that apply at PF—see
section ...):
3
See Bouchard () for a similar analysis of French movement verbs such as venir ‘come’ or aller ‘go’.
Latin as a satellite-framed language
() Sue danced into the room.
vP
v PathP
vMotion danceCo-event Path = to PlaceP
SueFigure Place’
Place the roomGround
Place inConformation
Roots too can be merged as Compl-Place. This is the case of change-of-state
predicates, like () (note that the verb is correspondingly interpreted as change
and not as motion), or motion predicates involving a single verb, like ():
() The sky cleared.
[vP [v vChange ] [PathP Path [PlaceP The skyFigure [Place’ Place CLEARGround ]]]]
() Catalan
En Joan eixí.
the Joan went_out
[vP [v vMotion ] [PathP Path [PlaceP En JoanFigure [Place’ Place EIXGround ]]]]
As for the Vector component, I shall assume that, at least when PathP appears
embedded within vP, it is always of value TO. In that sense the head Path is
significantly different from Talmy’s Path: it instantiates a transition into a final
location or state. In other words, Core Schemas are always goals, and not sources.
In a predicate such as She danced out of the room, hence, out of the room
corresponds to a goal of motion, describing where the dancing event will end
up. There is evidence for this position. For instance, change-of-state predicates
always describe a final, resultant state, and not an initial or medial state. In the
same way, there is no verb in any language that I know of lexicalizing the meaning
‘stop being’, which would correspond exactly to a predicate of change of state
expressing the source state. This is partly illustrated by the following paradigm
from Gehrke (), where turn must appear with a goal PP and cannot appear
with a source PP alone:
() Gehrke (: )
a. The frog turned from green to blue.
b. The frog turned to blue.
c. *The frog turned from green.
Talmy’s (2000) theory of change events
... A morphological account of the s-/v-framed difference The main aim of this
book is to explain cross-linguistic variation as depending on idiosyncratic morpho-
logical properties of functional heads. The s-/v-framed distinction will be dealt with
from this post-syntactic perspective. This means that the syntactic construction
of events of change, which are the locus of the distinction, and their interpretation
at LF are common to all languages, and that it is how those structures are
interpreted morphologically, at PF, which can vary from language to language.
I introduce the discussion in this chapter, although it will be of great importance in
Chapters and .
Observationally, the s-/v-framed distinction has to do with how morphs, in the
structuralist sense of the term, relate to morphemes, as Talmy’s definition of confla-
tion suggests: in s-framed constructions the same morph corresponds to (or con-
flates) the Motion and the Co-event components (here, v and a root adjoined to it,
respectively). In v-framed constructions the same morph corresponds to the Motion
and the Core Schema (here v and PathP). Since we know that s-framed languages
admit the v-framed strategy, but v-framed languages do not admit the s-framed
strategy (see section ..), there has to be a more restrictive mechanism in v-framed
languages than in s-framed ones, accounting for this asymmetry. Using the theoret-
ical tools introduced in Chapter , I propose that in v-framed languages like
Romance, the Path and v must be strictly adjacent to each other, which is accom-
plished through a Vocabulary Item like the following (e stands for exponent):
() Path ⟷ e / _-v
Assuming that Path is exponent-defective in these languages, that is, that it does not
have an elsewhere Vocabulary Item (a Vocabulary Item with no insertion frame), this
head can only receive an exponent if the contextual condition expressed in the above
Vocabulary Item is satisfied. In order to meet the insertion condition, Path must raise
to v in v-framed languages:
() Path-to-v Raising in v-framed languages
vP
v PathP
Path v PlaceP
In the case of predicates of change of state and change of location like Catalan eixir
‘go out’, in which all the nodes of the v-cycle raise to v, the Vocabulary Item for
Path establishes that it is realized with a null exponent when sandwiched between
Place and v:
() Path ⟷ ∅ / Place-_-v
Latin as a satellite-framed language
I illustrate with the derivation of En Joan eixí ‘Joan went out’. I ignore (here and
elsewhere) the terminal nodes higher than v and Voice that also belong in the vP
cycle (T, etc.):
() Catalan; PF-derivation of En Joan eixí
a. Structure delivered by syntax
[vP v [PathP [Path’ Path [PlaceP [Place’ Place EIX]]]]]
b. Raising
[v [Path [Place EIX Place] Path] v]
c. Linearization
EIX-Place-Path-v
d. Vocabulary Insertion
eix-∅-∅-∅
()a is the structure that arrives at PF. Raising forms a complex head out of the root,
Place, Path, and v. After linearization the nodes receive an exponent according to
their Vocabulary Items. I assume that Place and v also receive a null exponent in this
context. In particular, I assume that the following is one of the Vocabulary Items for
Place in v-framed languages:
() Place ⟷ ∅ / _-Path
As for v, its exponence may not be null, as pointed out by Oltra-Massuet ().
Thus, there are cases of predicates of change of state, involving the same structure as
in the above example, in which v is overtly realized:
() Catalan; non-null exponent for v
agud-itz-a-r
acute-ize-TH-INF
‘make more acute’
In these cases, the Vocabulary Item for v specifies an overt exponent in the context of
a particular root:
() v ⟷ itz / AGUD-v
Allomorphy of v can be triggered across Place and Path since these heads, although
linearly intervening, are null (see Embick ; Marantz ).
A slightly different derivation is involved in simple go-predicates involving a PP
like En Joan anà a la botiga ‘Joan went to the shop’:
() Catalan
En Joan anà a la botiga.
the Joan went at the shop
‘Joan went to the shop.’
Talmy’s (2000) theory of change events
a. Structure delivered by syntax
[vP v [PathP [Path’ Path [PlaceP [Place’ Place {}]]]]]
b. Path-to-v Raising
[vP [v Path v] [PathP [Path’ [PlaceP [Place’ Place {}]]]]]
c. Linearization
Path-v > Place
d. Vocabulary Insertion
an-∅ > a
In this case, in which Place does not raise to Path, the latter receives the exponent an
‘go’, according to the following Vocabulary Item:
() Path ⟷ an / _-v
Since Place does not raise to Path, it is endowed with a default exponent a ‘at’ by
virtue of the following Vocabulary Item:
() Place ⟷ a
The preposition a ‘at’ is a pure Place head, without the Conformation component
that is encoded as a root adjoined to Place. That a encodes Place is seen in the
following stative (by hypothesis, Path-less) example:
() Catalan
En Joan és a la botiga.
the Joan is at the shop
‘Joan is at the shop.’
There is evidence that in go-predicates in Romance, unlike in other unaccusative
change-of-state and change-of-location predicates, Place does not raise to the v-
complex head. Thus, it is only in these predicates that suppression of the goal PP,
without replacement by the appropriate pronoun in the languages that possess it
(Cat. hi, Occitan i, French y, Italian ci), is severely ungrammatical:
() Catalan
a. En Joan *(hi) anà.
the Joan there went
‘Joan went there.’
b. En Joan (hi) entrà.
the Joan there went_in
‘Joan went in.’
The first of the examples above shows that the position of PlaceP, represented by Cat.
hi ‘there’, does not form a complex head with the verb.
Importantly, how does the above scenario deal with the fact that v-framed
languages do not allow transition predicates involving a co-event? Recall that within
Latin as a satellite-framed language
the present account (as well as in Embick ; McIntyre ; Zubizarreta and Oh
; or Mateu b), typical s-framed constructions are analysed as involving the
adjunction of a root to v, being interpreted as a Manner Co-event. In v-framed
languages, this adjunction structure is not compatible with the Vocabulary Item
proposed above for Path. Assuming, as in Embick (), that roots must raise to the
upper functional head, the raising of the Co-event root to v, followed by linearization,
yields a sequence in which the root linearly intervenes between Path and v. Since the
root is overt, it blocks insertion of any exponent in Path. I illustrate below with the
derivation of ungrammatical *Ella ballà a l’habitació ‘She danced into the room’:
() Catalan
*Ella ballà a l’habitació. (In the directional sense.)
she danced at the=room
a. Structure delivered by syntax
[vP [v v BALL] [PathP [Path’ Path [PlaceP [Place’ Place {l’habitació}]]]]]
b. Raising
[vP [v Path [v BALL v]] [PathP [Path’ [PlaceP [Place’ Place {}]]]]]
c. Linearization
Path-BALL-v > Place
d. Vocabulary Insertion
?-ball-∅ > a
At Vocabulary Insertion, the strict adjacency condition imposed on Path is not met.
Since Path is exponent-defective, there is no other Vocabulary Item that can interpret
the Path node. This explanation can be extended to cover the disallowance of
complex AP resultative constructions in v-framed languages:
() Catalan; Mateu (: )
*El gos va bordar els pollastres desperts
the dog PST.SG bark.INF the chickens awake.M.PL
In this case the root BORD ‘bark’ provides the Co-event component by virtue of
its being associated with v. At Vocabulary Insertion it too intervenes between
Path and v, blocking the insertion of the exponent for Path. However, in the case
of AP resultative constructions with no independent co-event component (i.e. simple
AP resultative constructions; see section ...), the derivation is predicted to be
possible. Thus, in Romance languages we find AP resultative constructions with
causative verbs with no entailment of a co-event component:
() Catalan
El gos va deixar els pollastres marejats a lladrucs.
the dog PST.SG leave.INF the chickens dizzy.M.PL at barks
‘The dog got the chickens dizzy with its barking.’
Talmy’s (2000) theory of change events
In this case, I assume that the verb deixar ‘leave’ corresponds to a realization of Path
in the context of (active) Voice. Both v and Voice are realized with a null exponent:
() Path ⟷ deix / _-v-Voice
I also assume that the adjective is a PlaceP endowed with phi-features, and corres-
ponds to a PF cycle of its own (see section ..):
() Analysis of ()
a. Structure delivered by syntax
VoiceP
El gos Voice’
Voice vP
v PathP
els pollastres Path’
Path PlaceP
els pollastres Place’
Place marejat
b. Raising
[Voice [v Path v] Voice]
c. Linearization
Path-v-Voice
d. Vocabulary Insertion
deix-∅-∅
In s-framed languages, by contrast, the realization of Path is not necessarily
dependent on that of v. In English, for instance, there is a default Vocabulary Item
for Path, with an overt exponent:
() Path ⟷ to
Thus, a simple motion sentence like John went to the room is derived as follows (I am
ignoring, as usual, the presence of T, which of course triggers the allomorph wen):
() Derivation of John went to the room
a. Structure delivered by syntax
[vP v [PathP [Path’ Path [PlaceP [Place’ Place {}]]]]]
Latin as a satellite-framed language
b. Linearization
v > Path > Place
c. Vocabulary Insertion
wen(-t) > to > ∅
In English there is no strict adjacency condition on Path and v. Path may be realized
independently of v (above, in the absence of any Co-event root, realized as go/wen—
see Zubizarreta and Oh : ff. for discussion). The independence of Path and v
at Vocabulary Insertion makes it possible for a root, interpreted as Co-event, to
appear adjoined to v:
() PF-derivation of John tiptoed to the room
a. Structure delivered by syntax
[vP [v v TIPTOE] [PathP [Path’ Path [PlaceP [Place’ Place {}]]]]]
b. Raising of the root to v
[vP [v TIPTOE v] [PathP [Path’ Path [PlaceP [Place’ Place {}]]]]]
c. Linearization
TIPTOE-v > Path > Place
d. Vocabulary Insertion
tiptoe-∅ > to > ∅
On the other hand, nothing precludes, in a language like English, the generation of
predicates of change of state in which all heads of the vP cycle end up forming a
complex head:
() PF-derivation of The sky cleared
a. Structure delivered by syntax
[vP v [PathP [Path’ Path [PlaceP [Place’ Place CLEAR]]]]]
b. Raising
[v [Path [Place CLEAR Place] Path] v]
c. Linearization
CLEAR-Place-Path-v
d. Vocabulary Insertion
clear-∅-∅-∅
() presents the derivation of a v-framed construction in an s-framed language. It is
a v-framed construction, in Talmy’s terms, since the Core Schema is expressed within
the verb, and not independently of it. There is nothing in the morphophonological
specification of v or Path in English impeding the derivation of these cases. See
section . for more details.
The surface shape of PathP in Latin
Finally, recall from section .. that the requirement of a common phonological
realization of v and the Core Schema in v-framed languages is not effective when the
construction is stative, non-dynamic. In the present terms, this follows automatically
from the fact that the constructions in question do not feature a Path head. I illustrate
below with the analysis of (), repeated here as (), which follows the spirit of that
proposed by Hoekstra and Mulder () and Mateu (: ):
() Catalan
En aquesta coral hi canten nens.
in this choir LOC sing.PL children
‘Children sing in this choir.’
() PF-derivation of ()
a. Structure delivered by syntax
vP
v PlaceP
v cant nens Place’
Place DEICTIC
b. Raising
[vP [v CANT V] [PlaceP [Place’ [Place DEICTIC Place]]]]
c. Linearization
CANT-v > DEICTIC-Place
d. Vocabulary Insertion
cant-∅ > hi-∅
In () the root is adjoined to v and is interpreted, consequently, as a Co-event. v is,
in this case, interpreted as a stative non-externally originated event, since neither
Voice nor Path is projected. The DP nens ‘children’ is a Figure and enters into a
predicative relation with an abstract deictic element merged as Compl-Place. Since
there is no Path head, there is no requirement for v to be linearly adjacent to it, and
the adjunction structure [v v CANT] may morphologically survive.4
. The surface shape of PathP in Latin
The difference between s- and v-framed languages is primarily concerned with the
expression of PathP. In this section I outline the different morphosyntactic expre-
ssions of directionality in Latin, elucidating whether they in fact correspond to PathP
4
In the analysis I am ignoring cliticization of hi onto T.
Latin as a satellite-framed language
or not. To sum up, directionality can be expressed through (i) a verbal prefix, (ii) a
PP, (iii) a combination of both prefix and PP, (iv) a combination of a prefix and a DP,
and (v) finally, and marginally, a (case-marked) DP. I will show that APs are not
possible encoders of the PathP, which is well attested in other s-framed languages like
Germanic. Finally, I will analyse how case is assigned in DPs and PPs in predicates of
directed motion.
.. Verbal prefixes
A verbal prefix very frequently expresses directionality in Latin:
() Latin; Liv. , ,
Inspectum vulnus
examine.PTCP.PFV.NOM.N.SG wound(N)NOM.SG
abs-terso cruore.
away-wipe.PTCP.PFV.ABL.M.SG blood(M)ABL.SG
‘That the wound had been examined after wiping the blood off ’.
() Latin; Lucr. ,
Flatus [ . . . ] arbusta e-volvens.
gust(M)NOM shrub.ACC.PL out-roll.PTCP.PRS.NOM.SG
‘A gust of wind rolling shrubs out.’
In these exemples the prefixes abs- ‘off, away’ and e- ‘out’ indicate a resulting location
of an externally caused motion event. Note that in both examples the verb expresses
a Co-event, but not the Core Schema, which is encoded by the prefix:
() Analysis of ()
VoiceP
VoicePass vP
v Path
v terg cruor(e) Path’
Path PlaceP
cruor(e) Place’
Place abs
The surface shape of PathP in Latin
() Analysis of ()
VoiceP
Flatus Voice’
Voice vP
v PathP
v volv arbusta Path’
Path PlaceP
arbusta Place’
Place ex
The prefix originates as a root at Compl-Place. It is through Raising, at PF, that it ends
up pronounced in the same word as the verb (see section ..). In its original position
it is interpreted as a Terminal Ground, since PlaceP is embedded within a PathP. In the
case of (), for instance, the blood, cruore, ends up being off (the wound). On the
other hand, the root TERG, adjoined to v, is interpreted as Co-event: it is through wiping
that the blood ends up off the wound. Observe that in both cases I posit movement
from Spec-Place to Spec-Path, where the internal argument is interpreted as a Meas-
urer: the wiping and rolling events are over when the blood and the shrubs are off and
out, respectively. Recall from section ..., that Compl-Place is to be read necessarily
as a final location/state when PlaceP is embedded under a PathP projection. This might
seem counter-intuitive in the case of preverbs like ab(s)- ‘away’ or ex- ‘out’, which are
traditionally classified as ablative, that is, indicating a departing point (García
Hernández ). However, things become clearer when the distinction between
conceptual and structural meaning is seriously taken into account: EX, as a root,
expresses ‘outness’, a notion involving a reference point which, by inference, is taken
as the departure point—for instance, in () this departure point is the ground whence
the shrubs are uprooted; on the other hand, when embedded in a PlaceP, it comes to be
identified as a final location. So the final location is ‘outside the reference point’.
Analyses such as Molinari’s (: ), which do not acknowledge this difference
(or a syntactic theory of prefixation, for that matter), must resort to a double definition
for ‘ablative’ verbal prefixes such as ex-, namely, that they identify a departing point
and an arrival point that is outside the departing point.
In line with claims by Gehrke () for ambiguous prepositions in Germanic and
Real Puigdollers (, ) for ambiguous prepositions in Germanic and Romance,
I suggest that Latin verbal prefixes are never directional per se: the directionality is the
effect of their being merged as Compl-Place within a PathP. Evidence that this is the
right analysis is the fact that prefixes that may head directional, change predicates can
Latin as a satellite-framed language
also appear in stative predicates, combined with sum ‘be’. This is shown below, where
prefixes de- ‘away; down’ and ab(s)- ‘away’ are found in a transition predicate in the a
examples and in a stative, Pathless predicate in the b examples:
() Latin; Caes. Civ. , , and Ter. Phorm.
a. Ad naves de-currunt.
at ship.ACC.PL down-run.PL
‘They run down towards the ships.’
b. Argentum de-erat.
silver.NOM away-was.IPFV
‘Money was lacking.’
() Latin; Liv. , , and Plaut. Cas.
a. Inspectum vulnus
examine.PTCP.PFV.NOM.N.SG wound(N)NOM.SG
abs-terso cruore.
away-wipe.PTCP.PFV.ABL.M.SG blood(M)ABL.SG
‘That the wound had been examined after wiping the blood off ’.
b. Senex ab-est.
old_man.NOM away-is
‘The old man is missing.’
.. Directional PPs
A PP can express directionality, as shown below:
() Latin; Suet. Otho ,
Ac repente omnes in palatium cucurrerunt.
and suddenly all.NOM.PL in palace.ACC run.PRF.PL
‘Then suddenly everybody hastened into the palace.’
In (), the PP in Palatium apparently represents the PathP, with the root IN, in this
case, being merged as an adjunct to Place, and Palatium being merged as Compl-Place:
() Analysis of ()
vP
v PathP
v curr omnes Path’
Path Place
omnes Place
Place palatium
Place in
The surface shape of PathP in Latin
As shown in section .., there is evidence that a different analysis is preferable for
this kind of predicate. See also section .. for more considerations on the syntax of
directional PPs.
.. Verbal prefixes in combination with directional PPs
Prefixed verbs sometimes appear combined with a directional PP. The prefix may
coincide with the preposition—a phenomenon often referred to as duplication
(Lehmann ; Acedo-Matellán b, among others) or it may be different from
the preposition, as shown, respectively, below:
() Latin; Caes. Gall. , ,
Ex castris [ . . . ] copias suas e-duxit.
out camp.ABL troop.ACC.PL his.ACC.PL out-lead.PRF.SG
‘He led his troops out of the camp.’
() Latin; Cic. Caecin. ,
Ne in aedis ac-cederes.
lest in house.ACC at-march.SBJV.IPFV.SG
‘Lest you should come into the house’.
A natural way to analyse these ‘doubly marked’ predicates is by assuming that the
prefixal root is first merged at Path, while the PP is PlaceP. An analysis along these
lines for () would look like the following:
() Analysis of ()
vP
v PathP
v ced (tu) Path’
Path PlaceP
Path ad (tu) Place’
Place aedis
Place in
The root AD, by being merged as an adjunct to Path, is merely interpreted as directional.
The final location is specified by PlaceP in aedis ‘into the house’. As for cases in which the
prefix coincides with the preposition, Acedo-Matellán (, b) and Oniga (:
ff.) propose that the prefix and the preposition correspond to two pronounced copies
of a movement chain originating in the site of the preposition. This analysis cannot
account for the derivation of cases in which the prefix and the preposition do not
Latin as a satellite-framed language
coincide.5 If the preposition and the prefix are taken to be copies of the same object, it is
not clear why they should possess different phonological and semantic properties. On the
other hand, both this analysis and that put forward in () in section .. leave
unexplained the fact that PPs specifying final location in prefixed predicates are omissible
without the fundamental transition interpretation of the predicate being altered:
() Latin; Cic. Verr. , , ,
Subito ipse ac-currit.
suddenly self.NOM at-run.SG
‘Suddenly he himself arrives in haste.’
() Latin; Ov. Met. ,
Tergo velamina lapsa re-liquit.
back.ABL.SG veil.ACC.PL slip.PTCP.PFV.ACC.N.PL back-leave.SG
‘She left behind the veil which had slipped off her back.’
() Latin; Cato, Agr. ,
Tenuissimas radices ex-arabit.
slender.SUPERL.ACC.F.PL root.ACC.PL out-plough.FUT.SG
‘He will plough out the most slender roots.’
For these reasons I propose that the prefix originates at Compl-Place, and is therefore
interpreted as a Terminal Ground, and that the PP is in fact an adjunct to PathP or to
PlaceP, that is, a modifier of the direction component of the transition predicate or of
the final location. I exemplify with the analysis of ():
() Analysis of ()
vP
v PathP
v ced pP PathP
p aedis (tu) Path’
p in Path PlaceP
(tu) Place’
Place ad
5
See Lehmann () and López Moreda (). Interestingly, Biskup and Putnam () propose that
the German preposition/particle aus ‘out’ and the prefix ent- correspond to the same morpheme, spelled
out with different exponents according to its position:
(i) German; Biskup and Putnam (: )
Sie ent-steigt dem Auto. /Sie steigt aus dem Auto.
she ent-climbs the.DAT car she climbs out the.DAT car
‘She gets out of the car.’
This analysis does not seem adequate, however, for cases such as that in (), in which the prefix and the
preposition are not semantically equivalent.
The surface shape of PathP in Latin
In this representation the root AD, which is merged as Compl-Place, is inter-
preted as the final location of the motion event: somewhere near the speaker
(default interpretation) or some prominent discourse participant. The direction
of the event is identified with the PP in aedis ‘into the house’, an adjunct to
PathP. The analysis derives from the fact that the PPs are omissible in prefixed
predicates, since they are not properly a part of the argument structure config-
uration.6 Also, it explains away naturally the cases of prefixed predicates featur-
ing a PP headed by a preposition that is not homophonous with the prefix. See
section .. for the mechanism that explains the prefixation of the material in
Compl-Place.
.. Verbal prefixes in combination with DPs
The verbal prefix may alternatively appear with a DP specifying the final
location in a change-of-location event. That DP may appear in the same case
as that governed by the homonymous preposition: the accusative (as in ()) or
the ablative (as in ()). In some cases it may appear in the dative case (as
in ()):
() Latin; Tac. Ann. ,
Novissimos in-currere.
rear.ACC in-run.PRF.PL
‘They charged against the rear.’
() Latin; Caes. Gall. , ,
Omnes copias castris e-duxit.
all.ACC.PL troop.ACC.PL camp.ABL out-lead.PRF.SG
‘He led all the troops out of the camp.’
() Latin; Plin. Nat. ,
Caprarumque uberibus ad-volant.
goat.GEN.PL=and udder.DAT.PL at-fly.PL
‘And they fly onto the udders of the goats.’
For cases like those above, Miller (: –) and Oniga (: –) consider the
application of an analysis in terms of preposition incorporation (Baker ) to pre-
verbation: the preverb is a preposition incorporated from its base position into the verb,
stranding, in some cases, a case-marked DP. However, there are data not predicted by this
proposal, notably, the cases where no DP is stranded, as seen in section .., or those
featuring a PP that is headed by a preposition different from the preverb, as seen in section
6
According to Spencer and Zaretskaya (: ), in the Slavic languages, which, as I show in
Chapter , also make extensive use of verbal prefixes, PPs are omissible in predicates headed by a prefixed
verb—see, for Czech, Filip (: ), and for Russian, Rojina (: ). These authors too arrive at the
conclusion that in this kind of predicate the PP is an adjunct, the directional and resultative interpretation
stemming from the prefix itself.
Latin as a satellite-framed language
... Although these authors point these problems out, they do not provide a syntactic
analysis of this kind of predicate, and simply state that they correspond to lexicalizations,
opaque to the syntax. As we will see in sections to come, this cannot be the case, since
prefixed verbs, although they do not involve stranding, show systematic syntactic and
semantic properties which set them apart from their non-prefixed counterparts.
Taking into account the remarks above and the discussion in section .., which
made it clear that the transitional interpretation of the event depends exclusively on
the prefix, I propose that, at least for cases like (), the prefix originates at Compl-
Place, and that the ablative castris ‘out of the camp’ is an adjunct to PlaceP:
() Analysis of ()
Voice
pro Voice’
Voice vP
v PathP
v duc omnes copias Path’
Path PlaceP
pP PlaceP
p castris
omnes copias Place’
Place ex
Cases like (), involving a prefixed verb and an accusative-marked DP encoding
goal of motion are not amenable to the same analysis. In particular, the accusative DP
seems to be really an argument, since it can become a passive subject. Thus, in the
next example the nominative praetores ‘praetors’ is interpreted as the Ground of the
preverb ad ‘at’:
() Latin; Cic. ad Q. fr. , ,
Cum neque praetors [ . . . ] ad-iri possent.
since not-even praetors.NOM at-go.INF.PASS can.IPFV.SBJV.PL
‘Since not even the praetors could be addressed’.
At the end of section ... I will present an analysis of predicates like () and ()
in which the accusative DP is merged as Compl-Place and there is in fact no
Figure argument. Directional datives like that in example () will be treated in
section .., along with a more general discussion of case assignment in directional
PPs and DPs.
The surface shape of PathP in Latin
.. Directional DPs
In some cases of motion events the PathP can correspond to a DP marked either in
the accusative or in the ablative, in the absence of a prefixed verb. Examples of so-
called directive accusative are found in both () (Syracusas) and () (Hennam), and
an example of source ablative is found in () (Assoro):
() Latin; Cic. Verr. Actio secunda, ,
Veniunt Syracusas.
come.PL Syracuse.ACC
‘They come to Syracuse.’
() Latin; Cic. Verr. Actio secunda, ,
Assoro itur Hennam.
Assorum.ABL go.PASS.SG Henna.ACC
‘One goes from Assorum to Henna.’
DPs with unprefixed verbs cannot be used freely to express directionality, however.
On the contrary, they show restrictions of an encyclopaedic nature. Thus, the DP
must refer to a town or a small island (and not a country) or must contain one of a
small set of nouns: accusative domum ‘home’ (directional, as in Sue went home), rus
‘to the country’, and ablative domo ‘from home’, rure ‘from the country’, and humo
‘from the ground’ (Ernout and Thomas : ff.; Lavency : ; Hofmann
and Szantyr : –, ). Furthermore, Hofmann and Szantyr (: )
point out that the prepositionless ablative is licensed also by names of towns or
islands, crucially, when there is no specification whether movement takes place from
the inside or from the surroundings of the relevant location. If specification is
needed, prepositions ex ‘out’ and ab ‘away’ are respectively used. A striking evidence
of the encyclopaedic nature of the restrictions operating on the availability of
directional accusatives is the fact that, as observed by Echarte Cossío (: ),
the names of Greek cities are less frequent in prepositionless accusatives. All these
restrictions seem to point out that the absence of both prefixes and prepositions in
these constructions demands that the DPs be easily identifiable, by the very nature of
their reference, with proper Grounds in a motion schema.7 The phenomenon seems
similar to that found in other languages such as Greek (Terzi ).
Another important fact about these constructions is that the verbs that license
them must unequivocally express directionality, as do eo ‘go’ and venio ‘come’, in the
examples above. Finally, accusative DPs with verbs like eo ‘go’ and venio ‘come’
cannot receive the same analysis as those accusative DPs found with prefixed verbs of
directed motion, illustrated in section .. (example ()), since in the former case
7
With prefixed verbs the frequency of directional DPs grows considerably, as Hofmann and Szantyr
(: –) point out. This is what I would expect, under present assumptions, since it is the prefix
together with the DP that is structuring the PathP.
Latin as a satellite-framed language
these DPs cannot become passive subjects. These facts bring the cases of directional
DPs with unprefixed verbs together with the cases of directional PPs discussed in
section .., as we will see in section .., where I will provide a unified analysis.
.. APs
We have seen that s-framed languages like English or German admit an AP as
expression of the Core Schema, as shown by the following s-framed construction:
() German; Talmy (: )
Der Hund hat [den Schuh]Figure [kaputt]Core schema -[gebissen]Event+Co-event
the dog has the shoe in_pieces bite.PST.PART
‘The dog bit the shoe to pieces.’
In the present account, the resultative AP is, therefore, the manifestation of the vP
internal PathP. This is a natural consequence of assuming Mateu’s () reduction
of the argument structure of adjectives to that of adpositions (and recall that in Hale
and Keyser’s theory the A and P lexical heads display different projecting properties
and head different argument structures; see sections .. and ..):
() Analysis of ()
VoiceP
Der Hund Voice’
Voice vP
v PathP
v beiss den Schuh Path’
Path PlaceP
den Schuh Place’
Place kaputt
In Latin this option does not seem to be available, at least for s-framed constructions,
where the verb is independently associated with a Co-event root. Thus, for instance,
an example such as the following, with vacuum ‘empty’ being interpreted as the Core
Schema and with a v bundled together with the root BIB ‘drink’, is not found in this
language:
() Latin; Acedo-Matellán (: )
*Poculum vacuum bibere.
goblet.ACC.SG empty.ACC.SG drink.INF
‘To drink the globlet empty’.
The surface shape of PathP in Latin
In Chapter I will provide empirical evidence that the made-up example above
reflects a general fact of Latin—and of other similar languages like the Slavic
languages. I will also attempt an explanation of the lack of s-framed constructions
based on APs in these languages in terms of the morphological properties of v and
Path and of the adjective (section ..). Finally, I will show that APs can be part of
the PathP, as adjuncts to PlaceP, in predicates involving a non-complex change-of-
state event, that is, an event with no co-event associated:
() Latin; Plaut. Capt.
Eam [servitutem] lenem [ . . . ] reddere.
that.ACC.F.SG serfdom(F)ACC.SG mild.ACC.F.SG render.INF
‘To make that serfdom mild’.
In () the AP lenem ‘mild’ codifies the Core Schema, in that the Figure DP
servitutem ‘serfdom’ is entailed to end up in the state described by lenem. See section
.. for an analysis of these constructions.
.. Case and directional PPs and DPs
In Latin there is overt case marking on the DP. DPs interpreted as Grounds in predicates
expressing directed motion bear an accusative case mark if the motion approaches
their referent and an ablative case mark if the motion involves separation from their
referent.8 This is illustrated with the following examples, already shown above:
() Latin; Suet. Otho ,
Ac repente omnes in palatium cucurrerunt.
and suddenly all.NOM.PL in palace.ACC run.PRF.PL
‘Then suddenly everybody hastened into the palace.’
() Latin; Caes. Gall. , ,
Omnes copias castris e-duxit.
all.ACC troop.ACC.PL camp.ABL out-lead.PST.SG
‘He led all the troops out of the camp.’
8
See below for the question whether the dative does also express a Ground too. On the other hand,
there is a case of apparently prefix-governed accusative that is not directional:
(i) Latin; Caes. Gall. , , , in Bortolussi (: )
Flumen [ . . . ] exercitum tra-ducere maturavit
river.ACC army.ACC over-lead hasten.PRF.SG
‘He hastened to lead the army to the other side of the river.’
Predicates such as the one in (i) show two accusative DPs, one interpreted as a Figure (exercitum ‘army’),
and the other interpreted as the Ground of the preverb (flumen ‘river’). Crucially, this latter accusative is
not a directional accusative like those discussed in section ... See Lehmann (), Miller (),
Acedo-Matellán (), Bortolussi (), among others, for a description of traduco-predicates and
Acedo-Matellán () for a DM analysis.
Latin as a satellite-framed language
() Latin; Caes. Gall. , ,
Ex castris [ . . . ] copias suas e-duxit.
out camp.ABL troop.ACC.PL his.ACC.PL out-lead.PRF.SG
‘He led his troops out of the camp.’
Traditionally, assignment of case in Ground DPs has been considered the result of
a selection property of the preposition or prefix with which the case-marked
DP coappears—see, among others, Pinkster (: ff.), Luraghi (), and
Oniga () and the references cited therein. In particular, there is a series of
prepositions that exclusively select either the accusative (see ()a) or the ablative
(see ()b):
() Latin accusative- and ablative-taking prepositions; Ernout and Thomas
(: –)
a. ad ‘at, beside’, praeter ‘beyond’, ob ‘in front of, against’, ante ‘in front of ’,
post ‘behind, after’, per ‘through’, inter ‘between’, circum ‘around’, contra
‘against’.
b. a/ab/abs ‘off, away’, coram ‘in the presence of ’, cum ‘with’, de ‘away,
downward’, e/ex ‘out (of)’, prae ‘before, in front of ’, pro ‘before, in front
of, forth’, sine ‘without’.
There are other prepositions that may select either accusative or ablative, the most
frequent ones being in ‘in’, sub ‘under, below’ and super ‘over, on’ (Ernout and
Thomas : ). In these cases while the accusative expresses approach, the
ablative indicates location. The contrast is shown in the following examples involving
in ‘in’ and sub ‘under’:
() Latin; Suet. Otho ,
a. Ac repente omnes in palatium cucurrerunt.
and suddenly all.NOM.PL in palace.ACC run.PRF.PL
‘Then suddenly everybody hastened into the palace.’
b. Fuit certe contentio in senatu.
be.PRF.SG certainly struggle.NOM in senate.ABL
‘There was in fact a struggle in the senate.’
() Latin; Plaut. Curc. and Plaut. Epid.
a. Omnis sub-dam sub solum.
all.ACC.PL under-give.FUT.SG under sole.ACC
‘I will put them all under the sole of my foot.’
b. Sub vestimentis [ . . . ] habebant retia.
under clothes.ABL have.IPFV.PL net.ACC.PL
‘They had nets under their clothes.’
The surface shape of PathP in Latin
Case assignment in these examples cannot depend on the choice of prepositions.
Rather, it must be purely structural, as proposed by Gehrke () and Real Puig-
dollers (, ) for Germanic languages:9
() German; Gehrke (: )
a. Diana schwamm in den See.
Diana swam in the.ACC lake
‘Diana swam into the lake.’
b. Diana schwamm im See.
Diana swam in.the.DAT lake
‘Diana swam in the lake.’
As these German examples make clear, the directional/stative contrast, reflected in
the accusative/ablative contrast in Latin, cannot be ‘immediately recoverable from
the context’, as claimed by Luraghi (: ). Thus, the presence of a motion verb,
at least one that does not express inherent direction, does not necessarily induce
accusative in in-PPs, since they are perfectly compatible with a static location
expressed as an ablative-marked in-PP:
() Latin; Plaut. Curc.
In foro infumo boni homines
in forum.ABL lowest.ABL good.NOM.PL man.NOM.PL
atque dites ambulant.
and rich.NOM.PL walk.PL
‘The men of good standing and the rich walk in the lowest part of the forum.’
On the contrary, as I have said, it must be a structural condition that brings about the
case distinction. I hypothesize that accusative is assigned to a PP associated with PathP
(as an adjunct) and ablative is assigned elsewhere, i.e. to adjuncts of either PlaceP or vP:
() Latin; in silvam (acc.) ‘into the forest’ vs in silva (abl.) ‘in the forest’
a. [PathP [pP [p p IN] silvam] [PathP Path . . . ]]
b. [PlaceP [pP [p p IN] silva] [PlaceP Place . . . ]]
c. [vP [pP [p p IN] silva] [vP v . . . ]]
When only one case is allowed, as with prepositions ad and per (accusative) and ex and
ab (ablative), the pure choice of preposition overrides the structural factor. For instance,
ad silvam ‘at forest.ACC’ can either mean ‘at the forest’ or ‘to/towards the forest’:
9
Caha () has proposed an analysis of this accusative/oblique alternation, and an analogous one in
Dutch, based exclusively on the internal structure of the PP, which hosts two different heads providing
either accusative or oblique case. In this sense, the empirical fact is missed that accusative case in
directional PPs headed by simple locative prepositions in German correlates with the fact that the PP is
merged VP-internally (Gehrke : ff.).
Latin as a satellite-framed language
() Latin; ad silvam (acc.) ‘to the forest’/‘at the forest’
[pP [p p AD] silvam]
() Latin; ex silva (abl.) ‘out of the forest’/‘outside the forest’
[pP [p p EX] silva]
This could be modelled as a case of allomorphy exerted by the root involved in the
prepositional head on the case morpheme borne by the DP.
I address, finally, the dative used directionally. It has often been observed that the
dative case, which, unlike the accusative and the ablative, is not selected by any
preposition, can nevertheless present a directional meaning:
() Latin; Verg. Aen. ,
It clamor caelo.
go.SG clamor.NOM Heaven.DAT
‘A clamor rises to Heaven.’
The dative expressing the Ground, which I will here call directional, is particularly
frequent with prefixed verbs (Lehmann ; Pinkster ; Echarte Cossío ;
Serbat ), as illustrated here:
() Latin; Plaut. Most.
Tibi ad-duxi hominem.
you.DAT at-lead.PRF person.ACC
‘I have brought the man to you.’
In () the dative tibi may be interpreted as the final location of the spatial schema
involving the prefix ad-. Since there is not a single preposition assigning this case in
Latin, if the dative in () is really somehow governed by the prefix, this phenomenon
would be a problem for the hypothesis that the prefix originates as a preposition and
assigns case (accusative or ablative) as such to the DP. Ernout and Thomas (:
–) and Rubio Fernández and González Rolán (: –), among other authors,
argue that these ‘p-governed’ datives (that is, datives apparently selected by the prefix)
are in fact run-of-the-mill benefactive datives (the so-called dativus commodi/incom-
modi in, for instance, Landgraf ), expressing goal or interest. Ernout and Thomas
() support their claim by pointing out that most of these allegedly p-governed
datives involve a human referent, alternating with semantically equivalent PPs with a
non-human referent. Thus, () above contrasts with the next example, where inani-
mate urbem ‘city’ is interpreted as a final location in the presence of the preposition ad:
() Latin; Cic. Phil. ,
Ad urbem [ . . . ] exercitum maximum ad-duceret.
at city army.ACC biggest.ACC at-lead.IPFV.SBJV.SG
‘That he led the biggest army near the city’.
The surface shape of PathP in Latin
However, as Ernout and Thomas (: ) later point out, there are examples where
the correlations human/dative and non-human/PP do not hold. Thus, in () a
human goal is expressed as a PP headed by ad ‘at’ and in () a non-human goal is
expressed as a dative DP in a predicate involving the prefix in- ‘in’:
() Latin; Plaut. Epid.
Illum [ . . . ] ad-ducam huc ad te.
him.ACC at-lead.FUT.SG to_here at you.ACC
‘I will bring him to you here.’
() Latin; Caes. Gall. , ,
Aggeri ignem in-ferebant.
rampart.DAT fire.ACC in-carry.IPFV.PL
‘They were carrying fire to the rampart.’
In particular, as regards the prose example in (), it is difficult to maintain the view
that the dative expresses ‘interest’, as interpreted by Rubio Fernández and González
Rolán (: ) for the following poetic example:
() Latin; Verg. Aen. ,
Silici scintillam ex-cudit.
stone.DAT spark.ACC out-beat.SG
‘He beat a spark out of the stone.’
Lehmann (: ff.) claims that the directional dative is somehow governed by
the prefix, and that the dative case is assigned to avoid the coexistence, in transitive
predicates, of two accusatives: one corresponding to the object and one introduced by
the prefix. This ‘double accusative filter’ is theoretically dubious and seems to be also
empirically incorrect, in that it fails to predict the use of the dative in cases in which
no double-accusative scenario emerges: motion predicates without a prefix (see ())
and transitive predicates with an ablative-governing prefix—cf. () above, and see
Ernout and Thomas (: –) for more examples.
Here I propose an analysis of the dative DP being merged above the projection
where directionality is encoded (PathP) but at the same time capturing the fact that it
is somehow interpreted as being governed by the prefix—as the Ground of the
motion event. My analysis is inspired by that proposed by Oya () for some
constructions in German involving a goal dative and the particle zu ‘to’:
() German; Oya (: )
Ich warf dem Kind den Ball zu.
I threw the.DAT child the.ACC ball to
‘I threw the ball to the child.’
In this sentence the dative dem Kind ‘the child’ is interpreted as the Ground of the
preposition/particle zu ‘to’. Oya (), following Olsen (), points out that
Latin as a satellite-framed language
although this particle assigns dative in German, the directional dative does not
originate as its complement, and that it is in fact higher than the accusative (cf.
den Ball ‘the ball’), as in transfer of possession predicates headed by verbs like geben
‘give’ or schenken ‘give as a gift’. Following this line of thought, Oya () hypothe-
sizes that the dative is interpreted as inalienably possessing a null nominal merged as
the complement of the particle zu ‘to’ and referring to someone’s body. The inter-
pretation of the dative as the Ground of the preposition is nothing more than an
effect of the possessive relation between the dative and this null nominal. I think that
Oya’s analysis can be made more general. In particular, datives in motion construc-
tions can be modelled as inalienably possessing a particular region of the goal object,
by virtue of which they come to be understood as the goal object itself. Interestingly,
languages that allow datives to express possession also feature a dative used in
combination with locative expressions specifying the region of a Ground object:
() German
a. Gestern hatte ein Barbier ihr das Haar geschnitten.
yesterday had a.NOM barber DAT.F.SG the.ACC hair cut
‘Yesterday a barber had cut her hair.’
b. Die schwere Tür ist ihm hin-auf-gefallen.
the.NOM heavy.NOM door is DAT.M.SG hither-on-fallen
‘The heavy door has fallen on top of him.’
In Latin we also find this correlation, the possessive dative being well established in
the literature (cf., for instance, Fay ; Löfstedt ). In the next example the
reference of the dative Ascanio ‘Ascanius’ is actually understood as both the possessor
of the PP per membra ‘through the limbs’ and as the Ground of in-rigat ‘in-pours’:
() Latin; Verg. Aen. ,
Venus Ascanio placidam per membra quietem
Venus.NOM Ascanius.DAT placid.ACC.F.SG through limbs.A rest(F).ACC.SG
in-rigat
in-pour.SG
‘Venus pours a placid tranquillity through Ascanius’ limbs.’
This example suggests both the possessive nature of the dative used with a prefixed verb
and also that it must be merged higher than Path, provided that an accusative-marked
PP (per membra ‘through the limbs’), by hypothesis an adjunct to PathP, is interpreted
under its scope. In order to implement the analysis I assume that the dative is
introduced as the specifier of an applicative head (Cuervo ; Pylkkänen )
merged above PathP but under v:10
10
This analysis echoes that proposed by Cuervo (: ) for so-called Affected Applicatives, in
which ‘[t]he dative DP is applied to the end state of the object DP [ . . . ]. The dative DP is the “possessor” of
The surface shape of PathP in Latin
() Analysis of ()
VoiceP
Venus Voice’
Voice vP
v ApplP
v rig Ascanio Appl’
Appl PathP
pP PathP
p membra placidam quietem Path’
p per Path PlaceP
placidam quietem Place’
Place in
For the cases in which the directional dative is used with an unprefixed verb, I assume
that it takes scope too over the final location represented by the verbal root merged as
Compl-Place:
() Analysis of ()
vP
v ApplP
caelo Appl’
Appl PathP
clamor Path’
Path PlaceP
clamor Place’
Place i
the end state of the object.’ From a localistic point of view, as advocated here, the ‘end state of the object’
(i.e., the Figure) is equivalent to its final location.
Latin as a satellite-framed language
The analysis of the directional dative presented in this section explains the directional
interpretation of the dative while avoiding the problems of both Lehmann’s ()
account, on the one hand, and Ernout and Thomas’s () and Rubio Fernández
and González Rolán’s (), on the other. Thus, the dative is neither governed by
the preverb, in any sense, nor is it a benefactive/malefactive dative, since it is merged
below v, and is therefore not understood as being related with the event externally
(see Pylkkänen ). See section ..., for an extension of this analysis to the
dative of transfer predicates like do ‘give’.
. S-framed constructions in Latin
In this section I will present evidence that Latin behaves as an s-framed language, by
exploring a range of constructions that are amenable to an analysis in terms of a
transition predicate. The discussion is not limited to constructions that have an overt
motional semantics, Complex Directed Motion Constructions, but will encompass
also Unselected Object Constructions, constructions participating in the Locative
Alternation and so-called Pseudoreversatives (McIntyre ), that is, constructions
where the result expressed by the verb is cancelled by virtue of the element expressing
the Core Schema.
.. Complex Directed Motion Constructions
Complex Directed Motion Constructions (CDMCs) are constructions that express a
directed motion event with a Manner Co-event encoded in the verb. The following
English examples illustrate this:
() Zubizarreta and Oh (: )
a. John danced to the kitchen.
b. The bottle floated under the bridge.
c. They danced out of the room.
d. The horse galloped into the barn.
Note that, while ()a, ()c, and ()d involve directional predicates, ()b, con-
taining a Place preposition, is ambiguous between a directional and a non-directional
sense, respectively made evident through the addition of an in- and a for-adverbial:
() The bottle floated under the bridge {in a few minutes/for hours}.
Being overtly motion constructions, CDMCs most evidently show the pattern of an
s-framed language: the Core Schema is not expressed by the verb, but by some other
piece of the structure, and the verb, instead, expresses a Manner Co-event. As one
might expect, if we want to render the expressions of () literally in a Romance
language, we obtain, at most, expressions which, unlike those of English, are unam-
biguously non-directional:
S-framed constructions in Latin
() Catalan renditions of ()
a. *En John ballà a la cuina.
the John danced at the kitchen
b. *L’ampolla surà sota el pont {durant hores/ *en uns minuts}
the=bottle floated under the bridge during hours in some minutes
c. *Ballaren fora de l’habitació.
danced.PL out of the=room
d. *El cavall galopà a dins del graner.
the horse galloped at in of=the barn
... CDMCs and situation aspect It has been claimed that one of the hallmarks
of CDMCs is the fact that these constructions, unlike other constructions involving
non-directed motion, correspond to telic predicates, that is, to achievements or
realizations (Tenny ; van Hout ; Borer , among others). This contrast
is exemplified in () through the well-known for/in adverbial test:
() Atelic/telic contrast in motion constructions
a. Sue danced for/*in an hour.
b. Sue danced into the room in/*for five minutes.
In prima facie contradiction, we do find constructions indicating both directed
motion and manner of motion—thus qualifying as CDMCs as defined above—that
are nonetheless atelic:
() Sue danced towards the room for/*in some minutes.
There is a difference between ()b and (), however: while in the former the PP
expresses a bounded trajectory, entailing that Sue is at some stage in the room, in the
latter the trajectory is unbounded, and no such entailment is licensed. The difference
in the (un)boundedness properties of the PP are automatically mapped onto the
aspectual properties of the whole predicate: telic in ()b and atelic in (). More
importantly, that difference seems to be directly relevant to the s-/v-distinction, as
first pointed out in Aske () and incorporated in Talmy (). Indeed, while
s-framed languages allow CDMCs with a bounded directional PP, therefore, telic,
v-framed languages only allow atelic CDMCs, featuring an unbounded goal PP, such
as the Spanish one exemplified in () (where I have added the durative PP durante
cinco minutos ‘for five minutes’):11
11
Beavers, Levin, and Tham (: ) claim that languages acknowledged as v-framed, such as
Spanish or Japanese, do allow CDMCs with bounded paths, specifically by using elements meaning ‘until’
or ‘up to’, like Spanish hasta, to convey a bounded path and successfully combine with a manner-of-
motion verb. For space reasons I cannot refute their arguments here. See Narasimhan, Di Tomaso, and
Verspoor () and Real Puigdollers () for relevant discussion. See Inagaki () for an early
proposal of Japanese made ‘until’ as encoding final location.
Latin as a satellite-framed language
() Spanish; adapted from Aske (: )
Corrieron hacia adentro de la cueva (durante cinco minutos).
run.PRF.PL towards inward of the cave during five minutes
‘They ran towards the inside of the cave for five minutes.’
A perspective that has proved fruitful in accounting for the different properties of
bounded and unbounded directional PPs is that assuming that the former are
argumental and the latter are adjuncts. Direct evidence thereof comes from languages
like Dutch, as discussed by Hoektra and Mulder (), Hoekstra (), and
Zubizarreta and Oh (). In this language a directional PP like naar Groningen
‘to/towards Groningen’ can be interpreted as bounded or unbounded, licensing a
telic (‘to’) and an atelic (‘towards’) interpretation of the predicate, respectively.
Crucially, these two interpretations correlate with other facts. The former correlates
with a position closer to the verb and with BE-auxiliary selection, while the latter
correlates with a position farther from the verb and with HAVE-auxiliary selection:
() Dutch; Hoekstra and Mulder (), in Zubizarreta and Oh (: )
a. dat Jan in twee uur naar Groningen is gewandeld.
that Jan in two hours to Groningen is walked
‘That Jan walked to Groningen in two hours’.
b. dat Jan naar Groningen twee uur lang heeft gewandeld.
that Jan to Groningen two hours long has walked
‘That Jan has walked towards Groningen for two hours’.
c. ??dat Jan twee uur lang naar Groningen heeft gewandeld.
that Jan two hours long to Groningen has walked
Hoeskstra and Mulder interpret, on the basis of these facts, that in its bounded
intepretation the PP is merged VP-internally, as a predicate of the unaccusative
subject (Jan). This accounts for its position with respect to the verb, for the use of BE
as the auxiliary for the perfect tense, and for the telic interpretation of the predicate.
By contrast, the PP is interpreted as unbounded when it is merged as an adjunct to
the VP, whence its relative position farther from the verb, the selection of HAVE for the
perfect tense, and the interpretation of the predicate as an activity.12
The assumption that bounded directional expressions are merged VP-internally
and unbounded ones are merged as VP-external adjuncts straightforwardly explains
two facts: on the one hand, only bounded directional PPs, as VP-internal material,
may change the situation aspect of the motion event, as illustrated above, and, on the
other hand, cross-linguistic variation in argument structure expression involves only
bounded directionality. In this work, therefore, I will assume Aske’s () revision
12
See Zubizarreta and Oh (: ff.) for more related evidence from Korean.
S-framed constructions in Latin
of Talmy’s typology as correct, and I will take the difference between bounded and
unbounded directional expressions to be configurational in nature. Since my aim in
this section is to show the relevance of CDMCs in characterizing Latin as s-framed,
I will restrict that name to constructions involving a bounded directional element.13
... CDMCs and non-directed motion constructions in Latin In Latin CDMCs
can in principle be found as predicates headed by an unprefixed verb and a direc-
tional expression (cf. ()), by a prefixed verb and a directional expression (cf. ()
and ()), or as predicates with a prefixed verb in combination with no independent
directional phrase (cf. () and ()):
() Latin; Cic. Verr. , , ,
Subito ipse ac-currit.
suddenly self.NOM at-run.SG
‘Suddenly he himself arrives in haste.’
() Latin; Cic. Att. , ,
Se statim ad te navigaturum esse.
REFL.SG.ACC at_once at you.ACC sail.INF.FUT.M.ACC be.INF
‘That he was on the point of setting sail at once to join you’.
() Latin; Cic. Verr. , ,
Simulatque e navi e-gressus est dedit.
as_soon_as out ship.ABL out-walk.PRF.SG give.PRF.SG
‘As soon as he walked out of the ship, he handed it over.’
() Latin; Suet. Diuus Augustus ,
Draconem repente ir-repsisse ad eam
snake(M)ACC.SG suddenly in-glide.INF.PFV at her.ACC
pauloque post e-gressum.
a_little=and after out-walk.PTCP.PFV.ACC.M.SG
‘That, suddenly, a snake glided in towards her and glided away soon after’.
() Latin; Val. Max. , ,
[Vires atque opes humanae] ad-fluunt subito,
strength.NOM.PL and wealth.NOM.PL human.NOM.PL at-flow.PL suddenly
repente di-labuntur.
suddenly apart-slip.PL
‘The vigour and the wealth of humans come suddenly in a flow, and suddenly
slip asunder.’
13
See Folli and Harley () for the view that (transitive) CDMCs do not necessarily involve telicity.
Latin as a satellite-framed language
All of the above examples involve a predicate the telicity of which is made evident by
the licensing of a specific expression: subito ‘suddenly’, simul ‘at once’, simulatque, ‘as
soon as’, and repente ‘suddenly’. These adverbials are not possible in predicates
expressing a simple activity, which, on the other hand, license durative adverbials
such as per-phrases with a time measure expression, corresponding to English for-
adverbials, as shown in the following examples through per aliquot dies ‘for some
days’ and diu ‘for a long time’:
() Latin; Plin. Nat. ,
Per aliquot dies vagari.
for some days.ACC wander.INF
‘That it wanders for some days’.
() Latin; Ov. Am. , ,
Diu lacrimae fluxere per ora.
For_long tears.NOM flow.PRF.PL through face.ACC
‘Tears flowed down her face for a long time.’
I assume that the difference between examples such as those in () to () and
examples such as those in () and () is configurationally represented. In par-
ticular, I claim that CDMCs are unaccusative predicates. I illustrate this with an
analysis of ():
() Analysis of ()
[vP [v v CURR] [PathP ipse [Path’ Path [PlaceP ipse [Place’ Place AD]]]]]
The subject of the construction originates as a Figure in Spec-Place. Here it enters
into a predicative relation with the root AD, which refers to a place coreferent with
one already present in the discourse (as is also understood in the English rendition
‘He arrives in haste’). The entailment that the Figure effectively ends up in the
location encoded by PlaceP is licensed by the fact that the predicate incorporates a
PathP projection, which introduces a transition in the event. In turn, the quantity DP
ipse ‘he himself ’ rises to Spec-Path and is interpreted as a Measurer of that transition,
which is not over until ipse ‘he himself ’ is at the location referred to by AD. Telicity is
licensed thereby, as evidenced by the adverbial subito ‘suddenly’. Since no Voice head
is projected, ipse ‘he himself ’ is not assigned accusative case, and raises to T, where it
is assigned nominative case. Finally, the root of the verb is here an adjunct to the
eventive head v, and is interpreted, as such, as a Manner Co-event. The English
translation provided faithfully reflects this fact, since the celerity of the motion event
is expressed there as an adjunct (in haste).14 In turn, the predicates in () and (),
14
Cf. also Serbat’s (a: ) proposal that a prefixed verb like re-gredior, literally “back-walk” should
be glossed as ‘revenir en marchant’, that is ‘go back while walking’, rather than as ‘aller vers l’arrière’, that is,
‘go towards the back’.
S-framed constructions in Latin
which express activities, rather than accomplishments, are claimed to have the
following unergative structure:
() Analysis of ()
[VoiceP pro [Voice’ Voice [vP v NAVIG]]]
Here the subject (a pro, in this case) is not a Figure, but an Originator, since it
originates at Spec-Voice. The root of the verb is not adjunct to v, but a complement,
and is interpreted as an Effected Object.
... The unaccusative nature of CDMCs Do we have evidence that CDMCs, as
proposed above, are unaccusative predicates? In this section I show that at least two
tests can be invoked to show the unaccusative character of CDMCs: the disallowance
of cognate objects and measure phrases and the failure to yield agent nouns.
The results presented in this section emerge from an investigation of a wide range
of manner-of-motion verbs. I have searched for both unprefixed and prefixed verbs,
as shown in () and (), respectively. Specifically, () contains a list of the
combinations of each one of the verbs in () with the prefixes a(b)- ‘off, away’, ad-
‘at’, ex- ‘out’, and in- ‘in’, whenever the resulting compositum is entered in Gaffiot
() and retrievable from the Antiquitas corpus of the BTL:
() Unprefixed manner-of-motion verbs
ambulo ‘walk’, curro ‘run’, equito ‘ride’, erro ‘wander, stray, roam’, festino
‘make haste, hurry’, fluo ‘flow’, fugio ‘flee’, labor ‘slide, slip’, navigo ‘sail’, nato
‘swim, float’, no ‘swim, float’, propero ‘hasten, make haste’, repo ‘creep, crawl’,
salio ‘jump’, salto ‘dance’, vado, -as ‘wade’, vago ‘wander’, volo ‘fly’
() Manner-of-motion verbs prefixed with a(b)-/au- ‘off, away’, ad- ‘at’, ex- ‘out’
and in- ‘in’
ad-ambulo “at-walk” ‘walk up to’, ex-ambulo “out-walk”, in-ambulo “in-walk”,
ac-curro “at-run” ‘run up to’, ex-curro “out-run”, in-curro “in-run”, ab-equito
“away-ride”, ad-equito “at-ride” ‘ride up to’, in-equito “in-ride”, ab-erro “away-
wander”, ad-erro “at-wander” ‘wander up to’, ex-erro “out-wander”, in-erro
“in-wander”, ad-fluo “at-flow” ‘flow up to’, ef-fluo “out-flow”, in-fluo “in-flow”,
au-fugio “away-flee”, ef-fugio “out-flee”, ad-labor “at-slip” ‘slip up to’, il-labor
(in þ labor) “in-slip”, e-labor “out-slip”, ad-navigo “at-sail” ‘sail up to’, e-navigo
“out-sail”, in-navigo “in-sail”, ab-nato “away-swim”, ad-nato “at-swim” ‘swim
up to’, e-nato “out-swim”, in-nato “in-swim”, ad-no “at-swim” ‘swim up to’,
e-no “out-swim”, in-no “in-swim”, ap-propero “at-hasten” ‘hasten up to’,
im-propero “in-hasten”, ad-repo “at-crawl” ‘crawl up to’, e-repo “out-crawl”,
in-repo “in-crawl”, ab-silio (ab þ salio) “away-jump”, as-silio (ad þ salio) “at-
jump” ‘jump to’, ex-silio (ex þ salio) “out-jump”, in-silio (in þ salio) “in-jump”
‘jump onto’, e-vago “out-wander”, a-volo “away-fly”, ad-volo “at-fly” ‘fly onto’,
e-volo “out-fly”, in-volo “in-fly”
Latin as a satellite-framed language
I shall not attempt an analysis of the results obtained. In particular, I shall not
provide an account of the relation between unaccusativity and unergativity and the
relevant tests. My only (modest) aim is to show that two unaccusativity/unergativity
diagnostics which have been applied in other languages—the cognate object and
measure phrase diagnostics and the agent noun diagnostics—also work for Latin.
First, I show how CDMCs do not allow a certain class of ‘objects’ that have been
independently shown to be allowed only with unergative predicates: cognate objects
and measure phrases.
Cognate objects, or internal objects in the Latin linguistics tradition (Hofmann and
Szantyr : ; Bortolussi ; Pinkster : ; Serbat b), are objects that
share the same root as the verb with which they appear. For instance, in the predicate
of the following example the accusative object vitam ‘life’ shares the same root as the
verb vivo ‘live’:
() Latin; Ter. Ad.
Vitam duram [ . . . ] vixi.
life.ACC hard.ACC live.PRF.SG
‘I have lived a hard life.’
Several authors (Larson ; Massam ; Levin and Rappaport Hovav ) have
proposed that cognate objects are only allowed with unergative verbs. Unaccusative
verbs do not license them, as shown in the following examples:15
() Levin and Rappaport Hovav (: , )
a. *The glass broke a crooked break.
b. *The apples fell a smooth fall.
c. *She arrived a glamorous arrival.
Measure phrases (the so-called accusative of extension; cf., for example, Ernout and
Thomas : ) are quantified NPs that behave, partly, as standard objects.
Importantly, as pointed out by Real Puigdollers (), measure phrases also resist
appearing in unaccusative predicates (see ()a), but are perfectly normal in
unergative ones (see ()b):
() Catalan; Real Puigdollers (: )
a. *El Pere arriba tres metres del seu poble.
the Pere arrives three metres from=the his village
b. El Josep camina quatre quilòmetres.
the Josep walks four kilometres
‘Josep walks four kilometres.’
15
See Kuno and Takami (: ff.) for a different view.
S-framed constructions in Latin
If this generalization is on the right track, we do not expect CDMCs in Latin to
appear with cognate objects or measure phrases. This seems to be true at least for the
CDMCs based on the prefixed manner-of-motion verbs in (). On the one hand,
the search for constructions including a cognate object was based on combinations
of each prefixed verb with its corresponding nominal, in the accusative. The list
of nominals listed in Gaffiot () and present in the Antiquitas corpus is
displayed in ():
() Nominalizations based on the verbs in ()
in-ambulatio “in-walk”, ex-cursus “out-run”, ex-cursio “out-run”, in-cursio
“in-run”, in-cursus “in-run”, ab-erratio “away-wandering”, ef-fugium “out-
flee”, ad-lapsus “at-slip”, in-sultura (from in-silio: in ‘in’ þ salio ‘jump’),
e-vagatio “out-wandering”, ad-volatus “at-flight”, in-volatus “in-flight”
This search did not produce any positive results. The other search involved combin-
ations of the same prefixed verbs with the measure nouns displayed in the
following list:
() Measure nouns
pes ‘foot’, passus ‘pace’, stadium ‘stadium’, mille passuum ‘a mile (a thousand feet)’
This search produced two alleged cases of CDMCs combined with a measure phrase:
() Latin; combinations of prefixed manner-of-motion verbs with measure phrases
a. Liv. , ,
Inde lingua in altum mille passuum
thence tongue(F)NOM.SG in sea.ACC thousand pace.GEN.PL
ex-currens medium fere sinum [ . . . ] distinguit.
out-run.PTCP.PRS.NOM.F.SG middle.ACC almost bay.ACC divide.SG
‘Thence a tongue of land stretching out about a mile into the sea divides
the bay nearly in the middle.’
b. Plin. Nat. , ,
Mons [ . . . ] a planitie ex-currit in maria
mountain.NOM off plain.ABL out-run.SG in sea.ACC.PL
LXXV passuum.
pace.GEN.PL
‘The mountain extends from the plain into the sea a distance of seventy-
five paces.’
Both examples involve the verb ex-curro “out-run”. Crucially, however, they both
involve a static description: that of the extension of a tongue of land in ()a and of a
mountain in ()b. No movement is involved in either case, nor an activity
interpretation, for that matter. I suspect that this circumstance may account for the
Latin as a satellite-framed language
availability of the measure phrase. I will return to prefixed verbs interpreted
statically—involving so-called fictive motion—in section ....
On the other hand, some of the simple manner-of-motion verbs in () are found
to combine either with cognate objects or with measure phrases. This is in conform-
ity with the unergative status I have ascribed to them in (). I illustrate this with the
following three verbs:
() Latin; cognate objects and measure phrases with some of the verbs in ()
a. Cic. Pro P. Q. ,
Possit [ . . . ] septingenta milia passuum ambulare
can.SBJV.SG seventy thousand pace.GEN.PL walk.INF
‘He could walk seventy thousand paces.’
b. Liv. , ,
Cursus rigida ceruice [ . . . ] currentium.
run.ACC.PL stiff.ABL neck.ABL run.PTCP.PRS.GEN.PL
‘Of those who were running the race with a stiff neck’.
c. Plin. Nat. , ,
Proditur Alexandrum nullo die
tell.PASS.SG Alexander.ACC no.ABL day.ABL
minus stadia DC navigasse.
less.ACC stadium.ACC sail.INF.PFV
‘It is said that Alexander would never sail less than stadia a day.’
Another test to prove the unaccusative status of a predicate involves checking
whether it fails to yield agent nouns. Levin and Rappaport () argue that
unaccusative verbs, unlike unergative verbs, may not yield agent nouns based on
the suffix -er in English. Thus, while such formations as runner, walker, or swimmer,
based on unergative verbs, are acceptable, *arriver, *appearer, or *disappearer, based
on unaccusative verbs, are not.16
If we submit the simple and prefixed manner-of-motion verbs of () and ()
to this test, we find, respectively, the following agent nouns, as listed by Gaffiot ()
and present in the Antiquitas corpus:
() Agentive nominalizations based on unprefixed manner-of-motion verbs in ()
ambulator/ambulatrix ‘walker’ (m./f.), cursor ‘runner’, eques ‘rider’, errator
‘wanderer’, fugitor ‘fugitive’, navigator ‘sailor’, natator ‘swimmer’, saltor ‘dan-
cer’, saltator/saltatrix ‘dancer’ (m./f.)
16
An anonymous reviewer points out that this test has counterexamples such as early arriver, based on
unaccusative arrive. Crucially, however, the expression is felicitous to the extent that an adverb (early) is
used. I hypothesize that the structure involved in the -er nominalization imposes an agentive interpretation
with which arrive early (intentionally), but not arrive, is compatible.
S-framed constructions in Latin
() Agentive nominalizations based on prefixed manner-of-motion verbs in ()
ex-cursor “out-runner” ‘scout, spy’
As we can see, while Gaffiot () records eight agent nouns based on unprefixed
manner-of-motion verbs, only one agent noun, excursor, is found related to a
prefixed verb, excurro “out-run”. However, excursor does not mean ‘one who runs
out’, as we would expect from the structural meaning of excurro, but ‘scout’ or ‘spy’.
I conclude, therefore, that the test involving the failure to yield agent nouns also
selects the set of prefixed manner-of-motion verbs as unaccusatives.17
.. Figure Unselected Object Constructions
Unselected Object Constructions (UOCs) are constructions involving internal argu-
ments (direct objects or unaccusative subjects) not semantically selected by the verb
and not omissible in the construction. Crucially, a predicative element in the UOC,
whether a particle, an AP, or a PP, is the licenser of the unselected argument, as the
following cases show:
() Unselected direct objects; McIntyre (: )
a. Sue shouted *(John) deaf./ Sue shouted John *(deaf).
b. Sue wrestled *(John) to the floor./ Sue wrestled John *(to the floor).
c. Sue worked *(her debt) off./ Sue worked her debt *(off).
() Dutch: unselected unaccusative subjects; Hoekstra (: )
a. Dat mijn jas *(nat) regent.
that my coat wet rains
‘That my coat rains wet’.
b. Dat het papiertje *(in de sloot) waait.
that the paper in the ditch blows
‘That the paper blows into the ditch’.
c. Dat de plant *(onder) sneewt.
that the plant under snows
‘That the plant snows under’.
17
Interestingly, an unaccusativity test standard in Romance languages, namely, the licensing of past
participles in absolute constructions (see Burzio for Italian, Legendre for French, and
Mendikoetxea for Spanish), does not seem to pick out the class of unaccusatives in Latin, but, rather,
that of intransitive deponent verbs like morior ‘die’ (in addition, of course, to that of transitive non-
deponent verbs like mitto ‘send’). Thus, as far as I have been able to check for the verbs in () and (),
only the deponent labor ‘slip, slide, fall’ and its prefixed variants allow participles used as adjectives. See
Gianollo (: ff.) for the same observation. See, for constructions involving absolute participles in
Latin, Bolkestein (, ); Lavency (); and Coleman (), among others.
Latin as a satellite-framed language
UOCs, in particular those involving objects, have been analysed by Mateu (a) as
s-framed constructions, after attesting the fact that they are not allowed in v-framed
languages. I illustrate with the anomalous Catalan renditions of the predicates above:
() Catalan renditions of ()
a. *La Sue cridà en John sord.
b. *La Sue lluità en John a terra.
c. *La Sue treballà el seu deute fora.
() Catalan renditions of ()
a. *El meu abric (es) plou humit.
b. *El paperet (es) bufa al forat.
c. *La planta (es) neva a sota.
In a nutshell, Mateu (b) adopts a Hoekstrian (Hoekstra ) analysis in terms
of a Small Clause. UOCs, then, involve an abstract causative V and a Small Clause
complement whose subject is the unselected object of the UOC and whose predicate
is the piece of the UOC licensing the unselected object: a particle, a PP, or an
AP. The Small Clause is headed by a prepositional head. Since in s-framed languages
(like English and Dutch above) this prepositional head is realized independently
from the eventive V head, V may host an independent unergative structure codifying
the accompanying Co-event (a shouting event, for instance, in ()a). On the other
hand, v-framed languages, like Catalan, do not license the constructions, since the
prepositional head is conflated into V and conflation of an independent element is
incompatible with this circumstance. I assume a similar analysis:
() Analysis of ()a
[VoiceP Sue [Voice’ Voice [vP [v v SHOUT] [PathP John [Path’ Path [PlaceP John [Place’
Place DEAF]]]]]]]
Motivation for this analysis is mainly based on the semantic interpretation of the
predicative piece that licenses the unselected object. In (), for instance, the
inference is licensed that as the result of some event originated by Sue, which is
identified with a shouting event (see the adjunction relation of SHOUT with v), John
ends up deaf. John, interpreted as Figure in the predicative relation structured around
Place, is, in turn, interpreted as a Measurer of the event in Spec-Path position.
In this book I will analyse two types of UOCs in Latin: Figure UOCs (this section)
and Ground UOCs (section ..). In the former type the internal argument corres-
ponds to the Figure, that is, to the DP merged as Spec-Place. In the latter type the
internal argument corresponds to (and is interpreted as) the Ground, that is, the DP
merged as Compl-Place. What unifies both types is that, whether Figure or Ground,
this DP is internally merged as Spec-Path, where it is interpreted as a Measurer.
Different constructions will be shown to be UOCs in confronting them with
S-framed constructions in Latin
constructions involving the same verb but in the absence of a special context. The set
of semantic and syntactic differences between both types of constructions will be
established and shown to be naturally derived from the status of UOCs as change
predicates involving a PathP. Additionally, UOCs are presented as an optimal case
study to show how a neo-constructionist view of argument structure naturally
predicts that the licensing conditions in predicates—in the current case, the licensing
of objects—depend on the syntactically assembled pieces they are made of, and not
on a single projecting nucleus (the verb).
As just mentioned, Figure UOCs feature an internal argument interpreted as
Figure. Figure UOCs are very commonly represented in Latin in the form of predi-
cates headed by a prefixed verb and accompanied, sometimes, by a directional DP or
PP. I will be illustrating Figure UOCs through the prefixes ex-, ab-, and in-, and I will
show the semantic and syntactic differences between the unprefixed and prefixed
predicates. Then I will focus on a series of particular properties of these Figure UOCs:
the licensing of null objects, case and situation-aspect properties, and scopal
relations between prefix and verb. I shall argue that these properties naturally derive
from a syntactic neo-constructionist perspective on argument structure and word
formation.
... The syntax and semantics of prefixed vs unprefixed verbs In this section
I illustrate Figure UOCs in Latin through predicates headed by ex-, ab-, and in-
prefixed verbs. I show the great elasticity of verbs by pointing out the semantic and
syntactic differences between the prefixed verbs and their unprefixed counterparts.18
The prefix ex- (with the variant e-) has the core meaning of ‘out’. This is shown by
the following UOCs:
() Latin; Cato, Agr. ,
Qui oletum saepissime et altissime miscebit,
who.NOM olive-tree.ACC often.SUPERL and deeply.SUPERL mix.FUT.SG
is tenuissimas radices ex-arabit.
he.NOM slender.SUPERL.ACC.PL root.ACC.PL out-plough.FUT.SG
‘He who works his olives very often and very deep will plough out the most
slender roots.’
() Latin; Verg. Aen. ,
Immanisque columnas rupibus ex-cidunt.
huge.ACC.PL=and column.ACC.PL rock.ABL.PL out-cut.PL
‘And they hew huge columns out of rocks.’
18
For more in-depth studies of the semantics of Latin verbal prefixes, see Pottier () and García
Hernández ().
Latin as a satellite-framed language
() Latin; Plaut. Capt.
Tum pistores scrofipasci qui alunt
then miller.NOM.PL sow-breeding.NOM.PL who.NOM.PL feed.PL
furfuribus sues, [ . . . ]: eorum si quoiusquam scrofam in
bran.ABL.PL sow.ACC.PL them.GEN if anyone.GEN sow.ACC in
publico conspexero, ex ipsis dominis meis
public.ABL spot.FUT.PFV.SG out own.ABL.PL master.ABL mine.ABL.PL
pugnis ex-culcabo furfures.
fist.ABL.PL out-tread.FUT.SG bran.ACC
‘Then those sow-breeding millers who feed their swine with bran, [ . . . ]: if
I see the sow of any of them out in the street I will stamp the bran out of their
very masters with my fists.’
() Latin; Plaut. Capt.
HEGIO: Quid diuitiae? Sunt ne opimae?
what.ACC richness(F)NOM.PL are PART.INTER abundant.NOM.F.PL
PHILOCRATES: Vnde ex-coquat sebum senex.
whence out-boil.SBJV.SG tallow.ACC old_man.NOM
‘HEGIO: What about his riches? Are they abundant?—PHILOCRATES: So much
that the old rascal could melt out the tallow.’
() Latin; Varro, Rust. , ,
Apud alios ex-teritur grege iumentorum
by other.ACC.PL out-grind.PASS.SG herd(M)ABL.SG cattle.GEN.PL
in-acto [ . . . ], quod ungulis e spica
in-drive.PTCP.PFV.M.ABL.SG so_that hoof.ABL.PL out ear.ABL
ex-teruntur granae.
out-grind.PASS.PL grain.NOM.PL
‘Others cause it to be trodden out with a herd of cattle driven thereupon,
[ . . . ] so that the grains are trodden out of the ear under their hoofs.’
() Latin; Plin. Nat. ,
[Serpentes] [ova] solida hauriunt, [ . . . ] atque
snake(M)NOM.PL egg.ACC.PL whole.ACC.PL swallow.PL and
putamina ex-tussiunt.
shell.ACC.PL out-cough.PL
‘Snakes swallow the eggs whole and expel the shells through coughing.’
The Core Schema expressed in the above sentences is sometimes made specific by
overt directional PPs, as is the case in () with ex ipsis dominis ‘out of their very
masters’ or () with e spica ‘out of the ear’. In both cases the prefix coincides with
the preposition. In some other cases, however, the Core Schema is inferred context-
ually or through world knowledge. Thus, in () the Ground must be the earth or
S-framed constructions in Latin
tilling ground and in () it is the snakes’ own bodies. All the cases involve the
projection of a PathP, the adjunction of a Manner root to v, introducing the Co-
event, and root EX ‘out’, merged at Compl-Place, where it is interpreted as a Ground
(a Terminal Ground, in fact, since PlaceP is embedded under PathP, encoding a
transition), and predicates such as that of () emerge, where the specific reference
of the Ground is calculated contextually:
() Analysis of ()
[VoiceP is [Voice’ Voice [vP [v v AR] [PathP tenuissimas radices [Path’ Path [PlaceP
tenuissimas radices [Place’ Place EX]]]]]]]
The abstract final location expressed by the prefixal root may be further specified by
an adjunct, as in ():
() Analysis of ()
VoiceP
pro Voice’
Voice vP
v PathP
v cid immanis columnas Path’
Path PlaceP
pP PlaceP
p rupibus
immanis columnas Place’
Place ex
My analysis of prefixed verbs in Latin can claim to formalize Serbat’s (a: ff.)
important insight that in a prefixed predicate the verb is a secondary predicate, while
the prefix is the main predicate,19 although it clashes with most of Latin linguistics
tradition, in which the prefixed verb is the result of the affixation of a prepositional
element to a pre-existing simple verb20 (Donatus, th century AD; Priscian, th century
19
I reject, however, Serbat’s (a: ) implementation, where the verbal inflectional morphology
changes the prefix into a verbal predicate, since I dissociate the notion of morphosyntactic category from
that of predicate.
20
Note, also, that my analysis of Latin prefixation is orthogonal to the question of whether it is a case of
composition or of derivation, since these are not primitive concepts in the theory of Distributed
Latin as a satellite-framed language
AD; Brachet : ; Romagno : ). The analysis also assumes that whatever
semantic relation is established between the object and the verb, it is the result of, on
the one hand, the interpretation of the structure in which they appear (the structural
semantics) and the roots merged within that structure (the encyclopaedic semantics):
there is no (direct) thematic relation between the object and the verb. Thus, the object
is always interpreted as a subject of a predication established by an abstract head,
Place. In turn, it is also interpreted in the structures above as a Measurer of the event.
In (), for instance, the event is over when the huge columns are literally out of the
rocks: there is a direct relation between the quantity expressed by columnas ‘columns’
and the quantification of the event itself. In turn, the interpretation of the verb relies
on the existence of an event introduced by v and a Manner Co-event expressed by the
root adjoined to it. Note, then, that the conceptual dimensions of the verb and of
the object are completely severed from each other. In (), for instance, there is no
direct conceptual relation between the hewing activity and the columns, nor are they
affected thereby. Although this might seem counter-intuitive at first sight, it is
supported by cases of UOCs where the simple verb, outside the UOC, does not
usually take any object. In the above examples there is actually one such case: that of
(), headed by ex-tussio “out-cough” ‘expectorate (something) through coughing’.
Simple tussio ‘cough’ is not registered to allow for any objects (Gaffiot ). The
meaning of extussio is furthermore not licensed with an independent ex-PP. The rest
of the examples constitute cases of ‘weak’ UOCs, in that their simple counterparts
can be transitive but do not license the same type of object. These UOCs, however,
are also able to refute the inference that makes objects in prefixed-verb predicates
seem to be affected by the action conveyed by the verb. Thus, in () the seeds,
necessarily, are not damaged by a rubbing or grinding action (conveyed by tero
‘grind’), as are, naturally, the husks out of which they come. The same happens with
ex-aro “out-plough” in (): the roots are not ploughed (AR), but extracted (EX) upon
ploughing (AR).
Under the present assumptions, the verbs are expected to show, on the surface, a
great elasticity, since roots may, in principle, be merged in any context (where they
are structurally admitted, that is: as complements or adjuncts of functional heads).
This elasticity is, I argue, restricted by clashes between the encyclopaedic content of a
root and the interpretation of the position it occupies in the structure. Analyses
attributing grammatical features to verbs (roots) fail to predict this elasticity. For
instance, Lehmann (: ) proposes that the argument structure of the base verb
should be kept in the prefixed counterpart, that is, added to that of the preverb—see
also Carvalho (: ). However, this hypothesis cannot explain why an obliga-
torily transitive verb like rumpo ‘break’ may be used as an intransitive when prefixed:
Morphology. For phonological arguments in favour of the former view, see Heslin (). For phono-
logical, semantic, and syntactic arguments in favour of the latter view, see Oniga ().
S-framed constructions in Latin
() Latin; Ov. Met. ,
Virago [ . . . ] rupit [ . . . ] vestes
maiden.NOM break.PST.SG robes.ACC
‘The vigorous maiden tore the robes.’
() Latin; Cic. Verr. , ,
[Ignes] qui ex Aetnae vertice e-rumpunt.
fire(M)NOM.PL which.NOM.M.PL out Aetna.GEN summit.ABL out-break.PL
‘Fires that spurt out of the summit of the Aetna’.
In the second example the argument structure of rumpo ‘break’, as shown in the
previous example, seems to have vanished: neither the Agent nor the Patient are
required. Instead, the prefixed verb appears with a Theme argument (ignes ‘fires’). In
sections to come I will provide more illustrations of the loss of argument structure
properties of the base verb when it is prefixed.
The verbs appearing in () to () are licensed, as is to be expected, in other
contexts, where they receive other interpretations. For instance, they appear in
intransitive environments with an activity interpretation:
() Latin; unergative uses of some of the simple verbs in () to ()
a. Cic. Cato ,
Aranti L. Quinctio Cincinnato
plough.PTCP.PRS.DAT.M.SG L. Quinctius.DAT Cincinnatus.DAT
nuntiatum est eum dictatorem esse
announce.PTCP.PFV.NOM.N.SG is he.ACC dictator.ACC.M.SG be.INF
factum.
PTCP.PFV.NOM.M.SG
‘L. Quinctius Cincinnatus was ploughing when it was announced that he
had been designated dictator.’
b. Cato, Agr. ,
Ubi radices bene operueris, calcare bene,
when root.ACC.PL well bury.FUT.PFV.SG tread.INF well
ne aqua noceat.
lest water.NOM damage.SBJV.SG
‘When you have buried the roots well, tread well, lest water should damage
them.’
c. Plaut. Aul.
In nonum diem solet ire coctum
In ninth.ACC day.ACC use.SG go.INF cook.SUP.ACC
‘He usually cooks every nine days.’
Latin as a satellite-framed language
d. Hor. Sat. , ,
Si quis forte coheredum senior male tussiet.
if any.NOM haply co-heir.GEN.PL old.COMPAR.NOM badly cough.SBJV.SG
‘If haply any of your co-heirs, being advanced in years, should have a
dangerous cough’.
In all the above structures the verb conveys an activity; as such, it is the surface
manifestation of a root that is the complement of a v head. The root is, thus,
interpreted as an Effected Object. The analysis of ()d illustrates:
() Analysis of ()d
[VoiceP quis [Voice’ Voice [vP v TUSS]]]
The structure in () cannot license any object, as does that in (): there is no
place in the structure for any additional argument. The only argument is external,
quis ‘any’, the Originator, merged as Spec-Voice.
Of course, these verbs may also appear in transitive constructions without a prefix
(or any other directional expression):
() Latin; transitive uses of some of the simple verbs in () to ()
a. Varr. Ling. , ,
Omnes qui terram arabant.
all.NOM.PL who.NOM.PL earth.ACC plough.IPFV.PL
‘All those who ploughed the earth’.
b. Cato, Agr. ,
Frondem populneam, ulmeam, querneam caedito
foliage.ACC of_poplar.ACC of_elm.ACC oaken.ACC cut.IMP.FUT.SG
per tempus.
through time.ACC
‘Cut the foliage of the poplar, elm, oak, each at a time.’
c. Colum. ,
Cum uvam legeris et calcaveris [ . . . ]
when grape.ACC collect.FUT.PFV.SG and tread.FUT.PFV.SG
mustum in cor<ti>nam de-fundas.
must.ACC in caldron.ACC downward-throw.SBJV.SG
‘You should pour the must into the caldron after you have selected and
trodden the grapes.’
d. Hor. Ars,
Ne [ . . . ] humana palam coquat exta
not human.ACC.N.PL openly cook.SBJV.SG entrail(N)ACC.PL
nefarius Atreus.
nefarious.NOM Atreus.NOM
‘Let nefarious Atreus not cook human entrails openly.’
S-framed constructions in Latin
e. Varro, Rust. , ,
Id secundum aream faciendum,
it.ACC depending threshing_floor.ACC do.PTCP.FUT.PASS.NOM.N.SG
ubi triturus sis frumentum.
where thresh.PTCP.FUT.NOM.M.SG be.SBJV.SG wheat.ACC
‘This is to be done depending on the threshing floor where you are to
thresh the wheat.’
Some of the above sentences are amenable to an analysis where no PathP is projected
and some to an analysis where PathP is projected and a transition is interpreted. For
instance, ()b seems a good candidate for an analysis in terms of a PathP encoding
a change and a final state, namely the state of the foliage being cut:
() A (plausible) analysis of ()b
[VoiceP (tu) [Voice’ Voice [vP v [PathP frondem [Path’ Path [PlaceP frondem [Place’
Place CAED]]]]]]]
On the other hand, ()a seems a good candidate, at least in one of its possible
interpretations, for an analysis where there is no PathP, and hence, the object cannot
induce a telic reading of the event:
() A (plausible) analysis of ()a
[VoiceP qui [Voice’ Voice [vP [pP p terram] [vP v AR]]]]
() yields an interpretation where the ploughing activity is exerted on terram ‘the
earth’ merged as an adjunct to vP, with no resulting state entailed (see section ...).
Certain ex-verbs exist which head predicates where the Ground, rather than a physical
entity, is someone’s spiritual dimension or their possessions. They thus imply that
something (the Figure object) is obtained from someone by some activity, specified by
the root merged as an adjunct to the eventive v head. For instance, in the following
examples things are obtained through flattery, enchantment, and caresses, respectively:
() Latin; Liv. , ,
Neque enim omnia emebat aut e-blandiebatur
nor in_fact all.ACC.N.PL buy.IPFV.SG or out-flatter.IPFV.SG
‘Nor did he acquire everything by money or flattery.’
() Latin; Sen. Nat. b, ,
Ne quis alienos fructus
lest anybody.NOM of_another.ACC.PL fruit.ACC.PL
ex-cantassit.
out-enchant.PLUPRF.SBJV.SG
‘Lest anyone should obtain someone else’s fruits through enchantment’.
Latin as a satellite-framed language
() Latin; Plaut. Vid. Frag. ,
Nunc seruos argentum a patre ex-palpabitur.
now slave.NOM.SG money.ACC off father.ABL out-caress.FUT.SG
‘Now the slave will obtain the money out of the father through caresses.’
The semantic relation between the verb and the accusative object is, again, com-
pletely different in () to (), on the one hand, and in the unprefixed cases of
(), on the other. For instance, it is quite evident that the accusative DP argentum
‘the money’ in () is not the object of the flattery, but the accusative quem ‘whom’
(referring to Delator ‘betrayer’) is the object of the caresses in ()b.
() Latin; transitive uses of some of the simple verbs in () to ()
a. Ov. Met. ,
Cantato densetur carmine caelum.
sing.PTCP.PFV.ABL.N.SG darken.PASS.SG spell(N)ABL.SG sky.NOM
‘As her spell is sung out, the sky darkens.’
b. Iuv. ,
Delator [ . . . ] quem munere palpat Carus.
betrayer.NOM whom.ACC present.ABL caress.SG Carus.NOM
‘A betrayer, whom Carus flatters with his present.’
These verbs may also appear in an unergative construction, as shown below:
() Latin; unergative uses of some of the simple verbs in () to ()
a. Plaut. Men.
Meretrix tantisper blanditur,
courtesan.NOM so_long flatter.SG
dum illud quod rapiat uidet.
as_long_as that.ACC which.ACC seize.SBJV.SG see.SG
‘A courtesan flatters about as long as she sees what she may seize.’
b. Verg. Ecl. ,
Frigidus in pratis cantando
cold.M.NOM.SG in meadow.ABL.PL sing.GERUND.ABL
rumpitur anguis.
break.PASS.SG snake(M)NOM
‘The cold snake in the fields is ripped apart by the enchantment.’
c. Apul. Met. ,
Verbis palpantibus stimulum
word.ABL.N.PL caress.PTCP.PRS.ABL.N.PL torment.ACC
doloris obtundere.
sorrow.GEN calm_down.INF
‘(He tried) to calm down the torment of her sorrow with caressing words.’
S-framed constructions in Latin
In the examples seen so far the Ground is a concrete entity. But it can also be
understood more abstractly, as a general ‘here and now’, facilitating a ‘disappearance’
sense for the prefix:
() Latin; Cic. Phil. ,
E-dormi crapulam, inquam.
out-sleep.IMP.SG intoxication.ACC say.SG
‘Sleep off the intoxication, I said.’
() Latin; Cato, Agr. ,
Usque coquito, dum dimidium ex-coquas.
until cook.IMP.FUT.SG until half.ACC out-cook.SBJV.SG
‘Boil it until you boil half of it away.’
In () there is an example of simple dormio ‘sleep’. Since it cannot take objects in
the accusative, predicates involving e-dormio “out-sleep” constitute one of those
cases of UOCs where the prefix is ostensibly facilitating the projection of an accusa-
tive object:
() Latin; Ov. Rem.
Thalamo dormimus in illo.
bridal bed(M)ABL.SG sleep.PL in that.ABL.M.SG
‘We slept in that bridal bed.’
The verb ex-coquo “out-boil”, on the other hand, already appeared in () (repeated
below as ()) as an example of UOC that, although hyperbolically used, involves a
concrete entity as Ground (someone’s fortune):
() Latin; Plaut. Capt.
HEGIO: Quid diuitiae? Sunt ne opimae?
what.ACC richness(F)NOM.PL are PART.INTER abundant.F.NOM.PL
PHILOCRATES: Vnde ex-coquat sebum senex.
whence out-boil.SBJV.SG tallow.ACC old_man.NOM
‘HEGIO: What about his riches? Are they abundant?—PHILOCRATES: So much
that the old rascal could melt out the tallow.’
The semantic difference between () and () consists, then, in the fact that in the
former the object undergoes disappearance, while in the latter it happens to appear
out of somewhere, this location being identified by the pronoun unde ‘whence’. In
both cases the root COQU ‘boil’ is merged as an adjunct modifier of the change-of-state
predicate headed by the Path head: the boiling/melting event is in both cases a
manner co-event. The difference lies, I argue, in the nature of the element identified
as the location: non-referential in () and referential in (). Specifically, a
plausible analysis involves merging the root of the prefix as Compl-Place in both
Latin as a satellite-framed language
cases, while having the relative adverb unde ‘whence’ an adjunct to PlaceP in (), as
shown below:
() Analyses of () and ()
a. [VoiceP (tu) [Voice’ Voice [vP [v v COQU] [PathP dimidium [Path’ Path [PlaceP
dimidium [Place’ Place EX]]]]]]]
b. [VoiceP (tu) [Voice’ Voice [vP [v v COQU] [PathP sebum [Path’ Path [PlaceP unde
[PlaceP sebum [Place’ Place EX]]]]]]]]
A specialization of the disappearance sense is found in verbs where the Figure goes
away through expenditure. The way that expenditure is carried out is, as is to be
expected, expressed by the verbal root:
() Latin; Hor. Sat. , ,
Filius [ . . . ] haec [ . . . ] ut e-bibat [ . . . ] custodis?
son.NOM this.ACC.PL that out-drink.SBJV.SG guard.SG
‘You guard [these possessions] to the end that thy son guzzles them all up?’
() Latin; Plaut. Trin.
LESBONICUS: Quid factumst eo [minas quadraginta]?
what.NOM made=is it.ABL mina.ACC.PL forty
Stasimus: Com-essum, ex-potum,
with-eat.PTCP.PFV.NOM.N.SG out-drink.PTCP.PFV.NOM.N.SG
ex-unctum, e-lotum
out-anoint.PTCP.PFV.NOM.N.SG out-wash.PTCP.PFV.NOM.N.SG
in balineis.
in bath.ABL.PL
‘LESBONICUS: What has been done with it (forty minas)?—STASIMUS: It has
been eaten, drunk up, spent away in unguents, washed away in baths.’
The kind of object appearing with the simple counterparts of these verbs is quite
different. I capitalize here on the differences between bibo ‘drink’ and e-bibo “out-
drink” ‘drink up’ (see Vendryès and Brachet : ) in an attempt to provide
new evidence in support of a syntactic analysis of these phenomena. The following is
an example of simple bibo ‘drink’ where there is no specific entailment that the water
is exhausted through drinking. On the contrary, bibo ‘drink’ expresses an activity:
() Latin; Cat. Agr. ,
Per aestatem boues aquam bonam
through summer.ACC cow.NOM.PL water.ACC good.ACC
et liquidam bibant semper curato.
and clear.ACC drink.SBJV.PL always care.IMP.FUT.SG
‘One must always see to it that cows drink good and clear water all through
the summer.’
S-framed constructions in Latin
The contrast is particularly dramatic in the next example. I have included the whole
paragraph, since it involves both bibo ‘drink’ and e-bibo “out-drink”:
() Latin; Petr. Sat. ,
‘Quid? ego’ inquit ‘non sum dignus qui bibam?’
what I say.SG not am worthy.NOM who.NOM drink.SBJV.SG
ancilla risu meo prodita complosit
serf.NOM.F.SG laugh.ABL my.ABL betray.PTCP.PFV.NOM.F.SG clap.PRF.SG
manus et ‘apposui quidem adulescens, solus
hand.ACC.PL and serve_up.PRF.SG certainly youth.VOC alone.NOM.M.SG
tantum medicamentum e-bibisti?’
so_much.ACC medicine.ACC out-drink.PRF.SG
‘ita ne est?’ inquit Quartilla ‘quicquid satyrii
thus PART.INTERR is say.SG Quartilla.NOM whatever.ACC satyrion.GEN
fuit, Encolpius e-bibit?’
be.PRF.SG Encolpius.NOM out-drink.PRF.SG
‘ “Well, then, why should I not deserve to drink?” The serf, betrayed by my
laugh, clapped her hands and (said) “I have served you up already, youth. By
the way, have you drunk up such an amount of medicine all by yourself?”
“Really?”, said Quartilla, “Has Encolpius drunk up all the satyrion21?” ’
Simple bibam ‘I drink’ is interpreted as an atelic activity. Here the root is merged as a
Compl-v, and is interpreted as an Effected Object:
() Analysis of simple bibo
[VoiceP qui [Voice’ v BIB]]
However, the two instances of e-bibo “out-drink” express the exhaustion of the liquid,
as reflected on the translations. I propose that they correspond to a different structure:
() Analysis of tantum medicamentum ebibisti (in ())
[VoiceP (tu) [Voice’ Voice [vP [v v BIB] [PathP tantum medicamentum [Path’ Path
[PlaceP tantum medicamentum [Place’ Place EX]]]]]]]
The prefix originates as a root merged as Compl-Place; here it is understood as a
Terminal Ground, expressing the final state of the Figure tantum medicamentum: the
state of disappearance (akin to the one encoded by up in English drink the wine up).
The DP in Spec-Place rises to Spec-Path. There it is interpreted as Measurer: when
the amount described by tantum medicamentum ‘so much medicine’ reaches the
state described by the root EX ‘out’, the event, specified as a drinking event by the
adjunct root BIB ‘drink’, is over. I shall come back to the bibo/ebibo difference in
relation to the licensing of null objects in section ....
21
An aphrodisiac drink.
Latin as a satellite-framed language
I turn now to the prefix ab- (with variants a- and abs-) ‘off, away’, which presents a
central meaning of ‘separation from a surface’. This prefix is widely used in prefix-
ation to surface-contact verbs indicating the way in which the separation takes place:
() Latin; Tac. Hist. , ,
Is balineas ab-luendo
he.NOM bath.ACC.PL off-wash.PTCP.FUT.PASS.DAT.M.SG
cruori propere petit.22
blood(M)DAT.SG hastily head.SG
‘He hastened to the baths to wash off the blood.’
() Latin; Colum. Arb.
Sarmenta [ . . . ] arida [ . . . ] dolabra ab-radito.
shoot.ACC.PL dry.ACC.PL hatchet.ABL off-razor.IMP.FUT.SG
‘The dry vine shoots are to be razored off with a hatchet.’
() Latin; Liv. , ,
Inspectum vulnus
examine.PTCP.PFV.NOM.N.SG wound(N)NOM.SG
abs-terso cruore.
off-wipe.PTCP.PFV.ABL.M.SG blood(M)ABL.SG
‘That the wound had been examined after wiping the blood off ’.
() Latin; Hor. Sat. , ,
Cara piscis a-verrere mensa.
expensive.ABL.F.SG fish.ACC.PL off-sweep.INF stand(F)ABL.SG
‘To sweep away the fish from an expensive stand’.
The Ground in the above examples, corresponds to a surface whose identity is
discursively retrieved in (), (), and () (coreferent with vulnus ‘wound’);
in () the prefix coexists with an overtly expressed Ground in the ablative (cara
[ . . . ] mensa ‘an expensive stand’).
As is to be expected, the roots we find in the prefixed verbs above may appear in
other syntactic environments. In the following examples, tergo ‘wipe’ and verro ‘sweep’
appear in an unergative form, without any object, and with an activity interpretation:
() Latin; Cic. Parad. ,
Qui tergent, qui ungunt, qui verrunt.
who.NOM.PL wipe.PL who.NOM.PL anoint.PL who.NOM.PL sweep.PL
‘Those who wipe, those who anoint, those who sweep.’
The roots can appear in transitive predicates headed by simple verbs:
22
Abluendo cruori ‘to wash off the blood’ is a so-called gerundive construction, with a passive verbal
adjective abluendo agreeing with dative cruori, which is the logical object of the construction.
S-framed constructions in Latin
() Latin; transitive uses of some of the simple verbs in () to ()
a. Cic. Leg. ,
Mulieres genas ne radunto.
woman.NOM.PL cheek.ACC.PL not razor.IMP.FUT.PL
‘Do not let the women scratch their cheeks.’
b. Verg. Aen. ,
Clipeos [ . . . ] tergent arvina pingui.
shield.ACC.PL wipe.PL grease.ABL thick.ABL
‘They polish the shields with thick grease.’
c. Plaut. Merc.
Nil opust nobis ancilla,
nothing.NOM is_needed us.DAT slave_girl.NOM
nisi quae [ . . . ] aedis uerrat.
except who.NOM.F.SG house.ACC.PL sweep.SBJV.SG
‘We need nothing but a slave girl who can sweep the house.’
While the objects in predicates headed by prefixed verbs are understood as entities
which, through different process, become separated from a surface (explicit or not),
the ones in ()a to ()c refer, on the contrary, to surfaces on which the action
portrayed by the verb is exerted. Note, for instance, that genas ‘cheeks’, in ()a, are
not cut off from anywhere, as is the case with sarmenta ‘vine shoots’ in (). Instead,
genas ‘cheeks’ in ()a are understood as surfaces where a scratching action
takes place.
As is the case with ex- ‘out’, the sense of ab- as ‘separation from a surface’ meaning
is easily extended to a disappearance meaning, including the ‘spend by X-ing’ sense
we saw before. In this case, the Ground is understood deictically, as a general ‘here
and now’:
() Latin; Apul. Met. ,
Iucundiora [ . . . ] ab-ligurribam dulcia.
delicious.COMPAR.ACC.N.PL away-lick.IPFV.SG sweet.ACC.PL
‘I used to lick away rather delicious sweets.’
() Latin; Cat. Agr. ,
Omne caseum cum melle ab-usus eris.
whole.ACC cheese.ACC with honey.ABL away-use.FUT.SG
‘You will have used up all the cheese with honey.’
() Latin; Ter. Eun.
Patria qui ab-ligurrierat bona.
paternal.ACC.N.PL who.NOM away-lick.PLUPRF.SG good(N)ACC.PL
‘Who had wasted the paternal goods luxuriously.’
Latin as a satellite-framed language
I return to example () and a discussion of the utor/ab-utor contrast in section
....
Finally, the prefix in- ‘in’ is present in predicates where there is motion into an
enclosure, as the following examples make clear:
() Latin; Cat. Agr. ,
[Sarmenta] con-cidito minute et ibidem
twig.ACC.PL together-cut.IMP.FUT.SG minutely and right_there
in-arato aut in-fodito.
in-plough.IMP.FUT.SG or in-dig.IMP.FUT.SG
‘Chop the twigs up minutely and plough them into the same place, or dig
them in.’
() Latin; Cic. Phil. ,
[Acta] quae ille in aes in-cidit.
act.ACC.PL which.ACC.PL he in brass.ACC in-cut.PRFSG
‘The acts which he engraved on brass’.
() Latin; Cato, Agr.
Ubi coctum erit, [ . . . ] papauer in-friato.
when cook.PTCP.PFV.NOM.N.SG be.FUT.SG poppy seed.ACC in-crumble.IMP.FUT.SG
‘When it is cooked, crumble some poppy seeds into it.’
() Latin; Apul. Met. ,
Quasi soporiferum quoddam uenenum
as_though soporific.ACC certain.ACC poison.ACC
cantharis im-misceret illis.
jar.DAT.PL in-mix.IPFV.SBJV.SG those.DAT
‘As though he were mixing some sort of soporific poison into those jars’.
() Latin; Cato, Agr. ,
Eodem silpium in-radito.23
there silphium.ACC in-grate.IMP.FUT.SG
‘Grate silphium into it.’
() Latin; Ov. Met. ,
Purpureasque notas filis
purple.ACC.PL=and motif.ACC.PL yarn(N)DAT.PL
in-texuit albis.
in-weave.PRF.SG white.DAT.N.PL
‘And she wove purple motifs into white yarns.’
23
Silphium, -ii: a plant. Cato is describing a recipe for cabbage, into which silphium must be grated.
S-framed constructions in Latin
The objects in the above examples are quite evidently semantically unselected by the
base verbs. For instance, in () a vine cannot be ploughed, but introduced
somewhere by ploughing. Likewise, in () acta ‘acts’ cannot be caesa ‘cut’, but
can be in-cisa “in-cut” ‘engraved’. In () the plant silphium is not the Patient of
a scraping event, rather it is a Figure which changes location through scraping.
Similarly in () notae ‘motifs, designs’ cannot be woven, but they can be woven
into the fabric, that is, introduced into the fabric by weaving. The unprefixed
counterparts of these verbs show completely different semantic relations with their
objects (see (); the first two examples are passives) and some of them are found in
unergative environments (see ()):
() Latin; transitive uses of the simple counterparts of some of the verbs in () to ()
a. Cato, Agr. ,
[Posse] hortum fodiri.
can.INF yard.NOM dig.INF.PASS
‘The garden may be dug.’
b. Varr. Rust. , ,
Terra [ . . . ] facile frietur.
earth.NOM easily grind.SBJV.PASS.SG
‘Earth crumbles easily.’
c. Ter. Haut.
Texentem telam studiose ipsam offendimus.
weave.PTCP.PRS.ACC.F.SG cloth.ACC painstakingly her.ACC find.PRF.PL
‘We found her painstakingly weaving a cloth.’
() Latin; unergative use of the unprefixed counterparts of two of the verbs in
() to (), Ter. Haut.
Te in fundo conspicer fodere aut arare.
you.ACC in farm.ABL spot.INF dig.INF or plough.INF
‘(I see) you digging or ploughing on your farm.’
I make a final observation on im-misceo “in-mix”, in (). This case is interesting
because one of the usual arguments of simple misceo ‘mix’ is missing, namely, that
referring to the substance or set of things with which the object is mixed, which may
appear in the dative, ablative, or as a PP (see, respectively, ()a, ()b, and ()c);
alternatively, misceo ‘mix’ may appear with two coordinated DPs referring to the
substance being mixed together (see ()d):
() Latin; simple misceo ‘mix’
a. Ov. Met. ,
Fletumque cruori miscuit.
tear.ACC=and blood.DAT mix.PRF.SG
‘She mixed her tears with his blood.’
Latin as a satellite-framed language
b. Hor. Sat. , ,
Surrentina [ . . . ] miscet faece Falerna vina.
Surrentine.ACC.N.PL mix.SG dregs.ABL Falernian.ABL wine(N)ACC.PL
‘He mixes Surrentine wines with Falernian dregs.’
c. Cato, Agr.
Caseum cum alica [ . . . ] misceto.
cheese.ACC with spelt.ABL mix.IMP.FUT.SG
‘Mix the cheese with spelt.’
d. Plin. Nat. , ,
Vinum et aquam miscent.
wine.ACC and water.ACC mix.PL
‘They mix wine and water together.’
The syntactic environment in which im-misceo “in-mix” is found is different, and
highly predictable: it is the syntactic environment of any UOC. It features, on the one
hand, a DP, soporiferum venenum ‘soporific poison’, interpreted as Figure and as a
Measurer of the event, since the quantity of poison determines the temporal span of
the mixing event. On the other hand, a directional dative DP expresses the final point
of a spatial transition: illis cantharis ‘those jars’. The main event, then, is a transition
whereby the poison (venenum soporiferum) ends up in the jars (illis cantharis)
through a mixing event (encoded in the root MISC ‘mix’). As I discussed in section
.., I assume that the directional dative is not a Ground, syntactically, but the
specifier of an applicative head interpreted as inalienably posessing the referent of the
Ground, in this case the root IN ‘in’:
() Analysis of ()
VoiceP
pro Voice’
Voice vP
v ApplP
v misc illis cantharis Appl’
Appl PathP
soporiferum venenum Path’
Path PlaceP
soporiferum venenum Place
Place in
S-framed constructions in Latin
Importantly, the substance, with which the venenum soporiferum ‘soporific poison’ is
mixed and which is presumably contained in the jars is not expressed; in fact, it cannot be
expressed in (), at least not as a part of the argument structure configuration in ().
This discussion shows that whatever event participants roots require as part of their
idiosyncratic content (in this case, the ‘second’ substance in a mixing event) can and must
be overriden if the structure demands it. The syntactic configuration, therefore, imposes
a certain interpretation on the root: while the unprefixed verb may identify a final state
(see ()), the prefixed counterpart must be interpreted as a co-event (see ()).24
... Conditions on the licensing of null objects: bibo ‘drink’ vs ebibo ‘drink up’ In
this section I point out a crucial syntactic difference between bibo ‘drink’ and ebibo
“out-drink” ‘drink up’, which, within the present account, receives a natural explan-
ation. Specifically, bibo ‘drink’ may appear without an object, focusing merely on a
process (often of drinking wine), as has been shown above in () and is further
shown in () (in the usage referred to in traditional grammars as absolute—cf.
Ernout and Thomas : ff.):
() Latin; Object-less bibo ‘drink’
Andr. Commoediarum fragmenta in aliis scriptis servata,
Edi bibi lusi.
eat.PRF.SG drink.PRF.SG play.PRF.SG
‘I ate, I drank, I played.’
On the contrary, the rare object-less instances of e-bibo “out-drink” found in the
corpus and shown in () appear after an entity is introduced in the discourse that
provides the reference of the object. An objective null category (represented by ei in
the examples) corresponds to the object of e-bibo “out-drink” in these instances:25
() Latin; e-bibo “out-drink” with null objects
a. Plaut. Curc.
Propino [magnum poculum]i:
bring_forth.SG big.ACC goblet.ACC
ille ei e-bibit, caput de-ponit, con-dormiscit.
he.NOM out-drink.SG head.ACC downward-put.SG together-sleep.SG
‘I bring forth a big cup to him: he gulps it empty, lays his head down, and
falls asleep.’
24
See Zeller (b) and McIntyre () for data and accounts of how Germanic particles—which
behave similarly to Latin prefixes—can neutralize the usual argument structure displayed by a verb and
oblige its internal argument to be demoted as an adjunct. I will return to this ‘demotion of arguments’ in
sections ... and ...
25
I note that ()a is not a Figure UOC. It is, rather, a Ground UOC, since the object (coindexed with
magnum poculum ‘big cup’) is interpreted as a Ground (the container out of which the wine is drunk).
However, the Figure/Ground UOC distinction is not crucial for the current purpose, namely showing the
syntactic differences between simple and prefixed verbs. I will deal with Ground UOCs in section ...
Latin as a satellite-framed language
b. Cat. Agr. ,
Postridie caputi ulpici con-terito cum
the_next_day head.ACC onion.GEN together-grind.IMP.FUT.SG with
hemina uini facitoque ei e-bibat.
half-mina.ABL wine.GEN make.IMP.FUT.SG=and out-drink.SBJV.SG
‘The next day mash the head of an onion, mix it with a half-mina of wine,
and make it drink it up.’26
c. Gell. , ,
Artemisia [ . . . ] ossa cineremquei eius
Artemisia.NOM bone.ACC.PL ashes.ACC=and he.GEN
mixta odoribus con-tusaque
mix.PTCP.PFV.ACC.N.PL spice.ABL.PL together-grind.PTCP.PFV.ACC.N.PL=and
in faciem pulveris aquae in-didit
in form.ACC powder.GEN water.DAT in-give.PRF.SG
ei e-bibitque.
out-drink.PRF.SG=and
‘Artemisia [ . . . ] mixed his bones and ashes with spices, ground them into
the form of a powder, put them in water, and drank them up.’
The situation shown in () is amenable to the general fact that Latin licenses an
empty category as object, as shown in the next example taken from Luraghi ():27
() Latin; Cic. Cat. , , in Luraghi (: )
Ad hanc tei amentiam natura peperit,
at this.ACC.F you.ACC insanity(F)ACC nature.NOM bear.PRF.SG
voluntas ei exercuit, fortuna ei servavit.
will.NOM train.PRF.SG destiny.NOM preserve.PRF.SG
‘For such insanity nature bore you, your will trained you, and destiny
preserved you.’
This difference between bibo ‘drink’ and e-bibo “out-drink” emerges, I argue, from
the configurational properties of the predicates they represent: e-bibo “out-drink”
represents a structure, including a PathP, which needs a DP at its specifier interpreted
as a Measurer. By contrast, object-less bibo ‘drink’, expressing an activity, corres-
ponds to a simple unergative structure without any specifier (except for Spec-Voice,
where the external argument is merged). Thus, a structural difference, which also
accounts for the interpretational differences between the simple and the prefixed
verb, explains why e-bibo “out-drink”, unlike bibo ‘drink’, cannot appear without an
object (null or not).
26
Hemina, -ae: a measure of wine.
27
Discussions on null objects in Latin include Panhius (); Mulder (); Wurff (); and
Sznajder ().
S-framed constructions in Latin
... Case alternations, situation aspect, and the merging of roots I focus now on
the way prefixation changes the case-assigning properties of the predicate, and how
that change is related to the inner-aspectual interpretation of the predicate. I take the
utor/abutor ‘use’/ “away-use” ‘use up’ contrast (see example ()) as a case study.
Importantly, while ab-utor “away-use” licenses an accusative in (), repeated here
as (), the ‘object’ of utor ‘use’ appears in the ablative (see ()):
() Latin; Cat. Agr. ,
Omne caseum cum melle ab-usus eris
whole.ACC cheese.ACC with honey.ABL away-use.FUT.SG
‘You will have used up all the cheese with honey.’
() Latin; Caes. Gall. , ,
Minus idoneis equis utebantur.
less suitable.ABL.M.PL horse(M)ABL.PL use.IPFV.PL
‘They were using less suitable horses.’
In fact, my prediction is that the object of ab-utor “away-use” should appear always
in the accusative in UOCs: it sits at Spec-Path, as evidenced by the fact that it behaves
as a Measurer. In (), for instance, the event is over only when the whole amount of
cheese is used up. As a matter of fact, ab-utor “away-use” does sometimes take the
ablative case in Classical Latin. The next example, for instance, involves ablative
sagacitate ‘sagacity’ instead of accusative sagacitatem:
() Latin; Cic. Nat. deor. ,
Sagacitate canum ad utilitatem nostram ab-utimur.
sagacity.ABL dog.GEN.PL at benefit.ACC our.ACC away-use.PL
‘We (abusively) use the sagacity of dogs to our benefit.’
However, a look at Gaffiot’s () entry for abutor “away-use” reveals a possible
explanation for this double case-selection. The first sense in the entry, the only
transitive one, reads ‘use until the object disappears’. Gaffiot furthermore marks it
as archaic, providing examples from Cato, Plautus, Terence, and Sallust. This is the
sense illustrated in (). The second sense is intransitive, taking the ablative, and is a
more modern one. The definition here reads differently, however: ‘use fully, freely’ or
‘make a deviant use of something’. This is the usage relevant in (). Observe, in
addition, that the ablative, as expected, does not license a Measurer interpretation for
sagacitate ‘sagacity’ in (); in fact, as the famous Ciceronian sentence of ()
indicates, this sense of abutor as ‘make an improper use of, abuse’ is atelic, since it
licenses the durative adverbial quo usque ‘until when’:28
28
In light of these facts, claims such as Oniga’s (: ) that transitivization through prefixation of
ablative-selecting prepositions—like ab ‘away’—must be due to a phonological or semantic analogy with
verbs which are prefixed with accusative-selecting prepositions cannot be on the right track. In particular,
Latin as a satellite-framed language
() Latin; Cic. Catil. ,
Quo usque tandem ab-utere,
which.ABL.SG up_to finally away-use.FUT.SG
Catilina, patientia nostra?
Catilina.VOC patience.ABL our.ABL
‘Until when will you abuse our patience, Catilina?’
Within the present account, it is clear that this particular abutor, unlike the one in
(), cannot correspond to a vP embedding a PathP projection. One possible
solution to the ablative-selecting abutor is to consider that the verb embeds a
‘complex’ root, a combination of both AB ‘away’ and UT ‘use’, yielding a predicate
composition semantics conforming with the negative semantics of ‘improper use’
already mentioned. In section ... I made the assumption that roots cannot
project structure, and that, hence, there cannot be a RootP. However, Merge must
arguably be distinguished from projection: two elements yield a syntactic object if
one of them has the ability to project. The combination of two roots is, therefore,
expected, as it is also expected that neither of them will project: they behave as a
single root.29 Thus, in the combination of AB ‘away’ and UT ‘use’ no root projects: the
category of the whole emerges from the eventive v head within which the complex
is embedded. If we take into account the atelic nature of abutor-predicates such as
() and (), a possible analysis is one involving an unergative structure with
the complex root embedded at Compl-v and the ablative DP merged within
an adjunct to vP:30
() Analysis for ()
[VoiceP (nos) [Voice’ Voice [vP [pP p sagacitate canum] [vP v ABUT]]]]
Another possible solution is to consider that this adverbial ab- of ablative-taking
abutor ‘abuse’ is not merged inside PathP, but outside the vP, as an external prefix.
The external position of ab- would also explain the lack of effects on argument
structure and inner aspect of the predicate. See section .. for a summary of the
properties of internal and external prefixes in Slavic and some data suggesting that
the distinction may also exist in Latin.
he cites the case of transitive abutor as an analogy of ad-sumo “at-take” ‘take to oneself ’ (with accusative-
selecting ad ‘at’). This would imply a greater antiquity for the intransitive abutor, contrary to what Gaffiot
() documents.
29
See Zhang () for an application of this idea to the analysis of compounds in Chinese.
30
Wurmbrand () advocates a complex predicate approach for idiomatic particle-verb combinations
in German, while reserving a Small-Clause approach for cases of transparent particle-verb combinations.
I am sympathetic to her analysis, but I do not think that idiomaticity (here the possibility of retrieving
particular meanings for roots within the phase) is restricted to direct association of roots. See McIntyre ()
for a critique of Wurmbrand’s () dychotomic approach to particle-verbs.
S-framed constructions in Latin
... Scopal relations between prefix and verb Scopal effects have traditionally
been dealt with at the sentence level in discussions of configurationality. However,
within an account, such as the present one, where words are created by the syntax, we
expect there to be scopal effects within the word. I will now show that there is a group
of ab-prefixed Figure UOCs in Latin that show scopal effects affecting the prefix and
the verb. Importantly, these effects follow naturally from an account of UOCs where
the prefix is c-commanded by the v head and the root is merged as an adjunct to v.
I refer to a group of ab-verbs where the base is a communication verb and the prefix
is interpreted as a negation of sorts. I call them ab-verbs of denial:
() Latin; Pacuv. Trag.
[Eam] consanguineam esse ab-dicant.
her.ACC consanguineous.ACC be.INF away-proclaim.PL
‘They proclaim her not to share the same blood.’
() Latin; Cic. Div. ,
Cumque in quattuor partis vineam
since=and in four part.ACC.PL vine.ACC
divisisset trisque partis aves
divide.PLUPRF.SBJV.SG three.ACC=and part.ACC.PL bird.NOM.PL
ab-dixissent, quarta parte [ . . . ] mirabili
away-say.PLUPERF.SBJV.PL fourth.ABL part.ABL admirable.ABL
magnitudine uvam [ . . . ] invenit.
size.ABL grape.ACC find.PRF.SG
‘And after he had divided the vine into four parts and the birds had refused
[lit. ‘had said away’] three of them, in the fourth part he found a grape of
admirable size.’
() Latin; Plaut. Rud.
In iure ab-iurant pecuniam.
in court.ABL away-swear.PL money.ACC
‘In court they deny by oath that they have debts.’
() Latin; Plaut. Capt.
‘Ubi cenamus una?’ inquam: atque illi ab-nuont.
where sup.PL together say.SG and they away-nod.PL
‘I say, “Where shall we sup together?” And they refuse with a nod.’
These verbs involve the negation of the proposition expressed by the object (which
may take the shape of a whole proposition, as in the Exceptional Case Marking
construction of () (with accusative eam ‘her’ as the subject of the embedded
infinitive esse ‘be’) or the elided object proposition of (), or a propositionally
interpreted DP, as in () and ()). In the examples above, the base verb is,
Latin as a satellite-framed language
respectively, dico (infinitive dicare) ‘proclaim, declare’, dico (infinitive dicere) ‘say’,
iuro ‘swear’, and nuo ‘nod’. Take the case of ab-iuro “away-swear” ‘deny by oath’, in
(). Crucially, as García Hernández (: ) early observed, the negation is
understood as having narrow scope with respect to the swearing event introduced by
the root IUR, and not the other way around. Thus, () does not imply that they do
not swear that they have debts. The scopal properties of these verbs come for free in a
syntactic model, if we assume that the negation component alluded to above is
nothing but an inference from the general meaning of the prefix ab ‘away’: the v
introducing the event and being associated with the root IUR ‘swear’ is above the
PathP including the object of the predicate and the prefix. Importantly, the root of
the prefix is c-commanded by the v head introducing an event, and is predicted,
correctly, to show narrow scope with respect to that v head:
() Analysis of ()
[VoiceP pro [Voice’ Voice [vP [v v IUR] [PathP pecuniam [Path’ Path [PlaceP
pecuniam [Place’ Place AB]]]]]]]
Scope: v > ab, *ab > v
An approach conceding a preponderant role to configurationality and separating the
encyclopaedic from the structural meaning of expressions derives both the denial
interpretation of these verbs and the precise scopal effects straightforwardly.
.. Ground Unselected Object Constructions
... Case and situation aspect when the object is a Ground In Ground UOCs the
internal argument (either a surface object or a subject) is interpreted as a Terminal
Ground, that is, as a final location in a transition event. Consider the following
examples from Danish, German, and English:31
() Danish; Svenonius (: )
Tjeneren tørket {*af} bordet {af}.
waiter.DEF wipes off table.DEF off
‘The waiter wipes the table off.’
() McIntyre (: )
Pour the bucket *(out).
31
Studies on these constructions in other languages include Svenonius (: ff.) on Scandinavian,
Zeller (a) on German, McIntyre () on German, McIntyre (, ) on English and German,
Svenonius () on English, Svenonius () on Russian, Mateu (b) on German, Levin and Sells
() on English (calling the particles in these constructions unpredicated particles), and Oya () on
English, German, and Dutch.
S-framed constructions in Latin
() German; Stiebels (: ), in Mateu (b: )
Sie *(unter)-keller-ten das Haus.
they under-cellar-PST.PL the.ACC house.ACC
‘They put a cellar under the house.’
In these examples the object of the construction corresponds to the Ground of a
motion event that is specified by the particle/prefix. Thus, in () there is an event of
wiping something off the table, in () an event of emptying something out of the
bucket, and in () an event of putting a cellar under the house. Crucially, the
Figure argument does not appear in the structure. Observe that this scenario, where
an argument (the Ground) is promoted to a different syntactic position (as object)
and the argument that usually occupies that position is demoted, is reminiscent of
that of passives or unaccusatives, where the object surfaces as a subject and the
external argument (for passives) appears, at most, as an adjunct. It is for these reasons
that Svenonius (: ff.) calls the particles licensing these constructions unac-
cusative particles, since the constructions where they appear presumably involve the
kind of phenomenon referred to by Burzio’s () generalization: in the absence of
an external argument, objective case is unavailable. In this case the missing external
argument is the Figure, and the objective case is the one standardly assigned by the
Ground-taking particle. For instance, in () the particle af ‘off ’ does not project a
Figure argument and, hence, cannot assign case to bordet ‘table’, which raises to the
position where it may receive (objective) case. In particular, Svenonius (: )
proposes that adpositional projections contain a lexical preposition that selects the
Ground and a functional p-layer that selects the Figure, which qualifies, then, as a
true external argument. In the next example, the lexical preposition on selects the
Ground wagon and the functional preposition p selects, as a specifier, the Figure hay:
() Svenonius (: )
We loaded hay on the wagon.
[pP [DP hay] [p’ p [PP on [DP the wagon]]]]
In constructions such as () to (), Svenonius claims, p is missing, the Figure is
not selected, and the Ground cannot receive case from the lexical preposition.
In my view, there is a fact about Ground UOCs that has been neglected in this
case-based analysis and that might constitute the key to understanding how they
work: the fact that the Ground makes an aspectual contribution in Ground UOCs but
not when it appears ‘in situ’, in Figure UOCs. Specifically, the Ground is clearly
interpreted as a Measurer for the event, in the present terms. Thus, in () the event
is over only when the whole surface of the table is completely wiped off. Note that
when the Ground is not promoted to object it does not possess this interpretational
status. Thus, in Sue wiped the dust off the table, the table does not measure out the
event in any sense, as the dust effectively does. Specifically, Sue need not wipe the
Latin as a satellite-framed language
whole surface of the table for the event of wiping the dust off the table to be true. This
effect in Ground UOCs is observed by McIntyre (), who notes the contrast
between Read through a book and the Ground UOC Read a book through:
reading through a book is less thorough than reading a book through. Although the former
could exhibit the bounded reading of through in the sense that the reading encompasses
the beginning and end of the book, it is compatible with skim-reading or leaving out some
sections because there is no holistic effect to ensure that the whole book is involved.
McIntyre (: )
Observe that in Reading through a book there is apparently no Figure. In fact, McIntyre
() proposes that the whole event of reading is a Figure traversing the Path
expressed by through a book (he calls this kind of construction Event Path). What is
worth noting here is that the different position of the Ground determines the above-
mentioned holistic effect or measuring-out effect. Crucially, McIntyre’s () obser-
vation can be made stronger, by setting it in terms of (a)telicity: while read through a
book may be atelic, read a book through is necessarily telic. Similar observations on the
measurer role of Grounds in these constructions are to be found in Levin and Sells
(). As the next examples show, the quantity or non-quantity status of the object
Ground is what determines, respectively, telicity and atelicity in the resultant predicate:
() Levin and Sells (: )
a. She wiped the counter off in/*?for ten minutes.
b. She wiped glass off *in/for two hours.
These facts are easily accounted for in my theory, since the Figure and Ground
interpretations of a DP are dissociated from its role in the calculation of situation
aspect. Thus, we expect either one of them to be available to be merged as Spec-Path
and to be interpreted as Measurer. I illustrate with the analysis of ():
() Analysis of ()
VoiceP
Sue Voice’
Voice vP
v PathP
v pour the bucket Path’
Path PlaceP
Place the bucket
Place out
S-framed constructions in Latin
The DP the bucket is originally merged as Compl-Place, and is, therefore, interpreted
as a Ground. Path raises the nearest DP in its c-command domain to its specifier (see
section ...). This DP is usually the one sitting at Spec-Place, the Figure, by a
minimality condition, but when the Figure is missing, there is no DP available other
than the Ground. It is at Spec-Path that the bucket is interpreted as a Measurer. As for
case, the bucket receives the same treatment as any other DP at Spec-Path: it gets
accusative case if Voice is projected, as in this instance. However, we will see in
section ... that in unaccusative Ground UOCs, with no Voice head projected, the
Ground ends up receiving nominative case, as one would expect.
Note that a case-account, such as that of Svenonius (), is unable to explain
why the Ground is interpreted as a Measurer only when it appears as the object of the
verb. Indeed, this Measurer interpretation cannot be attributed to the accusative case
itself, since there are accusative-marked DPs that are not interpreted as Measurers
(such as Peter in the next sentences):
() Non-measuring accusatives
a. John loved Peter (for years).
b. John considered Peter intelligent (for years).
c. John thought Peter to be loyal (for years).
... Transitive Ground UOCs in Latin The following are examples of Ground
UOCs in Latin:
() Latin; Ov. Met. ,
Uberaque e-biberant
breast.ACC.PL=and out-drink.PLUPRF.PL
avidi [ . . . ] nati.
eager.NOM.M.PL born.NOM.M.PL
‘And her babes had drunk her breasts to exhaustion.’
() Latin; Plin. Nat. ,
Dracones esse tantos ut totum
snake.ACC.PL be.INF so_many.ACC.PL that whole.ACC.M.SG
sanguinem capiant, itaque elephantos ab iis
blood(M)ACC.SG take.SBJV.PL therefore elephant.ACC.PL by them.ABL
e-bibi.
out-drink.INF.PASS
‘That the snakes are so large that they can take all the blood, and therefore the
elephants are drunk dry by them’.
Note that these examples involve the prefixed verb e-bibo “out-drink”, which we have
already seen heading Figure UOCs (see section ...). Ground UOCs with e-bibo
“out-drink” present accusative objects referring to the container of the liquid, instead
Latin as a satellite-framed language
of the liquid itself. In the examples above the object is ubera ‘breasts’ and elephantos
‘elephants’, respectively, and, although they correspond to the Ground in the event
schema, they can be said to license a complete affection interpretation, as is clear
from the translation of ()—see the relevant entries for this verb in Lewis
and Short (), and Gaffiot (). It is worth observing that while simple bibo
‘drink’ may be used with container-naming objects, as in () below, I have
not found any such example (in a search of all the occurrences of simple bibo
‘drink’ in the Antiquitas corpus) with a non-standard container, such as those in
() and ():
() Latin; Plaut. Stich.
Vide quot cyathos bibimus.
see.IMP.SG how_many goblets.ACC drink.PRF.PL
‘See how many goblets we have drunk.’
This fact strongly suggests that cases such as () involve a metonymical reading of
the object, precisely because it refers to a canonical container holding a standard
quantity of liquid. The predicates in () and (), however, do not involve
metonymy. Neither the breasts nor the elephants are taken as standard measures
for the liquids they contain, nor are they, for that matter, conceived of as containers
of milk and blood, respectively. Rather, they seem to be really interpreted as the
Grounds in the motion schema. This is unexpected in, for instance, Lehmann’s
(: ) account of preverbation, in which the subject and the object of an
intransitive and a transitive unprefixed verb, respectively, emerge as the locatum—
in our terms, the Figure—of the preverb when the verb is prefixed.
The difference between Figure UOC e-bibo “out-drink” and Ground UOC e-bibo
is easily grasped: in () and (), for instance, the objects are not brought out or
made to disappear by virtue of a drinking event, as is the case in instances of e-bibo
“out-drink” in Figure UOCs. I repeat an example from section ... for the sake of
comparison:
() Latin; Petr. Sat. ,
Tantum medicamentum e-bibisti?’
so_much.ACC medicine.ACC out-drink.PRF.SG
‘Have you drunk up so much medicine?’
In this example, the prefix e- ‘out’, encoding, as was discussed in section ..., a
‘state of disappearance’, is predicated of tantum medicamentum ‘so much medicine’,
which is, thereby, a Figure. This is clearly not the interpretation of ubera ‘breasts’ and
elephantos ‘elephants’ in () and (), respectively.
The same difference is appreciated when contrasting the Figure UOC ab-luo “away-
wash” of (), repeated here as ()a, with the Ground UOC ab-luo of ()b:
S-framed constructions in Latin
() Latin; Figure and Ground UOCs based on ab-luo “away-wash”
a. Tac. Hist. , ,
Is balineas ab-luendo
he.NOM bath.ACC.PL off-wash.PTCP.FUT.PASS.DAT.M.SG
cruori propere petit.
blood(M)DAT.SG hastily head.SG
‘He hastened to the baths to wash off the blood.’
b. Cic. Tusc. , ,
Anticlea [ . . . ] Ulixi pedes ab-luens.
Anticlea(F)NOM.SG Ulysses.DAT feet.ACC away-wash.PTCP.PRS.NOM.F.SG
‘Anticlea, as she washes Ulysses’s feet clean’.
Observe that in the predicate of ()b Ulysses’s feet do not disappear by virtue of a
washing event, as is the case with cruori ‘blood’, in ()a. Rather, we understand that
Ulysses’s feet are a surface off which (= ab-) the dirt is washed. In turn, pedes ‘feet’ in
()b, ubera ‘breasts’ in (), and elephantos ‘elephants’ () are not only the
Grounds in their corresponding predicates, but are also Measurers: the events in
which they are involved are completed according to the physical extension of the
entities which they denote—a volume in () and (), and a surface in ()b.
These facts emerge naturally from the following analysis:
() Analysis of ()
VoiceP
Dracones Voice’
Voice vP
v PathP
v bib elephantos Path’
Path PlaceP
Place elephantos
Place ex
The root BIB ‘drink’ is merged as an adjunct to v and is interpreted, accordingly, as a
Manner Co-event. The DP elephantos ‘elephants’ is first merged as Ground. The root
E ‘out’, adjoined to Place, specifies the Conformation. A rough paraphrase could be
‘to cause something to go out of the elephants’. However, a Figure is not merged and,
therefore, when Path is merged, it raises to its specifier the only DP available in its
Latin as a satellite-framed language
c-commanding domain, namely elephantos ‘elephants’. This is why this DP is
interpreted both as a Ground and as a Measurer.32
The syntactic analysis proposed reveals itself as useful in accounting for other
apparently different values of the same prefix, as sometimes proposed in lexicalist
analyses. For instance, Van Laer (: ff.) contends that the prefix ob- ‘against’
in the verb ob-duco “against-lead” may show two different values: its ‘straight’
meaning of ‘against, facing, in front of ’ and a second meaning which she dubs as
that of ‘recouvrement’ (‘covering’). These two values are related to two different
argument structures, as shown, respectively, in the following examples:
() Latin; Plaut. Pseud. , in Van Laer (: )
Post ad oppidum hoc vetus continuo meum
afterwards at town.ACC this.ACC old.ACC directly my.ACC
exercitum protinus ob-ducam.
army.ACC straight_on against-lead.FUT.SG
‘Afterwards, I shall directly lead my army straight on against this old town.’
() Latin; Prop. , , , in Van Laer (: )
Terra tuum spinis ob-ducat, lena,
earth.NOM your.ACC thorn.ABL.PL against-lead.SBJV.SG bawd.VOC
sepulcrum.
grave.ACC
‘May the earth cover your grave with thorns, you bawd.’
Interestingly, the covering sense of ob- ‘against’ correlates with an argument struc-
ture where the accusative argument may be interpreted as a Ground (tuum sepulcrum
‘your grave’ in the example above) and an ablative argument may be interpreted
as a Figure (spinis ‘thorns’ in the example above), both in relation to the prefix ob-
‘against’. An analysis in terms of a Ground UOC for () would be as follows:
() Analysis of () as a Ground UOC
[VoiceP Terra [Voice’ Voice [vP [v v DUC] [PathP sepulcrum [Path’ [PlaceP [Place Place
OB] sepulcrum]]]]]]
This analysis proposes the same basic argument structure for the second ob-duco
“against-lead” and the first obduco of (), where the Figure emerges as an accusative
DP and the Ground is the root of the prefix ob- ‘against’. The difference lies in the fact
that in () no DP has been merged as Spec-Place, so there is no Figure argument in
the argument structure—although an ablative adjunct such as spinis ‘thorns’ in ()
32
Zeller (a: ) also observes the fact that a same particle-verb in German may head a predicate
where the object is interpreted as Figure or a predicate where it is interpreted as Ground. This is the case
with German ab-spülen “off-rinse”, used either as a removal verb (‘rinse off the grease’) or as a change-of-
state verb (‘rinse off the dish’). See also McIntyre ().
S-framed constructions in Latin
could be interpreted as a ‘demoted figure’. This is why the Ground (tuum sepulcrum
‘your grave’) emerges as accusative and is moreover understood as a Measurer for the
event (since it is pulled up by Path onto its specifier), providing thereby what Van Laer
interprets as a ‘covering’ sense in the resulting predicate. In Van Laer’s () analysis
both senses of ob-duco “against-lead” remain unrelated.
There are cases of Ground UOCs where the Ground is not physical, but meta-
phorical. Thus, for instance, we find predicates of utterance where the addressee is
realized as the accusative object. The verb is marked with the prefix ad- ‘at’:
() Latin; Verg. Aen. ,
Talibus ad-fata Aenean.
such.ABL.PL at-say.PTCP.PFV.NOM.F.SG Aeneas.ACC
‘Having addressed Aeneas with those words’.
() Latin; Plaut. Cist.
Ad-hinnire equolam possum ego hanc
at-neigh.INF mare(F)DIM.ACC.SG can.SG I this.ACC.F.SG
‘I can well neigh at this little mare myself.’
() Latin; Plaut. Amph.
Obsecro ut [ . . . ] liceat te al-loqui
beseech.SG that be_allowed.SBJV.SG you.ACC at-speak.INF
‘I beseech you to let me address you.’
These cases help us further illustrate how the syntactic structure dictates the number
and interpretation of the arguments of a verb, overriding whatever information is
contained in the encyclopaedic entry of its root. In particular, if it is assumed that the
prefixed predicates in () to () involve movement of the Ground DP to Spec-
Path position, there is predictedly at most and at least one overt argument per
prefixed predicate, since, on the one hand, there is no position left for any other
argument in PlaceP (since, by hypothesis, Spec-Place is not filled), and, on the other
hand, each PathP must have its specifier. This is what happens in the above examples,
with only an accusative object naming the addressee, and the utterance argument
being expressed, at most, as an instrumental adjunct in the ablative, as is the case of
talibus ‘with such (words)’ in (). Descriptively, it could be said that the utterance
argument is ‘demoted’ to adjunct-status.33 The unprefixed counterparts to al-loquor
“at-speak”, af-for “at-say”, or ad-hinnio “at-neigh” display, as expected, a different
syntax. Notably, they cannot link an addressee as object. They are either unergative
(see ()a, () and ()a), or take an accusative object, which is, however,
33
These ad-verbs are strikingly similar to an-prefixed verbs in German, like an-lügen “at-lie” ‘lie to’ or
an-motzen “at-whinge” ‘whinge to’, discussed by Stiebels () and McIntyre (), where the addressee
is expressed as the accusative object DP.
Latin as a satellite-framed language
interpreted as the utterance. In either case, they may optionally appear with a dative
or a PP expressing the addressee (see ()a for the former option and ()c and
()c for the latter):
() Latin; simple for ‘say’
a. Liv. , ,
Mihi ita Iuppiter fatus est.
me.DAT thus Jupiter.NOM say.PRF.SG
‘Jupiter has talked to me thus.’
b. Verg. Aen. ,
Ea fatus erat.
those_things.ACC say.PLUPRF.SG
‘He had said that.’
c. Cic. Tim.
Ad eos is deus [ . . . ] fatur haec.
at them.ACC that.NOM god.NOM say.SG this.ACC.N.PL
‘To them that god says these words.’
() Latin; simple hinnio ‘neigh’, Ps. Apul. Herm.
Proprium est equi hinnire.
typical.NOM.N.SG is horse.GEN neigh.INF
‘It is typical of the horse to neigh.’
() Latin; simple loquor ‘speak’
a. Ov. Rem.
Illa loquebatur.
She.NOM speak.IPFV.SG
‘She was speaking.’
b. Cic. Tusc. , ,
Pugnantia te loqui non vides?
contradiction.ACC.PL you.ACC speak.INF not see.SG
‘Are you not aware that you are saying contradictions?’
c. Ov. Pont. , ,
Certus eras [ . . . ] numen
sure.NOM.M.SG be.IPFV.SG divine.ACC
ad Augustum [ . . . ] loqui.
at Augustus.ACC speak.INF
‘You were resolute to speak to divine Augustus.’
It seems, once again, that verbs (in fact, roots) have little to say on the realization of
arguments. Rather, it is the syntactic structure that determines the number and
quality of the arguments.
S-framed constructions in Latin
I finish this section by reconsidering the examples of CDMCs with a prefixed verb
and a directional accusative-marked DP, as seen in section ..:
() Latin; Tac. Ann. ,
Novissimos in-currere.
rear.ACC in-run.PRF.PL
‘They charged against the rear.’
As was shown in section .., the accusative DP in these predicates seems to be an
argument, since it can become a passive subject. These cases are amenable to an
analysis in terms of a Ground UOC in which the accusative DP is merged as Compl-
Place and in which there is no Figure argument:
() Analysis of ()
VoiceP
pro Voice’
Voice vP
v PathP
v curr novissimos Path’
Path PlaceP
Place novissimos
Place in
If a passive Voice is projected, the Ground DP must raise up to Spec-T to obtain
nominative case, as in the next example:
() Latin; Cic. ad Q. fr. , ,
Cum neque praetores [ . . . ] ad-iri possent.
since not-even praetors.NOM at-go.INF.PASS can.IPFV.SBJV.PL
‘Since not even the praetors could be approached.’
... Unaccusative Ground UOCs McIntyre () discusses a class of construc-
tions where the Ground, rather than being realized as the object, is realized as the
derived subject of an unaccusative predicate. Predicates of this type are found in
Germanic:
() German; McIntyre (: )
Die Wanne fliesst schlecht *(ab).
the.NOM bath.NOM flow.SG badly off
‘The bath empties badly.’
Latin as a satellite-framed language
In Latin there are cases analogous to this:
() Latin; Plaut. Most.
Venit imber, lauit parietesi: ei per-pluont.
come.SG rain.NOM wash.SG wall.ACC.PL through-rain.PL
‘The rain comes, it washes the walls: they let the rain filter through.’
() Latin; Plaut. Pseud.
Senapis [ . . . ] oculi ut ex-stillent facit.
mustard.NOM eye.NOM.PL that out-drip.PL make.SG
‘Mustard makes the eyes drip out (with tears).’
In the above examples, the Ground appears as the subject (in () it happens to be a
pro subject, coreferential with accusative parietes ‘the walls’) of the sentence. The holistic
effect shown by the other cases of Ground UOCs obtains also in these unaccusative
Ground UOCs: thus, parietes ‘the walls’ and oculi ‘the eyes’ refer to entities completely
affected by the respective process. The analysis of the predicates in () and () is
essentially no different from that proposed for the cases of transitive Ground UOCs. In
unaccusative constructions, the Ground, after moving onto Spec-Path raises to Spec-T
and is provided with nominative case. The analysis of () is sketched in () below:
() Analysis of ()
vP
v PathP
v plu parietes Path’
Path PlaceP
Place parietes
Place per
.. The Locative Alternation
... Approaches to the LA The Locative Alternation (LA) is a widely known and
certainly not understudied phenomenon that can be illustrated by the following pair
of sentences:
() The LA in English
a. Sue loaded apples into the basket.
b. Sue loaded the basket with apples.
The sentences in () contain the same verb and correspond to the same conceptual
scene: by virtue of Sue’s action, apples end up in some basket. However, each sentence
has syntactic and semantic properties of its own. Thus, in ()a, the change-
S-framed constructions in Latin
of-location (COL) alternant, the object is the thing being located in a place, which is
expressed via a PP. In ()b, the change-of-state (COS) alternant, the syntax of
those two participants in the event is reversed, so that the object expresses the
location and the PP encodes the thing being moved. Moreover, it has very often
been observed that while ()b entails that the basket ends up full of apples, ()a
does not. ()b exhibits, therefore, the phenomenon known as ‘holistic effect’.34
Many studies have been devoted to the LA and a division can be made into two
basic types of approach. On the one hand, there are approaches where the COS
alternant is derived from the COL alternant, which is, thus, more ‘primitive’ (see
Larson ; Damonte ; Wunderlich , among others). These approaches,
based on classical theta-roles such as Theme and Location, aim at preserving a
privileged linking relation between the Theme role (apples in ()) and the syntactic
position of the object. On the other hand, there are approaches where the alternation
is not seen as a phenomenon to be explained in terms of a derivational relation
between both alternants (see Pinker ; Mulder ; Baker ; Mateu c;
Borer b, among others). These approaches adopt a significantly more abstract
view of theta-roles, which allows them an isomorphic mapping between the object
and its thematic interpretation without resort to a derivational mechanism. In
particular, for these approaches both apples and the basket receive the same ‘theta-
role’, so it comes as no surprise that they are both realized as objects. Here I will
follow a hybrid approach to the LA: although I believe that the non-derivational
approach is basically right for most cases of the LA, I will propose that some instances
of the LA do involve, at least in Latin, the derivation from one alternant to the other.
... The LA and the s-/v-framed distinction Importantly, the LA is the locus of
cross-linguistic variation, being quite rare in v-framed languages. Specifically, COL
alternants are hard to obtain in these languages—see Mateu (c) for Catalan and
Spanish, Rosselló () for Catalan, and Lewandowski () for a quantitative
study of Spanish. I illustrate this cross-linguistic asymmetry by the following
failed alternations in Catalan, which are perfectly acceptable in English. Note that
the a-sentences are COL alternants and the b-sentences are COS alternants:
() Catalan ruixar ‘spray’
a. *En Marc va ruixar aigua sobre la planta.
the Marc PRF.SG spray.INF water on the plant
‘Marc sprayed water onto the plant.’
b. En Marc va ruixar la planta {d’/amb} aigua.
the Marc PRF.SG spray.INF the plant of/with water
‘Marc sprayed the plant with water.’
34
See Anderson (); Dowty (); and Beavers (), among others. See also section ....
Latin as a satellite-framed language
() Catalan untar ‘smear’
a. */?La Maria va untar mantega a la llesca de pa.
the Maria PRF.SG smear.INF butter at the toast
‘Maria smeared butter onto the toast.’
b. La Maria va untar la llesca de pa {de/amb} mantega.
the Maria PRF.SG smear.INF the toast of/with butter
‘Maria smeared the toast with butter.’
In light of this fact, Mateu (c) entertains the hypothesis that COL alternants of
the Germanic kind (see ()a) are s-framed constructions, and hence, unavailable in
v-framed languages. For instance, ()a and ()a would be ungrammatical in
Catalan because they involve the combination of a verb naming the manner in which
the event takes place (spraying, smearing) and a PP specifying the final location of the
entity encoded by the object (sobre la planta ‘on the plant’, a la llesca de pa ‘onto the
toast’).35 This analysis has the advantage of accounting for the fact that verbs that can
be construed as directional, such as Spanish cargar ‘load’, do allow the alternation in
Romance:
() Spanish; Lewandowski (: –)
a. Cargó sus libros en varias cajas.
loaded her books in several boxes
‘She loaded her books into several boxes.’
b. Juan cargó el carro con heno.
Juan loaded the cart with hay
Both the COL and the COS alternants above are v-framed constructions. See
section ... for discussion of Latin COL alternants as involving a v-framed
configuration.
If the availability of the LA is related to s-framedness, we expect Latin to display
the LA freely. This prediction is borne out, as is shown by examples () through
(), where a-sentences are COL alternants and b-sentences are COS alternants:
35
As pointed out by Pinker (), Marantz (), and Borer (b), among others, the verb fill is a
conspicuous example of a verb that does not admit a COL alternant, even though English is an s-framed
language: *Fill water in(to) the glass. I assume that fill is an idiom in English, relating a root to a specific
position (Compl-Place). This is suggested by the fact that other languages do admit fill as a manner verb,
e.g. German füllen (Ambridge and Brandt ). On the other hand, acquisition studies such as that of
Gleitman and Landau () have shown that ‘English three year olds say “Fill water into the glass” almost
per cent of the time’ (Gleitman and Landau : ). This is compatible with the idea that roots are
freely inserted into structures, as shown by the more liberal grammar of children, and also that in the adult
grammar of English fill is fixed as a result verb through idiomaticity.
S-framed constructions in Latin
() Latin spargo ‘scatter’
a. Cato, Agr.
Stercus columbinum spargere oportet
manure(N)ACC of_pigeon.ACC.N scatter.INF be_necessary.SG
in pratum.
in meadow.ACC
‘Pigeon manure must be scattered onto the meadow.’
b. Cato, Agr.
Pabulum [ . . . ] amurca spargito.
fodder.ACC dregs_of_oil.ABL scatter.IMP.FUT.SG
‘Scatter the fodder with dregs of oil.’
() Latin sterno ‘spread’
a. Ov. Fast. ,
Sternitur in duro vellus utrumque solo.
spread.PASS.SG in hard.ABL.N fleece.ACC either.ACC floor(N)ABL
‘Both fleeces are spread on the hard floor.’
b. Cic. Mur.
Stravit pelliculis haedinis lectulos.
spread.PRF.SG skin(F)DIM.ABL.PL of_goat.ABL.F.PL bed.DIM.ACC.PL
‘He covered the little beds with goatskins.’
() Latin lino ‘smear’
a. Ov. Medic.
Medicamina [ . . . ] lini per corpora possint.
makeup.NOM.PL smear.INF.PASS through body.ACC can.SBJV.PL
‘Such make-up as may be smeared on the body’.
b. Ov. Pont. , ,
Vipereo spicula felle linunt.
of_viper.ABL.M arrow.ACC.PL bile(M)ABL smear.PL
‘They smear their arrows with viper bile.’
() Latin farcio ‘stuff ’
a. Sen. Dial. --, , ,
In os farciri pannos imperavit.
in mouth.ACC stuff.INF.PASS rag.ACC.PL order.PRF.SG
‘He ordered the rags to be stuffed into his mouth.’
b. Plin. Nat. , ,
Medios parietes farcire fractis
central.ACC.M.PL wall(M)ACC.PL stuff.INF broken.ABL.N.PL
caementis.
quarry stone(N)ABL.PL
‘To stuff the central part of a wall with fragments of quarry stones’.
Latin as a satellite-framed language
() Latin stipo ‘cram’
a. Varro, Ling. ,
Asses [ . . . ] in aliqua cella stipabant.
coin.ACC.PL in some.ABL.F.SG room(F)ABL.SG cram.IPFV.PL
‘They used to cram the coins in some room.’
b. Cic. Phil. ,
Senatum stiparit armatis.
senate.ACC cram.PRF.SBJV.SG armed.ABL.M.PL
‘(That) he had crammed the senate with armed men.’
I provide, below, a non-derivational analysis of the LA in (). I analyse the COL
alternant in ()a and the COS alternant in ()b, with the caveat that in section
... I will propose a different analysis for unprefixed COL alternants:
() Analysis of ()
a. VoiceP
PRO Voice’
Voice vP
v PathP
v sparg stercus Path’
Path PlaceP
stercus Place’
Place pratum
Place in
b. VoiceP
(tu) Voice’
Voice vP
v PathP
pabulum Path’
Path PlaceP
pabulum Place’
Place sparg
S-framed constructions in Latin
Note that in this non-derivational analysis the LA presented by spargo ‘scatter’ boils
down to the possibility of associating the same root with different positions of one
basic abstract configuration encoding an externally originated transition. Specifically,
in the COL alternant the root SPARG ‘scatter’ is merged as an adjunct to v, and is
interpreted, consequently, as a Co-event of the transition (change-of-location) event.
The COL alternant is, according to this analysis, an s-framed construction. In section
... I will argue that this type of unprefixed COL alternant involves, in fact, a
verbal root merged as Compl-Place. In the COS alternant the root is merged at
Compl-Place, and is interpreted as a Terminal Ground, as the final state of a
transition (change-of-location) event. As regards the object, it is a Figure in both
cases, since it is first merged at Spec-Place. However, since in the COL alternant it
appears in a predicative relation with a location, codified by in pratum ‘onto the
meadow’—with a root IN ‘in’ specifying the head Place and inducing a spatial reading
thereof—it is interpreted as an entity that changes location. By contrast, in the COS
alternant it holds a predicative relation with the verbal root, and is therefore
interpreted as an entity that enters into a specific state (a state of being ‘scattered’,
identified with SPARG). Observe, importantly, that I am positing the projection of a
PathP for both COL and COS alternants, and that in both cases the Path head raises
the nearest DP in its c-commanding domain, the Figure, to Spec-Path, where it is
interpreted as a Measurer. This means that in both cases the so-called holistic effect
must emerge, as seems to be the case: in ()a stercus ‘manure’ measures out the
event as much as pabulum ‘fodder’ does in ()b. This is in tune with Dowty’s
() observation that the objects of either COL or COS alternants are interpreted
as Incremental Themes, and that, if possessing the appropriate quantificational
properties, they might induce telicity in the predicate:36
() Dowty (), in Baker (: )
a. John sprayed this whole can of paint onto subway cars in an hour.
b. John sprayed this wall with paint in an hour.
Thus, the fact that pratum ‘meadow’ in ()a is not interpreted holistically (the field
need not end up covered with manure) is a syntactic effect: it cannot raise to Spec-
Path, and, hence, cannot be interpreted as a Measurer.
Note, finally, that I am treating the ablative amurca ‘dregs of oil’ in the COS
alternant of ()b as an adjunct to vP, as also proposed by Rappaport and Levin
(), Mateu (c), and Borer (b).37
36
See also Pinker (: ) and Borer (b: ). The latter capitalizes on this fact to show that in
both COL and COS the object is a Subject-of-quantity, sitting at Spec-AspQ (see section ..).
37
By contrast, and specifically for Latin, Pinkster (: ) considers these ablatives as arguments
(complements in his terminology).
Latin as a satellite-framed language
... The LA and prefixation. The heterogeneity of the LA The LAs shown in
examples () to () do not exhaust the exploration of the LA in Latin. Rather, it
has been observed (Hofmann and Szantyr ; Lemaire ) that this form of
argument structure alternation is very frequently mediated through prefixation. In
the following sections I capitalize, therefore, on the patterns of prefixation shown
by both alternants in the LA in Latin, and put them in relation both to other
constructions of the language and to similar patterns in other languages.
I purport to show that the different morphological manifestations of the LA in
this language suggest that it might be a rather heterogenous phenomenon, calling
for a non-uniform account.
One first prefixal pattern shown by the LA in Latin involves the presence of a prefix
in the COL alternant. The verbs laedo ‘hit, harm’ and quatio ‘shake, agitate’ illustrate
this pattern (I present the COL alternant first):38
() Latin in-lido “in-hit” ‘thrust against’ and laedo ‘hit’
a. Verg. Aen. ,
Notus [naves] in-liditque vadis.
south_wind.NOM ship.ACC.PL in-hit.SG=and sandbank.DAT.PL
‘The south wind thrusts the ships against the sandbanks.’
b. Plaut. Bacch.
Lembus ille mihi laedit latus.
boat.NOM that.NOM me.DAT hit.SG side.ACC
‘That boat hits my side.’
() Latin; quatio ‘shake, agitate’ and in-cutio “in-shake” ‘stamp against’
a. Quint. Inst. , ,
Terrae pedem in-cutere.
earth.DAT foot.ACC in-shake.INF
‘To thrust the foot against the earth’.
b. Hor. Carm. , ,
Terram quatiunt pede.
earth.ACC shake.PL foot.ABL
‘They shake the earth with their feet.’
38
Laedo ‘hit, harm’ and quatio ‘shake, agitate’ can be said to enter, in () and (), what Levin and
Rappaport Hovav () call the with/against alternation, exemplified below, which involves impact verbs:
(i) Levin and Rappaport Hovav (: )
a. Kerry hit the stick against the fence.
b. Kerry hit the fence with the stick.
S-framed constructions in Latin
The prefixed counterparts in-cutio “in-shake” ‘stamp against’ and in-lido “in-hit”
‘thrust against’ are not found as COS alternants (Lewis and Short ). In the
analysis given here, the prefix corresponds to a root merged as an adjunct to Place,
where it is thus interpreted. The verbal root is merged as an adjunct to v, specifying
the kind of transition undergone by the Figure.
As shown by Lemaire (), many cases of the LA involve the same prefix for
both alternants. I illustrate with circum-icio “around-throw” ‘surround’ and in-duco
“in-lead” ‘smear’:
() Latin circum-icio “around-throw” ‘surround’
a. Liv. , ,
Fossam [ . . . ] uerticibus iis, quos
ditch.ACC peak(M)DAT.PL those.DAT.M which.M.ACC.PL
in-sederant, circum-iecere.
in-sit.PLUPRF.PL around-throw.PRF.PL
‘They put a ditch around the peaks where they had settled down.’
b. Tac. Ann. , ,
Planitiem saltibus circum-iectam.
plain(F)ACC forest.ABL.PL around-throw.PTCP.PFV.ACC.F
‘A plain surrounded by forests’.
() Latin in-duco “in-lead” ‘smear’
a. Cels. ,
Ulceri medicamentum [ . . . ] in-ducatur.
ulcer(N)DAT.SG medicament.NOM.N.SG in-lead.SBJV.PASS.SG
‘Let the medicament be smeared into the ulcer.’
b. Plaut. Most.
Postes [ . . . ] sunt in-ducti pice.
doorpost.NOM.PL be.PRS.PL in-lead.PTCP.PFV.NOM.M.PL pitch.ABL.SG
‘The doorposts have been smeared with pitch.’
In the COS alternants of these instances of the LA, the objects, which happen to be
passive in both examples, hold a Ground semantic relation with the prefixes. Thus, in
()b the forests (saltibus) are around (circum-) the plain (planitiem), and in ()b
the pitch (pice) is smeared into (in-) the doorposts (postes). Thus, these cases of COS
alternants can be treated as Ground UOCs, with no DP merged at Spec-Place and
with the Ground raising to Spec-Path:
Latin as a satellite-framed language
() Analysis of ()b
vP
v PathP
v ic
planitiem Path’
Path PlaceP
Place planitiem
Place circum
Therefore, in these cases of COS alternants endowed with a spatial prefix, I argue for
a derivational approach to the LA: these COS alternants are derived from structures
where the object is first merged as a Ground and there is no Figure merged at Spec-
Place.
Finally, I point out that many verbs that are prefixed with co(m)- ‘together’ are
only interpreted as COS alternants. Thus, in the following examples the object
(passivized or not)—ora ‘face’ and me ‘me’, respectively—seems to be interpreted
as an entity that changes state through a locating event (of covering with make-up or
tears, respectively):
() Latin; Ov. Rem.
Con-linit ora venenis.
together-smear face.ACC make-up.ABL.PL
‘She covers her face completely with make-up.’
() Latin; Cic. Planc.
[Me] con-spersitque lacrimis.
me.ACC together-scatter=and tear.ABL.PL
‘And he covered me with tears.’
By contrast, the absence of the com- ‘together’ prefix licenses a COL reading. Thus,
the following predicates feature an unprefixed verb and present a COL reading and
a COL syntax. Thus, the object (again, passivized or not) is interpreted as a
Figure and per corpora ‘on the whole body’ and in pratum ‘onto the field’ are the
Grounds:
() Latin; Ov. Medic.
Medicamina [ . . . ] lini per corpora possint.
makeup.NOM smear.INF.PASS through body.ACC can.SBJV.PL
‘Such a makeup as may be smeared on the body’.
S-framed constructions in Latin
() Latin; Cato, Agr.
Stercus columbinum spargere oportet
manure(N)ACC of_pigeon.ACC.N scatter.INF be_necessary.SG
in pratum.
in meadow.ACC
‘Pigeon manure must be scattered onto the field.’
Likewise, a spatial prefix licenses, as we saw above for in-cutio “in-shake” ‘stamp
against’ and in-lido “in-hit” ‘thrust against’, a COL reading of the predicate. Thus, in
the following examples, the farina ‘flour’ and quidquid pingue secum tulit ‘whatever
richness it has carried along’ are interpreted as Figures, while arentibus locis ‘dry
places’ and ulceribus ‘ulcers’ are interpreted as Grounds:
() Latin; Sen. Nat. a, ,
[Nilus] quidquid pingue secum tulit,
Nile.NOM whatever.ACC rich.ACC with_it carry.PRF.SG
arentibus locis ad-linit.
dry.DAT.PL place.DAT.PL at-smear.SG
‘The Nile smears onto the dry places whatever richness it has carried along.’
() Plin. Nat. , ,
[Farina] in-spergitur ulceribus.
flour in-scatter.PASS.SG ulcer.DAT.PL
‘The flour is sprinkled into the ulcers.’
In line with Hoekstra and Mulder’s () and Mulder’s () analysis of the
be-prefix in Dutch to be shown below, I propose that, in fact, the COS variants
with a com-prefix are a case of s-framed constructions where the root is, again,
merged as an adjunct to v, and where the Ground is the root which will end up as
prefix. This root is interpreted, in combination with the root of the verb, as inducing
a complete affection of the entity encoded by the Figure DP.39 For instance, in ()
the face is entailed to be completely covered with make-up. Thus, in these com-
prefixed COS alternants what is predicated of the Figure argument is the prefix itself
(its root, to be precise), and not the verbal root. On the other hand, they are not cases
of Ground UOCs, that is, the object is not a promoted Ground. I illustrate this with
the analysis of ():
39
Lemaire (: ) also observes a contrast between pairs like con-scribo “with-write” ‘cover with
inscriptions’ and in-scribo “in-write” ‘inscribe, write in(to) or upon’, interpreting an opposition between a
‘contact’ sense of com- and an ‘insertion’ sense of in-, which introduces the sense of insertion. On the other
hand, Meillet (: ff.) and Barbelenet (: –, –) had already noticed a grammatical-
ization of the prefix com- ‘together’ when they argued that it constituted a morphological way to induce a
‘perfective’ interpretation in an otherwise ‘durative’ verb (as in specio ‘watch’ / con-spicio “with-watch”
‘spot’). See also Moussy ().
Latin as a satellite-framed language
() Analysis of ()
VoiceP
pro Voice’
Voice vP
v PathP
v lin ora Path’
Path Place
ora Place’
Place com
Note that, as usual, the merging of PathP as a sister to v brings about movement of
the highest DP, ora ‘face’, onto its specifier. A paraphrase for this predicate would be
‘to affect the face completely through a making-up event’.
That these com-prefixed COS alternants are s-framed constructions is not surpris-
ing when we consider that they mirror analogous predicates in other languages
claimed to be s-framed. Thus, in the following sentences the particles be (Dutch),
be (German), and meg (Hungarian) induce a complete affection interpretation:
() Dutch; Hoekstra and Mulder (: )
Hij be-hing de muur met posters.
he be-hang.PST.SG the wall with posters
‘He covered the whole wall with posters.’
() German; Wunderlich (: )
Er be-giesst die Blumen mit Wasser.
he be-pour.SG the.ACC.PL flower.ACC.PL with water.DAT
‘He waters the plants (with water).’
() Hungarian; Ackerman (: )
A paraszt meg-rakta a szekeret (szénával).
the peasant meg-load.PST.SG the cart.ACC hay.INSTR
‘The peasant loaded the cart full with hay.’
Specifically for Dutch, Hoekstra and Mulder (: –) and Mulder (:
–) claim that the prefix be-, inducing complete affection, is in fact a predicate
heading a Small-Clause-like structure, since, it happens to be in complementary
distribution with a resultative AP (vol ‘full’ in the example):
() Dutch; Mulder (: )
*Hij be-hangt de muur vol me foto’s.
he be-hangs the wall full with photos
S-framed constructions in Latin
As will become clear in Chapter , I cannot apply this test to Latin, since Latin does
not license complex AP resultative constructions. However, com- ‘together’ can
change the argument structure properties usually displayed by the unprefixed verb,
and, in that sense, it is amenable to an analysis along the lines of those proposed
above for other prefixes which induce changes in argument structure. I underpin this
claim with the contrast between mingo ‘piss’, an intransitive creation verb (see ()
a) and com-mingo “together-piss” ‘piss all over’ (see ()b):
() Latin mingo ‘piss’ and com-mingo “together-piss” ‘piss all over’
a. Mart. , ,
Minxisti currente semel, Pauline, carina.
piss.PRF.SG run.PTCP.PRS.ABL.F once Paulinus.VOC boat.ABL
‘You pissed once, Paulinus, while the ship was sailing along.’
b. Hor. Sat. , ,
Com-minxit lectum.
together-piss.PRF.SG bed.ACC
‘He pissed the bed.’
As usual, I treat the unselected object lectum in ()b as a Figure, while the prefix
originates as a predicative root in Compl-Place and the verbal root is an adjunct to v:
() Analysis of ()b
[VoiceP pro [Voice’ Voice [vP [v v MING] [PathP lectum [Path’ Path [PlaceP lectum
[Place’ Place COM]]]]]]]
From this discussion a scenario emerges in which the LA might be more heterogenous
than previously considered. Specifically, COS alternants may respond to different
syntactic strategies based on the type of element merged as the Terminal Ground at
Compl-Place. They can be change-of-state predicates with the verbal root merged as a
Terminal Ground (see ()b), they can correspond to Ground UOCs, with the object
first merged as a Terminal Ground (see ()b), and they can correspond to predicates
with the prefix com- ‘together’, inducing a complete affectedness semantics, merged as
a Terminal Ground (see ()). In the second case, crucially, the COS alternant can be
said to derive from a basically COL structure that lacks, however, a Figure. I summarize
the scenario for the LA in both v- and s-framed languages in the table below:40
40
Hofmann and Szantyr (: ) document a kind of the LA built around adjectival predicates:
(i) Latin; based on Hofmann and Szantyr (: )
a. flores plenae in campo (COL alternant)
flower(F)NOM.PL full.NOM.F.PL in field.ABL
b. campus floribus plenus (COS alternant)
field(M)NOM flower.ABL.PL full.NOM.M
Latin as a satellite-framed language
() The LA in v- and s-framed languages
V-framed languages41 S-framed languages
With or without a prefix (cf. ()a,
COL alternants */?
()a)
Derived: Ground UOCs (cf. ()b)
Underived, with no Underived, with an independent pre-
COS alternants independent predicative dicative particle (cf. (), ())
particle (cf. ()b)
Underived, with no independent pre-
dicative particle (cf. ()b)
.. Pseudoreversatives
The last constructions I would like to deal with are the ones McIntyre (: )
calls Pseudoreversatives, which, to my knowledge, have not been dealt with before in
the literature on Latin.42 These are constructions where ‘the result expressed or
implied by the base verb gets reversed by adding a particle which contradicts this
result’ (McIntyre : ).43 The following German particle-verbs illustrate:
() German; McIntyre (: )
a. aus-parken
out-park.INF
‘Drive (a car) out of a parking space’.
b. ab-schwellen
down-swell.INF
‘Swell down, become less swollen’.
c. los-binden
free-tie.INF
‘Untie (a horse, etc.)’.
These constructions once again exemplify the s-framed pattern: the verb indicates the
nature of the process involved and a morphologically different element encodes the
These examples show that plenus ‘full’ could be predicated both of the entity that is full of something (see
(i)b) and of the matter or objects of which something is full (see (i)a). I leave this striking kind of the LA for
future research.
41
See Munaro () for Italian cases of the LA involving a contrast between an unprefixed verb and a
prefixed verb.
42
Although, for a diachronic remark on cases like aperio ‘open’/operio ‘close’, which had arguably been
cases of Pseudoreversatives, see Turcan (: –).
43
See also Stiebels ().
S-framed constructions in Latin
Core Schema. Thus, in ()a the conceptual scene evoked is the same as that evoked
by the verb parken ‘park’, the driving of a vehicle, but the result part of the event
usually entailed by parken ‘park’ is missing: the car is not in the parking space by the
end of the event. The addition of the particle aus- ‘out’ imposes a different result
state: the car ends up out (of the parking space). Pseudoreversatives are, therefore, a
particularly interesting probe into the nature of the semantic contribution of the verb
in s-framed constructions: it is truly understood as an adjunct, a modifier of the
event, the result being codified by an independent element (the particle, in the
examples above). Unsurprisingly, Latin features Pseudoreversatives, as exemplified
below:
() Latin; Verg. Aen. ,
Quid moror? an mea Pygmalion
what.ACC delay.SG whether my.ACC.N.PL Pymalion.NOM
dum moenia frater de-struat [ . . . ]?
until wall(N)ACC.PL brother.NOM down-build.SBJV.SG
‘What am I waiting for? Maybe for my brother Pygmalion to destroy my
walls?’
() Latin; Plaut. Curc.
Valetudo de-crescit, ad-crescit labor.
health.NOM down-grow.SG at-grow.SG work.NOM
‘Health wanes; work increases.’
() Latin; Ov. Met.
Dis-iunxisse iuvencos.
asunder-yoke.INF.PFV oxen.ACC
‘Having unyoked the oxen’.
() Latin; Ov. Fast. ,
Dis-suto [ . . . ] sinu.
asunder-sew.PTCP.PFV.ABL.M pleat(M)ABL
‘With an unsewn pleat’.
() Latin; Plaut. Cist.
Ex-pungatur nomen, nequid debeam.
out-puncture.SBJV.PASS.SG name.NOM nothing.ACC owe.SBJV.SG
‘Let my name be erased (from the register of debtors), so that I’m left with no
debts.’
() Latin; Colum. ,
Ne ventis [pampini] ex-plantentur.
lest wind.ABL.PL shoot.NOM.PL out-plant.SBJV.PASS.PL
‘Lest the vine shoots be uprooted by the wind’.
Latin as a satellite-framed language
In all these examples the result inferred from the unprefixed verb is superseded by
that conveyed by the prefix. Thus, in () nomen ex-pungo “name out-puncture”
refers to the action opposite to nomen pungo ‘puncture a name’, that is, ‘write a name
by puncturing’, on a wax tablet with a sharp instrument. The name is, in effect, ‘taken
out of the tablet’, and this is conveyed by ex- ‘out’. The effect expressed by ex-pungo
“out-puncture” is, thus, that of erasing.
I propose that these constructions receive the same analysis as Figure UOCs. They
involve a PathP with a prefix encoding the result, and a root adjoined to v. The object
is merged as Spec-Place, and is interpreted as a Figure. In the following example
involving de-struo “down-build” ‘destroy’ (cf. the German literal correspondence
ab-bauen), the walls (mea moenia) are predicated to end up down (de-). The
Figure raises then to Spec-Path and is interpreted as a Measurer of the event:
() Analysis of ()
VoiceP
Pygmalion frater Voice’
Voice vP
v PathP
v stru mea moenia Path’
Path PlaceP
mea moenia Place’
Place de
The verbal root is merged as an adjunct to v, and whatever resulting state it may
convey as part, of course, of its encyclopaedic content, in this case that of being
built, is overriden by DE ‘down’. More clearly: the root DE, by virtue of its position at
Compl-Place, must be interpreted as a final state (a Terminal Ground) and the root
STRU ‘build’, by virtue of its position as an adjunct to v, cannot be interpreted as a
final state.
Crucially, as pointed out by McIntyre (), Pseudoreversatives (hence the name)
are not equivalent to predicates endowed with a reversative particle, which are to be
found in Romance or Germanic: Catalan des-fer ‘un-do’, English un-lock, etc. These
particles only furnish the reversative meaning. By contrast, the Latin prefixes
involved in the above Pseudoreversatives preserve a spatial meaning.44 In particular,
44
As regards de- ‘down, downwards’, Brachet (: ff.) points out that it is found as a pure
‘opérateur d’inversion’. However, he acknowledges (Brachet : ff.) that in the first attestations of
S-framed constructions in Latin
the reversative interpretation is a secondary effect derived from a clash between the
semantics of the prefix and the semantics of the verb, as has been shown for ().
The examples () t () are illustrative of the fact that the prefixes do have a
locational meaning. In () the verb dis-suo “asunder-sew” is secondarily interpreted
as the opposite of suo ‘sew’, but the final state encoded by the prefix is specifically that
of separation (of two pieces of fabric, in this case). This semantic nuance is different
from that conveyed by the prefix ex- ‘out’ in ex-planto “out-plant” (see ()),
where the final result is for the plant to be out of the earth. Note, finally, the contrast
obtained by the combination of two different prefixes with the same verb in ().
Pseudoreversatives make it particularly evident that some of the meaning compo-
nents traditionally attributed to roots, such as ‘state’ are, in fact, derived from the
structure. Thus, run-of-the-mill change-of-state verbs like iungo ‘yoke’ or planto
‘plant’ simply cannot be interpreted as such if their root is not inserted as Compl-
Place. In theories such as Harley’s () and Levinson’s () roots are typed
depending on what they encode: entity, state, or event. For instance, Levinson (:
), following Harley’s () classification of roots, takes a root such as OPEN as
being typed as <ss, t>, that is, a state. This explains the adequacy of this root in
change-of-state predicates:
() The archaeologist opened the sarcophagus.
However, a typing approach such as this one, when applied to cases of verbal
elasticity such as the one at hand, is forced to propose different groups of homoph-
onous roots distinguished by the semantic type. For instance, the root IUNG ‘yoke’ in
() must be of type <e, t>, entity (akin to that of the cognate noun iugum ‘yoke’),
since the (end) state in that predicate is codified by the prefix dis- ‘asunder’, and not
by the root. However, in the following example the root is interpreted as a final state
(that of ‘being yoked’), which would require type <ss, t>:
() Latin; Hyg. Fab. , ,
Equum cum boue iunxit ad aratrum.
horse.ACC with ox.ABL yoke.PST.SG at plough
‘He yoked the horse with the ox to the plough.’
The scenario in which roots, like DP arguments, receive an interpretation dictated by
their position in the configuration (see section ...) does away with this redun-
dancy problem.
de-prefixed verbs exhibiting a reversative meaning, the prefix retains the ‘downward’ nuance: de-scendo
“down-ascend” ‘go down’ (from scando ‘climb’), de-cresco “down-grow” ‘diminish’ (from cresco ‘grow’),
de-molior “down-construct” ‘demolish’ (from molior ‘construct’).
Latin as a satellite-framed language
. Summary
In this chapter I have shown that Latin is an s-framed language, in Talmy’s ()
sense, since the Core Schema, that is, the component specifying a transition into a
final state or location, and the eventive component of transition events are realized
independently. In addition, I have endeavoured to show the adequacy of a neo-
constructionist model in dealing with the constructions which make Latin an s-
framed language. I have introduced Talmy’s theory, and I have adapted it to the
theory introduced in section ., introducing a correspondence between the semantic
components in Talmy’s theory of transition events and the syntactic-semantic terms
of my theory. I have shown that the s-/v-framed distinction can be explained as a
result of a different interpretation of the structure at PF: in v-framed languages, v and
Path must be strictly adjacent to each other, which disallows the linear intervention
of any root previously adjoined to v. This mechanism explains why v-framed
languages do not feature constructions involving a manner-naming verb and an
expression encoding the Core Schema. In s-framed languages there is no such
adjacency requirement between v and Path, and, hence, v can be associated with a
root merged as an adjunct. This analysis, where v-framed languages are more
complex than s-framed ones with respect to the PF derivation, makes the welcome
prediction that s-framed languages allow v-framed constructions, that is, predicates
where the verb encodes the Core Schema: there is nothing in s-framed languages
precluding these constructions. I have provided an overview of the expression of
directionality in Latin, in the form of a prefix, a PP, or a DP, or a combination of the
former and the latter. I have pointed out that APs cannot express the PathP in Latin.
I have presented the evidence that Latin is an s-framed language by approaching a set
of constructions that conform to the s-framed schema: CDMCs, Figure UOCs,
Ground UOCs, constructions involved in the LA, and Pseudoreversatives. All these
constructions have been argued to involve a verbal root merged as an adjunct to v
and interpreted, consequently, as a Co-event. In turn, the PathP is expressed through
an independent element. In the discussion of all these constructions I have tried to
show how the facts naturally derive from a neo-constructionist account, where it is
the syntactic structure, independently of the roots inserted therein, that determines
the structural semantics and the argument structure properties of the constructions.
5
Weak satellite-framed languages
In Chapter I showed that Latin is an s-framed language, through the examination of
a range of different constructions that conform to the s-framed pattern: the PathP
realized as an element morphologically different from v, which is therefore able to
associate with a root expressing Co-event. In this chapter, I set Latin in relation to
other s-framed languages with respect to the type of s-framed constructions it allows.
In particular, I focus on the fact that Latin does not feature s-framed constructions
based on adjectival predicates, i.e. complex adjectival resultative constructions. In
section . I show that corpus research reveals that this disallowance exists, and
I discuss why it is a puzzle in the present framework. After considering Slavic, a
group of languages established as s-framed, and exhibiting the same ban on adjectival
resultatives, I make the observation that both Latin and the Slavic languages in fact do
not allow PP resultatives either if they are not headed by a prefixed verb. The
generalization is then formulated that complex resultative constructions are always
prefixed in these languages, which I call weak satellite-framed, and the hypothesis is
put forward that the prefixation requirement is at the basis of the non-existence of
AP resultatives. In section . I provide evidence that the generalization holds for
both Latin and Slavic. In section . I offer an approach to the cross-linguistic
allowance of complex adjectival resultative constructions based on the consideration
of two factors: the affixal relation between v and Path, implemented as an instance of
Raising (of Path to v) and the inflectional marking of predicative adjectives. An
overall summary is presented in section ..
. The non-existence of complex adjectival resultatives
in Latin and Slavic
.. Resultative constructions: initial clarifications
Before dealing with the non-existence of complex adjectival resultative constructions
in s-framed Latin and Slavic we need to clarify the term complex resultative con-
struction: first, by distinguishing the phenomenon it labels from simple resultative
constructions; second, by making a distinction between strong resultatives and weak
The Morphosyntax of Transitions. First edition. Víctor Acedo-Matellán.
© Víctor Acedo-Matellán . First published by Oxford University Press.
Weak satellite-framed languages
resultatives, the former being found only in s-framed languages, the latter being
found in both s- and v-framed languages; and, finally, by discussing the situation
aspect that complex resultative constructions usually involve.
... Complex and simple resultative constructions I take complex resultative
constructions to be constructions that depict a complex event (see Levin and Rappa-
port Hovav : ) involving the attainment of a resulting state/location but, also,
a differentiated activity leading to that state/location.1 The constructions referred to
in section . as s-framed are all in fact complex resultative constructions in this
sense. For instance, in () the location expressed by the prefixed ad- ‘at’ (assimilated
as ac- in the example), understood as the vicinity of a reference point already
introduced in the discourse, is attained as the result of a running event encoded in
the verb currit ‘runs’:
() Latin Complex Directed Motion Constructions; Cic. Verr. , , ,
Subito ipse ac-currit.
suddenly himself.NOM.M.SG at-run.SG
‘Suddenly he himself arrives in haste.’
() Latin Figure and Ground Unselected Object Constructions
a. Liv. , ,
Neque enim omnia emebat aut e-blandiebatur.
nor hence everything.ACC buy.IPFV.SG or out-flatter.IPFV.SG
‘Nor did he acquire his object in all cases by money or flattery.’
b. Plin. Nat. ,
Elephantos ab iis e-bibi.
elephant.ACC.PL by them.ABL out-drink.INF.PASS
‘That the elephants are drunk dry by them’.
() Latin prefixed COL alternants of the Locative Alternation; Cels. ,
Ulceri medicamentum [ . . . ] in-ducatur.
ulcer(N)DAT.SG medicament.ABL.N.SG in-lead.SBJV.PASS.SG
‘Let the medicament be smeared into the ulcer.’
() Latin Pseudoreversatives; Plaut. Cist.
Ex-pungatur nomen, nequid debeam.
out-puncture.SBJV.PASS.SG name.NOM anything.ACC owe.SBJV.SG
‘Let my name be erased, so that I’m left with no debts.’
1
The term resultative construction has almost always been applied to complex resultative constructions
where the XP expressing the result state is an AP—see Halliday (); Simpson (); Levin and
Rapoport (); Hoekstra (); Carrier and Randall (); Levin and Rappaport Hovav (,
); Neeleman and van der Koot (); Mateu (); Boas (); Kratzer (); and Tomioka
(), among others. Crucially, I use the term in a wider sense.
The non-existence of complex adjectival resultatives in Latin and Slavic
The complex component of the formula complex resultative construction is crucial.
Indeed, the term resultative has sometimes been applied to any construction
implying a resulting state, as in Nedjalkov (). Thus, sentences such as He
made the table clean or He cleaned the table could be called (simple) resultative
constructions, but not complex resultative constructions, since they do not involve
any differentiated activity event leading to the resulting state. In the first case, the
result state is encoded by the AP, while the verb expresses an abstract change of
state, but no differentiated process leading thereto. In the second case, the result
state is encoded by the deadjectival verb clean. These constructions are perfectly
possible in v-framed languages, as the well-formedness of the next Catalan sen-
tences shows:
() Catalan
La Sue deixà la taula neta.
the Sue leave.PRF.SG the table clean
‘Sue made the table clean.’
() Catalan
La Sue netejà la taula.
the Sue clean.PRF.SG the table
‘Sue cleaned the table.’
In the discussion central to this chapter I will focus almost only on complex
resultative constructions, but see section .. and, particularly, section .., for
simple resultative constructions in Latin based on a light change-of-state verb and an
AP, as in ().
... Strong and weak resultative constructions A second initial clarification that
needs be made is about the difference between so-called strong and weak (com-
plex) resultative constructions. Importantly, Washio (), in his comparison of
English and Japanese adjectival resultative constructions, makes a distinction
between these two types of resultative constructions, illustrated by the following
examples:
() Washio (: )
a. John hammered the metal flat.
b. John painted the wall blue.
In the strong resultative construction of () the activity event expressed by ham-
mered, in the absence of the adjective flat, does not necessarily lead to any result
state: John could hammer indefinitely on a diamond-hard metal without the
slightest flattening thereof being attained. This, of course, changes when the
adjective is added. In contrast, the verb painted in the weak resultative construction
Weak satellite-framed languages
of ()b entails the attainment of a result state, namely, that of being painted, and the
AP blue is a specification of that result state. Washio observes that the adjectival
resultatives allowed in Japanese are always of the weak type (see the Japanese rendition
of ()b in ()b), the strong type being disallowed (see the Japanese rendition of ()a
in ()a):
() Japanese; Washio (: )
a. ??John-ga kinzoku-o petyanko-ni tatai-ta.
John-NOM metal-ACC flat-ni pound-PST
b. John-ga kabe-o buruu-ni nut-ta.
John-NOM wall-ACC blue-ni paint-PST
As Washio (: ff.) himself observes, the contrast of () is to be found, to a
certain extent, in other languages, like Romance:
() Italian; Napoli (), in Washio (: )
a. *Gianni ha martellato il metallo piatto.
Gianni has hammered the.M metal(M) flat.M
b. Gli operai hanno caricato il camion pieno.
the workers have loaded the.M truck(M) full.M
‘The workers have loaded the truck full.’
The division of languages allowing and disallowing strong resultative constructions
seems amenable to the s-/v-framed distinction: the former allow strong resultatives
because in them the Core Schema expressing a result state can be completely
independent from the verb, which expresses a pure process. Since in the latter
the Core Schema must be expressed through the verb, the only type of resultatives
that they may allow are those in which the verb already entails a result state (the
Core Schema) further specified by an adjective (weak resultatives). From this
perspective, weak resultative constructions turn out to be simple resultative con-
structions in the sense expressed in the previous section: they imply no differen-
tiated process leading to a result state; rather, the result state encoded by the AP is
a mere specification of the one already encoded (entailed) by the verb, pretty much
in the sense of Tortora’s () Further Specification Constraint—see also Tenny’s
() Single Delimiting Constraint. To capture this idea, I propose that the AP in
weak resultative constructions is an adjunct to PlaceP, and must be thus inter-
preted as a modifier of the (final) state encoded by PlaceP. Thus, an analysis of the
Japanese weak resultative of ()b, ignoring—for clarity’s sake—word order and
tense and case morphology, would be as follows:
The non-existence of complex adjectival resultatives in Latin and Slavic
() Analysis of ()b
VoiceP
John-ga Voice’
Voice vP
v PathP
kabe-o Path
Path PlaceP
AP PlaceP
buruu-ni kabe-o Place’
Place nut
In the present discussion the weak/strong distinction, as applied to adjectival
and, crucially, non-adjectival resultative constructions, will be important in
section ...
... Situation aspect in complex resultative constructions. The AP as a result
predicate The last preliminary qualification I would like to make refers to the
situation aspect of complex resultative constructions. This qualification is important,
since I will be using telicity in distinguishing true complex resultative constructions
from other atelic constructions which resemble them. This is why I will examine here
the cases involving a mismatch between resultativity and telicity. In addition, I will
defend the view that in (complex) AP resultative constructions the AP is to be
analysed as a resultative predicate.
(Complex) resultative constructions are standardly assumed to be accom-
plishments, involving a process, expressed by the verb, incrementally leading
to a result state expressed by the AP (or the respective XP) and predicated of
the internal argument. In this sense they are generally taken to be telic, featuring
a telos or culmination point: the result state (see Dowty ; Rappaport
Hovav and Levin ; Rapoport ; Rappaport Hovav and Levin ;
and Mateu , , among others). I myself have adopted this view in
defining complex resultative constructions in section .... However, Borer
(b: ff.), building on Wechsler (), observes that complex resultative
constructions are not necessarily telic, both when the internal argument is
a bare plural or mass NP, as in (), and, more surprisingly, when it is
a quantity DP, as in () (see also MacDonald : ff., who advocates a
Weak satellite-framed languages
dissociation between resultativity and telicity on the basis of examples such as
(), not ()):
() Wechsler (), in Borer (b: )
a. John hammered metal/cans flat (for an hour/*in an hour).
b. Kim sang babies asleep (for an hour/*in an hour).
() Wechsler (), in Borer (b: )
a. You can paint these walls white for hours, and they won’t become white
(e.g. because something in the plaster oxidizes the paint).
b. We yelled ourselves hoarse (for ten minutes).
On the basis of these facts, Borer rejects a syntactic analysis of resultatives
interpreted as incremental processes leading to a result state; instead, she puts
forward an account in which the verb and the adjective in a resultative construc-
tion are two listemes forming a complex head (a complex predicate)—such as
paint-white for ()a—which, as such, is neither telic nor atelic, like any other
listeme in her framework. This complex listeme may be embedded under an AspQP
projection, giving rise to telicity if a quantity DP is merged as the specifier (as in
You can paint these walls white in a few days) or under the semantically vacuous,
structural counterpart of AspQP, what she calls a shell Functional Projection FSP
(Borer b: ff.), giving rise to atelicity (as in () and ())—see section
... The accomplishment reading of telic resultatives would, therefore, be the
result of imposing a telic structure, the AspQP projection, on a complex listeme
such as paint-white.
Whatever the right approach is to the aspectual data in () and (), I believe
that, pace Borer (b), there are reasons to think that the AP in resultative
constructions does not form a complex predicate with the verb. One of them is the
incompatibility of result APs with the telicity signalling particle up in English.
Borer (b: ) suggests that this particle does not directly induce telicity. In
her terms, it does not assign range to the open value heading AspQP; rather, up is
an adjunct forcing the projection of and modifying AspQP, which is still in need of
range assignment from a quantity DP sitting in the Spec-AspQ position. Thus, up is
only possible if AspQP is projected, but its presence cannot by itself license the
projection of AspQP. This explains the following paradigm, in which the letters, a
quantity DP, is an appropriate range assigner for the head of AspQP, and the non-
quantity DP letters is not:
() Borer (b: –)
a. Kim wrote the letters up.
b. *Kim wrote letters up. (Single-event interpretation.)
The non-existence of complex adjectival resultatives in Latin and Slavic
However, if up is an (optional) adjunct in telic predicates and the adjective in
resultative constructions merely forms a complex listeme with the verb, it is not
clear why these two elements are not always compatible with each another:2
() English informants
a. ?John hammered the metal up flat.
b. *?John hammered up the metal flat.
c. *John hammered the metal flat up.
d. *John hammered flat the metal up.
e. *John hammered flat up the metal.
Importantly for the present perspective, the most deviant cases in () are, crucially,
those where the particle follows the adjective, namely ()c to ()e. I suggest that in
the other cases, where the adjective follows the particle, the adjective can actually be
analysed as a low adjunct, further specifying the result state encoded by the particle (a
state identified with complete affectedness—see section ..., for a discussion of this
value in the Latin prefix com-). In ()c to ()e, however, it is the particle that is forced
to be analysed as an adjunct, by virtue of its relative position with respect to the verb
and the adjective. But this interpretation is extremely awkward, since the content of the
adjective is of course more specific than that of the particle. If it is on the right track,
this will constitute further evidence in favour of a result predicate approach to
resultative constructions. On the other hand, from Borer’s perspective, the only
distinction between John hammered the metal flat in two hours and Kim wrote the
letters in two hours is reduced to whether the listeme embedded under AspQP is
complex (hammer-flat) or not (write), a distinction not capable, in Borer’s (b)
system, of generating a difference in grammaticality. So it comes as a surprise that only
the latter is fully compatible with the adjunct up.3 Facts similar to those of () are
found in Dutch, illustrated by well-known examples from Hoekstra and Mulder
()—see also section ..., for a comparison between Dutch be- and Latin com-:
() Dutch; Hoekstra and Mulder (: , )
a. Dat Jan de tuin {vol/be-}plant.
that Jan the garden full/be-plant.SG
‘That Jan fills the garden with plants’.
2
See Den Dikken (: ) for discussion of similar data from Jackendoff (: ).
3
Another problem for Borer’s account is the fact that the adjective and the verb may appear separate,
which is unexpected if they form a complex listeme. Moreover, it is not clear, within her account, why strong
resultative constructions are systematically ungrammatical in v-framed languages like Romance (Mateu ;
Acedo-Matellán and Mateu ). Indeed, if the peculiarity of these constructions boils down to the
embedding of two listemes, rather than one, within the functional structure, why are such languages unable
to combine them? Of course, the combination of two listemes into a complex one could be stipulated as
unavailable in their grammars. However, Borer rejects any account of cross-linguistic variation that is not
based on morphophonological properties of the functional lexicon (see Borer b: ff.).
Weak satellite-framed languages
b. Dat ik de tuin (*vol) be-plant.
that I the garden full be-plant.SG
‘That I fill the garden with plants’.
In ()b the result AP vol ‘full’ seems to be also incompatible with the prefix be-, which,
like up, also signals complete affection (see McIntyre : ff.; see also Mulder for
a further analysis of the be-/vol complementary distribution). I think that the comple-
mentary distribution of these elements further argues against an analysis in terms of
complex predicate formation. For all the above reasons, I favour an analysis of complex
resultatives where the AP really is a result predicate and is, therefore, in complementary
distribution with any other element that also arguably qualifies as a result predicate (cf. up
for English and be- for Dutch). As we will see, this position will be crucial in the analysis of
complex resultatives constructions based on prefixes in Latin and Slavic, and also in
understanding why these languages do not allow complex AP resultative constructions.
Turning back to the (a)telicity of complex resultative predicates, I will adopt the
strong position that these constructions are telic, unless the object is a non-quantity
DP, namely a mass DP. In these cases, I argue for the coexistence of both resultativity
and atelicity. For example, in ()a, although PathP is projected and resultativity
obtains, the non-quantity DP metal, which raises from Spec-Place to Spec-Path,
cancels a telic interpretation of the predicate. In particular, a transition of becoming
flat is entailed to have taken place: it must be true that some metal has become flat.
However, since the quantity of metal is not determined, the end of the flattening
event cannot be determined either. Thus, atelicity arises.
As for the example in ()a, where one cannot appeal to a non-quantity status of
the object to explain the atelic reading, I note that not all authors agree on its
grammaticality. For instance, MacDonald (: ) observes that ‘AP resultatives
are incompatible with durative phrases on a single event interpretation’, providing
the following examples:
() MacDonald (: )
a. John wiped the table clean (#for an hour).
b. Bill hammered the metal flat (#for an hour).
c. They painted the barn yellow (#for an hour).
Cases such as ()b are residual. Wechsler (: ) points out that the result
adjective is in these cases reinterpreted as an intensifier, so that the whole sentence
We yelled ourselves hoarse comes to mean something like We yelled a lot. I grant that
this analysis does not constitute a possible avenue within the present framework,
where structural semantics, including (a)telicity, cannot be overridden. However,
pending a better solution, I leave the problem at that.
There are other (apparent) mismatches between resultativity and telicity worth
commenting on. The first is seen in predicates where a durative adverbial measures
The non-existence of complex adjectival resultatives in Latin and Slavic
the final resulting state encoded by the result predicate. This is illustrated by the
following example:
() German; Kratzer (: )
Wir werden das Boot für ein paar Stunden auf-pumpen.
We will.PL the boat for a pair hours up-pump.INF
‘We will pump up the boat for a few hours.’
As Kratzer herself notes, the sentence above implies that ‘the boat will remain inflated
for a few hours’, and not that there will be any pumping-up event that will last two
hours. Accordingly, the adverbial für ein paar Stunden ‘for a few hours’ can be
claimed to be measuring the result state incarnated as PlaceP, where the root AUF
‘up’ is embedded as Compl-Place.
I want to make a final remark about the situation aspect of CDMCs, since I have
assumed that these constructions are also complex resultative constructions. As
I pointed out in section ..., CDMCs are usually taken to be telic. Cases in
point are the examples in ():
() Dutch and German; Randall, van Hout, Weissenborn and Baayen (: )
a. John is in twee seconden de kamer in gedanst.
John is in two seconds the room in dance.PTCP.PFV
‘John has danced into the room in two seconds.’
b. John ist in zwei Sekunden in-s Zimmer getanzt.
John is in two seconds in-the.ACC room dance.PTCP.PFV
‘John has danced into the room in two seconds.’
However, it seems that atelic CDMCs based on an unbounded directional expression
are possible in some languages. This is shown in the contrast between the following
two examples from Dutch and German. The German example of ()b seems to be a
CDMC, that is, an unaccusative construction in which the subject is a Figure and the
unbounded PP is its predicate, as suggested by the use of the auxiliary BE. In contrast,
in the Dutch correlate of ()a, HAVE, and not BE, is selected, which argues for the
adjunct (to vP) status of the PP door de saal ‘through the room’ and the unergative
status of the predicate (see section ...):
() Dutch and German; Randall et al. (: )
a. John heeft door de saal urenlang rond-gedanst.
John has through the room for hours around-dance.PTCP.PFV
‘John has danced around the room for hours.’
b. John ist stundenlang durch den Saal herum-getanzt.
John is for hours through the.ACC room around-dance.PTCP.PFV
‘John has danced around the room for hours.’
Weak satellite-framed languages
That the sentence in ()b does not involve an unergative predicate expressing non-
directed motion is further suggested by the fact that unquestionably non-directed
motion predicates, featuring either no spatial PP or a locative PP at most, present a
HAVE-auxiliary:
() German; Randall et al. (: )
John hat stundenlang auf dem Tisch getanzt.
John has for hours on the.DAT table dance.PTCP.PFV
‘John has danced on the table for hours.’
The German data suggest, therefore, that CDMCs cannot be claimed to be universally
telic. However, on the basis of the fact that in many languages, like Dutch, CDMCs
must be telic, I will use telicity as a criterion for detecting CDMCs in Latin (and Slavic).
In sum, if provisions are made for the resultativity-telicity mismatches we have
seen, mainly induced by the cases of non-quantity direct objects and of durative
adverbials measuring the duration of the resulting state, I think that telicity can be
taken as a quite reliable criterion in distinguishing complex resultative constructions,
structured around a PathP, from unergative vPs encoding an activity.
.. No complex AP resultatives in Latin
Complex resultative constructions typically feature a predicative element expressing
the resulting state. The result predicate may correspond to different categories: an AP
(see the German example of ()), a PP (see the Norwegian example of ()) or a
particle/prefix (see the English and Latin examples of () and (), respectively):
() German; Kratzer (: )
Die Teekanne leer trinken.
the teapot empty drink
‘To drink the teapot empty’.
() Norwegian; Tungseth (: )
Jon syklet til byen på en time.
Jon bike.PST into town in one hour
‘Jon biked into town in an hour.’
() Hale and Keyser (: )
He slept the hours away.
() Latin; Plin. Nat. ,
[Serpentes] [ova] solida hauriunt, [ . . . ] atque
snake(M)NOM.PL egg.ACC.PL whole.ACC.PL swallow.PL and
putamina ex-tussiunt.
shell.ACC.PL out-cough.PL
‘Snakes swallow the eggs whole and expel the shells through coughing.’
The non-existence of complex adjectival resultatives in Latin and Slavic
As I already pointed out in section .., the adjectival type is included by Talmy
() himself in the range of constructions possible in s-framed languages and
impossible in v-framed ones. This is illustrated in () through a contrast between s-
framed German and v-framed Spanish: while German encodes the resulting state of
the complex event as an AP, leaving the verb to express manner, Spanish encodes the
resulting state as the verb, and the manner has to be expressed as an adjunct. A literal
Spanish translation of () is not well-formed:
() German and Spanish; Talmy (: )
a. Der Hund hat den Schuh [kaputt]Core schema -gebissen
the dog has the shoe in_pieces -bite.PTCP.PFV
‘The dog bit the shoe to pieces.’
b. El perro [destrozó]Event + Core schema el zapato a mordiscos
the dog destroy.PRF.SG the shoe to bites
As was pointed out in section .., anyone acquainted with Latin recognizes the type
represented by () or () as not possible or general in this language, although it is
an s-framed one. That is, Latin does not seem to feature complex resultative con-
structions where the result predicate is an AP, as illustrated by the next made-up
example:
() Latin made-up ungrammatical example
*Ovidia poculum vacuum bibit.
Ovidia.NOM goblet(N)ACC.SG empty.ACC.N.SG drink.SG
Pinkster (), in a discussion on resultative secondary predication in Latin, points
out that this type of predication is only productive with ago ‘lead, drive’:
() Latin; Cic. Ver. ., in Pinkster (: )
Agunt eum praecipitem poenae
drive.PL him.ACC headlong.ACC.M.SG punishment.NOM.PL
civium Romanorum.
citizen(M).GEN.PL Roman.GEN.M.PL
‘The punishments of Roman citizens drive him crazy.’
However, it is pretty evident that this type of resultative is not complex in the
sense intended in the present discussion, since it does not involve a co-event
leading to the causation of the final state described by the AP. Contrary to
Pinkster’s claim above, there seem to be other verbs in Latin licensing simple
resultative constructions: the prefixed verbs red-do “back-give” ‘return’ and re-
linquo “back-leave” ‘leave’, and the simple verb facio ‘do’. I will deal with simple
adjectival resultative constructions headed by these verbs and also by ago ‘lead,
drive’ in section ...
Weak satellite-framed languages
Interestingly, Kühner and Stegmann (: –) provide examples of apparent
complex AP resultative constructions in Latin. They call the phenomemon ‘proleptic
use of the adjective’, since the AP does not seem to be a pure modifier of the noun
with the exclusion of the verb:
() Latin; Verg. A. , , in Kühner and Stegmann (: –)
Submersas ob-rue puppas.
sink.PTCP.PST.ACC.F.PL against-hurl.IMP ship(F)ACC.PL
‘Overwhelm the ships so that they sink.’
() Latin; Verg. A. , , in Kühner and Stegmann (: –)
Tectosque per herbam dis-ponunt enses
cover.PTCP.ACC.M.PL=and through grass.ACC separate-put.PL sword(M)ACC.PL
et scuta latentia condunt.
and shield(N)ACC.PL be_hidden.PTCP.PRS.ACC.N.PL lay.PL
‘They arrange the swords in different places, hidden in the grass, and they lay
the shields out of sight.’
() Latin; Verg. A. , , in Kühner and Stegmann (: –)
Premit placida aequora pontus.
press.SG quiet.ACC.N.PL surface(N)ACC.PL sea.NOM
‘The sea presses its surface calm.’
() Latin; Ov. Met. , , in Kühner and Stegmann (: –)
Laniataque pectora plangens.
lacerate.PCTP.ACC.N.PL=and chest(N)ACC.PL beat.PTCP.PRS
‘And beating his chest in agony’.
There are two reasons to doubt that we are dealing with complex AP resultative
constructions of the type encountered in the Germanic languages. The first is that the
AP in the above constructions is almost always a participle (in out of the cases
provided by Kühner and Stegmann : –). We know that participles do not
make good secondary resultative predicates in languages like English (Green ;
Carrier and Randall ):
() Embick (: )
John kicked the door open/*opened.
As pointed out by Embick (: ), the above restriction has to do with the fact
that the participle is in itself already resultative, that is, it presupposes a previous
event whose result it identifies.
Second, many of the examples provided by these authors are headed by a prefixed
verb. We have seen evidence in the preceding chapter that prefixed verbs in transition
predicates in Latin already encode the result of the event, specifically that the prefix
The non-existence of complex adjectival resultatives in Latin and Slavic
can be claimed to express it. Therefore, it seems that, if anything, the APs in the above
examples are further specifying a result state already encoded by the prefix, which
makes these examples cases of a different kind of adjectival resultative construction.
I will turn to the derivation of this kind of AP prefixed resultative at the end of section
... In light of these two empirical differences and the fact that, as pointed out by
the same authors, this phenomenon is very rare in prose, I consider that the above
evidence does not support a claim that complex AP resultatives are productive in
(Early and Classical) Latin.
To see whether () represents a general situation in Latin, that is, whether this
language in fact cannot generate complex AP resultative constructions, I have
performed a corpus search, based on the one designed by Boas () for English.4
Boas selected a set of adjectives recurrently used in the literature on resultatives as
result predicates, such as dead, clean, awake, etc. He then conducted a search to
find out which verbs were most often used in resultative constructions with those
adjectives.5
In applying this methodology to Latin, I have first established the correspondences
of the English adjectives in Boas’s set and then the correspondences of (some of) the
verbs he established as more collocative for each adjective, wherever possible.
I present below the list of the combinations I have searched for:
() Adjectives and verbs used in the search for adjectival resultatives in Latin
Adjectives Verbs
aeger ‘ill’ bibo ‘drink’
cassus/inanis/vacuus
bibo ‘drink’, haurio ‘scoop’, poto ‘drink’
‘empty’
calcitro ‘kick’, clamo ‘scream’, figo ‘prick’, grunnio
experrectus ‘awake’ ‘grunt’, osculor ‘kiss’, plaudo ‘clap’, quatio ‘shake’,
‘jerk’, terreo ‘induce terror’
amens/demens/insanus
clamo ‘scream’, loquor ‘talk’, strideo ‘yell’
‘insane, mad’
caedo ‘cut, knock’, calcitro ‘kick’, cudo ‘knock’, occido
mortuus ‘dead’
‘kill’, tundo ‘strike, knock’, verbero ‘smite’
4
Analogously, Whelpton () uses Boas’s () appendix of examples of adjectives, taken from the
British National Corpus, as the starting point for his own investigation of Icelandic resultatives (see
section ..).
5
See Boas (: ff.).
Weak satellite-framed languages
Adjectives Verbs
frico ‘rub’, lambo ‘lick’, lavo, ‘wash’, luo ‘wash’, polio
mundus/nitidus ‘clean’ ‘scour, polish’, rado ‘scrape’, sorbeo ‘suck’, sugo ‘suck’,
tergeo ‘wipe’, verro ‘sweep’
amburo ‘burn’, aspergo ‘spray’, cremo ‘burn’, maculo
niger ‘black’ ‘stain’, pingo ‘paint’, spargo ‘spray’, tingo ‘dye’, tundo
‘beat’, uro ‘burn’, verbero ‘smite’
opimus/pinguis ‘fat’ cibo ‘feed’, pasco ‘pasture’
aro ‘plough’, caedo ‘cut, knock’, cudo ‘knock’, premo
planus ‘flat’ ‘press’, sicco ‘dry’, sorbeo ‘suck’, sugo ‘suck’, tero ‘grind’,
tundo ‘beat’, volvo ‘roll’
farcino ‘stuff ’, farcio ‘stuff ’, saturo ‘stuff ’, sorbeo ‘suck’,
plenus ‘full’
stipo ‘cram’, sugo ‘suck’
quietus/tranquillus
cano ‘sing’, lallo ‘lull’
‘calm’
raucus ‘hoarse’ clamo ‘scream’, loquor ‘talk’, strideo ‘yell’
amplector ‘hug’, bibo ‘drink’, bullio ‘boil’, clamo
‘scream’, complector ‘hug’, ferveo ‘boil’, fleo ‘weep’, flo
‘blow’, frico ‘rub’, mulceo ‘caress’, mulgeo ‘milk’, ploro
aridus/siccus ‘dry’
‘cry’, premo ‘squeeze’, rado ‘scrape’, sanguino ‘bleed’,
sorbeo ‘suck’, strideo ‘yell’, stringo ‘squeeze’, sugo ‘suck’,
tergeo ‘wipe’, verro ‘brush’
caedo ‘cut, knock’, clamo ‘scream’, cudo ‘knock’, strideo
surdus ‘deaf ’
‘yell’, tundo ‘beat’
tortus ‘crooked’ caedo ‘cut, knock’, cudo ‘knock’, tundo ‘beat’
I have dismissed some of the verbs in Boas’s subcorpus. For instance, verbs such as
get, render, or make, which head simple resultative constructions. In some cases
I have added verbs which I imagined could be possible with the adjective. This is the
case of the verbs combining with aeger ‘ill’, or pinguis/opimus ‘fat’. The subcorpus
obtained was composed of all the sentences in which each adjective combined with at
least one of the verbs of the same row in the box. Despite the ample range of
adjectives and verbs used and their high absolute frequency in the Antiquitas corpus
(and in Latin in general), the results have been utterly negative. Therefore, my
conclusion is that Latin disallows this type of complex resultative construction.
The non-existence of complex adjectival resultatives in Latin and Slavic
This situation constitutes a puzzle within the perspective adopted here, where
adjectives are expected to be able to fulfil the role of result predicates in s-framed
languages in general. However, as I shall show in section .., Latin is not the only
s-framed language to ban the formation of AP resultatives.
.. No complex AP resultatives in Slavic
Slavic languages are considered by Talmy (: ) to be s-framed, since they
typically convey the Core Schema as an element different from the verb:
() Russian; Talmy (: )
Ptica [v]Core Schema -letela.
bird.NOM in -flew
‘The bird flew in.’
Gehrke (: ff.) disputes this claim by arguing that Slavic languages (Czech and
Russian, in particular), like v-framed languages, do not allow the integration of a non-verbal
predicate into an activity VP to derive an accomplishment structure. In these languages, the
author argues, resultativity is to be expressed in the verb, as shown with prefixed verbs.
However, by using a strictly Talmian perspective, this author misses the point that verbal
prefixes are precisely the kind of non-verbal predicates ‘integrated into an activity structure’
that are allowed in these languages (and, as I shall show, in Latin). At the basis of Gehrke’s
argument lurks the use of the word as a syntactic-semantic unity: Slavic (and Latin) can be
taken to be v-framed since the verb, that is, a word which may include, for instance, a verbal
prefix, is the privileged unit where resultativity is expressed. We note, however, that Talmy’s
() typology is constructed on considerations about morphemes, and not about words—
see section . for relevant discussion of the status of words as syntactic units. Gehrke’s
observation, however, does account for the fact that, as in Latin, when the result predicate is
an AP, the construction turns out to be ungrammatical:
() Russian; Strigin (: )
*Ona mylila men’a skolzkim.
she soap.PST me slippery
‘She soaped me slippery.’
Importantly, the contrast between () and () is not to be stated in terms of change
of location versus change of state. As it turns out, Russian (and Slavic, in general)
seems to succeed in mimicking typical adjectival resultative constructions found in
English, expressing a change of state, and even featuring unselected objects, with the
use of adpositional prefixes:
() Russian; Spencer and Zaretskaya (: )
a. Oni na-ezdili ètu dorogu.
they on-drive this road.ACC
‘They’ve made this road nice and smooth (by driving over it).’
Weak satellite-framed languages
b. Ona pere-igrala ruku na pianino.
she pere-played hand on piano
‘She’s hurt her hand playing the piano.’
We must conclude that a formal factor, categorial or otherwise, is responsible for the
contrast between () and (), on the one hand, and (), on the other hand.
Svenonius () points out the lack of AP resultative constructions in Slavic
languages. Thus, Svenonius (: ) states that ‘Slavic languages do not allow the
free formation of resultatives like shoot Dillinger dead, the way Germanic languages
do (Spencer and Zaretskaya , Strigin and Demjjanow )’. Other authors have
mentioned this state of affairs for other Slavic languages. Snyder (: ) includes
Russian and Serbo-Croatian in his list of languages disallowing AP resultatives.
Angelina Markova and Wojciech Lewandowski (personal communication) respect-
ively report that Bulgarian and Polish do not feature these constructions. I illustrate
with Bulgarian:
() Bulgarian
a. Te go za-streljaha (*umrial).
they him za-shot dead
‘They shot him (dead).’
b. Toj iz-chisti masata (*chista) (ot prah).
He out-wiped table.the clean (of dust)
‘He wipes the table clean of dust.’
Angelina Markova (p. c.) reports that AP resultatives expressing change of colour are
possible in Bulgarian:
() Bulgarian
Bojadisah stenata chervena.
paint.PST.SG wall.the red
‘I painted the wall red.’
However, as discussed in section ..., resultatives of this type are typical examples
of weak resultatives. If, as I have argued, weak resultatives are in reality simple
resultative constructions, the resultative AP merely specifying the result state already
encoded by the verb, data such as () do not constitute counterexamples to the
claim that Slavic does not admit complex adjectival resultative constructions.
.. No complex PP resultatives without a prefix. Internal and external prefixes
I conclude this section by presenting an empirical observation which may shed light
on why s-framed languages like Latin or Slavic do not allow resultative constructions
based on APs. It may also help us maintain a syntactic modelling of the Talmian
typology as the basic explanation for the availability of resultative constructions
The non-existence of complex adjectival resultatives in Latin and Slavic
cross-linguistically, regardless of the category of their resultative secondary predi-
cates. The observation is that of () and is first illustrated in () and () for Latin
and Bulgarian (Angelina Markova, p. c.), respectively:
() Neither Latin nor Slavic, both disallowing AP resultatives, seems to allow the
expression of complex PP resultatives without a result-conveying prefix
attached to the verb.
() Latin; Caes. Liv. , ,
Qui ubi *(ad-)equitavit portis . . .
who.NOM.SG as_soon_as at-ride.PRF.SG doors.DAT
‘This one, as soon as he had ridden up to the gates . . .’
() Bulgarian
*(Iz-)kopah sukrovishte (iz dupkata).
out-dig.PST.SG treasure.the out hole.the
‘I dug a treasure out of the hole.’
The observation in () involves the conception of prefixes as conveyors of the
resulting state or location of a complex event (the result predicate). In Chapter
I adopted this as the right analysis for prefixes in Latin: the sequence we identify as a
prefix in the surface is the result of the affixation of phonological material coming
from PlaceP, the projection codifying states and locations. Here, I add that a
resultative analysis is also proposed for Slavic verbal prefixes by Arsenijević (),
Gehrke (), and Žaucer (). This view holds well for the examples in () and
(), where the prefixes ad- ‘at’ and iz- ‘out’ clearly express the final location of the
complex directed motion constructions they are involved in. Importantly, the pre-
fixes involved in the generalization of () are merged vP-internally, and so they are
an instance of internal prefixes, which must be distinguished from external prefixes,
merged, by hypothesis, outside of the vP.6 The different merging site of both types of
prefixes accounts for a series of properties distinguishing them, as pointed out by
various authors (Svenonius ; Arsenijević ; Gehrke ). I briefly present
the most important differences before continuing the exploration of the generaliza-
tion of (). Morphologically, internal prefixes attach to stems, while external
prefixes may appear prefixed to already prefixed verbs, that is, they can stack:
() Czech; Gehrke (: )
a. Po(EXT)-od(INT)-stoupit.
a_little-from-step.INF
‘To step aside a little’.
6
The division, first proposed for Romance prefixes by Di Sciullo (, ), corresponds, roughly, to
that made between outer and inner prefixes (Padrosa and Markova ) and superlexical and lexical
prefixes (cf., for instance, Svenonius b and the other articles in the same volume on Slavic prefixes).
Weak satellite-framed languages
b. *Od(INT)-po(EXT)-stoupit.
from-a_little-step.INF
Syntactically, internally prefixed verbs show different argument structure properties
from those of their unprefixed counterparts. This is not the case with externally
prefixed verbs. For instance, in the first of the examples following, the internal prefix
makes the object obligatory. In the second example, the externally prefixed verb
allows the omission of the object, as does the unprefixed verb:
() Russian; Gehrke (: )
a. Na-pisat’P *(pis’mo).
on-write.INF letter.ACC
‘To write (up) *(a letter)’.
b. Po-pisat’P (pis’mo).
po-write.INF letter.ACC
‘To write (a letter)’.
Semantically, both internal and external prefixes induce (outer-aspectual) perfectiv-
ity, but only the former necessarily induce (inner-aspectual) telicity, according to
Gehrke (: ):
() Russian; Gehrke (: )
a. Ja na-pisalP pis’mo *(za) dve minuty.
I on-wrote letter.ACC in two minutes
‘I wrote a letter in/*for two minutes.’
b. On po-spalP (*za) dve minuty.
he po-slept in two minutes
‘He slept for/*in two hours.’
Finally, internal prefixes may trigger a special meaning of the base verb (see ()a),
while external prefixes only introduce aspectual (quantificational) modifications of
the whole event (see ()b):
() Serbo-Croatian; Arsenijević (: )
a. biti u-biti raz-biti pro-biti od-biti do-biti
beat in-beat around-beat through-beat away-beat to-beat
‘beat’, ‘kill’, ‘break’, ‘make a hole in’, ‘bounce’, ‘get’
b. kuvati na-kuvati iz-kuvati pro-kuvati pre-kuvati
cook on-cook out-cook through-cook over-cook
‘cook’, ‘cook many’, ‘cook all/fully’, ‘cook a bit’, ‘overcook’
There is reason to believe that the distinction between internal and external prefixes
holds in Latin too. Specifically, some prefixes act more like adverbs, rather than as
resultativity markers. For instance, the prefix sub- ‘under’ may be added to a simple
Latin and Slavic complex resultatives with an internal prefix
verb with an attenuative value, as in sub-rideo “under-laugh” ‘smile’, sub-bibo
“under-drink” ‘drink a little’, sub-luceo “under-shine” ‘glow’, etc. (see García
Hernández : ). Importantly, when this is the value of sub-, rather than its
locative value of ‘under’, it induces no changes in argument structure or inner aspect.
For instance, sub-blandior “under-flatter” ‘flatter a little’ is an activity predicate that
optionally takes a dative corresponding to the person being flattered, much like the
unprefixed counterpart blandior ‘flatter’. Thus, attenuative sub- ‘under’ contrasts
with the prefix e- ‘out’ in e-blandior “out-flatter” ‘obtain through flattery’, which,
as was shown in section ..., shows quite different argument structure properties
with respect to those of blandior ‘flatter’. Moreover, sub- ‘under’ shows some cases of
stacking, always outside the inner prefix, as in sub-ad-moveo “under-at-move” ‘to
bring near surreptitiously’, where sub- ‘under’ also conveys an adverbial value.
After this important proviso, I go back to the observation in (). If it is on the
right track, which I will attempt to show later, a possible way to make sense of it is
through an implication construed in the following terms:
() In some (s-framed) languages, there is a morphological requirement on the
element expressing the result predicate and the verb: they have to form a single
(prosodic) word. This requirement prevents those languages from featuring
complex adjectival resultative constructions.
If the morphological packaging or, to borrow Pinault’s () term, the univerbation
affecting the result predicate and the verb is taken as obligatory, possible, or impos-
sible, and if no other factor is taken into account, the implication in () yields a
certain distribution of s-framed languages with respect to their allowance of complex
resultatives based on APs:
() Relation between univerbation of the result predicate and the verb and avail-
ability of AP resultatives
Univerbation of the result predicate and the verb
impossible possible obligatory
AP resultatives available available unavailable
I will come back to this typology in section .. But first it should be shown that the
observation in () is empirically correct for Latin and Slavic. This is done in section ..
. Latin and Slavic complex resultatives with an internal prefix
.. Latin complex resultative constructions
In the case of Latin, I have conducted three corpus searches in order to ascertain
whether this language expresses complex resultative constructions always through
Weak satellite-framed languages
the aid of a prefix. The first two searches aim at finding constructions formed by a PP
and a surface-contact verb, in the first search, and a sound-emission verb, in the
second search, taking into account only unprefixed verbs. The type of resultative
constructions I am looking for in these searches is respectively illustrated by the
English constructions in () and ():7
() Rappaport Hovav and Levin (: )
Terry swept the crumbs into the corner.
() Folli and Harley (: )
Mary whistled Rover to her side.
The search involving sound-emission verbs, did not produce any results whatsoever,
confirming () for Latin. That involving surface-contact verbs yielded a few appar-
ent examples, those in () to ():
() Latin calco ‘tread, press’
a. Cato, Agr.
[Oleas] in orculam calcato.
olive.ACC.PL in vessel.ACC press.IPV.FUT.SG
‘Press [the olives] down into an earthenware vessel.’
b. Stat. Theb. ,
Clipeumque in pectora calcat.
shield.ACC=and in chest.ACC press.SG
‘He stands/presses his shield against his chest.’
() Latin tero ‘rub, grind; thresh’
a. Petron. ,
Sparserunt [ . . . ] ex lapide speculari pulverem
sprinkle.PRF.PL, out mica.ABL powder(M)ACC.SG
tritum.
grind.PTCP.PFV.ACC.M.SG
‘They sprinkled powder ground out of mica.’
7
The material used for these two searches is displayed in (i) and (ii), respectively:
(i) Latin; Search for complex PP resultatives with unprefixed surface-contact verbs
a. Verbs: calco ‘tread, press’, frico ‘rub’, rado ‘scrape, scratch; razor’, tergeo ‘wipe’, tero ‘rub, grind;
thresh’, verro ‘sweep’.
b. Prepositions: ab ‘off, away’, ad ‘at, beside, by’, de ‘downward; from, away’, ex ‘out of ’, in ‘in’.
(ii) Latin; Search for complex PP resultatives with unprefixed sound-emission verbs
a. Verbs: fremo ‘roar’, strideo ‘yell’, rideo ‘laugh’, sibilo ‘whistle’, latro ‘bark’, ululo ‘howl’, mugio
‘moo’, hinnio ‘neigh’, strepo ‘make a lot of noise’, grunnio ‘grunt’, rudo ‘bray’, balo ‘bleat’.
b. Prepositions: ab ‘off, away’, ad ‘at, beside, by’, de ‘downward; from, away’, ex ‘out of ’, in ‘in’.
Latin and Slavic complex resultatives with an internal prefix
b. Plin. Nat. ,
Astragali radix in pulverem trita.
astragalon.GEN root(F)NOM.SG in powder.ACC grind.PTCP.PFV.NOM.F.SG
‘Astragalon root ground into powder’.
c. Plin. Nat. ,
[Caseum] veterem [ . . . ] in farinam tritum.
cheese(M)ACC old.ACC.M in flour.ACC grind.PTCP.PFV.ACC.M.
‘Cured cheese, ground into flour’.
() Latin verro ‘sweep’
a. Hor. Sat. , ,
Piscis hiberno ex aequore verris.
fish.ACC.PL stormy.ABL.N.SG out sea(N)ABL.SG sweep.SG
‘You sweep the fish from a stormy sea.’
b. Hor. Carm. , ,
Proprio condidit horreo quidquid
own.ABL.M.SG hide.SG barn(M)ABL.SG whatever.ACC
de Libycis verritur areis.
off Libyan.ABL.PL sweep.PASS.SG threshing-floor.ABL.PL
‘He hides in his own barn whatever is swept out of Libyan threshing-floors.’
Example ()a is an excerpt from Cato’s De agricultura in which the procedure for
pickling green olives is described. After the olives have been soaked in vinegar
and mixed with other spices, they have to be pressed in order to extract the liquid.
This is expressed by in orculam calcato ‘into an earthenware vessel’. But note that the
olives need not end up in the vessel as a result of a pressing event (calcato); rather, it
seems, they are first put in and then they are pressed. This is the interpretation
chosen by Nisard () in his translation: ‘foulez-les avec vos mains bien sèches
dans un vase de terre’ ‘press them with well dry hands in an earthenware vessel’. That
this must be the right interpretation is supported by the other example involving
calco ‘press’, ()b, where the PP in pectora ‘against his chest’ clearly represents an
unbounded directional PP, since there is, of course, no entailment that the shield
end up within the soldier’s chest. As for the examples in (), they could be taken as
cases of weak resultatives, the PP merely modifying a result state encoded in the verb
tero ‘grind’. Independent evidence of this is that in v-framed languages, with no
allowance for complex resultative constructions, predicates analogous to those in
() are fine:
() Spanish; examples from a Google search
a. Triturar en polvo fino las nueces.
crush.INF in powder fine the nut.PL
‘To grind the nuts into fine powder’.
Weak satellite-framed languages
b. El arroz crudo se puede moler en harina.
the rice raw REFL.SG can.SG grind.INF in flour
‘Raw rice can be ground into flour.’
Notwithstanding other examples, such as ()a, which really seem to imply a final
location, it seemed to me that another search was needed that took into account the
inner-aspectual dimension of the complex predicate in Latin. So, I searched for
combinations of, on the one hand, prefixed and unprefixed manner-of-motion
verbs, and, on the other, a series of telicity-signalling expressions which could
guarantee the resultativity of the predicate. Here, the results were significant: out of
the telic predicates yielded by the search, , listed in (), are headed by
unprefixed manner-of-motion verbs, while , represented by the sample in (),
are headed by prefixed ones—see the Appendix for the totality of telic cases of
prefixed verbs:8
() Latin; telic predicates headed by unprefixed manner-of-motion verbs
a. Cic. Quinct.
Non statim ad C. Aquilium [ . . . ] cucurrisses?
not at_once at C. Aquilius.ACC run.PRF.SBJV.SG
‘Wouldn’t you have run up to C. Aquilius at once?’
8
The search for telic complex directed motion constructions with unprefixed and prefixed verbs
involved the following criteria:
(i) Prefixed and unprefixed verbs (‘p-’ represents any prefix):
(p-)ambulo ‘walk’; (p-)curro ‘run’; (p-)equito ‘ride’; (p-)fluo ‘flow’; (p-)gredior ‘walk, step’; (p-)labor
‘slip’; (p-)navigo ‘sail’; (p-)repo ‘crawl’; (p-)salio ‘jump’; (p-)volo ‘fly’
(ii) Telicity-signalling expressions (cf. Pinkster : ff., : ff.)
a. Adverbs
extemplo, repente, repentino, statim, subito or subitum, ‘suddenly’
b. Prepositions
intra ‘in’ (as in intra tres dies ‘in three days’)
c. Complementizers
simul ac, simul atque, ubi or ut primum, ‘as soon as’
d. Ablative forms of nouns and adjectives encoding periods of time
dies ‘day’, hora ‘hour’, nox ‘night’, mensis ‘month’, annus ‘year’, diurnus ‘of the day’, diutinus
‘lasting’, diuturnus ‘lasting’, nocturnus ‘of the night’, menstruus ‘which lasts a month’, menstrualis
‘which lasts a month’, annuus ‘which lasts a year’, annalis ‘of a year’, annualis ‘a year old’, aestas
‘summer’, hiems ‘winter’, ver ‘spring’, autumnus ‘autumn’, mane ‘morning’, vesper ‘evening’,
vesperus ‘of the evening’, calendae/kalendae ‘calends’, idus ‘ides’, nonas ‘nones’, lustrum ‘lustrum’,
meridies ‘noon’, vigilia ‘time of keeping watch by night’, hibernus ‘of the winter’, saeculum/
seculum/saeclum ‘century’, saecularis ‘of a century’, aestivus ‘of the summer’, aestivalis ‘of the
summer’, vernus ‘of the spring’, vernalis ‘of the spring’, autumnus ‘of the autumn’, autumnalis ‘of
the autumn’, horalis ‘which lasts an hour’, matutinus ‘of the morning’, postmeridianus ‘of the
afternoon’, vespertinus ‘of the evening’, spatium ‘time span’
e. Ablative form of adjectival suffixes indicating a period of time
-duus ‘of X days’, -ennius ‘of a year’, -noctius ‘of X nights’, -menstruus ‘of X months’, -menstris ‘of
X months’
Latin and Slavic complex resultatives with an internal prefix
b. Petron. ,
Statimque, ad re-ficiendum ignem in viciniam cucurrit.
at_once=and, at re-make.GER.ACC fire.ACC in neighbour’s.ACC run.PRF.SG
‘And immediately, he ran to the neighbor’s to kindle the fire.’
c. Suet. Otho ,
Ac repente omnes in palatium cucurrerunt.
and suddenly all.NOM.PL in palace.ACC run.PRF.PL
‘Then suddenly everybody hastened into the palace.’
d. Sil. ,
Subito vilis rubenti fluxit mulctra mero.
suddenly, worthless.NOM red.ABL flow.PFV.SG milk pail.NOM wine.ABL
‘Suddenly, the worthless milking pail flowed with red wine.’
e. Cic. Att. , ,
Se statim ad te navigaturum esse.
REFL.SG.ACC at_once at you.ACC sail.INF.FUT.ACC.M be.INF
‘That he was on the point of setting sail at once to join you’.
f. Cic. Fam. , , ,
Si statim navigas, nos Leucade consequere.
if at_once sail.SG us.ACC Leucas.ABL follow.FUT.SG
‘If you sail off at once, you will overtake me at Leucas.’
g. Lucr. ,
E terraque ex-orta repente arbusta
out earth.ABL=and out-rise.PTCP.PFV.NOM.N.PL suddenly tree(N)NOM.PL
salirent.
leap.IPFV.SBJV.PL
‘And trees would suddenly leap out of the turf.’
h. Enn. Ann. ,
Simul ex alto longe pulcherruma praepes
suddenly out high.ABL far beautiful.SUPERL.NOM nimble.NOM
laeva volavit avis.
left.NOM fly.PFV.SG bird.NOM
‘Suddenly there appeared to the left, in the distance, out of high heaven,
a most beautiful bird flying with good omen.’
() Latin; a sample of telic predicates headed by prefixed manner-of-motion verbs
a. Liv. , ,
Deinde subito ad arma dis-currerunt.
then suddenly at weapon.ACC.PL apart-run.PRF.PL
‘Then, suddenly, they ran in all directions for the weapons.’
Weak satellite-framed languages
b. Liv. , ,
Qui ubi ad-equitavit portis.
who.NOM.SG as_soon_as at-ride.PFV.SG doors.DAT
‘This one, as soon as he had ridden up to the gates’.
c. Val. Max. , ,
[Vires atque opes humanae]
strength.NOM.PL and wealth.NOM.PL human.NOM.PL
ad-fluunt subito, repente di-labuntur.
at-flow.PL suddenly suddenly apart-slip.PL
‘The vigour and the wealth of humans come suddenly in a flow, and
suddenly slip asunder.’
d. Liv. , ,
Triduo a-scenderat biduo est
three_days.ABL at-climb.PLUPRF.SG two_days.ABL be.SG
de-gressus.
downward-walk.PTCP.PFV.M.NOM.SG
‘He walked down in two days, though he had climbed up in three.’
e. Plin. Nat. ,
XXX dierum spatio prae-navigaverint.
day.GEN.PL span.ABL before-sail.PLUPRF.SBJV.PL
‘It took thirty days to sail past their territory.’
f. Suet. Diuus Augustus ,
Draconem repente ir-repsisse ad eam
snake(M)ACC.SG suddenly in-glide.INF.PFV at her.ACC
pauloque post e-gressum.
a_little=and after out-walk.PTCP.PFV.ACC.M.SG
‘That, suddenly, a snake glided in towards her and glided away soon after’.
g. Liv. , ,
Repente ex equis de-siliunt.
suddenly out horses.ABL down-leap.PL
‘Suddenly they leapt down from their horses.’
The many examples found where a prefixed verb is used support the observation
made in (). It can be argued that these examples represent complex resultative
constructions. As such, they admit the following analysis:
() Latin; analysis of ()c (Vires ad-fluunt “vigour at-flows”)
[vP [v v FLU] [PathP vires [Path’ [Path [PlaceP [PlaceP vires [Place’ Place AD]]]]]]]
The root AD ‘at’ is merged as Compl-Place, and is later associated morphologically
with the verb at PF. The v head is independently associated, by adjunction, with a
Latin and Slavic complex resultatives with an internal prefix
root expressing the Manner component, FLU ‘flow’, identifying a co-event. The
s-framed pattern is preserved, since the Path and the verb correspond to different
phonological realizations.
The predicates in (), on the other hand, seem to be counterexamples to ().
Interestingly, however, out of the eight predicates headed by unprefixed verbs that
are shown in (), five are headed by curro ‘run’, salio ‘jump’, and volo ‘fly’. These
verbs exhibit a special behaviour in v-framed languages like Italian or French, a
behaviour which can be, despite appearances, compared with that of the Latin
cognate verbs of (). Specifically, these verbs can head constructions that at first
glance could be taken as CDMCs, which, as we know, are not possible in v-framed
languages. First, it has been observed that Italian correre ‘run’ and French courir ‘run’
may appear in predicates of bounded directed motion, as respectively illustrated in
()a and ()b. ()a additionally shows that bounded predicates with correre ‘run’
trigger selection of BE as auxiliary for the perfect tense and must, therefore, be
considered as heading an unaccusative predicate expressing a resultative event rather
than an activity:
() Italian and French; Folli and Ramchand (: ) and Pourcel and Kopecka
(: )
a. Gianni è corso in spiaggia in/*per un secondo.
John is run.PTCP.PFV.M.SG in beach in/for one second
‘John ran to the beach in a second/*for one second.’
b. Il court dans le jardin.
he runs in the garden
‘He runs into the garden.’
An analogous scenario can be described for verbs like jump and fly. Mateu (), for
instance, shows that these verbs display unaccusative behaviour in Italian—
specifically, BE-selection in the perfect—if accompanied with a PP:
() Italian; Mateu (: )
a. Gianni è/*ha volato a Roma.
Gianni is/*has flown to Rome
‘Gianni has flown to Rome.’
b. Gianni è/*ha saltato dalla finestra.
Gianni is/*has jumped from.the window
‘Gianni has jumped from the window.’
Crucially, not all verbs in Romance behave in this way. For instance, Folli and
Ramchand () show that Italian camminare ‘walk’ and galleggiare ‘float’, are
unable to license unaccusative predicates (with BE-selection) even in the presence
of a goal PP:
Weak satellite-framed languages
() Italian; Folli and Ramchand (: )
a. *Gianni è camminato in spiaggia.
John is walk.PTCP.PFV.M.SG in beach
‘John walked to the beach.’
b. *La barca è galleggiata sotto il ponte.
the boat is float. PTCP.PFV.F.SG under the bridge.
‘The boat floated under the bridge.’
Within the neo-constructionist framework adopted here, we cannot appeal, as a first
resource, to alleged formal lexical properties of run, jump, fly as opposed to other
verbs of manner of motion such as the ones in () to explain the behaviour of either
type. Rather, the fact that both classes of verbs tend to be the same in different
languages argues for an account in terms of grammar-concept compatibility, rather
than for one in terms of idiosyncratic lexical marking. Thus, it is the case that verbs
such as run, fly, or jump accept with more ease a linguistic construal in terms of telic
change of location than other manner-of-motion verbs such as float or walk. First, it
has to be noticed that the type of motion described by these three verbs is usually
directed (though not necessarily bounded) motion, that is, it usually involves a
change of position (as opposed to float or dance, for instance). Second, at least run
and jump can be standard ways of attaining a goal of motion, as can fly in the case of
descriptions of trips to distant destinations, such as in ()a. Third, there is a
conceptual component of rapidity in the three of them, as opposed to, say, walk
(see ()a), which makes them apt to be used as change-of-location predicates, that
is, to describe a transition from one place to a different one. These conditions,
I suggest, allow run, jump, and fly to be used as change-of-location verbs. Specifically,
I claim that they head weak resultative (unaccusative) constructions, analogous to the
ones seen in section .... I illustrate this with an analysis of ()a:
() Italian; analysis of ()a
vP
v PathP
Gianni Path’
Path PlaceP
pP PlaceP
in spiaggia Gianni Place’
Place corr
Latin and Slavic complex resultatives with an internal prefix
The root CORR ‘run’ is first merged as Compl-Place, where it is interpreted as a
terminal Ground (since PlaceP is embedded within PathP), and the subject, Gianni,
is merged at Spec-Place, where it is interpreted as Figure. As for the PP in spiaggia
‘at the beach’, I take it to be an adjunct to PlaceP: it provides a specification of the
result location, much as the AP in other weak resultatives serves as a specifier of
the result state encoded by the verb. Claiming an adjunct status for the goal PP in
these constructions, where it seems to be compulsory, is at first sight contradictory.
I take this obligatoriness to be also a pragmatic effect, active only in motion
constructions, where absence of the goal produces straightforward awkwardness.
An indication that this might be on the right track are examples such as the following
from Italian, where correre ‘run’ might appear with the auxiliary BE in the absence
of a goal PP:
() Italian; Sorace (: )
Maria {ha corso/ è corsa} velocemente.
Maria has run.PTCP.PFV.M.SG/ is run.PTCP.PFV.F.SG fast
‘Maria ran fast.’
Note that an analysis of these predicates as weak resultatives is in conformity with the
fact that the languages in which they are found are v-framed and, hence, do not
feature strong resultative constructions. These weak resultatives are licensed precisely
because they do not involve a root adjoined to v.9
I now return to the Latin data in (). A neo-constructionist analysis of
the Romance data considered so far, whereby the construal of run-verbs as
change-of-state verbs hangs on a compatibility of the conceptual dimension
of these roots with a resultative PathP structure, and not on lexical marking of
the roots/verbs within independent lexicons, predicts that roots with similar
conceptual content in other languages will also be amenable to the same con-
strual. I argue that this is what happens with the Latin verbs in (). Thus, a
predicate such as ()c would also receive a weak resultative analysis, within this
view, with the accusative-marked PP in palatium ‘into the palace’ as an adjunct
to Path:
9
The analysis is inspired by a similar one in Mateu (). The difference between Mateu’s ()
analysis and mine is the treatment of the PP that typically appears in these Romance constructions of directed
motion. While I take it to be an adjunct to PlaceP, Mateu, assuming Late Insertion of roots, proposes that it is
the phonological realization of PlaceP (his Ploc projection), after the root has been incorporated into v (see
Haugen ). See Zubizarreta and Oh (), Den Dikken (), Gehrke (), Ramchand (), and
Real Puigdollers () for other analyses of similar phenomena in different languages.
Weak satellite-framed languages
() Analysis of ()c
vP
v PathP
pP PathP
p palatium omnes Path’
p in Path PlaceP
omnes Place
Place curr
In this light, the cases of unprefixed predicates of directed motion with directional
DPs presented in section .., can be provided with the same analysis, on the basis of
the fact that the verbs licensing the directional DP (eo ‘go’, venio ‘come’) also show a
strongly directional character:
() Latin; Cic. Verr. Actio secunda, ,
Veniunt Syracusas.
come.PL Syracuse.ACC
‘They come to Syracuse.’
() Latin; analysis of ()
vP
v PathP
pP PathP
p Syracusas pro Path’
Path PlaceP
pro Place’
Place ven
In this analysis the verbal root VEN ‘come’ is merged as Compl-Place, identifying the
final location of the motion event, a deictic corresponding to the speaker. As is to be
expected, venio ‘come’, and also eo ‘go’, can also be found without a goal DP or PP:
() Latin; Caes. Gall. ,
Venisse tempus victoriae demonstrat.
come.INF.PRF time.NOM victory.GEN show.SG
‘He makes (them) see that the time of victory has come.’
Latin and Slavic complex resultatives with an internal prefix
() Latin; Verg. Aen. ,
Per artus sudor iit.
along limbs.ACC sweat.NOM go.SG
‘Sweat spreads over his limbs.’
Last, I point out that there is evidence that the weak resultative analysis of simple
directional verbs can be extended to the Slavic languages. For instance, the counter-
part of jump in these languages may head change-of-location predicates without a
prefix, as the following Czech example illustrates:
() Czech; Filip (: ), in Gehrke (: )
a. Skočil metr od okna.
jumped metre.ACC from window.GEN
‘He jumped a metre away from the window.’
b. Skočil metr k oknu.
jumped metre.ACC to window.DAT
‘He jumped a metre to the window.’
If the analysis presented here for most of the examples in () is on the right track,
they cannot be taken as counterexamples to the observation in (), namely, that
complex resultative constructions are always prefixed in Latin. The reason is that,
under this analysis, the alleged counterexamples turn out to be weak resultative
constructions, and, hence, simple resultative constructions (see sections ... and
...). See section ... for discussion of other cases of transition predicates not
involving an internal prefix.
I shall assume, then, that the expression of complex resultative predicates in Latin
involves the univerbation of the verb and the result predicate, in the form of a prefix.
I note here that Van der Heyde’s (: ) view on Latin preverbation, far from the
outer-aspectual theories of Meillet’s (), Barbelenet’s (), or Hofmann and
Szantyr’s (: ff.), comes close to the present one in that he considers preverbs
to be resultativity markers.10 In particular, within the realm of verbs of movement,
he claims that prefixed verbs of movement such as ad-eo “at-go” ‘go to, address’ or
af-fero “at-bring” ‘bring to’, with some rare exceptions, are never used in a context
representing the referent of the subject as attempting to arrive at a local goal,
signalled by the preverb. In this case, it is the construction with the preposition
that is used (that is, eo ad ‘go towards’ fero ad ‘bring towards’). The presence of the
prefix entails, then, that the goal has been reached. Observe, however, that the
hypothesis I am working with here is stronger than Van der Heyde’s (), in
10
See also Haverling (, a, b, , ), who claims that the prefixes provide bounds for
the event, and Romagno (, ), who considers them to be telicizers.
Weak satellite-framed languages
that it implies that all complex resultativity in Latin must involve prefixation. I also
do not support Van der Heyde’s (: ) claim that prefixes only attach to
‘determined’ verbs like venio ‘come’, highlighting the goal of motion inherent to
these verbs. It is evident from the discussion in this section and elsewhere in this
work, that preverbs may readily attach to verbs like blandior ‘flatter’ or equito ‘ride’,
which, in the absence of the preverb, do not imply a goal.
.. Slavic complex resultative constructions
Let us turn now to Slavic. First of all, Slavic complex motion predicates are similar to
Latin analogous predicates in that they typically feature a prefix encoding the final
location of the motion path, as shown in ():
() Slavic complex directed motion predicates bear a prefix
a. Bulgarian; Padrosa-Trias and Markova (: )
Do-bjagah do bolnitsata.
to-run.PST.SG to hospital.the
‘I ran to the hospital.’
b. Czech; Filip (: )
Při-nesl ze sklepa uhlí.
to-carry.PST.SG from.PREP basement.GEN coal
‘He brought (some) coal from the basement.’
c. Polish; Lindvall (), in Svenonius (: )
Dzieci w-skoczyly do wody.
children in-jump.PST.SG to water
‘The children jumped into the water.’
d. Russian; Babko-Malaya (: )
Ivan vy-kopal klad.
Ivan out-dig.PST.SG treasure.ACC
‘Ivan dug out the treasure.’
e. Serbo-Croatian; Arsenijević (: )
Jovan je od-gurao prikolicu od prskalice za vodu.
Jovan AUX from_at-push.PST.SG trailer from sprinkler for water
‘Jovan pushed the trailer away from the sprinkler.’
Specifically within the realm of CDMCs, the role of the prefix in triggering the
resultative bounded interpretation of the predicate is evident, as the following Polish
examples show. Note, crucially, the difference between the directional but
unbounded reading of the (b), unprefixed, examples as opposed to the directional
and bounded interpretation of the (c), prefixed, examples:
Latin and Slavic complex resultatives with an internal prefix
() Polish; Kopecka (: )
a. Anna biegła.
Anna.NOM run.PST
‘Anne has run.’
b. Anna biegła do szkoły.
Anna.NOM run.PST to school.GEN
‘Anna has run towards the school.’
c. Anna w-biegła do szkoły.
Anna.NOM in-run.PST to school.GEN
‘Anna has run into the school.’
() Polish; Kopecka (: )
a. Paweł płynął.
Paweł.NOM swim.PST
‘Paweł has swum.’
b. Paweł płynął do brzegu.
Paweł.NOM swim.PST to riverbank.GEN
‘Paweł has swum towards the riverbank.’
c. Paweł do-płynął do brzegu.
Paweł.NOM to-swim.PST to riverbank.
‘Paweł has swum to the riverbank.’
A number of researchers have pointed out that the prefixation in this type of
predicate must in fact be considered a general rule. For Russian, Rojina (: )
observes that prefixation is obligatory:
() Russian; Rojina (: )
*(Vy-)brosit’ kota iz okna.
out-throw.INF cat from window
‘To throw the cat out of the window.’
The same situation is found in Bulgarian (Angelina Markova, p. c.):
() Bulgarian
a. *(S)-nesoh topkata ot durvoto
down-carry.PST.SG ball.the from tree.the
‘I carried down the ball from the tree.’
b. *(Iz)-kopah sukrovishte (iz dupkata)
out-dig.PST.SG treasure.the out hole.the
‘I dug a treasure out (of the hole).’
c. *(V)-karah kolata (v garaja).
in-drive.PST.SG car.the in garage.the
‘I drove the car into the garage.’
Weak satellite-framed languages
According to Svenonius () and Gehrke (), the prefixation requirement is
general in Slavic, as stated in () to ():
() Svenonius (: )
‘Selected PPs often occur with prefixes (see Rojina for extensive discus-
sion and examples), in fact they are often obligatory.’
() Gehrke (: )
‘[T]here seems to be some morphological requirement to express resultativity
on the verb in these languages. Indicative of this approach is that these Slavic
languages lack adjectival resultatives of the English type (e.g. hammer the
metal flat) but generally have to use accomplishment/achievement verbs
(that are additionally marked for resultativity by an internal prefix) and/or
PPs in such constructions.’
() Gehrke (: , footnote )
‘In fact, it seems like resultativity is always expressed morphologically by an
internal prefix on the verb.’
It has to be clear that, as was the case for Latin, we are dealing here with true complex
resultative constructions, since the prefix may induce telicity. This has been pointed
out by Arsenijević (), Gehrke (), Ramchand (), and Svenonius (),
among many others before, and is illustrated in () for Russian by the test of
delimiting/durative adverbials:
() Russian; Gehrke (: )
a. On pri-exal v Moskvu *(za) den’.
he to-drove.PST in Moscow.ACC in day
‘He arrived in Moscow in/*for a day.’
b. On ot-kryl okno *(za) dve minuty.
he from-cover.PST window.ACC in two minutes
‘He opened the window in/*for two minutes.’
That Slavic predicates headed by verbs featuring goal or source prefixes are telic is
argued for at length by Gehrke (). This author proposes that directional predi-
cates are headed by prefixes originating as independent heads within the vP domain.
A possible counterexample to the tight relation between internal prefixation and
telicity is the fact that certain predicates containing a bounded path may license
durative adverbials, as the example ()b from Polish illustrates:
() Polish; Wojciech Lewandowski (p. c.)
a. Jan w-biegl do pokoju w/ *przez sekundy.
Jan in-run to room in for seconds
‘Jan ran into the room in seconds.’
Latin and Slavic complex resultatives with an internal prefix
b. Jan w-biegal do pokoju przez/ *w sekundy.
Jan in-run.SI to room for in seconds
‘Jan ran into the room for seconds.’
In ()b, in spite of the prefix w- ‘in’ and the consequent bounded reading of the path,
the predicate allows the presence of a for-adverbial, a usual signal of atelicity. However,
the durative adverbial PP przez sekundy ‘for seconds’ here does not signal
atelicity, nor does it measure the duration of an event of Jan running into the room,
but the duration of a collection of several identical and, crucially, telic events: a
Sequence of Similar Events Interpretation (see section ...). This effect is due to
the fact that w-biegal “in-ran” is a Secondary Imperfective (SI). The SI is a case of outer
aspect (see Ramchand ; Borer a, b; Gehrke ), and hence orthogonal
to the telic/atelic distinction. It may be interpreted as a collection of events, in this case,
of different telic events. Ramchand (: ), for instance, suggests that the Second-
ary Imperfective may introduce ‘a superevent consisting of habitual repetitions of e’,
where e refers to the event introduced at the VP level. A similar contrast is observed in
the following Serbo-Croatian examples extracted from Arsenijević (), where his
English translations confirm the outer-aspectual nature of the Secondary Imperfective:
() Serbo-Croatian; Arsenijević (: )
a. Jovan je od-gur-ao kolica.
Jovan AUX away-push-PTCP cart
‘Jovan pushed the cart away.’
b. Jovan je od-gur-av-ao kolica.
Jovan AUX away-push-SI-PTCP cart
‘Jovan was pushing the cart away.’
Arsenijević clarifies the two readings of the verb simultaneously featuring the spatial
prefix and the SI suffix:
() Arsenijević (: )
[I]t is also possible to translate (c) [my ()b] as ‘Jovan pushed the cart
away’, but allowing only for the iterative reading of this translation. The lack of a
singular telic reading for the S-C example makes this translation incomplete.
The translation that is provided in the example, which uses the English present
continuous form, has both the readings of the S-C sentence. The readings are a)
that there is a singular eventuality of pushing the cart away in a progressive
interpretation (i.e. only its initiating subevent is really entailed), and b) that there
is an unbounded set of iterations of a full telic eventuality (bare plural reading).
Importantly, in each of the two readings the telicity of the inner predicate (the
predicate interpreted as ‘putting away a cart through pushing’) is preserved, since
the SI is an instance of outer aspect, located outside the vP.
Weak satellite-framed languages
In sum, there is evidence that Slavic may indeed share with Latin the morpho-
logical requirement that I assumed in ().
.. The relation between resultativity and prefixation
In this section I present some observations on the relation between resultativity and
prefixation in both Latin and Slavic. First, I point out that the present system allows for
resultativity to emerge in the absence of an internal prefix, and that it is only for the
expression of complex resultativity that the prefix seems required. Second, I explore a
prediction made on the assumption that prefixation is effected as a Raising operation and
is, therefore, sensitive to syntactic structure, that is, that predicates expressing non-
resultative (i.e. unbounded) directionality do not bear an internal prefix. Third, I discuss
the fact that the use of internal prefixes seems to be much more widespread in Slavic than
in Latin, relating it to a difference in the expression of viewpoint aspect in both languages.
... Resultativity without an internal prefix The empirical observation made in
(), repeated below, seems to be adequate for Latin and Slavic. In particular, I claim
that any complex resultative construction in these languages has to be prefixed, even
if provided with a directional PP.
() Neither Latin nor Slavic, both disallowing AP resultatives, seems to allow the
expression of complex PP resultatives without a result-conveying prefix
attached to the verb.
In these constructions the verb embodies the event leading to the resulting state
expressed by the prefix. Furthermore, these predicates have been shown to express a
final result, both in Latin and Slavic. Crucially, though, the relation between resulta-
tivity and prefixation is not bidirectional: predicates headed by (internally) prefixed
verbs are resultative and may thus be telic, but not all resultative predicates contain a
prefixed verb, as also pointed out by Carvalho () for Latin and Gehrke ()
and Arsenijević () for Slavic. In other words, (internal) prefixation is a sufficient,
but not necessary, condition for resultativity. Specifically, resultative predicates not
representing a complex resultative construction, such as those based on the Latin
verbs capio ‘take, conquer’ or neco ‘kill’ are not necessarily prefixed:
() Latin; Bell. Afr. ,
Cirtamque oppidum [ . . . ] paucis diebus [ . . . ] capit.
Cirta.ACC=and town.ACC few.ABL.PL day.ABL.PL take.SG
‘And he conquers the town of Cirta in a few days.’
() Latin; Plin. Nat. , ,
Gelatio [ . . . ] paucis diebus necat.
frost.NOM few.ABL.PL day.ABL.PL kill.SG
‘The frost kills them [the trees] in a few days.’
Latin and Slavic complex resultatives with an internal prefix
The predicates above receive an analysis in which it is the root of the verb that is
merged as Compl-Place, identifying the final state of the event. Since, as we have
argued, this is also the position in which the prefixal root is first merged, a
natural explanation arises for the lack of obligatory prefixation in the above
predicates:
() Analysis of ()
VoiceP
Gelatio Voice’
Voice vP
v PathP
arbores Path’
Path PlaceP
arbores Place’
Place nec
The PF derivation of these predicates is straightforward. As was first proposed
in section .., Path is specified as being linearly adjacent to Place and v in Latin.
This triggers Raising of all the heads to v, which, in the cases at hand, has no root
adjoined to it:
() PF-derivation of ()
a. Structure delivered by syntax
[VoiceP [Voice’ Voice [vP v [PathP [Path’ Path [PlaceP [Place’ Place NEC]]]]]]]
b. Raising
[v [Path [Place NEC Place] Path] v]
c. Linearization
NEC-Place-Path-v
d. Vocabulary Insertion
nec-∅-∅-∅
In section .. the analysis in () was applied to telic directed motion constructions
based on verbs such as Latin curro ‘run’ () or venio ‘come’ () or Czech skočit
‘jump’ (). In this section I extend it to unprefixed change-of-location (COL)
alternants of the Locative Alternation and give-verbs.
In section ..., I proposed that COL alternants of the Locative Alternation
featuring an unprefixed verb like spargo ‘scatter’ involve an s-framed pattern, the
Weak satellite-framed languages
verbal root being merged as an adjunct to v and the directional PP being merged
directly as PlaceP:
() Cato, Agr.
Stercus columbinum spargere oportet
manure(N)ACC of_pigeon.ACC.N scatter.INF be_necessary.SG
in pratum.
in meadow.ACC
‘Pigeon manure must be scattered onto the meadow.’
() Analysis of ()
VoiceP
PRO Voice’
Voice vP
v PathP
v sparg stercus Path’
Path PlaceP
stercus Place’
Place pratum
Place in
In an alternative analysis that explains the absence of the prefix in these constructions
the verbal root is first merged as Compl-Place and the directional PP is an adjunct to
either PathP (when the embedded DP is marked with accusative) or PlaceP (when
the embedded DP is marked with ablative):
() Analysis of ()
VoiceP
PRO Voice’
Voice vP
v PathP
pP PathP
p pratum stercus Path’
p in Path PlaceP
stercus Place’
Place sparg
Latin and Slavic complex resultatives with an internal prefix
The rationale for the above analysis, in which the verbal root occupies the position of
Compl-Place is the same as that adopted in the case of directed motion constructions
with unprefixed verbs like curro ‘run’ in section ... Roots like SPARG ‘scatter’ (and
also LIN ‘spread’ or STIP ‘cram’) are conceptually compatible with a syntactic-semantic
construal in terms of change of state of the thing scattered. A root like SPARG ‘scatter’
is expected, therefore, to be usable in the absence of any directional PP, as is the case:
() Latin; Plin. Nat. , ,
Densum spargi semen.
dense.NOM.N scatter.INF.PASS seed.NOM
‘The seed is to be scattered thick.’
By contrast, a strongly manner unprefixed verb like flo ‘blow’ does not license the
Locative Alternation, as shown in the entry for this verb provided by Gaffiot ()
and Lewis and Short (), since it is difficult to integrate its meaning as the final
state in a change-of-location event (the COL alternant). Its root FL ‘blow’ can,
however, be used as a Manner Co-event in COL alternants involving prefixed ad-flo
“at-blow” ‘blow upon’, which also license a COS alternant (Lemaire : ):
() Latin; Locative Alternation with ad-flo “at-blow” ‘blow upon’
a. Lucr ,
Calidum membris *(ad-)flare vaporem.
warm.ACC.M limb.DAT.PL at-blow.INF vapor(M)ACC
‘To blow a warm vapour upon our limbs’.
b. Verg. Aen. ,
[Sibylla] *(ad-)flata est numine [ . . . ] dei.
Sibylla at-blow.PTCP.NOM.F.SG is power.ABL god.GEN
‘The sibyl has been breathed upon with god’s power.’
The fact that COL alternants involving the root FL ‘blow’ must bear a prefix and are
not licensed by a mere directional PP is indicative of the fact that the prefix is really
structuring the part of the configuration expressing the result.
Another interesting case of resultativity without internal prefixation is that of give-
verbs. As the following examples from Latin and Russian show, give licenses a telic
interpretation with a quantity object:
() Latin (Caes. Gall. , ) and Russian (Gehrke : )
a. Partem statim dederunt, partem [ . . . ] paucis diebus
part.ACC.SG at_once give.PFV.PL part.ACC.SG few.ABL.PL day.ABL.PL
sese daturos dixerunt.
REFL.SG.ACC give.INF.FUT.ACC.M.PL say.PFV.PL
‘These gave part of it at once, and said they would give the rest in a few
days.’
Weak satellite-framed languages
b. On dal ženščine knigu *(za) dve minuty.
he gave woman.DAT book.ACC in two minutes
‘He gave a/the woman a/the book in/*for two minutes.’
I present an analysis of Latin predicates involving do ‘give’ and a dative argument,
based on the desire to provide a unified analysis of dative arguments in Latin. As was
shown in section .., the so-called directional dative can be analysed as merging in
an applicative projection above PathP and taking scope over PlaceP, which is
interpreted as possessed by the referent of the dative. I propose that in predicates
involving the verb do ‘give’ and a dative argument, the verbal root D ‘give’ is merged
as Compl-Place, expressing, in this case, possession:
() Latin; Plaut. Bacch.
Decumam partem ei dedit.
tenth.ACC.F.SG part(F)ACC him.DAT give.PRF.SG
‘He gave him a tenth.’
() Analysis of ()
VoiceP
pro Voice’
Voice vP
v ApplP
ei Appl’
Appl PathP
decumam partem Path’
Path PlaceP
decumam partem Place’
Place d
A paraphrase for the above example would be ‘cause a tenth to go to his possession’.
In any case, it is clear why we do not find a prefix in predicates of pure transfer: the
Compl-Place position is already occupied by the verbal root. In turn, this analysis, in
which the dative is an ‘added argument’ (Pylkkänen ) easily explains why Latin
do ‘give’ is very frequently encountered without the dative, as shown in ()a. This
analysis of Latin do ‘give’ and its Slavic counterparts is compatible with the fact that
these verbs may sometimes appear with prefixes, as shown by non-transfer verbs like
Russian iz-dat’ “out-give”, ‘publish’ (cf. its exact Latin counterpart e-do):
Latin and Slavic complex resultatives with an internal prefix
() Russian; Gehrke (: )
Iz-dat’ (*komu) čto.
out-give.INF who.DAT what.ACC
‘To publish something (*to someone)’.
For these cases, however, I assume that the verb involves an adjunction of the root to v,
as a Co-event:
() Analysis of ()
[VoiceP PRO [Voice’ Voice [vP [v DA] [PathP čto [Path’ Path [PlaceP čto [Place’ Place
IZ]]]]]]]
The resulting combination of the prefix and the verb bears an idiosyncratic meaning,
‘publish’, which I take to be registered in the Encyclopaedia, in the entries for DA ‘give’
and IZ ‘out’, and triggered in the environment depicted in (), within the local
domain represented by vP.
In sum, the above data can be taken as supportive of the view that resultativity
depends, at least in change-of-state predicates, on the projection of specific structure
(PathP), and not on the presence of specific elements, the prefixes, which here are
assumed to contribute only conceptual content. It is only in the case of complex
resultativity that the presence of the prefix seems required, both in Latin and Slavic,
since the verbal root is associated directly to v and the position encoding the final
result/location can be occupied by the prefixal root. See section .. for an analysis of
yet another kind of unprefixed resultative predicate in Latin: AP resultatives based on
unprefixed verbs like ago ‘lead’ and facio ‘do’.
... Atelic predicates and prefixation The analysis proposed in this work, where
prefixation in Latin and Slavic is taken to result from an application of Raising of
Path to v, yields interesting predictions as to the shape of some atelic predicates with
a directional PP. Crucially, note that prefixation of Path to v is to be understood in
structural terms: v forms one word with the head of its sister PathP. Positing a Path-
to-v Raising operation aims at capturing this fact, since Raising, as an operation
previous to Vocabulary Insertion, is sensitive to structure, and not to linear adja-
cency. As a result, we expect no morphological relation to be effected between v and
any p head merged outside vP, that is, as an adjunct to vP, as shown in the following
diagram (where EA stands for external argument):
() VoiceP
EA Voice’
Voice vP
pP vP
Weak satellite-framed languages
Crucially, on the semantic side, a vP-external p cannot induce telicity, since it
cannot act as a probe to raise a quantity DP to Spec-Path, where it would be
interpreted as Measurer (see section ...). Therefore, we expect configurations
such as the one above, featuring an unergative structure with an adjoined pP, to
reflect both lack of prefixation and lack of telicity, much as the p may be interpreted
as directional. Note, importantly, that prefixation is to be understood here as internal
prefixation, that is, as the affixation onto the verb of phonological material coming
from PathP. External prefixation is expected and attested in atelic predicates, since it
involves vP-external material and, hence, does not signal the presence of a vP-
internal PathP (see section ..).
Direct attestation of the prediction just made is the fact that, according to Gehrke
(), in Russian and Czech the only prepositions that do not have a prefixal
correspondence are those representing an unbounded path, that is, one incompatible
with telicity: k ‘toward’, in Russian, and k, vůči ‘toward’, in Czech. Similarly, in
Bulgarian, directional predicates with an unbounded directional PP and an atelic
reading feature an unprefixed verb (see ()a) contrasting with directional predicates
with a bounded directional PP (see ()b) (Angelina Markova, p.c.):
() Bulgarian
a. Toj pulzi kum vratata.
he crawls towards door.the
‘He crawls towards the door.’
b. Toj do-pulzia do vratata.
he to-crawled to door.the
‘He crawled up to the door.’
For Latin, the prediction is in consonance with an observation by Van der Heyde
(: –) that unprefixed motion verbs with (accusative) PPs are not resultative,
in that the goal is not entailed to be reached. To ascertain the empirical validity of this
observation I performed a search involving prefixed and unprefixed manner-of-
motion verbs and expressions signalling atelicity.11 In a subcorpus of sentences
11
The components for the search are as follows:
(i) Prefixed and unprefixed verbs (‘p-’ represents any prefix)
(p-)ambulo ‘walk’, (p-)curro ‘run’, (p-)equito ‘ride’, (p-)fluo ‘flow’, (p-)gredior ‘walk, step’, (p‑)labor
‘slip’, (p-)navigo ‘sail’, (p-)repo ‘crawl’, (p-)salio ‘jump’, (p-)volo ‘fly’
(ii) Atelicity-signalling expressions (cf. Vester : ff.; Torrego ; Pinkster : ff.; : ff.)
a. Adverbs
diu ‘for a long time’, diutule ‘for a little while’, paulisper ‘for a while’
b. Prepositions
per + quantified period of time ‘for’
c. Accusative forms of nouns and adjectives encoding periods of time
dies ‘day’, hora ‘hour’, nox ‘night’, mensis ‘month’, annus ‘year’, diurnus ‘of the day’, diutinus
‘lasting’, diuturnus ‘lasting’, nocturnus ‘of the night’, menstruus ‘which lasts a month’, menstrualis
Latin and Slavic complex resultatives with an internal prefix
containing each a motion verb, either prefixed or not, and an atelic adverbial
expression, ten sentences, shown in (), feature an unprefixed verb; in all these
sentences the durative expression is understood as temporally bounding the other-
wise unbounded (motion) activity expressed by the unprefixed verb:
() Latin examples with unprefixed manner-of-motion verbs and atelicity markers
a. Cels. ,
Tum diu ambulandum.
then for_long walk.PTCP.FUT.PASS.NOM.N.SG
‘Then one must walk for a long time.’
b. Plin. Epist. , ,
Diu iacui vel ambulavi.
for_long lie.PRF.SG or walk.PRF.SG
‘I have lain in bed or walked for a long time.’
c. Apul. Flor.
Ambulant diutule.
walk.PL for_a_while
‘They walk for a while.’
d. Ov. Am. , ,
Diu lacrimae fluxere per ora.
For_long tear.PL flow.PRF.PL through face.ACC
‘Tears flowed down her face for a long time.’
e. Liv. , ,
Nuntiatum est [ . . . ] sanguinis riuos
report.PTCP.PFV.NOM.N.SG is blood.GEN river.ACC.PL
per diem totum fluxisse.
through day.ACC whole.ACC flow.INF.PFV
‘It was reported that rivers of blood had flowed for a whole day.’
‘which lasts a month’, annuus ‘which lasts a year’, annalis ‘of a year’, annualis ‘a year old’,
aestas ‘summer’, hiems ‘winter’, ver ‘spring’, autumnus ‘autumn’, mane ‘morning’, vesper
‘evening’, vesperus ‘of the evening’, calendae/kalendae ‘calends’, idus ‘ides’, nonas ‘nones’,
lustrum ‘lustrum’, meridies ‘noon’, vigilia ‘time of keeping watch by night’, hibernus ‘of the
winter’, saeculum/seculum/saeclum ‘century’, saecularis ‘of a century’, aestivus ‘of the summer’,
aestivalis ‘of the summer’, vernus ‘of the spring’, vernalis ‘of the spring’, autumnus ‘of the
autumn’, autumnalis ‘of the autumn’, horalis ‘which lasts an hour’, matutinus ‘of the morning’,
postmeridianus ‘of the afternoon’, vespertinus ‘of the evening’, spatium ‘time span’
d. Adjectival suffixes indicating a period of time
-duus ‘of X days’, -ennius ‘of a year’, -noctius ‘of X nights’, -menstruus ‘of X months’, -menstris
‘of X months’
Weak satellite-framed languages
f. Sen. Nat. , ,
Ventus per multos dies fluxit.
wind.NOM.M.SG through many.ACC day.ACC.PL flow.PRF.SG
‘A stream of air flew for many days.’
g. Hyg. Fab. ,
Totumque diem nauigassent.
whole.ACC.SG=and day.ACC.SG sail.PLUPRF.SBJV.PL
‘That they had sailed the whole day’.
h. Ps. Quint. Decl. ,
Te iuvet diu navigare.
you.ACC please.SG for_long sail.INF
‘You like sailing for a long time.’
i. Plin. Nat. ,
Diuque ita navigatum est.
for_long=and thus sail.INF.PFV.PASS
‘And they have sailed thus for a long time.’
j. Ov. Met. ,
Diuque inter utrumque volat dubiis
for_long=and between either.ACC fly.SG uncertain.ABL.PL
Victoria pennis.
Victory.NOM wing.ABL.PL
‘And between both Victory flies for a long time with uncertain wings.’
The number of sentences featuring a prefixed manner-of-motion verb and a durative
adverbial is , shown in () to (). Nevertheless, there is reason to believe that
my hypothesis can be maintained, as I will now show. A particular set of sentences,
all from the same work, involve static descriptions of the trajectory of rivers:
() Latin; static description of rivers in present tense
a. Mela ,
Baetis [ . . . ] diu sicut nascitur uno amne de-currit.
Baetis.NOM for_long as be_born.SG one.ABL riverbed.ABL down-run.SG
‘The Baetis flows for a long time on one bed only and just as it is at its
origin.’
b. Mela ,
[Oxos] [ . . . ] aliquamdiu ad occasum ab oriente
Oxus.NOM for_a_while at west.ACC off east.ABL
oc-currens iuxta Dahas primum in-flectitur.
against-run.PTCP.PRS.NOM.M.SG beside Dahas.ACC first in-bend.SG
‘The Oxus, flowing for a while from east to west, bends first at Dahas.’
Latin and Slavic complex resultatives with an internal prefix
c. Mela ,
[Hypanis] [ . . . ] diu qualis natus est de-fluit.
Hypanis.NOM for_long how born.NOM.M.SG is down-flow.SG
‘The Hypanis flows down as it is in its spring for a long time.’
d. Mela, ,
Cyrus et Cambyses [ . . . ] [per] Hiberas et Hyrcanos
Cyrus.NOM and Cambyses.NOM through Iberian.ACC.PL and Hyrcanian.ACC.PL
diu [ . . . ] de-fluunt.
for_long down-flow.PL
‘The Cyrus and the Cambyses flow down through the lands of the Iberians and
Hyrcanians for a long time.’
In section ... we already saw cases like these, which are also pointed out by Van
der Heyde (: ) as problematic for an analysis of the Latin preverbs as
resultativity markers. I assume that these predicates display so-called fictive motion
(Jackendoff : ; Talmy ). They exploit the linguistic expression of motion
but they are interpretable as involving no motion at all:
() Talmy (: )
This fence goes from the plateau to the valley.
The pragmatically acceptable interpretation of () is one in which the fence is not
understood as undergoing a spatial transition from the plateau to the valley; rather, it
is understood to be as long as the space encompassed between the plateau and the
valley. As a result, the addition of an in-adverbial to () sounds odd, since it forces
the pragmatically aberrant interpretation whereby the fence is an entity actually
setting off from the plateau and arriving at the valley in a given amount of time:
() This fence goes from the plateau to the valley (#in an hour).
I understand that the Latin cases in () behave in the same way: the river, as a whole
entity, is not entailed to undergo a spatial transition. This licenses the atelic reading
signalled by the durative adverbials.
Another possible counterexample resulting from the search is ():
() Latin; Cat. Agr. ,
De-ambuletque horas IIII.
down-walk.SBJV.SG=and hour.ACC.PL four
‘He is to walk about for four hours.’
Note, first, that de- does not contribute here any spatial meaning, so the predicate is
not interpreted as ‘walk down’. Here I would like to suggest that de- is behaving as an
external prefix, licensing a quantification of the activity event, ‘to walk for an amount
of time’, much as does po- in the next Russian example:
Weak satellite-framed languages
() Russian; Gehrke (: )
On po-spalP (*za) dve minuty.
he po-slept in two minutes
‘He slept for/*in two minutes.’
I note that García Hernández (: ) considers deambulo ‘walk about’ to be a
bleached evolution of a former deambulo in which de- had an intensifying meaning,
recognizable in deamo ‘love passionately’ (cf. amo ‘love’), defatigo ‘exhaust’ (cf. fatigo
‘tire’), deposco ‘persistently ask for’ (cf. posco ‘ask for’). All these are optimal candi-
dates for an analysis in terms of external prefixation, since they bear an adverbial
meaning and they do not induce argument structure changes.
The availability of the rest of the examples must receive other explanations. First,
for the next uncontroversially directional motion predicates, I assume an atelic
reading is possible because of the fact that the Figure (pituita ‘rheum’, umor ‘liquid’)
is interpreted as a mass, and hence, non-quantity DP:12
() Latin; a mass noun as Measurer
a. Cels. ,
Tumore jam finito, diu lacrima
swelling.ABL already subside.PTCP.PFV.ABL for_long flow_of_tear.NOM
cum pituita pro-fluit.
with rheum.ABL forth-flow.SG
‘Even after the swelling has subsided, there continues for some time a flow
of tears mixed with rheum.’
b. Plin. Nat. , ,
Larici et magis abieti suc-cisis umor
larch.DAT and more fir.DAT under-cut.DAT.PL liquid.NOM
diu de-fluit.
for_long down-flow.SG
‘From the larch and still more the fir, after they have been cut into, liquid
flows down for a long time.’
When the non-quantity DP Figure is raised to Spec-Path, it is unable to yield a telic
reading.
The next example also expresses a directional movement with a bounded path.
However, I argue that the durative adverbial paulisper ‘for a while’, does not—in fact,
12
These and other atelic predicates endowed with an internal prefix are problematic for Romagno
(, ), who establishes an explicit link between the telicizing effect of the preverb and its power of
supporting both transitive and unaccusative UOCs. In particular, telicization of the predicate by the
preverb requires the presence of an internal argument, which emerges either as direct object or, in the
case of unaccusatives, as sentential subject. See Romagno () for the same analysis applied to
preverbation in Ancient Greek.
Latin and Slavic complex resultatives with an internal prefix
cannot—measure the temporal extent of the motion event, but, on the contrary, the
temporal extent of the resulting state, in this case the state of having descended from
the chariots:
() Latin; Gell. , ,
Quaeso [ . . . ] de-grediare paulisper curriculis
pray.SG down-step.IMP.SG for_a_little_while chariot.ABL.PL
istis disputationum vestrarum academicis.
those.ABL argumentation.GEN.PL your.GEN.PL academic.ABL.PL
‘Please descend for a while from those academic chariots of your
argumentation.’
Finally, we find cases where the telic event encoded by the prefixed verb is interpreted
as iterated because of the presence of the durative adverbial. Thus, in ()a we are
forced to understand that an event of sailing off into the sea (e-navigat “out-sails”) is
repeated identically for some successive days (per aliquot dies continuos); in the
same way, the events of leaping at someone (ad-silient “at-jump”) or flying
around someone (circum-volant “around-fly”) of ()b and ()c, respectively,
are repeated ‘day and night’ (noctesque diesque, dies noctesque):
() Latin; Sequence of Identical Events Interpretation
a. Plin. Nat. ,
Eademque hora per aliquot dies continuos
same.ABL=and hour.ABL through some day.ACC.PL successive.ACC.PL
piscator e-navigat certo spatio escamque pro-icit.
fisher.NOM out-sail.SG certain.ABL space.ABL bait=and forth-throw.SG
‘And at the same time for several successive days a fisher sets sail a pretty
way into the sea, and casts forth a bait.’
b. Stat. Theb. ,
Te volantes quinquaginta animae circum
you.ACC fly.PTCP.PRS.NOM.PL fifty spirit.NOM.PL around
noctesque diesque ad-silient.
night.ACC.PL=and day.ACC.PL=and at-leap.FUT.PL
‘The fifty spirits flying around you will leap at you day and night.’
c. Quint. Decl. ,
‘Dies’ inquit ‘noctesque miseranda patris
day.ACC.PL said night.ACC.PL=and pitiable.NOM.F father.GEN
umbra circum-volat’.
shadow(F)NOM around-fly.SG
‘ “Day and night,” he said, “does the father’s pitiable shadow fly around.” ’
Weak satellite-framed languages
In conclusion, there are reasons to think that the prediction made at the beginning of
this section is borne out for both Slavic and Latin: atelic predicates expressing a
directional but non-resultative, i.e. unbounded, motion are not internally prefixed.
Even if they sport some PP expressing the direction of the motion, since this PP is not
a sister to v, it cannot provide the material for the Raising-to-v operation to take
place, which gives rise to the prefixation effect.13
... A contrast between Latin and Slavic. The role of viewpoint aspect Notwith-
standing the unidirectional relation between prefixation and resultativity argued for
until now, it is only fair to acknowledge that telic predicates in Slavic languages
almost always bear an internal prefix, except those headed by a few verbs such as
Czech skočit ‘jump’ (see ()) or Russian dat’ ‘give’ (see ()b).14 There is controversy
about whether the simple imperfective forms of so-called incremental verbs like
Russian pisat’ ‘write’ or čitat’ ‘read’ may allow a telic interpretation, depending on
contextual factors, notably, when accompanied by an Incremental Theme. Berit
Gehrke (: , footnote ) contends that they may—see also Filip (:
). Crucially, however, the addition of the prefix cancels any atelic interpretation,
as pointed out by Gehrke (: ), in a way similar to the contrast involved in
pairs such as write/write down, read/read through and eat/eat up in English—see, for
instance, Borer (b: –). On the other hand, Krifka () and Borer
(b) champion the view that simple imperfectives in Slavic are always atelic.
The class of telic predicates in Slavic thus contrasts greatly with that of Latin,
where it is quite usual for an unprefixed verb still to head a telic predicate, as
illustrated in the following examples with capio ‘take’, facio ‘make’, neco ‘kill’, and
scribo ‘write’, which license the telic adverbial (in) paucis diebus ‘in a few days’:
() Latin; Bell. Afr. ,
Cirtamque oppidum [ . . . ] paucis diebus [ . . . ] capit.
Cirta.ACC=and town.ACC few.ABL.PL day.ABL.PL take.SG
‘And he conquers the town of Cirta in a few days.’
13
In this work I am not concerned with atelic, internally prefixed verbs displaying stative semantics,
like ab-sum “away-be” ‘be away’ or con-tineo “with-hold” ‘contain’. Assuming the theory of argument
structure configurations of section ..., these verbs, as states, involve a Place projection, which could
explain the prefixation if a vP-internal Place, as Path, were also marked as prefixal. Interestingly, Gibert
Sotelo (forthcoming), argues that these verbs are necessarily interpreted as stage-level and not individual-
level states (Carlson ; Kratzer ), which suggests that they may involve both a Path projection
encoding change and a stative BE verb encoding the state, as was also pointed out to me by Jaume Mateu
(p.c.). I leave this interesting hypothesis for future research.
14
See Filip (: ) and Bohnemeyer and Swift (: , footnote ) for more examples of
unprefixed telic verbs like Russian dat’ ‘give’, both in Russian and Czech.
Latin and Slavic complex resultatives with an internal prefix
() Latin; Cic. Phil. ,
Paucis diebus exercitum fecit.
few.ABL.PL day.ABL.PL army.ACC make.PRF.SG
‘He created an army in a few days.’
() Latin; Plin. Nat. , ,
Gelatio [ . . . ] paucis diebus necat.
frost.NOM few.ABL.PL day.ABL.PL kill.SG
‘The frost kills them [the trees] in a few days.’
() Latin; Fronto , ,
Tot [ . . . ] in tam paucis diebus epistulas
so_many in so few.ABL.PL day.ABL.PL letter(F)ACC.PL
scriptas.
write.PTCP.PFV.ACC.F.PL
‘So many letters written in so few days’.
I suggest a solution for this contrast between Latin and Slavic, in the spirit of
Bohnemeyer and Swift’s () hypothesis on default aspect. These authors, on the
basis of facts from German, Inuktitut, and Russian, argue that there are languages
displaying ‘telicity-dependent aspectual reference, [ . . . ] the phenomenon that
clauses or verbal projections not overtly marked for viewpoint aspect are assigned
semantic viewpoint-aspectual operators on the basis of the telicity of their event
predicates’ (Bohnemeyer and Swift : ). In a nutshell, Bohnemeyer and Swift
() claim that in languages lacking overt morphological marking for perfectivity a
telic predicate is by default perfective. Thus, internally prefixed verbs like Russian
pere-pisat’ ‘write down’ and inherently telic unprefixed verbs like Russian dat’ ‘give’
are perfective. If we add externally prefixed predicates to the picture, like Russian po-
pisat’ ‘write’, which is also perfective, we see that different types of predicates come to
have a perfective interpretation in these languages. On the other hand, as Bohne-
meyer and Swift (: ), following Klein (), point out, unprefixed verbs like
Russian pisat’ ‘write’, although atelic, allow both a perfective and an imperfective
interpretation. In this scenario the only way to express unambiguously a perfective
event of writing, as in I have written, is through prefixation.15 I take the strong
relation between telicity and prefixation in Slavic to be a side effect of the fact that
internal prefixes secondarily and by default express perfectivity. Crucially, a mech-
anism that primarily derives telicity is further exploited to derive perfectivity.16
15
See also Gehrke (: ) for the claim that internally prefixed predicates are telic and, by default,
perfective. An external prefix also marks perfectivity, but these prefixes carry an additional, quantificational
meaning of their own, such as ‘for a while’, ‘for a long time’, etc. (see section ..).
16
For the moment, I must remain agnostic about the actual implementation of the licensing of
perfective aspect by the prefixed verb. See Csirmaz () and É. Kiss (b) for an analysis of analogous
Weak satellite-framed languages
On the contrary, the marking of (im)perfectivity in Latin is rather different from
that in Slavic, much against a prominent part of the tradition of Latin linguistics
inaugurated by Meillet (), who first put forward the idea that Latin closely
mirrors Slavic in the establishment of a perfective/imperfective dichotomy through
the presence/absence of prefixes. This trait in the literature soon became controver-
sial because, as pointed out by Pinkster (: ) or Romagno (: ), more
often than not the authors failed to realize the difference between viewpoint and
situation aspect—see, for instance Jensen (: ) on the study of aspect by
Barbelenet (), and Reinhold (: ) on Stolz, Schmalz, Leumann, and
Hofmann (: }c). But the fact is that Latin, unlike Slavic, uses a paradigmatic
system of dedicated inflectional morphology to mark perfectivity and imperfectivity,
much as in both languages internal prefixation yields, as has been shown, transition
predicates. Thus, for instance, the prefixed verb ex-eo “out-go” ‘go out’, presents a
present (=imperfective) stem ex-e-/ex-i- and a perfect (=perfective) stem ex-i(v)-
which allows contrasts such as the following:17
() Latin; Pl. Aul. and Pl. Pseud.
a. Praesagibat mi animus [ . . . ],
have_a_presentiment.IPFV.SG me.DAT mind.NOM
quom ex-ibam domo.
when out-go.IPFV.SG house.ABL
‘My mind was having a presentiment as I was leaving my house.’
b. Inde huc ex-ii, crapulam
then hither out-go.PRF.SG intoxication.ACC
dum a-mouerem.
until away-move.IPFV.SBJV.SG
‘Then I came out here, to get rid of my intoxication’
The perfect ex-ii “out-went” expresses a bounded event of going out: the event is seen
from the outside, and, hence, it is presented as having temporal bounds, the right-
most (later) of which coincides with the telos identified by the prefix ex- ‘out’. By
contrast, ex-ibam “out-went” presents the same telic eventuality but there is no
entailment that the speaker actually got out of the house. The imperfect portrays
the event, be it telic or not, from the inside. This is why it is translated into English
facts in Hungarian, a language in which there is a similar relation between perfectivity and telicizing
particles.
17
For the perfective/imperfective tense distinction in Latin, see Reinhold (); Vairel (); Pinkster
(: ff.); Haverling (), among others. For the interpretation of the perfect and its relation to
Aktionsart, and particularly to the conclusive/non-conclusive (in our terms, telic/atelic) distinction, see
Jensen (); Pinkster (: ).
The role of morphology in the analysis of resultative constructions
with the progressive. Importantly, then, the expression of perfectivity in Latin does
not have to rely on anything but dedicated morphology.18
. The role of morphology in the analysis of resultative constructions
In this section I propose an analysis of complex resultative constructions cross-
linguistically, focusing on the generation of AP resultative constructions.
I introduce the Split S-framedness Hypothesis as a way to tackle the cross-linguistic
variation in the allowing of AP resultatives. Two independently needed factors will be
shown to conspire at the failure of Latin and Slavic to generate this type of complex
resultatives: the morphological relation of v and Path in Latin and Slavic-like
languages, where they are required to belong to the same complex head, and the
obligatory marking of the adjective with agreement morphology.
.. The morphological properties of Path. The Split S-framedness Hypothesis
In Chapter a general analysis was proposed for s-framed constructions in which
PathP codifies a change to a state or location and v is bundled together with a root
providing the conceptual dimension of the event. The overall semantic import of the
structure is that of a complex transition event: a change or transition to a new state/
location brought about through some differentiated event specified by the root
adjoined to v. The same basic analysis is put forward here for s-framed constructions
based on adjectival predicates, as shown in the following example for English She
hammered the metal flat. I add the PF-derivation of this sentence:
() PF-derivation of Sue hammered the metal flat
a. Structure delivered by syntax
[VoiceP [Voice’ Voice [vP [v v HAMMER] [PathP [Path’ Path [PlaceP [Place’ Place
FLAT]]]]]]]
b. Raising
[v HAMMER v] [Path [Place FLAT Place] Path]]
c. Linearization
HAMMER-v > FLAT-Place-Path
d. Vocabulary Insertion
hammer-∅ > flat-∅-∅
18
Needless to say, the morphological expression of the (im)perfective is obtained through different
morphophonological means, ranging from suffixal morphology to vowel changes, or even suppletion—see
Hewson () for a recent overview. What is important for my point, however, is that the expression of
(im)perfectivity is completely independent from internal (and external) prefixation.
Weak satellite-framed languages
In English-like languages there is no requirement for Path and v to find themselves in
the same complex head when Vocabulary Insertion takes place. In the case above,
Raising does not affect Path, although it affects Place, which raises to Path—and, as
usual, the two roots HAMMER and FLAT, which raise to their respective functional
heads. Independent evidence for Raising of Place to Path in English is provided by
the complex prepositions in-to and on-to. Within the realm of AP resultative
constructions such as that above, although English does not show any overt exponent
of Path in the adjective, we will see in section .. that languages like Finnish and
Hungarian do.
The PF-derivation in () is typical of the s-framed languages that I will call
strong s-framed languages. These languages allow the verb (v) and the Core Schema
(PathP) to be realized independently and, crucially, do not require them to be
realized as the same word. By contrast, in weak s-framed languages like Latin or
Slavic, the Path head and the v head must be within the same word—they must
undergo univerbation. Univerbation of v and Path in Latin (and Slavic) is achieved
through successive Raising from Compl-Place position up to v. I illustrate it in ()
with the PF-derivation of the complex prefixed resultative in ():
() Latin; Plin. Nat. ,
[Serpentes] [ova] solida hauriunt, [ . . . ] atque
snake(M)NOM.PL egg.ACC.PL whole.ACC.PL swallow.PL and
putamina ex-tussiunt.
shell.ACC.PL out-cough.PL
‘Snakes swallow the eggs whole and expel the shells through coughing.’
() PF-derivation of ()
a. Structure delivered by syntax
[VoiceP [Voice’ Voice [vP [v v TUSS] [PathP [Path’ Path [PlaceP [Place’ Place
EX]]]]]]]
b. Raising
[v [Path [Place EX Place] Path] [v TUSS v]]
c. Linearization
EX-Place-Path-TUSS-v
d. Vocabulary Insertion
ex-∅-∅-tuss-∅
Raising has the effect of forming a complex head with v, to which also the adjunct root
TUSS ‘cough’ raises, independently. In Latin and Slavic Raising affects also the Path
head, unlike in other s-framed languages like Germanic. The Vocabulary Item for Path
in these languages has the following shape (see section . for a examination of this):
() Path ⟷ ∅ / Place-_-[ . . . v . . . ]
The role of morphology in the analysis of resultative constructions
The insertion frame of this Vocabulary Item ensures that Path must be linearly
preceded by Place and must linearly precede a span containing v. Since roots,
independently, raise to their dominating functional heads, the above Vocabulary
Item has the effect of bringing all the material up to the v head. If Place does not raise
to Path and Path does not raise to v, the contextual conditions for the insertion of the
exponent are not met and the structure fails to be interpreted. Thus, the univerbation
requirement in weak s-framed languages aims at accounting for the lack of AP
complex resultatives and PP complex resultatives headed by a simple (unprefixed)
verb in these languages. The univerbation of v, Path, and Place is incompatible with a
result predicate that is itself a word. I formulate this difference as a typological
hypothesis on the morphological properties of the Path head, as stated in ():
() The Split S-framedness Hypothesis
There are two types of s-framed languages: those with a morphologically
independent Path—strong s-framed languages—and those with an affixal
Path—weak s-framed languages.
The terms strong and weak are chosen to depict the fact that strong s-framed
languages are s-framed languages in a strong sense, in that the Core Schema,
expressed independently from the verb, may adopt any morphosyntactic form. On
the other hand, languages like Latin and Slavic are s-framed languages in a weak
sense, in that they pose morphological restrictions on the expression of the Core
Schema, much as it also is expressed as an element phonologically independent from
v in these languages. The terminology also aims at hinting at a diachronic develop-
ment in the morphosyntactic expression of complex transition events from Proto-
Indo-European down to Romance: strong s-framed Proto-Indo-European yielded
weak s-framed Latin, which yielded, in turn, v-framed Romance.
Next, I will show that the characterization of an s-framed language as weak, in terms
of the Split S-framedness Hypothesis, conspires with other independent morphological
factors of the language either to allow or to ban the formation of AP resultatives.
.. The lack of complex AP resultatives in Latin and Slavic
I hypothesize that Latin and Slavic, both weak s-framed languages, do not allow
adjectival resultative constructions, since the univerbation of v and Path is incom-
patible with the fact that predicative adjectives in both languages always bear
agreement markers. This last fact is illustrated in () for Latin and Croatian:
() Agreement morphology on the predicative adjective in Latin and Slavic
a. Latin; Mart. –, , ,
Nigr-a est coma.
black-NOM.F.SG is hair(F)NOM.SG
‘Your hair is black.’
Weak satellite-framed languages
b. Croatian
Knjiga je plav-a.
book(F)NOM.SG is blue-NOM.F.SG
‘The book is blue.’
Following Mateu’s () and Kayne’s () proposal that adjectives and prepositions
involve the same basic category (see section ..), I take the (predicative) adjective to
be an instantiation of PlaceP. Specifically, in this case the head Place is endowed with
uninterpretable phi-features reflected in the agreement morphology, as shown in the
Latin and Croatian examples above. As described by Citko (: ff.), following
Chomsky () and Richards (), among others, uninterpretable features must be
deleted before arriving at the LF interface. However, since they may produce effects at
PF, as is the case in Latin and Slavic, they must also arrive at the phonological interface
first, which is implemented by positing that valuation of these features and transfer to
the interfaces is simultaneous. The PlaceP endowed with uninterpretable phi-features
(the adjective) is, thus, defined as a phase. This means, in turn, that such a PlaceP is not
computed within the cycle headed by v, given the theory of cyclic Spell-Out adopted in
section .. Let us see now what effect this scenario has on the derivation of complex
AP resultatives in Latin, such as that in the following made-up example:
() Latin; made-up ungrammatical example
*Ovidia poculum vacu-um bibit.
Ovidia.NOM goblet(N)ACC.SG empty-ACC.N.SG drink.SG
() PF-derivation of ()
a. Structure delivered by syntax
[VoiceP [Voice’ Voice [vP [v v BIB] [PathP [Path’ Path {[PlaceP [Place’ Place
VACU]]}]]]]]
b. Raising
[v Path [v BIB v]]
c. Linearization
Path-BIB-v
d. Vocabulary Insertion
?-bib-∅
At Vocabulary Insertion, after Raising and Linearization have taken place, Path
linearly precedes the root and the verb. However, Place does not precede it, since it
has been spelled out in a different cycle. Since the insertion frame in the Vocabulary
Item of Path specifies this condition, namely that Place should precede Path, the
exponent cannot be inserted, and the derivation fails:
() Path ⟷ ∅ / Place-_-[ . . . v . . . ]
The role of morphology in the analysis of resultative constructions
In section . I will examine the structure of this Vocabulary Item and I will try to
derive the same results on more principled grounds. For the time being, though,
consider that this Vocabulary Item would be compatible with a derivation in which
the adjective is an adjunct to PlaceP, which contains a prefixal root at Compl-Place.
Such an analysis could perhaps be applied to cases like the following, described by
Pinkster (: –), and other cases shown already in section ..:
() Verg. A. ., in Pinkster (: )
Tum sterilis ex-urere Sirius agros.
then sterile.ACC.PL out-burn.INF Sirius fields.ACC
‘Then did Sirius burn the fields sterile.’
() Analysis of ()
VoiceP
Sirius Voice’
Voice vP
v PathP
v ur agros Path’
Path PlaceP
{pP} PlaceP
p steril agros Place’
Place ex
In this case, the adjective does not correspond to the PlaceP of the spinal derivation,
since it is an adjunct. It is spelled out too in a different cycle from that headed by v.
Since the PlaceP containing the prefixal root EX ‘out’ does not constitute an inde-
pendent cycle, Raising can take place from this position and the insertion frame of
the Vocabulary Item for Path is satisfied. Although such cases are fully compatible
with the analysis developed here, Pinkster (: ) points out their scarcity, which
makes them suspect. In particular for the case in hand, sterilis ‘sterile’ could be an
AP modifying the object agros ‘fields’ directly. See section .. for more related
discussion.
In the next section I will show how Latin allows simple (i.e. non-complex)
adjectival resultative constructions. It remains also to be seen whether within the
class of s-framed languages there can be prefixation of a resultative adjective when
the adjective is not inflected. This is what we find in Icelandic and Mandarin Chinese,
as will be shown in section .. and ..
Weak satellite-framed languages
.. Simple adjectival resultatives in Latin
As I mentioned in sections .. and .., Latin allows the formation of simple
adjectival resultatives, that is, resultative constructions in which the verb encodes
pure transition, with no Co-event, and the AP specifies the result state. Besides
Pinkster’s (: ) example based on ago ‘lead, drive’ shown in section ..
and repeated here as (), I show simple AP resultatives based on facio ‘do’, red-do
“back-give” ‘render’, and re-linquo “back-leave” ‘leave’:
() Latin; Cic. Ver. ., in Pinkster (: )
Agunt eum praecipitem poenae
drive.PL him.ACC headlong.ACC.M.SG punishment.NOM.PL
civium Romanorum.
citizen(M)GEN.PL Roman.GEN.M.PL
‘The punishments of Roman citizens drive him crazy.’
() Latin; Cic. Phil. ,
Senatum [ . . . ] firmiorem [ . . . ] fecistis.
senate(M)ACC.SG firm.COMPAR.ACC.M.SG make.PRF.PL
‘You made the senate stronger.’
() Latin; Plaut. Capt.
Eam [servitutem] [ . . . ] lenem reddere.
that.ACC.F.SG serdom(F)ACC.SG mild.ACC.F.SG render.INF
‘To make that serfdom mild.’
() Latin; Cic. Catil. ,
Simul atque ad-sedisti, partem istam subselliorum
at once and at-sit.PRF.SG part(F)ACC.SG that.ACC.F.SG seat.GEN.PL
nudam atque inanem reliquerunt.
nude.ACC.F.SG and empty.ACC.F.SG leave.PRF.PL
‘At the moment you sat down among them, they left that part of the seats
nude and empty.’
The heading verb in the above examples does not encode a co-event leading to the
final state. Therefore, these constructions do not contradict the claim made in section
.. that Latin does not feature complex adjectival resultative constructions. In the
case of the resultatives headed by red-do “back-give” ‘render’ and re-linquo “back-
leave” ‘leave’, I assume that re- is the exponent of Path when it is strictly left-adjacent
to a v-Voice span:
() Path ⟷ re / _-v-Voice
Observe that the conditions for insertion of the above Vocabulary Item pre-empt the
insertion of the other null exponent of this functional head:
The role of morphology in the analysis of resultative constructions
() Path ⟷ ∅ / Place-_-[ . . . v . . . ]
Thus, whenever, Path is strictly left-adjacent to v, re, and not ∅, will be inserted.
I illustrate with the derivation of ():
() Analysis of ()
a. Structure delivered by syntax
VoiceP
PRO Voice’
Voice vP
v PathP
servitutem Path’
Path {PlaceP}
servitutem Place’
Place len
b. Raising
[Voice [v Path v] Voice]
c. Linearization
Path-v-Voice
d. Vocabulary Insertion
red-d-∅
The exponent of v in these cases is also predictable, like the light verb do ‘give’ or
linquo ‘leave’, and it is triggered directly by re. The contextual specification of the
Vocabulary Item for Path in these constructions has the power to preclude another
pattern of complex AP resultatives which is not found in Latin:
() Latin; made-up ungrammatical example
*Ovidia poculum vacu-um re-bibit.
Ovidia.NOM goblet(N)ACC.SG empty.ACC.N.SG re-drink.SG
In () v is associated with a root (BIB ‘drink’), which, as we know, ends up linearized
to its left. Thus, Path cannot receive its exponent, since it does not meet the
contextual condition of strict linear adjacency with v.
The data dealt with up to this point show that the head Path in Latin is always
prefixal, in that it attaches to the verb to the left. However, Path only emerges as a
prefix (re-) when it is strictly left adjacent to v. Otherwise, Path is realized as a null
exponent in a sequence including the root and Place to its left.
Weak satellite-framed languages
Importantly, this re must be distinguished from a homophonous prefix involving a
root RE with spatial meaning ‘back’ and not requiring any AP adjunct. Thus, in the
absence of the adjective red-do “back-give” ‘render’ and re-linquo “back-leave” ‘leave’
are interpreted very differently, as verbs of transfer and backwards motion,
respectively:
() Latin; Plaut. Amph..
Tibi habeas res tuas,
you.DAT have.SBJV.SG thing.ACC.PL your.ACC.F.PL
re-ddas meas.
back-give.SBJV.SG my.ACC.F.PL
‘You keep your things for yourself and give mine back.’
() Latin; Ov. Met. ,
Tergo velamina lapsa re-liquit.
back.ABL.SG veil(N)ACC.PL slip.PTCP.PFV.ACC.N.PL back-leave.SG
‘She left behind the veil which had slipped off her back.’
In these examples the prefix can be claimed to be interpreted as English back. Hence,
I assume that they involve a root RE merged as Compl-Place, while the root of the
verb (D, LINQU) is merged as an adjunct to v.
The analysis of AP resultative constructions headed by red-do “back-give” ‘render’
and re-linquo “back-leave” ‘leave’ can clearly not be carried over to those AP
resultative constructions headed by unprefixed verbs like ago ‘lead’ and facio
‘make’. For this type of unprefixed resultative constructions I explore an analysis
whereby the verb is a light creation verb taking a nominalized Small Clause as
object.19 The nominalized Small Clause is interpreted, therefore, as an Effected
Object:
19
The analysis is inspired by that proposed by Folli and Harley () for the so-called faire par
causative constructions (Kayne ) in Italian. On the basis of different syntactic and semantic facts these
authors propose that causative constructions with a causee introduced by da ‘by’, and not by a ‘to’, in
Italian, involve a nominalized VP (riparare la macchina ‘repair the car’, below) and a creation, rather than
causative, fare ‘do’ (Folli and Harley : ):
(i) Gianni ha fatto riparare la macchina a/da Mario.
Gianni has made repair.INF the car to/by Mario
‘Gianni got Mario to repair the car.’ / ‘Gianni got the car repaired by Mario.’
The role of morphology in the analysis of resultative constructions
() Analysis of ()
VoiceP
pro Voice’
Voice vP
v nP
n PlaceP
senatum Place’
Place ferm
The predicative relation between the object and the adjective is articulated through a
Place head. I hypothesize that PlaceP is taken as complement by a null n head, which
nominalizes it. Importantly, since the above structure does not involve the projection
of PathP, no prefix is expected, under the assumption that prefixation in Latin (and
Slavic) involves Raising of Path onto v.
There is independent evidence that in Latin a Small Clause combining a DP
subject and a participle, AP, or DP predicate can function wholly as an argument,
which reveals that it is in fact nominalized. This is shown by the following examples
from Pinkster (: ), in which I have italicized the nominalized Small Clause,
which in both cases is acting as the sentential subject:
() Latin; Tac. Ann. .., in Pinkster (: )
Augebat metum gnarus Romanae
increase.IPFV.SG fear.ACC aware.NOM.M.SG Roman.GEN.F.SG
seditionis et [ . . . ] in-vasurus hostis.
sedition(F)GEN and in-rush.PTCP.FUT.NOM.M.SG enemy(M)NOM
‘That the enemy was aware of the sedition in Rome and that he would quickly
enter increased the fear.’
() Latin; Tac. Ann. .., in Pinkster (: )
Filius legati orator publicae causae
son.NOM commander.GEN spokesman.NOM common.GEN.F.SG cause(F)GEN
satis ostenderet necessitate ex-pressa
sufficiently show.IPFV.SBJV.SG force.ABL out-wrest.PTCP.PFV.ACC.N.PL
quae per modestiam non obtinuissent.
what.ACC.N.PL by modesty.ACC not obtain.PLUPRF.SBJV.PL
‘The fact that their commander’s son was spokesman for their common cause
would clearly show that they had wrested by compulsion what they had not
obtained by their modesty.’
Weak satellite-framed languages
In the above examples the thematic relation is established between the main predi-
cate (augebat metum ‘increased the fear’, ostenderet ‘would show’) and the whole
italicized sequences, and not between the main predicate and the head of those
sequences (hostis ‘the enemy’, filius legati ‘their commander’s son’). This is evidence
that those italicized sequences, which contain, as I have said, a predication relation,
are functioning as fully fledged arguments.
. Summary
In this chapter I have located Latin within the wider cross-linguistic scenario with
respect to the way it syntactically builds complex events of change. Departing from
the results arrived at in Chapter , that is, that Latin qualifies as an s-framed
language, I have shown that, nevertheless, it differs from other s-framed languages
such as the Germanic languages in disallowing typically s-framed constructions
based on AP resultative predicates. Since neither Talmy’s typology nor the subse-
quent revisions thereof predict such a scenario, I have tried to seek out a possible
explanation for this behaviour. My first step has been to observe that Latin patterns in
this sense with the group of Slavic languages, also acknowledged for their s-framed
status, and I have focused on an additional feature that characterizes both: the fact—
arrived at for Latin through corpus searches—that complex resultative constructions
are always built on prefixed verbs. I have proposed that in these languages there is a
general univerbation requirement bringing Place and Path to form a complex head
with v. This requirement, encoded in the insertion frame of the head Path, is
incompatible with the situation in which the PlaceP is an adjective inflected for
agreement, since it forms an independent Spell-Out cycle. It is only when Path is
strictly left adjacent to v, in simple resultative constructions, that Path receives an
exponent (re) independently of the realization of PlaceP, which can thus be realized
as an adjective.
6
A revision of Talmy’s typology
In this chapter I explore the empirical coverage of the Split S-framedness Hypothesis,
both for strong and weak s-framed languages. Icelandic will be shown to provide data
underpinning the assumption that the morphological characterization of Path and
the inflectional morphology on the resultative adjective are the factors at stake in
triggering the split within the s-framed class of languages. As stated in the Split
S-framedness Hypothesis (section ..), there are two basic types of s-framed
languages: those where Path can be morphologically independent—strong s-framed
languages, and those where Path is always an affix and must then lean on another
head to be licensed—weak s-framed languages. If v-framed languages are taken into
account, a three-way typology emerges based on the phonological dependence of
Path with respect to the verb. At one extreme are strong s-framed languages, where
the Path is morphologically independent from the verb, both being expressed as
different morphemes and words. These languages allow the generation of PP,
particle, and AP resultatives. Next to these languages are weak s-framed languages,
in which the Path and the verb are different morphemes but one phonological word.
This allows resultatives based on affixal particles but precludes the formation of PP
resultatives and of AP resultatives if the predicative adjective is inflected. Finally,
I compare previous analyses of the issue of the expression of resultativity and change
of location cross-linguistically, and address particular problems.
. Another weak s-framed language: Ancient Greek
Ancient Greek shows the hallmarks of s-framedness: encodement of the Core
Schema as an element different from the verbal morpheme and the possibility of
expressing a manner co-event within the verb, as illustrated through the following
CDMCs:
() Ancient Greek; Thuc. , and , ; in Liddell and Scott ()
a. Tô:n andrô:n apo-kolumbe:sánto:n.
the.GEN.M.PL man(M)GEN.PL away-swim.PTCP.AOR.GEN.M.PL
‘The men having swum away’.
The Morphosyntax of Transitions. First edition. Víctor Acedo-Matellán.
© Víctor Acedo-Matellán . First published by Oxford University Press.
A revision of Talmy’s typology
b. [Hoì] hippê:s [ . . . ] pros-ippeúontes.
the.NOM.M.PL cavalryman(M)NOM.PL forth-ride.PTCP.PRS.NOM.M.PL
‘The cavalry, riding up to them’.
This language allows for resultative constructions based on prefixes, in the same way
as Latin and Slavic. The following examples involving transitive prefixed verbs, in
particular, are cases of UOCs, since the occurrences of the unprefixed counterparts
orkhéomai ‘dance’ and kubeúo: ‘play dice’ are intransitive—see more examples in
Romagno (, ) and Lavidas (: ff.):
() Ancient Greek; in Meillet and Vendryès (: )
a. Hdt. ,
*(Ap-)orkhé:saó [ . . . ] tòn gámon.
away-dance.AOR.MID.SG the.ACC.M.SG wedding(M)ACC.SG
‘You have danced your wedding away.’ (i.e. ‘You have ruined your wedding
by dancing.’)
b. Lys. ,
*(Kata-)kubeúsas tà ónta.
down-gamble.PTCP.AOR.NOM.M.SG the.ACC.N.PL possession(N)ACC.PL
‘Having gambled away his possessions’.
There is evidence of obligatory affixation of the Path onto the verb in Ancient Greek.
Horrocks and Stavrou (: –) and Horrocks (: –) contend that a
predicate headed by a manner-of-motion verb is always interpreted as directional and
implying a goal if the verb appears with a complement goal PP, a prefix, or both.
However, Horrocks and Stavrou () point out, against this position, that a resultative
interpretation of predicates of that kind, illustrated in (), might be only apparent:
() Ancient Greek; in Horrocks and Stavrou (: )
a. Thucydides ,
Es Himéran prô:ton pleúsantes.
(In)to Himera.ACC first sail.PTCP.AOR.NOM.M.PL
‘Having sailed first to Himera.’
b. Thucydides ,
Par-épleusan es Lókrous.
beside-sailed.AOR.PL (in)to Locri.ACC.
‘They sailed along (the coast) to Locri.’
c. Thucydides ,
Kata-pléontes [ . . . ] es tà pròs tò
Down-sail.PTCP.PRS.NOM.M.PL (in)to the.ACC.N.PL facing the.ACC.N
pélagos tê:s né:sou.
open_sea.ACC the.GEN island.GEN
‘Sailing down to the parts of the island facing the open sea’.
Another weak s-framed language: Ancient Greek
In particular, Horrocks and Stavrou (: ) claim that
there remains the further possibility that these PPs are actually to be understood as adjuncts
[ . . . ] used with verbs that retain their agentive manner-of-motion meaning, [ . . . ] (i.e. (para/
kata)-pleo e(i)s X = ‘go-sailing/go-on-a-sail (along/down) [to X]’). In support of this alterna-
tive analysis involving pseudo-unaccusativization, an exhaustive search for verbs meaning
‘walk’, ‘run’, ‘swim’ and ‘sail’ in the very large corpus of classical Greek literature contained in
the electronic database of the Thesaurus Linguae Graecae (http://tlg.uci.edu) revealed no
examples which also contained a time-within-which adverbial.
Unfortunately, Horrocks and Stavrou () do not specify whether the search they
performed included the prefixed counterparts of ‘verbs meaning “walk”, “run”,
“swim”, and “sail”’. Geoffrey Horrocks, in a personal communication, informs me
that, in fact, the search was carried out taking into account only unprefixed verbs.
According to my assumptions and the hypothesis that Ancient Greek is a weak
s-framed language, the results of Horrocks and Stavrou’s () search are unsurpris-
ing: unprefixed verbs in weak s-framed languages cannot support telic complex
resultative constructions, even if accompanied by an alleged goal PP.1
However, prefixed predicates are telic in Ancient Greek, even in the absence of
directional PPs, in conformity with present assumptions about weak s-framed
languages. Thus, in performing a search of prefixed motion verbs in a subcorpus
of the Thesaurus Linguae Graecae (Pantelia ), a non-trivial amount of unam-
biguously telic examples were found:2
() Ancient Greek; telic manner-of-motion predicates with a prefixed verb
a. Thucydides, Historiae, , , ,
Kéra:i tôn Athe:naío:n
wing(N)DAT.SG the.GEN.M.PL Athenian(M)GEN.PL
euthùs apo-bebe:kóti.
right_away away-step.PTCP.PFV.DAT.N.SG
‘The wing of the Athenians, which had just disembarked’.
1
See also Lavidas (: ), who acknowledges that the prefixes in Ancient Greek can mark ‘the
completion of the action denoted by the verb’.
2
Verbs searched for: apo-baíno: ‘walk, step away’, ek-baíno: ‘walk, step in’, em-baíno: ‘walk, step in’,
kata-baíno: ‘walk, step down’, án-eimi ‘go up’, áp-eimi ‘go away’, eís-eimi ‘go in’, kát-eimi ‘go down’, ap-
hippeúo: ‘ride away’, kat-hippeúo: ‘ride down, over’, ana-kolumbáo: ‘come up after diving’, apo-kolumbáo: ‘dive
and swim away’, eis-kolumbáo: ‘swim into’, ek-kolumbáo: ‘swim ashore, plunge into the sea from’, kata-
kolumbáo: ‘dive down’, ana-pléo: ‘sail upwards, go up-stream, rise to the surface’, apo-pléo: ‘sail away’, eis-pléo:
‘sail into a harbour’, ek-pléo: ‘sail out’, kata-pléo: ‘sail down, back’, ana-trékho: ‘run back’, apo-trékho: ‘run off,
away’, eis-trékho: ‘run in’, ek-trékho: ‘run out’, en-trékho: ‘run in, enter’, kata-trékho: ‘run down’. The subcorpus
of authors consisted of non-late (pre-Christian) authors (and corpora): Aeschylus, Alcidamas, Anonymi
medici, Antiphon, Pseudo-Apollodorus, Aristophanes, Aristotle and the Corpus Aristotelicum, Bacchylides,
Chariton, Demosthenes, Epicurus, Euclid, Euripides, Hesiod, Homer, Isocrates, Lysias, Plato, Plutarch,
Sophocles, Thucydides, Vettius Valens, Xenophon, and the Scholia in Aeschylum. The references for examples
provided here are those provided by the Thesaurus Linguae Graecae (Pantelia ). The transliterations of all
the Ancient Greek examples in this section are my own. I am grateful to Geoffrey Horrocks for suggesting the
kind of adverbial or case marked DP I should use as the telicity-signalling expression in Greek.
A revision of Talmy’s typology
b. Xenophon, Hellenica, , , ,
Alkibiáde:s [ . . . ] ap-ébaine mèn ouk euthús.
Alcibiades.NOM.SG away-step.IPFV.SG PART not right_away
‘Alcibiades didn’t disembark right away.’
c. Thucydides, Historiae, , , ,
Euthùs ek-bántes autoû
right_away out-step.PTCP.NOM.M.PL it.GEN.SG
áriston epoioûnto.
luncheon(N)ACC.SG make.IPFV.MID.PL
‘Right after disembarking, they prepared themselves luncheon.’
d. Thucydides, Historiae, , , ,
En ou pollô:i khróno:i kat-ébainen es
in not much.DAT.M.SG time(M)DAT.SG down-step.IPFV.SG in
tà sté:the: ho pónos.
the.ACC.N.PL chest(N)ACC.PL the.NOM.M.SG pain(M)NOM.SG
‘In a short time the pain descended into the chest.’
e. Thucydides, Historiae, , , ,
Tòn dè mè: ethélonta
the.ACC.M.SG PART not want.PTCP.PRS.ACC.M.SG
ap-iénai [ . . . ] pénte he:merô:n.
away-go.INF five day(F)GEN.PL
‘That those that didn’t want to (should) leave in five days’.
f. Xenophon, Hellenica, , , ,
Olígo:n mèn he:merô:n anágke: ésoito ap-iénai.
few.GEN.F.PL PART day(F)GEN.PL need(F)NOM.SG be.OPT.SG away-go.INF
‘That there was need that he should leave in a few days’.
g. Plutarchus, Agesilaus, , ,
Euthùs ap-épleusen.
right_away away-sail.AOR.SG
‘He sailed away immediately.’
h. Thucydides, Historiae, , , ,
Ho dè pròs mèn tè:n E:ióna katá te
the.NOM.M.SG PART facing PART the.ACC.F.SG Eion.ACC down and
tòn potamòn [ . . . ] áphno: kata-pleúsas.
the.ACC.M.SG river(M)ACC.SG immediately down-sail.PTCP.AOR.NOM.M.SG
‘He having sailed down the river immediately, towards Eion’.
Another weak s-framed language: Ancient Greek
i. Thucydides, Historiae, , , ,
Kaì hoi Athe:naîoi ou pollô:i
and the.NOM.M.PL Athenian(M)NOM.SG not much.DAT.N.SG
hústeron kata-pleúsantes.
later.ACC.N.SG down-sail.PTCP.AOR.NOM.M.PL
‘The Athenians having sailed up to there not much later’.
j. Thucydides, Historiae, , , ,
Ek-dramóntes áphno: ek tê:s póleo:s.
out-run.PTCP.AOR.NOM.M.PL suddenly out the.GEN.F.SG city(F)GEN.SG
‘Having run out of the city all of a sudden.’
k. Xenophon, Hellenica, , ,
hoi mèn psiloì euthùs ek-dramóntes
the.NOM.M.PL PART light(M)NOM.PL right_away out-run.PTCP.AOR.NOM.M.PL
e:kóntizon.
hurl_javelins.IPFV.PL
‘The light troops, having run out immediately, started hurling javelins.’
Note, importantly, that in the examples above the directional PP or the DP is
optional—see ()h and ()j for cases of the former and ()c for a case of the latter.
If in Ancient Greek, as the data seem to suggest, complex resultatives feature a prefix
representing Path, it should count as a weak s-framed language, within present
assumptions. Since the predicative adjectives in Ancient Greek are always inflected
for agreement, as shown below, the prediction emerges that this language will not
allow complex adjectival resultative constructions.
() Ancient Greek; Pseudo-Apollodorus, Bibliotheca (sub nomine Apollodori), ,
Toûto dè hupermégeth-és estin.
this.NOM.N.SG PART exceedingly_difficult-NOM.N.SG be.SG
‘This is exceedingly difficult.’
As far as my (limited) competence in Ancient Greek tells me, those constructions are
not found in Ancient Greek. This is also hinted at by Horrocks (: ) and, most
importantly, it is claimed as an empirical fact by Horrocks and Stavrou (: ),
who point out that a search ‘for predicate adjectives in the Thesaurus Linguae
Graecae electronic database of ancient Greek literature (www.tlg.uci.edu) produced
no examples of result-state readings, and only depictive ones’.
Ancient Greek thus turns out to be a weak s-framed language, since, being s-framed
(recall the UOCs in () above), it does not seem to accept a morphologically inde-
pendent Path, as suggested by an analysis of CDMCs in this language. As a result,
Ancient Greek is correctly predicted not to license adjectival resultative constructions.
A revision of Talmy’s typology
. Strong s-framed languages
.. German and Dutch
In Dutch and German we find particle verb constructions easily amenable to the
same analysis as that proposed here for Latin and Slavic prefixed predicates. First,
these particles can be shown to be also interpreted as resultative, that is, as
specifiying a state or location resulting from a particular event (encoded by the
verb). Thus, for instance, in () the German particles ein ‘in’, aus ‘out’, auf ‘up’, ab
‘off ’, hinein ‘therein’, and zurück ‘back’ describe the final location of the subject (Peter
and Hans in ()a and der Taucher ‘the diver’ in ()b) or the object (die Flasche ‘the
bottle’ in ()c):
() German particles
a. Zeller (b: )
Weil Peter ein-steigt und Hans aus.
because Peter in-climbs and Hans out
‘Because Peter gets in and Hans gets out’.
b. Lüdeling (: )
Dass der Taucher auf-taucht.
that the diver up-dives
‘That the diver surfaces’.
c. Lüdeling (: )
Der Prinz stellt die Flasche ab/hinein/zurück.
the prince puts the bottle away/in/back
Similarly, the Dutch particles in ‘in’ and af ‘off ’ describe the final location of the truck
and of the hat in ()a, and ()b, respectively, and in ‘in’ describes a more abstract
location or state (the state of being available) for the petition in ()c (crucially, in in
the first example must be understood as a particle, and not a postposition—see Van
Riemsdijk : ):
() Dutch particles
a. Van Riemsdijk (: )
Omdat Jan de vrachtwagen in reed.
because Jan the truck in drove
‘Because John drove the truck in’.
b. Broekhuis (: )
Jan zet zijn hoed af.
Jan puts his hat off
‘Jan takes his hat off.’
Strong s-framed languages
c. Gehrke (: )
Zij diende een aanvraag in.
she handed a petition in
‘She filed a petition.’
There are other instances of particles bearing a less evidently resultative interpret-
ation. McIntyre () makes a case for this claim through an analysis of the German
particle ein ‘in’, which indicates a final ‘state of readiness for the activity expressed by
the verb’ (McIntyre : ):
() German; McIntyre (: )
a. sich ein-spielen
REFL in-play.INF
‘get warmed up (in sport/music)’
b. sich ein-arbeiten
REFL in-work.INF
‘work one’s way into something’
c. sich ein-singen
REFL in-sing.INF
‘get warmed up in singing’
McIntyre points out that ein in these predicates is not a mere inchoative marker
whereby ein-spielen “in-play” for instance, should be rendered ‘begin playing’.
Rather, these predicates present an interpretation analogous to adjectival resultative
constructions such as the following:
() German; McIntyre (: )
sich warm spielen.
REFL warm play.INF
‘get warmed up (in sport)’
Second, German and Dutch particles are very capable of licensing unselected objects.
In () the particles ab ‘off ’ and an ‘on’ permit intransitive arbeiten ‘work’ and husten
‘cough’ to appear with an accusative object:
() German; Zeller (b: –)
a. Peter arbeitet seine Schulden *(ab).
Peter works his debts off
‘Peter works off his debts.’
b. Peter hustete ihr eine dicke Erkältung *(an).
Peter coughed her a thick cold on
‘She caught a heavy cold from Peter’s coughing.’
A revision of Talmy’s typology
Dutch particles also display the power of introducing an argument, as af ‘off ’ and toe
‘to’ illustrate in the following examples:
() Dutch; Booij (: )
a. lopen / de straten *(af-)lopen
walk.INF the streets off-walk.INF
‘walk’ / ‘tramp the streets’
b. juichen / iemand *(toe-)juichen
cheer.INF somebody to-cheer.INF
‘cheer’ / ‘cheer someone up’
Finally, German and Dutch particles may induce telicity in the predicate in which they
appear, which suggests, within the framework adopted here, the projection of PathP. For
instance, the atelic process portrayed by schmort ‘braises’ in ()a is turned into an
accomplishment (where the result state of the referent of the object is that of being partially
or lightly affected by the action—see McIntyre ) in ()b by virtue of the particle an ‘on’:
() German; Stiebels and Wunderlich (: )
a. Er schmort den Braten.
he braises the roast
‘He braises the roast.’
b. Er schmort den Braten an.
he braises the roast PART
‘He braises the roast to a certain degree.’
Analogously, Levin and Rappaport Hovav (: –) report that Dutch atelic
bloeien ‘bloom’ (see ()a) turns into a telic change-of-state predicate with the
addition of op ‘up’ (see ()b):
() Dutch; Levin and Rappaport Hovav (: –)
a. Deze bloem heeft het hele jaar gebloeid.
this flower has the whole year bloomed
‘This flower bloomed for the whole year.’
b. Het boompje is helemaal op-gebloeid
the little_tree is completely up-bloomed
toen ik het regelmatig mest gaf.
when I it regularly fertilizer gave
‘The little tree completely flourished when I regularly gave it fertilizer.’
It is important to note that the addition of the particle op ‘up’ also provokes a change
in auxiliary selection from hebben ‘have’ in ()a to zijn ‘be’ in ()b, which further
suggests a change in argument structure properties for the predicate.
Strong s-framed languages
As can already be gathered from the examples above, particles in these languages
may appear attached to the verb or separated from it. The following minimal pair in
German illustrates both possibilities:
() German; Zeller (b: )
a. Weil Peter in den Bus ein-steigt.
because Peter in the.ACC bus in-climbs
‘Because Peter gets on the bus’.
b. Peter steigt in den Bus ein.
Peter climbs in the.ACC bus ein
‘Peter gets on the bus.’
According to Zeller (b) and Lüdeling (), in subordinate clauses like
()a the particle appears adjacent to the verb, which sits in its original final
position. In matrix clauses, however, the verb undergoes movement to a ‘second’
position (the well-known phenomenon of V—see Haider and Prinzhorn ();
Weerman (); Vikner (), among others), stranding the particle. Dutch
particles are also separable from the verb through stranding under V
movement:
() Dutch; Booij (: )
a. . . . Hans zijn moeder op-belde.
Hans his mother up-called
b. Hans belde zijn moeder op.
Hans called his mother up
c. . . . de fietser neer-stortte.
the cyclist down-fell
d. De fietser stortte neer.
the cyclist fell down
A second proof of the morphological independence of the particle is the fact that it
may be fronted under topicalization:
() German; Zeller (b: )
Zu hat er die Tür gemacht.
to has he the door made
‘He locked the door.’
() Dutch; Bennis (: )
a. Op gaat de zon in het oosten.
up goes the sun in the east
‘The sun rises in the east.’
A revision of Talmy’s typology
b. Uit voert Angola veel koffie.
out takes Angola much coffee
‘Angola exports a lot of coffee.’
In light of these facts, we must conclude that, assuming that these particles appear
within a PathP structure, the Path cannot bear the status of a prefix in German or
Dutch. Accordingly, it comes as no surprise, under the assumptions made here, that
these languages allow AP resultatives:
() German; Kratzer (: –)
a. Die Teekanne leer trinken.
the teapot empty drink.INF
‘To drink the teapot empty’.
b. Die Tulpen platt giessen.
the tulips flat water.INF
‘To water the tulips flat’.
c. Er hat seine Familie magenkrank gekocht.
he has his family stomach_sick cooked
‘He made his family ill with his cooking.’
() Dutch; Hoekstra (: –)
a. Hij liep zijn schoenen scheef.
he walked his shoes worn_on_one_side
b. Hij schaatste het ijs kapot.
he skated the ice cracked
c. De boorhamer dreunde mij doof.
the pneumatic drill pounded me deaf
It is, of course, well known that both Dutch and German feature sets of so-called
inseparable prefixes, which must appear adjacent to the verb in all circumstances. As
a result, in V environments the prefix, unlike the particle, cannot strand, as is
illustrated by German be-:
() German; Zeller (b: )
a. Weil Peter den Berg be-steigt.
because Peter the.ACC mountain be-climbs
‘Because Peter climbs the mountain’.
b. Peter {be-steigt den Berg/ *steigt den Berg be.}
Peter be-climbs the.ACC mountain climbs the.ACC mountain be
Prefixes, like particles, suggest the projection of a PathP, since, according to Stiebels
() and Lüdeling (: ), they can also induce telicity; they can also induce
Strong s-framed languages
changes in argument structure properties. Thus, in the following example the
prefixed verb vergärtnern ‘spend away in gardening’ is a transitive, telic accomplish-
ment, contrasting with atelic intransitive gärtnern ‘garden’:
() German; Stiebels (: )
Er ver-gärtner-te sein gesamtes Vermögen.
he ver-garden-PST his.ACC whole.ACC fortune.ACC
‘In gardening, he used up all his fortune.’
Accordingly, I will assume that at least some prefixes, like particles, also originate as
roots in Compl-Place position. In this sense I follow Hoekstra and Mulder’s (:
–) and Mulder’s (: –) seminal analyses of Dutch prefixes as predicates
of a Small Clause, as was shown in section ..., and Mateu’s (b) analysis of
German vergärtnern ‘spend in gardening’ and other prefixed denominal verbs, where
the prefix originates as the head of a Small-Clause projection.
Prefixes, unlike particles, are endowed with a morphological requirement to be
prefixed onto the verb. I am claiming, crucially, that it is the root that is endowed
with this requirement, and not the functional head Path. This can be modelled by a
Vocabulary Item along the following lines:
() VER $ ver / _-[ . . . v . . . ]
This Vocabulary Item ensures that ver- can only be inserted as linearly preceding a
span containing the v head. A series of raising operations starting from VER, at
Compl-Place, bring this terminal to its final position.
.. English
As for Dutch and German, following a long-standing tradition in which particles are
merged as part of a Small Clause (Stowell ; Kayne ; Hoekstra ; Den
Dikken ; Svenonius ; and Hale and Keyser , among others), I assume
that English particles also signal the projection of a PathP structure, and that the
particle originates as a root in Compl-Place position. With that in mind, we can quite
safely claim that Path is not obligatorily prefixed onto the verb in this language, as the
following examples show:
() Svenonius (: )
a. The doorman threw the drunks out.
b. The firefighters hoisted the equipment up.
c. The police chased the demonstrators off.
This state of affairs is in accordance, under present assumptions, with the well-
known allowance of adjectival resultatives in this language:
A revision of Talmy’s typology
() Carrier and Randall (: –)
a. She pounded the dough flat as a pancake.
b. They ran their sneakers ragged.
c. The maid scrubbed the pot shiny.
d. The chef cooked the kitchen walls black.
e. The tourists walked their feet sore.
There is a productive prefixation mechanism in English that presumably involves the
projection of a PathP, namely out-prefixation:
() Irube ()
a. Mary outspends John.
b. The Brownies outguessed the Girl Scouts in the contest.
c. Outfielders must outthrow infielders.
Evidence that out is initially projected within the vP comes from the fact that it
licenses otherwise unselected objects, as in the examples in (). Marantz ()
argues that out-predicates are resultative predicates in which the internal argument is
understood as being exceeded by the external argument along some dimension
specified by the root of the verb. Thus, for instance, in John outran the bus, the bus
is exceeded by John in running. The caused final state is, then, that of being exceeded.
If out-predicates are instances of complex resultative predicates with a Path projec-
tion, and if the fact that this prefixation is obligatory is taken into account, my
proposal that Path in English is not affixal is in doubt. This obligatorily prefixed out
must be a root different from, but homonymous to, the out that appears in other
verb-particle combinations, such as that in ():
() (*Out-)put (out) the fire (out).
There is evidence that we are dealing with two outs here. On the one hand, the
semantics are clearly different, and, most notably, the prefixed out never delivers an
idiomatic meaning, such as the one we find in (). On the other hand, the
phonologies are also different, since, though segmentally identical, the out in ()
is a prefix, and that in () cannot be prefixed. I propose then, that out-prefixation
depends on the idiosyncratic properties of this out root and not on the properties of
Path in English, which, as discussed, is not required to be prefixed. Thus, as in the
case of verbal prefixes in Dutch and German seen in the previous section, the out- of
out-prefixation has the following Vocabulary Item:
() OUT $ out / _-[ . . . v . . . ]
This analysis can be applied, on the other hand, to other cases of prefixes in
predicates arguably involving a PathP. Marantz (), for instance, proposes that
in Latinate verbs like construct and destroy the segments con- and de- are actually
Strong s-framed languages
predicates of the internal argument and end up prefixed onto the verb. Harley ()
applies the same analysis to Latinate verbs in general (like com-pose, dis-sect, ex-hibit,
in-cise), accounting for the failure of these complex verbs to combine with particles,
with resultative predicates or to head double object constructions.3
.. Icelandic
Icelandic resultative constructions may feature non-prefixed particles:
() Icelandic; Den Dikken (: )
a. Ég gaf (*upp) Maríu (upp) símanúmerið mitt (*upp).
I gave up Maríu up phone_number my up
‘I gave Mary my phone number.’
b. Í gær hafa þeir sent (*upp) strákunum (?upp) peningana (upp).
yesterday have they sent up boy.the.PL up money.the up.
‘Yesterday the boys sent up the money.’
c. Ég hef rétt (*niður) Jóni (?niður) hamarinn (niður).
I have passed down John down hammer down
‘I have passed John the hammer.’
Moreover, resultatives may be licensed singly by a PP, without the aid of either
particle or prefix:
() Icelandic; Whelpton (: )
Báturinn flýtur undir brúna.
boat.the floats under bridge.the.ACC
‘The boat is going under the bridge floating.’
We expect, accordingly, that adjectival resultatives are allowed in Icelandic, as is
the case:
() Icelandic; Whelpton (: –)
a. Járnsmiðurinn barði málminn flatan.
blacksmith.the pounded metal(M)the.ACC.SG flat.ACC.M.SG
‘The blacksmith pounded the metal flat.’
b. Ég kýldi hann kaldann.
I punched him.ACC.SG cold.ACC.M.SG
‘I punched him out cold.’
3
As for prefix re- (rewrite, reopen, reconsider, etc.), which induces the presupposition that the state
codified by the predicate had previously existed, I will assume that it is an adjunct merged lower than v,
modifying PlaceP. See Marchand (); Keyser and Roeper (); Lehrer (); Lieber (); Marantz
(, ); and Acedo-Matellán () for discussion on the syntax and semantics of re-.
A revision of Talmy’s typology
c. Að nudda þá slétta.
to rub them.ACC.M.PL smooth.ACC.M.PL
‘To rub them smooth’.
d. [Þá] slengdi illi andinn honum flötum.
then slung evil spirit.the him.DAT.M.SG flat.DAT.M.SG
‘Then the evil spirit slung him down flat.’
e. Þeir dældu hana fulla af lyfjum.
they pumped her.ACC.F.SG full.ACC.F.SG of drugs
‘They pumped her full of drugs.’
f. Dóra æpti sig hás-a.
Dóra screamed herself.ACC.SG hoarse.ACC.F.S
‘Dóra screamed herself hoarse.’
g. [Hann] reif hurðina opna.
he tore door(F)the.ACC.SG open.ACC.F.SG
‘He tore the door open.’
h. Hann skrúbbaði pönnurnar hreinar.
he scrubbed pot(F)the.ACC.PL clean.ACC.F.PL
‘He scrubbed the pots clean.’
i. Þvo mig hreinan.
wash me.ACC.M.SG clean.ACC.M.SG
‘Wash me clean.’
However, what is most interesting about Icelandic for the present discussion is that it
presents two types of adjectival resultative constructions. We encounter those in
which the adjective is morphologically independent from the verb, as in the examples
in () above, and those in which it is prefixed to the verb, as the following examples
show:
() Icelandic, Whelpton (: , : )
a. Hann hvít-bæsti rammann.
he white-stained frame.the
‘He stained the frame white.’
b. Svart-litaður.
black-coloured.NOM.M.SG
c. Þunn-sneiddu sveppirnir.
thin-cut.NOM.M.PL mushroom(M)the.NOM.PL
‘Thin-cut mushrooms’.
d. Fín-muldu piparkornin.
fine-ground.NOM.N.PL peppercorn(N)the.NOM.PL
‘Fine-ground peppercorns’.
Strong s-framed languages
e. Hrein-skrúbbuðu pönnurnar.
clean-scrubbed.NOM.F.PL pot(F)the.NOM.PL
‘Clean-scrubbed pots’.
f. Mjúk-brædda súkkulaði.
soft-melted.NOM.N.SG chocolate(N)NOM.SG
‘Soft-melted chocolate’.
Crucially, the data in () show a correlation between prefixation of the adjective and
lack of agreement morphology, while the data in () show that when the adjective is
inflected, it is not prefixed. This could be interpreted in the following way: the
adjective in Icelandic (complex) resultative constructions is allowed to bear agree-
ment morphology. If it does, it cannot be attached to the verb, but this does not yield
a deviant ouput, since the Path is not specified as affixal in this language. When it
does not bear agreement morphology, however, it may be prefixed to the verb.4 Thus,
Icelandic subsumes two logical types of strong s-framed languages: those where the
adjective is inflected and those where it is not inflected. As one might expect, it is only
in the latter that adjectival resultatives of the English type are permitted. In this sense,
Icelandic lends support to the hypothesis that the allowance of this type of resultative
constructions depends, first, on the morphological features of Path and, second, on
the requirement that predicative adjectives bear agreement morphology.
.. Finno-Ugric
Outside Indo-European (at least) two Finno-Ugric languages are found that pattern
with Germanic in being strong s-framed languages: Finnish and Hungarian.
In Finnish the verb does not seem to require the appearance of a Path-signalling
affix in resultative constructions based on PPs or particles:
() Finnish; Fong (: ) and Kolehmainen (: )
a. Toini tanssi huonee-seen / huonee-sta.
Toini danced room-ILL room-ELA
‘Toini danced into/out of the room.’
4
Both Dutch and German feature particle verb constructions where the particle is an adjective and
appears adjacent to the verb in verb final environments. I illustrate this with Dutch:
(i) Dutch; Booij (: )
a. . . . Jan het huis schoon-makte.
Jan the house clean-made
b. Jan maakte het huis schoon.
Jan made the house clean
However, since in these languages the adjective is only inflected when used attributively, and not
predicatively (consider, for instance, Ger. Das weiss-e Buch, ‘The.NOM.N.SG white-NOM.N.SG book(N)NOM.SG’
vs Das Buch ist weiss(*e)), the formal dissociation shown by Icelandic adjectival resultatives does not
obtain.
A revision of Talmy’s typology
b. Uolevi asui täällä, mutta hän muutti pois.
Uolevi live.PST here but he move.PST away
‘Uolevi has lived here, but he has moved away.’
() Finnish; Kolehmainen (: )
Nyt sinä ammut yli!
now you shit over
‘Now you exaggerate!’
() Finnish; Kolehmainen (: )
Pekka laukoi pallot maaliin, Matti ampui yli.
Pekka fired ball.ACC.PL goal.ILL, Matti shot over
‘Pekka shot the balls into the goal, Matti, however, shot them away over it.’
Unsurprisingly, Finnish allows the formation of complex resultatives based on
adjectival predicates:
() Finnish; Levinson (: )
a. Mari joi teekannu-n tyhjä-ksi.
Mari.NOM drank teapot-ACC empty-TRANSL
‘Mari drank the teapot empty.’
b. Mari hakkasi metalli-n litteä-ksi.
Mari.NOM hammered metal-ACC flat-TRANSL
‘Mari hammered the metal flat.’
c. Mari nauroi itsensä käheä-ksi.
Mari.NOM laughed herself hoarse-TRANSL
‘Mari laughed herself hoarse.’
d. Joki jäätyi kiinteä-ksi.
river.NOM froze solid-TRANSL
‘The river froze solid.’
e. Tuuli jäädy-tti joe-n kiinteä-ksi.
wind.NOM freeze-CAUS river-ACC solid-TRANSL
‘The wind froze the river solid.’
As for Hungarian, this language possesses a set of particle-like elements that are
readily amenable to an analysis in terms of resultative particles analogous to those we
have described for other languages—see Perrot (: ) and É. Kiss (, a).
That these elements are good candidates of Path(P) is the fact that they describe the
final state of a motion event (see ()), affect the telicity of the predicate (see the
diagnostics with temporal modifiers in ()), and may introduce unselected objects
(see ()):
Strong s-framed languages
() Hungarian; Horvath (), in Julien (: ) and Hegedűs (: )
a. János ki-ment.
János out-went
‘János went out.’
b. János át-jött.
János over-came
‘János came over.’
() Hungarian; É. Kiss (: –)
a. János hétfőre *(el) olvasta a regényt.
János by_Monday PART read.PST the novel
‘János read the novel by Monday.’
b. János egész este (*el) olvasta a regényt.
János whole evening PART read.PST the novel
‘János read the novel the whole evening.’
() Hungarian; Bende-Farkas ()
A kutya *(fel-)ugatta a szomszédokat
the dog up-bark.PST the neighbour.ACC.PL
‘The dog woke the neighbours with its barking.’
Hungarian particles are not obligatorily affixed to the verb. It is true that, as shown by
É. Kiss (: ), when there is no logical operator in the sentence the particle must
form one and the same phonological word with the verb:
() Hungarian; É. Kiss (: )
János [ω fel olvasta] a verseit.
János up read.PST the poems
‘János read out his poems.’
However, there is a variety of syntactic conditions that may disrupt the morpho-
logical connection between the particle and the verb: the presence of negation (see
()a), contrastive topicalization of the particle (see ()b), or even movement into a
matrix clause (see ()c and ()d) (see also Farkas and Sadock ; Puskás :
ff.; and É. Kiss a):
() Hungarian; É. Kiss (: –)
a. Péter nem olvasta őket fel.
Péter NEG read.PST them up
‘Péter did not read them out.’
b. Fel csak János olvasta a verseit.
out only János read.PST the poems
‘Only John read his poems out loud.’
A revision of Talmy’s typology
c. János fel szeretné olvasni a verseit.
János up would_like read.INF the poems
‘János would like to read out the poems.’
d. János fel szeretném, hogy olvassa a verseit.
János up would_like.SG that read.SBJV.SG the poems
‘I would like János to read out his poems.’
In spite of these facts, É. Kiss (: ) considers the particle to originate within the
VP, a fact which is in no contradiction with the hypothesis that it originates within a
PathP, as assumed here. On this assumption, examples ()a to ()c show that Path
is not required to be prefixed to the verb. Hence, we expect adjectival resultative
constructions to be possible in Hungarian:
() Hungarian
a. Snyder (: )
A munkás lapos-ra kalapácsolta a fémet.
the worker flat-TRANSL hammer.PST the metal
‘The worker hammered the metal flat.’
b. Bende-Farkas (: )
Mari beteg-re ette magat.
Mari sick-onto eat.PST.SG self.ACC
‘Mari ate herself sick.’
c. Csirmaz (: )
János tisztá-ra mosta a ruhát.
János clean-onto washed the dress
‘János washed the dress clean.’
Note, finally, that both in Finnish and in Hungarian the adjective heading the
resultative predicate is marked with a special case: translative -ksi in Finnish (see
()) (Levinson ) and sublative -ra/-re in Hungarian in () (Marácz ). That
this case mark signals resultativity is shown by the fact that in Finnish depictive
secondary predication, unlike resultative secondary predication, requires the essive
case (-na):
() Finnish; Fong ()
a. Rakennus paloi vakuuttamattoma-na.
building burned uninsured-ESS
‘The building burnt down uninsured.’
b. Keitto nautitaan kuuma-na.
soup enjoy.PASS.PRS hot-ESS
‘The soup is to be enjoyed hot.’
A typology based on the morphology of Path. The case of Mandarin
This morphological fact fits nicely with the analysis put forward here in which the
resultative adjective, encoding final state, is embedded within a PathP: the translative
case of Hungarian and Finnish would correspond to the Vocabulary Item for the
Path head, to which Place raises—see section .. for the same analysis of AP
resultative constructions in English. I illustrate this with the PF-derivation of the
Finnish complex AP resultative of ()a:
() PF-derivation of ()a
a. Structure delivered by syntax
VoiceP
Mari Voice’
Voice vP
v PathP
v jo teekannun Path’
Path PlaceP
teekannun Place’
Place
b. Raising
[v JO v] [Path [Place TYHJÄ Place] Path]
c. Linearization
JO-v > TYHJÄ-Place-Path
d. Vocabulary Insertion
jo-∅ > tyhjä-∅-ksi
. A typology of languages based on the morphology of Path.
Mandarin as a weak s-framed language
To sum up, the Path head can be argued to be marked as affixal or not, and,
within the former category, as not strictly adjacent to v or as strictly adjacent to v.
This morphological specification produces a fine-grained typology that
captures the cross-linguistic variation in the expression of complex resultative
A revision of Talmy’s typology
predicates more precisely than Talmy (: –). It is schematized in the
table below:
() A revision of Talmy’s () typology
Strong s-framed languages: Germanic,
Non-affixal Path
Finno-Ugric
not strictly adjacent Weak s-framed languages: Latin, Slavic,
Affixal to v Ancient Greek
Path V-framed languages: Romance, Modern
strictly adjacent to v
Greek, Basque
Strong s-framed languages, with a non-affixal Path, do not require (although they
allow) the univerbation of v and PathP, so these two elements may be realized
independently. As a result, constructions in which the v head is associated with a
root encoding a Co-event are predicted as being possible, since this association does
not interfere with the morphological realization of the material in PathP.
In weak s-framed languages Path has generally a null exponent and must be affixed
to a span of nodes containing the verb and the rest of the material in PathP (see
section ..). This has the effect of ruling out structures in which the material in
PathP may be realized independently, such as adjectival resultative constructions. It
is only when Path is strictly adjacent to v, that is, in simple resultative constructions,
that, although prefixed onto v, it receives an exponent independently of the realiza-
tion of PlaceP, which can thus be realized as an adjective (see section ..).
Finally, in v-framed languages, Path has a null exponent and must be strictly adjacent
to v. This has the effect of precluding the presence of any Co-event root associated with v,
since such a root would be linearized between Path and v. It also has the effect of yielding
the effect that Path and v are always realized as the same morph. V-framed languages
admit constructions with PPs and APs encoding a resulting location or state provided
that the above-mentioned condition on the realization of Path is met. Thus, locative PPs
with simple verbs like Cat. anar ‘go’ and resultative APs with simple causative verbs like
Cat. deixar ‘leave’ are fine in these languages (see section ...).
Importantly, the properties of the Path head may interfere with independent properties
of the language to yield either allowance or disallowance of complex resultative construc-
tions based on APs. It has been argued that in Latin and Slavic the fact that adjectives are
inflected for agreement makes them constitute independent Spell-Out cycles. Thus, when
an AP is merged as PlaceP in a transition predicate, the univerbation requirement pointed
out above, and encoded in the Vocabulary Item of Path, cannot be met. Indirect evidence
of the role of agreement inflection in the adjective in licensing or not adjectival resultative
constructions comes from the strong s-framed language Icelandic, since, although the
adjective is allowed to be prefixed onto the verb, this happens only when it is not inflected.
A typology based on the morphology of Path. The case of Mandarin
Nothing precludes, of course, the existence of s-framed languages with an affixal
Path, i.e., weak s-framed languages, and no inflection in the predicative adjective.
These languages should not feature independent, i.e., non-affixal particles, although
they could feature resultative constructions based on affixed adjectives. Some var-
ieties of Mandarin Chinese seem to conform to this pattern. I base my observation on
the description carried out by Fan (). This author argues that both in directed
motion constructions with a manner component and in adjectival resultative con-
structions the element expressing the resulting location (jin ‘enter, in’) or state (bian
‘flat’) must be morphologically attached to the verb expressing a Co-event:
() Chinese; Fan (: )
Zhangsan pao-jin-le fangjian.
Zhangsan run-enter-le room
‘Zhangsan ran into the room.’
() Chinese; Fan (: )
Zhangsan qiao-bian guanzi.
Zhangsan hammer-flat can
‘Zhangsan hammered the can flat.’
Importantly, Fan (: ff.) argues that in CDMCs like that in the example above,
the locative element does not correspond to the main verb, which would conform
rather to a v-framed pattern. Therefore, both examples involve an s-framed pattern.
The crucial fact is that Mandarin adjectives do not show agreement morphology. Thus,
in the following copular sentence huang ‘yellow’ is not inflected for agreement with yezi
‘leaf ’—the particle hen, as pointed out by Fan (), is necessary for the construal of
huang ‘yellow’ as a state, but it certainly does not correspond to phi-features:
() Chinese; Fan (: )
Yezi *(hen) huang.
leaf hen yellow
‘The leave was yellow.’
That the adjective is not inflected for agreement would explain why it can be freely
attached to the verb, much in the same way as locative elements such as jin ‘in’.5
The typology emerging turns out to be a revision of Talmy’s (: –) in that
the word/non-word divide, that is, a divide based on whether the Path and v are
univerbated or not, comes to the fore as an important one. The univerbation factor
can be argued to be important from a diachronic point of view. Thus, the morpho-
logical specifications of Path as proposed in this typology fall into a phased sequence
5
See Talmy (), Peyraube (), Xu (), Mateu (), and Fan () for arguments that
Chinese is an s-framed language. For other divergent analyses see Tai (), Slobin (, ), and
Chen and Guo ().
A revision of Talmy’s typology
if, as discussed in Acedo-Matellán and Mateu (), Path starts out as an inde-
pendent element, then it optionally attaches onto the verb, afterwards the affixation is
obligatory, only allowing the stranding of PlaceP when Path is strictly adjacent to the
verb, and, lastly, it becomes phonologically indistinguishable from the verb, this last
option modelled here as emerging from a requirement of strict adjacency between
Path and v.6
. Previous approaches and possible counterexamples
In this section I summarize and revise some of the works that have, to different extents,
dealt with the cross-linguistically uneven availability of complex resultatives based on
APs and those based on PPs/particles. In addition, I confront some of the empirical
problems that their data pose for my own account, and try to propose a solution.
.. Snyder (, ), Beck and Snyder (a)
Snyder (, ) has proposed that a necessary—but, crucially, not sufficient—
condition for a language to admit complex predicates such as particle verb construc-
tions, (adjectival) resultative constructions, or double object constructions is the
availability of productive endocentric root compounding, regulated by the Com-
pounding Parameter:
() Compounding parameter: The grammar {disallows*, allows} formation of
endocentric compounds during the syntactic derivation. [*unmarked value]
(Snyder : )
Thus, languages like English, positively marked for this parameter, may generate all
the above complex predicates because they can also productively generate non-
idiosyncratically interpreted compounds like banana box. On the other hand, lan-
guages like Catalan, negatively marked for the Compounding Parameter (consider,
for instance, the ungrammaticality of *plàtan caixa/*caixa plàtan ‘banana box’/‘box
banana’) cannot form complex predicates either. On the basis of the idea developed
by Neeleman and Weerman () and Neeleman () for Dutch and by Le Roux
() for Afrikaans that in these languages verb particle predicates and adjectival
resultatives must be analysed as compounds of the verb and the particle or adjective,
Snyder proposes that all complex predicates, for him a natural class, must be treated
as compounds. The proposal is then underpinned by data from a wide range of
6
See Eythórsson () and Acedo-Matellán and Mateu (, ) for similar considerations, and
Verkerk () for discussion of the status of proto-Indo-European as an s-framed language on the basis of
a phylogenetic comparative study. See also Coleman (), Untermann (), Vincent (), and
Oniga () for other theories of the emergence of preverbs in Latin. See Stolova () for a lexical study
of Late Latin as an intermediate stage between s-framed Latin and v-framed Romance.
Previous approaches and possible counterexamples
languages, where a correlation is shown to obtain between availability of productive
root compounding and of complex predicates. In turn, Beck and Snyder (a)
extend this analysis to CDMCs, implying a correlation between the availability of
CDMCs and adjectival resultative constructions. Importantly, Beck and Snyder
(a) and Snyder (, ) introduce the qualification that languages allowing
the formation of complex predicates must feature, in addition to a positive setting of
the Compounding Parameter, some semantic rule allowing the semantic integration
of the elements of a complex predicate.
Several authors (Horrocks and Stavrou ; Son ; Son and Svenonius ;
Padrosa-Trias ) have pointed out the empirical problems of Snyder’s (, )
and Beck and Snyder’s (a) proposal. In particular, a cross-linguistically attested
double dissociation between compounding and complex predicate formation can be
shown to thwart the predictions of the analysis. On the one hand, languages like
Basque or Modern Greek allow productive root compounding and simultaneously
disallow complex predicates. In this case, it could be adduced that the relevant
languages do not possess the above-mentioned semantic operation allowing them to
construe complex predicates. On the other hand, however, there are languages like
Slavic or Latin, which do not productively generate root compounds but which admit
at least a set of complex predicates, namely those in which the result predicate is
prefixed. Other problems include the virtual unavailability of overt VV compounds in
Germanic, a group of languages well known for their allowance of complex predicates.
More importantly for the discussion to which the current chapter is dedicated,
there does not exist a compulsory correlation of complex predicates based on
particles or PPs and those based on adjectives, as we have seen in the case of Latin,
Slavic, and Ancient Greek, and as argued by Son () for Korean, Japanese,
Hebrew, Czech, and Javanese (see section ..).
.. Horrocks and Stavrou (, ) and Horrocks ()
Horrocks and Stavrou (, ) and Horrocks () propose an account of the
(un)availability of complex resultative constructions (as understood here) in terms of
the presence/absence of a grammaticalized opposition of perfective and imperfective
in viewpoint aspect. In particular, they observe that languages that allow complex
predicates do not possess a grammaticalized opposition of these two aspectual
interpretations (English), while languages that disallow them (Ancient and Modern
Greek, Romance) do possess such a grammaticalized opposition. They argue that,
although viewpoint aspect and inner aspect are independent dimensions of aspec-
tuality, the interpretation of a perfective or an imperfective form of a verb depends in
part on its inner aspect. In particular, perfective forms always describe complete
eventive wholes with initial and final bounds. However, the perfective form of an
atelic verb conveys arbitrary bounds and no reference to the internal contour of the
A revision of Talmy’s typology
event. The perfective form of a telic verb, on the other hand, presents a final temporal
bound which coincides with the telos inherent in the lexical semantics of the verb. As
regards imperfective forms, they imply viewing the event without bounds. But whereas
imperfective atelic verbs convey an event unfolding continuously or randomly, imper-
fective telic verbs entail an incrementality towards a goal which, crucially, is not
implied to be attained. Consequently, for Horrocks and Stavrou () inner aspect
has to be determined before grammatical aspect, in order to compute an overall
aspectual value for the verb. But, since, in languages like Ancient or Modern Greek,
either the perfective or imperfective form or stem—in the many cases of suppletion—
has to be chosen before inserting the verb into the tree (given that quite often one form
is not predictable from the other), it follows that the inner-aspectual value must have
also been determined before that insertion, and cannot interact with the syntactic
environment of the predicate. This is the reason why in languages where that mor-
phological/lexical choice is forced, the semantics of a non-terminative verbal lexeme
cannot interact with syntactic material, such as a goal PP or a resultative adjective, to be
rendered terminative. By contrast, in languages lacking such grammaticalized perfect-
ive/imperfective opposition the overall aspectual value of the verb is not fixed when it is
inserted in the tree and can, therefore, interact with the syntax in constructions such as
adjectival resultative constructions and CDMCs.
I detect two incompatibilities between these accounts and my own. The first
concerns the existence of languages with a grammaticalized opposition between
perfective and imperfective that do license, however, complex resultative construc-
tions, pace Horrocks and Stavrou (, ) and Horrocks (): Latin, Slavic,
and Ancient Greek. The fact that in these languages those complex resultative
constructions are always based on prefixed particles does not make them less
complex resultative constructions, with a resulting state/location encoded by the
prefix and a differentiated event leading to it encoded by the verb.
The second problem is restricted to the accounts in Horrocks and Stavrou ()
and Horrocks (), and not to that in Horrocks and Stavrou (). In these analyses
a dissociation is made between the availability of adjectival resultative constructions
and that of CDMCs. While the availability of the former depends, as we have seen, on
the absence of a grammaticalized perfective/imperfective opposition, the availability of
the latter is subject to the fact that the language in question possesses the formal means
to unambiguously express goals (that is, telic Paths) in PPs.7 In particular, Ancient
Greek and English are shown to be able to express bounded Paths with dedicated
prepositions (English to, Ancient Greek eis ‘into’) and, in the case of Ancient Greek,
(accusative) case. Thus, in Ancient Greek predicates headed by a manner-of-motion
verb and accompanied by a goal-encoding PP, the verbs are claimed to be reclassified
7
An analysis based on the lexical availability of particular prepositions is also that adopted by Folli and
Ramchand (, ), Son (), and Son and Svenonius ().
Previous approaches and possible counterexamples
as unaccusatives through a change in their lexical representation. That reclassification
can be carried out either through the addition of a ‘directional’ prefix—forming, as
Horrocks and Stavrou (: ) claim, ‘a different lexical item’—and/or the addition
of the complement goal PP. This reconversion is, crucially, not available in the case of
APs, since ‘adjectives are naturally stative, and so cannot in Greek force a directional/
transitional reading for what is basically a simple-activity verb’ (Horrocks : ).
But there is an inconsistency here in allowing a language such as Ancient Greek, with a
grammaticalized imperfective/perfective opposition, to have goal PPs interact with the
already determined aspectual value of the verb yielding telic CDMCs. On the other
hand, why should PPs in Ancient Greek be able to unaccusativize an unergative verb
and APs not? Clearly, an appeal to the presence/absence of the grammaticalized
division of aspects is unavailable, unless the dubious claim is made that adjectives
can be directional/eventive in languages like English but not in Ancient Greek. By
contrast, and as we have already seen in section ., Horrocks and Stavrou ()
propose that the availability of CDMCs is also subject to the absence of a grammat-
icalized imperfective/perfective opposition. However, although a unified explanation is
reached by Horrocks and Stavrou (), a problem remains, as I have already pointed
out above: the parallelism between English-type resultative constructions and Latin,
Ancient Greek, or Slavic prefixed resultative constructions is not accounted for.
.. Kratzer ()
Kratzer () presents an analysis of adjectival resultative constructions in terms of
a Small Clause formed by the object of the construction and the adjectival predicate.
Above the adjective an affixal null head of causative semantics, CAUSE, is merged,
accounting for the causative interpretation typical of these constructions. The adjec-
tive incorporates into CAUSE to satisfy its affixal needs, and this complex is subse-
quently merged onto the lexical verb in order to create a complex predicate and,
hence, to circumvent the lack of a selection relation between that lexical verb and the
subtree below. Her analysis of German Die Teekanne leer trinken ‘Drink the teapot
empty’, is shown in () below:
() German; Kratzer (: ) (adapted)
trinken
[cause]
die Teekanne leer
Crucially for the position expounded in this chapter and in Chapter , Kratzer also
appeals to inflectional morphology on the predicative adjective as a factor regulating
the licensing of adjectival resultatives. However, her use of this factor is different
A revision of Talmy’s typology
from mine: she contends that languages in which the predicative adjective obliga-
torily bears inflectional morphology cannot license adjectival resultatives because, for
an adjective to function as resultative, the null CAUSE morpheme must be affixed
onto it, a factor that precludes further affixation of the inflectional morphology.8
However, she herself already notes that Norwegian might be a counterexample to the
claim that the resultative adjective cannot be inflected:
() Norwegian; Åfarli (: footnote ), in Kratzer (: )
a. Vi vaska golvet rein-t/ *rein.
we washed floor.the.N.SG clean-N.SG/ clean
‘We washed the floor clean.’
b. Vi vaska rein(-t) golvet.
We washed clean floor.the.N.SG
‘We washed the floor clean.’
c. Golvet er rein-vaska/*reint-vaska.
floor.the.N.SG is clean-washed.
‘The floor is washed clean.’
In the above examples the adjective must bear inflection (see ()a) if it is not
adjacent to the verb. It optionally bears inflection when adjacent to a finite verb
(see ()b) and it cannot bear it when left-attached to the verb (see ()c). Kratzer
observes that when the adjective is overtly incorporated into the verb, as in ()b and
()c, the inflection disappears, and when it is—under her assumptions—covertly
incorporated, as in ()a, inflection is compulsory.9 This author takes the data as
suggesting that agreement morphology in (a) and (b) is a PF phenomenon, orthog-
onal to the incorporation of the adjective into CAUSE.
As was shown in section .. and already observed by Whelpton () in his
evaluation of Kratzer’s () proposal from the Icelandic perspective, Icelandic
resultatives are also built on obligatorily inflected adjectives when the adjective is
not prefixed to the verb. Furthermore, recall from section .. that resultative
adjectives in Finnish and Hungarian, although they do not bear agreement inflection,
must be endowed with a special case, which is translative in Finnish and sublative in
Hungarian. This would also be a problem for Kratzer’s proposal in the same way as is
8
Kratzer adopts Hay’s () contention that derivational affixes that can be easily parsed out should
never occur closer to the root than those that are less easily parsed out (see also Hay ). CAUSE, being
null and, hence, ranking lowest on the parsability scale, should always affix before any other (overt) affix is
added, least of all if the affix is inflectional, as agreement affixes are. But this condition can never be met
when the adjective already bears inflection before raising to CAUSE.
9
I recall that it is argued that the adjective incorporates into an upper null CAUSE head. In ()a it
remains overtly in situ, after the object; in ()b and ()c it overtly incorporates, but the linearization with
respect to the verb is different because of the presence of voice features in ()b versus their absence in ()c.
See Kratzer (: , footnote ), for a detailed explanation.
Previous approaches and possible counterexamples
inflectional morphology in the Scandinavian languages, unless the translative and
sublative suffixes were actually realizations of her CAUSE morpheme. However, as
the following example shows, Finnish translative appears in non-causative events,
which argues against an analysis as a realization of CAUSE and in further favour of
one as a realization of Path (transition):
() Finnish, Fong (: –)
Toini tuli sairaa-ksi.
Toini.NOM became ill-TRANSL
‘Toini got ill.’
Lastly, and also importantly for present concerns, Kratzer claims that adjectival
resultatives have to be studied as a phenomenon independent from similar construc-
tions involving a particle, or, as in German, a separable prefix:
() Kratzer (: –)
‘Resultatives built from verbs and adjectives must also be distinguished from
directional particle constructions, which have a causative interpretation, but
do not involve adjectives. [ . . . ] Inclusion of directionals in discussions of
resultatives has obscured important generalizations that emerge clearly once
we restrict our enterprise to resultatives built from adjectives.’
Illustration of the above claim is provided in a side discussion on the hybrid
behaviour of English open, where Kratzer points out two differences between par-
ticles and adjectives. Particles may undergo shift, while AP results may not (see ()),
and a particle-verb complex may appear separated from its object by another verb,
while that does not happen in the case of an AP resultative (see ()):
() Kratzer (: )
a. We threw (out) the documents (out).
b. They painted (?green) the barn green.
() Kratzer (: )
a. We threw out and shredded the documents.
b. ?They painted green and sold the barn.
Kratzer contends that these facts can be explained if we assume that ‘particles can,
but adjectives cannot that easily form PF-visible compounds with verbs’ (Kratzer
: ). Note that this explanation is perfectly compatible with the possibility that,
syntactically and semantically, adjectival resultative constructions and verb particle
constructions boil down to the same phenomenon. Thus, it has to be claimed that
‘inclusion of directionals in discussions of resultatives’ has sometimes brought to
light important generalizations. This is particularly striking in the work by Mateu
(a, b, ) and Mateu and Rigau (), who provide a unified
A revision of Talmy’s typology
explanation for both types of constructions in s-framed languages, and, simultan-
eously, an account of their non-existence in Romance. More strikingly still, Kratzer’s
dissociation of adjectival resultative constructions and verb particle construction fails
to explain why in (certain) v-framed languages a certain class of both adjectival
resultative contructions and a certain class of verb particle constructions are allowed,
namely those classes where the adjective/particle conveys a mere specification of the
result state encoded by the verb (see ()b and ()b). By contrast, equivalent
constructions in which this condition is not met are not allowed in the same
languages (see ()a and ()a)—see also section ...:10
() Italian; Napoli (), in Washio (: )
a. *Gianni ha martellato il metallo piatto.
Gianni has hammered the.M.SG metal(M)SG flat.M.SG
b. Gli operai hanno caricato il camion pieno.
the workers have loaded the.M.SG truck(M)SG full.M.SG
‘The workers have loaded the truck full.’
() Italian; Mateu and Rigau (: )
a. *Gianni è danzato via.
Gianni is dance.PTCP.PFV.M.SG away
‘Gianni danced away.’
b. Gianni è corso via.
Gianni is run.PTCP.PFV.M.SG away
‘Gianni ran away.’
.. Svenonius ()
Svenonius () proposes an analysis of the fact that Germanic allows AP and
PP/particle resultatives, while Slavic only allows the latter type. He assumes a vP
structure for resultative constructions along the lines of Ramchand (), as
shown in ():
() Ramchand’s () analysis of the resultative VP (adopted in Svenonius )
[VP V [RP R [XP Figure [X’ X Ground]]]]
The main feature of () is the head R, conveying resultative semantics. What
Svenonius proposes is that in Germanic languages R corresponds to a null morpheme
which may take an AP, PP, or particle as complement, while in Slavic, R is never null,
but is always instantiated as the prefix. Svenonius does not make it explicit, but
I assume that R, being always realized as one of the prefixes, has its selection
10
See also Horrocks and Stavrou (: –) for a revision of Kratzer (), fundamentally on the
basis of compounding in Modern and Ancient Greek.
Previous approaches and possible counterexamples
properties restricted, in this case to PPs. This would be why Slavic does not allow AP
resultative constructions. I see two problems with this account. First, if the prefix
originates as R, it is not easy to see how it may be interpreted as a final location or a
resulting state, as argued for by Žaucer (, ), Arsenijević (), and Gehrke
(), and in this book. Rather, it seems that the prefix should be first merged as an
adposition heading the phrase which is complement to R, or as the Ground itself. The
other objection has to do with one prediction made by Svenonius’s analysis. Sveno-
nius fails to link the obligatory prefixation of R in Slavic with the fact that these
languages do not allow AP resultatives. As a result, in principle, nothing impedes the
existence of languages with an R that selects only PPs, as does, by hypothesis, Slavic,
but which is not prefixed onto the verb. Those languages would present the gram-
maticality pattern illustrated in ():
() Complex resultative predicates in a language with a non-affixal, PP-selecting R
Sue danced into the room. (Complex resultative construction based on PPs)
Sue ran in. (Complex resultative construction based on (non-affixal) particles)
*Sue beat the metal flat. (Complex resultative construction based on APs)
If that pattern is not empirically attested, which to my knowledge it is not (see section
...), Svenonius’s analysis fails to predict it.
.. Son () and Son and Svenonius ()
The last two works I would like to consider pose direct counterexamples to the
predictions made by the present account of adjectival resultatives and PP/particle
resultatives, although in them it is claimed that there is no obligatory correlation
between the availability of both types of construction. Son () and Son and
Svenonius () argue for an anti-macroparametric account of the Talmian typ-
ology on the grounds of the alleged fact that the cross-linguistic variation involved in
that typology is greater and more complex than is commonly acknowledged. Accord-
ingly, they claim that a microparametric account based on a scrutiny of the lexical
features of the items involved (mostly adpositions) should be adopted, instead. In
particular for present concerns, they present a series of languages as demonstrating a
dissociation between the licensing of PP resultatives (CDMCs) and that of adjectival
resultative constructions. Thus, according to Son () and Son and Svenonius
(), Japanese and Korean allow adjectival resultative constructions while dis-
allowing CDMCs, whereas the licensing pattern in Hebrew and Javanese is reverse.
I examine their claims for some of these languages in turn, showing how they
jeopardize my account and proposing a possible solution.
... Korean: presence of complex adjectival resultatives, absence of CDMCs Accord-
ing to Son and Svenonius () Japanese and Korean, in spite of disallowing
A revision of Talmy’s typology
complex directed motion constructions, allow complex adjectival resultative con-
structions. In () there are relevant examples from Korean:
() Korean; Son and Svenonius (: , )
a. *Mary-ka cip-ey {ttwi/kel}-ess-ta.
Mary-NOM house-LOC {run/walk}-PST-DECL
‘Mary ran/walked to the house.’
b. Inho-ka kkangthong-ul napcakha-key twutulki-ess-ta.
Inho-NOM can-ACC flat-key pound-PST-DECL
‘Inho pounded the can flat.’
The data in () seem to go directly against the analysis I have proposed, which
predicts that if a language generates AP resultatives, it will also generate PP resulta-
tives, since the basic structure is the same and it is only the derivation of AP
resultatives that can be blocked by independent features of the language: obligatory
univerbation of PathP with the verb and obligatory inflectional morphology on
predicative adjectives. The analysis would not be endangered, though, if the AP
resultative in ()b were shown not to be a complex AP resultative of the Germanic
type, like those I have examined. Here I would like to appeal to Shim and Den
Dikken’s () work on resultatives in Korean and English. An in-depth presenta-
tion of their position being impossible here, I will limit myself to focusing on their
observation that the key-suffixed AP typical of these constructions in Korean, which
is to be found also in ()b, behaves as an adjunct to VP, and, therefore, does not
qualify as a true secondary predicate of the resultative type, which is most probably
inside the vP. Shim and Den Dikken () use the diagnostics illustrated in () to
show the adjunct-status of key-APs as opposed to the vP-internal status of result APs
in English. First, Korean key-APs may be stranded under VP-replacement by the
verbal proform kuleh ‘do so’, which, as pointed out by these authors, does not allow
stranding of non-adjunct material:
() Korean; Shim and Den Dikken (: )
Jim-i meli-lul nolah-key yemsaykha-ko
Jim-NOM hair-ACC yellow-key dye-CONJ
Susana-nun ppalkah-key kuleh-ess-ta.
Susana-TOP red-key kuleh-PST-DECL
Cf. English *‘Jim dyed his hair yellow, and Susana did so red.’
Second, key-APs, unlike APs encoding a resulting state in English, may be iterated:
() Korean; Shim and Den Dikken (: )
Jim-i patak-ul hayah-key panccaki-key chilha-ess-ta.
Jim-NOM floor-ACC white-key twinkle-key paint-PST-DECL
Cf. English *‘Jim painted the floor white shiny.’
Previous approaches and possible counterexamples
This evidence shows that the key-AP does not qualify as a true resultative secondary
predicate sitting inside the vP, and, hence, that the constructions claimed by Son and
Svenonius () as complex AP resultatives in fact are not. This state of affairs is
compatible with Korean being a v-framed language, as already stated by Talmy
(: ).11
... Hebrew and Javanese: presence of CDMCs, absence of complex adjectival
resultatives Hebrew and Javanese are presented by Son () and Son and
Svenonius (), respectively, as languages allowing CDMCs and disallowing adjec-
tival resultatives. At first sight, this scenario is not problematic for my present
account, since Latin and Slavic have also been correctly predicted to behave in that
way. However, Hebrew and Javanese, unlike Latin and Slavic, show no signs of a
morphological dependence of the verb and the element expressing the Core Schema.
If they do allow CDMCs and their Path is not affixal, they should behave as strong
s-framed languages, like Germanic, allowing adjectival resultatives, contrary to data
presented by Son () and Son and Svenonius ().
Beginning with Hebrew, Son () reports the following scenario:
() Hebrew; Son (: )
a. *Hu kara et ha-xavila ptuxa.
he tore ACC the-package open
‘He tore the package open.’
b. *Hu cava et ha-kir adom.
he painted ACC the-wall red
‘He painted the wall red.’
() Hebrew; Son (: )
a. David {rac/zaxal} {la-xeder/ el ha-xeder}.
David ran/crawled DAT.DEF-room/ ALL the-room
‘David ran/crawled to the room.’
b. Ha-bakbuk caf {la-me’ara/ el ha-me’ara}.
the-bottle floated DAT.DEF-cave/ ALL the-cave
‘The bottle floated (in)to the cave.’
As for the predicates in (), Son () does not provide explicit aspectual tests to
show that they are telic, that is, that they qualify as true CDMCs in the sense
described in section ..., and neither do Son and Svenonius (), although
they too consider Hebrew to license CDMCs. As it turns out, Horrocks and Stavrou
(: ) note that ‘Beck and Snyder (b) show that an in-PP modifier is not
11
See Son (), and references therein, for more considerations on resultatives in Korean. I shall not
consider her work here, since she does not provide a rebuttal of Shim and Den Dikken’s () claims on
the status of the suffix -key.
A revision of Talmy’s typology
allowed in such cases’ (that is, in cases of predicates headed by a manner-of-motion
verb and accompanied by an alleged bounded Path PP). Note that the counterexam-
ple involves only an el-PP, not a le-DP:
() Hebrew; Beck and Snyder (b), in Horrocks and Stavrou (: )
*Dan halax el ha-kfar tox Sa’a.
Dan walked to the-village in hour
‘Dan walked to the village in an hour.’
Similarly, my Hebrew informants report that the entailment of an attainment of a
final location is licensed by the dative mark le (cf. la in (), which is le plus
definiteness) but not by el. Specifically, I report that the use of el-PPs but not that
of le-PPs seems compatible with a durative adverbial like be meshex sha’a ‘for an
hour’ in examples like ()a. So, although el-PPs can be safely called directional, it
seems doubtful that they can be considered equal to to-PPs in English, which do
induce an entailment that the location has been reached.
These findings are, within my account, in full conformity with the fact that Hebrew
appears to possess the hallmarks of v-framedness, rather than those of s-framedness
(see Slobin ). Thus, Berman and Neeman (: ff.) report that the normal
expression of paths of motion in this language is of a verbal nature, that is, that
Hebrew, much like Catalan or Modern Greek, features a set of basic monomor-
phemic verbs encoding directional motion:
() Hebrew; Berman and Neeman (: )
Root Intransitive form Causative form
k-n-s nixnas ‘enter’ hixnis ‘introduce’
y-c-ʔ yaca ‘exit’ hoci ‘take out’
y-r-d yarad ‘descend’ horid ‘take down’
‘-l-y ala ‘ascend’ he’ela ‘take up’
n-p-l nafal ‘fall’ hipil ‘drop’
The same authors observe that motion verbs in Hebrew do not ‘have the equivalent
of motion verbs like idiomatic “run into”, “run around”, “run up”. The Hebrew
counterparts of such expressions have no etymological relation to the verb rac “run”
or to each other’ (Berman and Neeman : ). Thus, typically s-framed para-
digms of complex predicates sharing the same verb and differing only in a particle or
prefix seem to be absent from Hebrew. Note, importantly, that the v-framed status of
Hebrew would not be at odds with the construction in ()a, since it features verbs
that can be construed as change-of-state verbs in Romance (notably, Italian), as has
been shown in section ... Indeed, with respect to Hebrew rac ‘ran’ we already
know that Italian correre ‘run’ admits of being construed in such a way. With respect
to Hebrew zaxal ‘crawled’—which, tellingly, for some speakers sounds more odd
Previous approaches and possible counterexamples
than rac ‘ran’ in ()a—the Italian counterpart of Hebrew, è gattonato, is also allowed
in directed motion constructions:
() Italian; Folli and Ramchand (: )
Il bambino di Gianni è gattonato a casa.
the child of Gianni is crawl.PTCP.PFV.M.SG at home.
‘Gianni’s child crawled home.’
However, when the construction in ()a is used with the root r-k-d ‘dance’ most of
my informants find the construction very odd or straightforwardly ungrammatical:
() Hebrew
*/??David rakad la-xeder
David danced DAT.DEF-room
‘David danced to the room.’
This is what we expect under the conjecture that Hebrew is in fact more similar to v-
framed Italian than to s-framed English. Thus, Folli and Ramchand (: ) report
that the Italian equivalent (danzare) is not possible in a goal construction. The same
obtains in other Romance languages:
() Catalan
*En Joan ha ballat a l’habitació. (Directional.)
the Joan has danced at the=room
() Spanish
*Juan ha bailado a la habitación. (Directional.)
Juan has danced at the room
As for example ()b, Asaf Bachrach, in a personal communication, informs me that
it sounds strange when accompanied by a le-PP (not by an el-PP). Moreover, the
next example in Son (: )—which I have enlarged with an in-adverbial to
ascertain its telicity—was judged by most of my informants as ungrammatical, and
Noam Faust pointed out that he needed the directional el before mitaxat ‘under’ to
render it possible:
() Hebrew; an enlarged example in Son (: )
*Ha-bakbuk caf mitaxat le-gesher tox shloshim shniyot.
the-bottle floated under DAT-bridge in thirty seconds
‘The bottle floated under the bridge in thirty seconds.’
Interestingly, another native informant reported that the only interpretation com-
patible with this example is one in which the floating of the bottle under the bridge
starts after seconds have passed, and not that the bottle floats for seconds until
it reaches the position under the bridge. Specifically, she spontaneously construed a
A revision of Talmy’s typology
compatible scene in which the bottle is held under the water and then released, taking
seconds to come to the surface and begin floating. This shows that Hebrew cannot
construe an accomplishment reading with the manner-of-motion verb equivalent to
float and a PP like mitaxat le gesher ‘under the bridge’. Rather, the frame adverbial
tox shloshim shniyot ‘in seconds’ can only have a start-time interpretation,
revealing that the predicate behaves as an activity, and not as a transition event
(see MacDonald ). This fact is in accordance with the hypothesis that Hebrew in
fact behaves like v-framed Italian, where galleggiare ‘float’ is also strange in a directed
motion construction, presumably because the root GALLEGGI is difficult to coerce into
a change-of-state reading—in my terms, it does not fit well as a Terminal Ground, in
Compl-Place:
() Italian; Folli and Ramchand (: )
*La barca è galleggiata sotto il ponte.
the boat(F) is float.PTCP.PFV.F.SG under the bridge.
‘The boat floated under the bridge.’
To sum up, if the qualifications just made on Son’s () data are on the right track,
Hebrew would behave like v-framed Romance, and not like s-framed Germanic: it
displays a wide range of path-verbs (cf. ()), it may mimic CDMCs with what
probably may correspond to a toward-like preposition (el), and, finally, it features
directed motion constructions in which the root of the verb is not really inserted as a
Co-event component adjoined to v, but, rather, as Compl-Place and must thereby be
interpreted as a Terminal Ground. This, of course, is pragmatically not possible for
every root, as exemplified with caf ‘floated’ in () and rakad ‘danced’ in (). In
conclusion, if Hebrew really turns out to be a v-framed language, its disallowance of
adjectival resultative constructions is, within the current framework, expected
(see () above).
According to Son and Svenonius (), Javanese (and Indonesian) does not allow
AP resultatives (see ()a), but does apparently allow CDMCs where the manner-of-
motion verb does not bear any affix conveying the final location of movement
(see ()b):
() Javanese; Son and Svenonius (: )
a. Mary nyacah daging *(sampek) ajur.
Mary beat meat until flat
‘Mary beat the meat until it became flat.’
b. Tika fmlaku/mlayu/mbrangkangg ning ngisor jembatan.
Tika walk/run/crawl LOC bottom bridge
‘Tika walked/ran/crawled under the bridge.’ (Both locative and directional
readings.)
Summary
The problem these data represent is the same as that discussed above with reference
to Hebrew: if Javanese is a v-framed language, ()a is expected, but not ()b. On the
other hand, if it is an s-framed language it is not clear why ()a should be out, since
there does not seem to be any morphological requirement for the result-conveying
element to be attached to the verb, as happens in Latin and Slavic. As is also the case
with Hebrew, Javanese could turn out to be a v-framed language, despite appear-
ances. In particular, two of the manner-of-motion verbs in ()b, mlayu ‘ran’ and
mbrangkangg ‘crawled’ belong to the run-class, that is, to the class of verbs that can
head change-of-state predicates in other v-framed languages like Italian or Catalan.
. Summary
In this chapter I have explored the empirical validity of the Split S-framedness
Hypothesis stated in Chapter , characterizing Ancient Greek as a weak s-framed
language and different Germanic languages and also Finnish and Hungarian as
strong s-framed. A typology has thus emerged, more fine-grained than that put
forward by Talmy, and based on the morphological properties of the Path head,
which may or may not lead to a univerbation of the transition vP and which may
interfere with independent factors of the language, such as agreement morphology
on the element expressing the result of the complex event. Already pointed out in
Chapter , in relation to Latin and Slavic, this interference has been explored further
in this chapter. Thus, in a strong s-framed language such as Icelandic, the resultative
adjective may appear prefixed to the verb only if it does not bear inflection. Com-
plementarily, weak s-framed languages like some varieties of Mandarin Chinese,
although requiring univerbation, as in Latin or Slavic, admit all kinds of complex
resultative constructions, since the element expressing result is never inflected for
agreement. Finally, I have critically examined previous accounts that focus on the
issue of the category of the resultative predicate, and I have tried to solve some of the
puzzles they involve for my own account.
7
Challenges and prospects
In this chapter I concentrate on two of the principal endeavours to which I hope this
work has contributed. On the one hand, the endeavour of accurately describing the
cross-linguistic variation involved in the expression of complex transition events. On
the other hand, the endeavour to reduce cross-linguistic variation in argument
structure to how PF interprets the same syntactic output. For either case, after
pointing out the main results arrived at in the book, I discuss the challenges that
my account, together with the field in a broader perspective, faces. Thus, I examine
the properties of Complex Effected Object Constructions, whose distribution correl-
ates with Talmy’s () distinction between the class of s-framed languages and that
of v-framed languages, but which plausibly do not involve the functional head that
has been considered responsible for that typological divide, namely Path. On the
other hand, I point out a challenge in the theory of the syntax-morphology interface
adopted in this work and shared in the framework of Distributed Morphology,
namely how to handle the relationship between number of PF cycles and number
of words. In relation to this issue I examine one of the Vocabulary Items proposed for
Path in weak satellite-framed languages like Latin.
. The locus of cross-linguistic variation: Talmy or Snyder?
In this book I have provided evidence supporting Talmy’s theory of lexicalization
cross-linguistically, whereby languages opt to lexicalize the Core Schema either
together with the eventive head (v-framed languages) or as an independent, non-
verbal element (s-framed languages). The main contribution that this book makes to
the understanding of the cross-linguistic variation in the expression of events of
transition is found in Chapters and . It has been argued that the class of s-framed
languages can be further split into the subclass of those languages where the Core
Schema is not affixal—strong s-framed languages—and the subclass of those where it
is affixed to the verb—weak s-framed languages. This split has allowed me to explain
why only the former type of language allows complex AP and PP resultatives. I have
argued, in sum, that the head encoding transition, Path, is crucial in the cross-
linguistic variation in the expression of complex events of change, following a line
The Morphosyntax of Transitions. First edition. Víctor Acedo-Matellán.
© Víctor Acedo-Matellán . First published by Oxford University Press.
The locus of cross-linguistic variation: Talmy or Snyder?
of research represented by Klipple (), Mateu (), Mateu and Rigau (),
and Real Puigdollers (, ), among others. In particular, if the realization of
Path depends on that of v, this eventive head cannot be combined with a Co-event
root. However, as pointed out in section .., and developed at length in Acedo-
Matellán and Mateu (), there is another major line of research on the type of
variation dealt with in this book, namely that initiated by Snyder (). Indeed, a
range of authors such as McIntyre (), Zubizarreta and Oh (), and Mateu
() have adopted the idea that the free use of verbs expressing manner in
constructions encoding a transition in s-framed languages is due to the availability
of the compounding of the eventive head with a manner root. Thus, Mateu (:
), for instance, argues that languages like English are characterized by allowing
‘compounding of a root with a null light verb during the syntactic derivation’. The
main reason to adopt a Snyderian rather than a Talmian perspective on the cross-
linguistic variation at hand is the existence, only in languages allowing a ‘Co-event
conflation pattern’, of constructions showing this pattern but arguably not any
directional/resultative component:
() Mateu (: )
a. John smiled his thanks.
b. The factory horns sirened midday.
In the rest of this section I will show that these constructions were also available in
Latin, and I will examine their properties. I will then show, following Mateu (,
), that they are absent in Romance, and I will discuss the theoretical implications
involved.
.. Complex Effected Object Constructions in English and Latin
I use the label Complex Effected Object Construction (CEOC) to refer to constructions
such as those ones above, which involve an object interpreted as a created object and
a verb that specifies the way the event is carried out.1 In their seminal work, Levin
and Rapoport () included this type of construction in the class involving lexical
subordination, which, in the present terms, corresponds to the adjunction of a
root to v. In ()a the object his thanks is the entity resulting from or being expressed
by a smiling event, and in ()b the object midday is likewise expressed by the factory
sirens.
1
Levinson (: ) introduces the difference between explicit creation verbs and implicit creation
verbs. In the former, an example of which could be bake (a cake), the created object is expressed as an
argument of the verb, while in the latter the created object appears to be the very root of the verb. Thus, in
Mary braided her hair, a braid is entailed to be created when the event comes to conclusion, but an actual
braid is not expressed as an argument of the verb. Here I will only deal with explicit creation predicates.
Challenges and prospects
The Latin examples following are cases of CEOCs and are a further argument in
favour of aligning Latin with s-framed languages, rather than v-framed ones. The last
three of them are adapted from Lemaire ()—see also Haudry () for relevant
remarks on sterno ‘strew’:
() Latin; Cic. Fin. , ,
Qui alteri misceat mulsum.
who.NOM another.DAT mix.SBJV.SG honeyed_wine.ACC
‘He who makes honeyed wine for someone else’.
() Latin; Cic. Mil.
Vulnus [ . . . ] quod acu punctum.
wound(N)NOM.SG which.NOM.N.SG needle.ABL puncture.PTCP.PFV.NOM.N
videretur.
seem.IPFV.SBJV.SG
‘A wound that seemed to have been punctured with a needle’.
() Latin; Ov. Met. ,
[Serpens] volubilibus squamosos
snake.NOM looping.ABL.M.PL scaly.ACC.M.PL
nexibus orbes torquet.
writhing(M)ABL.PL coil(M)ACC.PL twist.SG
‘The snake twists his scaly coils in looping writhings.’
() Latin; Liv. , ,
Viam silice sternendam [ . . . ] locauerunt.
way(F)ACC flint-stone.ABL strew.PTCP.FUT.PASS.ACC.F establish.PRF.PL
‘They established that the way was to be paved with flint stone.’
() Latin; Stat. Theb. ,
Aeriam truncis [ . . . ] cumulare pyram.
high.ACC.F log.ABL.PL gather.INF pyre(F)ACC
‘To build a high pyre out of logs’.
In all these examples the verb is used as a manner modification of a creation
event. Thus, in (), the DP mulsum ‘mixed wine’ is not mixed with anything, but
is rather the result of a mixing process, and, hence, does not exist before that
process. It is crucial to bear in mind that mulsum refers to a mixture of liquids
(specifically, wine and honey), in contrast to merum, which means ‘pure,
unmixed wine’: mulsum denotes, undoubtedly, the result of the event specified
by the verb, namely, mixing. In the same way, a wound (vulnus) appears
through puncturing (see ()), the snakes’ coils (orbes) appear through twisting
(see ()), the way (viam) is created by a strewing action (see ()) and a pyre
(pyram) is created by accumulating (trunks) (see ()). Importantly, there is a
The locus of cross-linguistic variation: Talmy or Snyder?
non-creation use of all these five verbs, which does not elicit the effected object
interpretation:
() Change-of-state uses of the verbs in () to ()
a. Latin; Hor. Sat. , ,
Surrentina [ . . . ] miscet faece Falerna vina.
Surrentine.ACC.N.PL mix.SG dregs.ABL Falernian.ABL wine(N)ACC.PL
‘He mixes Surrentine wines with Falernian dregs.’
b. Latin; Cels. ,
Cutis debet [ . . . ] acu pungi.
skin.NOM must.SG needle.ABL puncture.INF.PASS
‘The skin must be punctured with a needle.’
c. Latin; Ov. Met. ,
Stamina pollice torque.
yarn.ACC thumb.ABL wind.IMP.SG
‘Spin the yarn with your thumb.’
d. Latin; Liv. , ,
Sternunt corpora.
strew.PL body.ACC.PL
‘They lay their own bodies down.’
e. Latin; Liv. , ,
Vivi mortuis [ . . . ] cumularentur.
alive.NOM.M.PL dead.DAT.M.PL heap.IPFV.SBJV.PASS.PL
‘Those alive would heap up onto the dead.’
This double use of the verbs can be argued to constitute a case of elasticity of
the verbal meaning, akin to that shown by English bake, also usable as a creation
and as a change-of-state verb (see Atkins et al. ; Pustejovsky , ; and
Mateu ).
Re-prefixation allows us to explore further the semantics of CEOCs in Latin, as will
be done for English in the next section. Consider the following example:
() Latin; Verg. Aen. ,
Re-coquont patrios fornacibus enses.
re-forge.PL paternal.ACC.M.PL furnace.ABL.PL sword(M)ACC.PL
‘They forge the forefathers’ swords anew in the furnaces.’
In this example—where COQU, referring to the submission of an object to the action of
fire, means ‘forge’—a repetitive reading of re- involving two forgings of the same
(token) swords is impossible. Specifically, we must understand that new tokens of the
same type of sword are created as a result of a forging event. The verb is, therefore,
not used as a change-of-state predicate.
Challenges and prospects
.. Absence of CEOCs in v-framed languages. Theoretical implications
CEOCs do not seem to obtain in Romance or v-framed languages in general, as
observed by Mateu (, ). In particular, while sentences such as John baked the
cake are ambiguous between a creation interpretation and a change-of-state inter-
pretation (awkward, for world knowledge reasons), their v-framed counterparts only
license the change-of-state interpretation:
() S-framed English: John baked the cake.
a. ! John created a cake through baking.
b. ! John submitted an already made cake to a baking action.
() V-framed Spanish: John horneó el pastel.
a. ¬ ! John created a cake through baking.
b. ! John submitted an already made cake to a baking action.
Importantly, Spanish el pastel ‘the cake’ necessarily refers to an entity that exists
before the process identified by hornear ‘bake’, and the sentence has a thematic
paraphrase in ():
() Spanish
Lo que le hizo John al pastel fue hornearlo.
what DAT.SG did.SG John to=the cake be.PRF.SG bake.INF=it.ACC
‘What John did to the cake was bake it.’
Similarly, as pointed out by Marantz () and tested in () for Latin, a creation
predicate like John baked the cake, when combined with a re- prefix in John rebaked
the cake, yields a reading where there is a creation of another token of the same type
of cake, rather than a double baking process exerted on the same (token) cake. Thus,
it is possible to say John baked the cake but he did not like it, so he threw it away and
rebaked it. A Spanish rendition of this sentence, involving the adverb de nuevo
‘again’, is very odd:
() Spanish
#John horneó el pastel pero no le gustó;
John baked.SG the cake but not DAT.SG pleased.SG
así que lo tiró y lo horneó de nuevo.
so it.ACC threw_away.SG and it.ACC baked.SG again
The reason for the oddness of () is precisely that in Spanish hornear ‘bake’, being
exclusively a change-of-state predicate, can only appear in predicates linked to a
presupposition that the entity referred to by the object (el pastel ‘the cake’) exists
before the event named by hornear ‘bake’. In other words, in Spanish the creation
reading, which allows the type reading of the object, is incompatible with a change-
of-state verb like hornear ‘bake’.
The locus of cross-linguistic variation: Talmy or Snyder?
If we assume, with Mateu (), that the cross-linguistic variation involved in
CEOCs is akin to that dealt with in previous chapters of this book, it is not clear, on
present assumptions, why it takes place, that is, why v-framed languages do not allow
CEOCs. Note, in fact, that the creation semantics make these constructions similar to
unergative ones such as John smiled, which would receive the following analysis
according to the theory of activity predicates expounded in section ...:
() [VoiceP John [Voice’ Voice [vP v SMILE]]]
Likewise, in the spirit of Levin and Rapoport () and Mateu (), we could
consider that in a sentence like John smiled his thanks or the Latin sentence in (), the
roots SMILE and MISC ‘mix’ are a kind of adjunct to the creation event. In our
vocabulary, the resulting analyses are as follows:
() [VoiceP John [Voice’ Voice [vP [v SMILE] his thanks]]]
() [VoiceP Qui [Voice’ Voice [vP [v v MISC] mulsum]]]
In this analysis the root is merged as an adjunct to the eventive head, and the object is
directly merged as an Effected Object at Compl-v. Crucially, these constructions do
not involve a Path projection, so it is unexpected, from a purely Talmian perspective,
that they are banned in v-framed languages like Romance. Indeed, as I argued in
section ..., in v-framed languages the head v cannot associate with an independ-
ent Co-event root in structures involving a PathP, since this association disrupts the
strict linear adjacency of Path and v at Vocabulary Insertion. Crucially, this strict
linear adjacency is required for the exponent of Path to be inserted in these lan-
guages. It is precisely for this reason that a Snyderian explanation of the cross-
linguistic variation involved, which does not take into account the properties of
Path, would seem to fare better here. For instance, Mateu () proposes an updated
version of Snyder’s () Compounding Parameter that is restricted to the com-
pounding possibilities of the null light verb (thus avoiding the problems of Snyder’s
account pointed out in section ..):
() Mateu’s (: ) version of Snyder’s () Compounding Parameter
The grammar {disallows*, allows} conflation/compounding of a root with a
null light verb during the syntactic derivation. [*unmarked value]
According to this parameter, the structures of () and () are impossible in languages
like Romance because they involve the combination of v with a root. However, this
account is not free of problems. From a theoretical point of view, and, particularly,
under the perspective that cross-linguistic differences stem solely from properties of the
lexicon, it is not clear why v-framed languages should disallow the combination of these
two elements. From an empirical perspective, this analysis precludes the possibility that
in v-framed languages there are instances of such combinations. Specifically, I have
Challenges and prospects
argued in section .., that such combinations are possible when there is no Path
involved, in existential predicates with a locative expression:
() Catalan; Mateu (: )
En aquesta coral n’hi canten molts, de nens.
in this choir PARTVE=LOC sing..PL many.PL of child.PL
‘There are many children who sing in this choir.’
The fact that it is in precisely these plausibly Path-less constructions that conflation is
allowed in v-framed languages suggests, on the contrary, that the v-/s-framed
distinction is linked to the presence of Path and its expression.
Future accounts of the cross-linguistic variation dealt with by Talmy and Snyder
must reconcile both views and find a unifying explanation that overcomes the
problems pointed out above.
. Reducing cross-linguistic variation to PF properties
The other main endeavour of this book has been to contribute to the project of
deriving cross-linguistic differences from differences in the PF interpretation of the
same syntactic structure, by focusing on the case of Talmian variation. On the basis of
the data from Latin, I have concurred with previous researchers in arguing that this
variation is systematic, despite apparent counterexamples like those dealt with in
section .., and must thus be handled through abstract properties of the languages
in question. In particular, I have related this variation to properties of functional
heads, and not to properties of particular lexical items with conceptual content. In
this sense the approach presented here contrasts with that developed in works by
Folli and Ramchand (), Fábregas (), Ramchand (), and Pantcheva
(), among others, in which the lexical items corresponding to particular verbs
are marked according to the nodes of the tree that they can lexicalize. For instance, in
Folli and Ramchand () it is proposed that Italian does not have a lexical item
corresponding to the node encoding goal (of motion), namely R(esult). This forces
some other lexical item to realize that node, such as verbs of directed motion like
andare ‘go’ or correre ‘run’, endowed with the feature þR. Other verbs like ballare
‘dance’ or galleggiare ‘float’ cannot realize this node, since they do not contain the
feature þR in their lexical representation. As a consequence, a verb like correre ‘run’,
but not a verb like ballare ‘dance’, can be used in a directed motion construction with
a locative PP (see section ..). However, in languages like English there is a
dedicated preposition for the node R, so constructions expressing directed motion
can feature manner-of-motion verbs, like dance, which clearly do not incorporate the
feature þR. In this lexical-marking type of approach there does not seem to be an
expressed relationship between the semantic content of the verb and its ability to
realize a particular node. Thus, it is left as a coincidence that Italian correre ‘run’ or
Reducing cross-linguistic variation to PF properties
rotolare ‘roll’ possesses a þR feature while galleggiare ‘float’ or ballare ‘dance’ does
not, which raises the question of whether this is a clear tendency in all languages and
not a quirk of the Italian lexicon.
There is another dimension in which the present approach contrasts with those
developed within the nanosyntactic framework (Starke ). As pointed out by
Borer (: ff.), in works such as the above-mentioned the very notion of
substantive terminal or root disappears. Rather, substantive lexical items are stored
chunks of structure to be inserted into a stretch of functional nodes at Spell-Out. In
my approach, however, there is a semantic motivation for locating the root contrib-
uting the relevant exponent at a given node. See the discussion in section .. and
also in Acedo-Matellán and Mateu ().
Another significant feature of the analysis of cross-linguistic variation presented in
section .. is the assumption that derivations can crash at PF. Such a crash stems
from the failure to insert the exponent of a given node, a failure caused by, first, the fact
that the node does not possess a default exponent and, second, by the fact that the
particular environment in which the node appears does not match the insertion frame
of (any of) its Vocabulary Items. The crashing character of the theory of the syntax-
morphology interface adopted in this work is not shared by other DM-based theories
but is common to Nanosyntax (cf. Fábregas’s Exhaustive Lexicalization Principle).
Maybe the main challenge to be addressed in theories assuming the realization of
individual terminals, like that adopted here, has to do with the relationship between
Spell-Out domains and words. Indeed, it does not seem the case that a given Spell-Out
domain, say that defined by the cyclic head v, always yields one single word. Embick
() does not address this problem, beyond pointing out that complex heads are
yielded either by syntactic head movement or by certain PF operations (Embick :
–). However, this cannot be the whole story, even in the cases discussed by this
author, such as the realization of inflectional material in the verb. In particular,
certain functional heads of the same v cycle may be realized as either affixes or words.
For instance, in Greek T is realized as an affix in the present tense (-o, -is, -i, etc.) but
as a non-affixal unit (a proclitic; cf. Oostendorp ) in the future tense: tha. To
compound the problem, it is not the case that v does not raise to T when T is future,
as shown by the fact that Greek tha, unlike English will/shall, does not allow the
intervention of adverbs between it and the verb:
() Greek; based on Rivero (: )
O Yánis tha (*akómi) milái (akómi).
the Yánis FUT still speak.SG still
(Cf. English Yánis will still speak.)
According to a standard analysis, the verb moves to the future T by head movement
(Alexiadou and Anagnostopoulou ). However, this movement clearly does not
feed affixation of T to v when T is future.
Challenges and prospects
The issue affects, of course, the kind of data dealt with in this book. For example,
I have claimed that in strong s-framed languages like English the same cycle headed
by v can yield either one or more than one word:
() One word vs more than one word for the v Spell-Out cycle
a. Sue [v-cycle flattened] the metal.
b. Sue [v-cycle hammered flat] the metal.
Strong s-framed languages have been characterized as not requiring the affixation of
the Path head onto v, as in weak s-framed and v-framed languages. This circum-
stance is precisely what explains the existence of complex AP resultatives such as that
in the above example. So, if Path is not specified to attach to v, the question arises
why it does so, seemingly, in ()a and not in ()b. One way to go is to stipulate that,
universally, roots raise as high as they can, that is, to the highest functional node that
is not associated itself with a root. This means that in ()a, but not in ()b, all the
nodes c-commanded by v must successively raise onto v. Place and Path receive a zero
exponent, so the exponent of v can be determined by that of the root: it is en when the
root is FLAT and it is a null exponent when the root is, for instance, MELT (yielding melt).
In ()b there is raising as far as Path, since v is already associated with a root that
raises onto it, yielding hammer. However, this explanation does not straightforwardly
account for the cases in which v is realized as a (non-affixal) light verb:
() Sue got the metal flat.
The stipulation introduced above forces the root FLAT to end up affixed onto v, which
precludes the insertion of the exponent get. If we maintain that there are, in fact, light
verbs, and that get does not involve a root, the above stipulation could be eliminated
in favour of a laissez-faire scenario in which the root can either raise as high as v or
not. If it does not raise, get is inserted into v, as a last resort option. If the root raises, it
yields the synthetic verb. In future research on the issue of the synthetic/analytic
expression of PF cycles it will be crucial, therefore, to explore the extent to which
roots universally raise to the c-commanding functional heads.2
The issue of the raising of roots to the upper functional heads is also involved in a
plausible, more natural derivation of Latin complex predicates headed by prefixed
verbs and of the ban on unprefixed complex resultative constructions in this lan-
guage. In section .. I argued for the next Vocabulary Item for Path in languages
like Latin and Slavic (weak satellite-framed languages):
() Path $ ∅ / Place-_-[ . . . v . . . ]
2
See Acedo-Matellán () for an analysis of French expressions like Avoir froid/faim ‘Have cold/
hunger’ in which the apparent noun (froid ‘cold’, faim ‘hunger’) is shown to be a non-incorporated root.
Reducing cross-linguistic variation to PF properties
This Vocabulary Item forces the raising of all the material in the PathP to v,
accounting for the shape of complex resultative constructions in these languages
(prefixed) and for the non-existence of complex adjectival or PP resultatives
based on unprefixed verbs. Thus, unprefixed resultative constructions are ruled
out in the language because they involve a PlaceP that, by virtue of agreement
morphology, is spelled out as an independent cycle. The insertion frame of the
Vocabulary Item for Path, which requires that Place linearly precede this head, is
not met and the derivation crashes. On the other hand, I have shown that at
least in Latin there is a case in which Path is prefixed onto the verb without the
rest of the material in PathP (Place and the root), namely, when this head is
strictly left-adjacent to v, in simple resultative constructions (see section ..).
In this case, Path receives an overt exponent (re) according to the following
Vocabulary Item:
() Path $ re /_ -v-Voice
However, there may be a more principled way to derive the same empirical results
while reducing the difference between these Vocabulary Items. Specifically,
I speculate that the Vocabulary Item in () can be simplified by eliminating from
the insertion frame the requirement that Place linearly precede Path:
() Path $ ∅ / _-[ . . . v . . . ]
If () and () are the two Vocabulary Items for Path in Latin, it can be stated that
Path must linearly precede v in this language, the difference being between an
immediate (re) or a not immediate (a null exponent) linear precedence. However,
it is true that, with no other proviso, () allows the generation of complex adjectival
resultative constructions in Latin, contrary to fact (see section ..):
() Latin
*Ovidia poculum vacu-um bibit.
Ovidia.NOM goblet(N)ACC.SG empty-ACC.N.SG drink.SG
() PF-derivation with the Vocabulary Item in ()
a. Structure delivered by syntax
[VoiceP [Voice’ Voice [vP [v v BIB] [PathP [Path’ Path {[PlaceP [Place’ Place
VACU]]}]]]]]
b. Raising
[v Path [v BIB v]]
c. Linearization
Path-BIB-v
d. Vocabulary Insertion
∅-bib-∅
Challenges and prospects
PlaceP, corresponding to the inflected adjective (vacuum ‘empty’), constitutes an
independent Spell-Out cycle. Path raises to v and, after linearization, it is realized
through a null exponent by virtue of (), yielding the overt simple verb bibit ‘drinks’.
It is at this point that a general principle must be invoked to predict the ungram-
maticality of expressions such as (). In particular, a prosodic word beginning with a
null exponent such as ∅-bib-∅ in the above example could arguably be ruled out on
more general grounds, which would allow us to keep the Vocabulary Item for Path as
stated in () and discard that in (). I draw an analogy with the degradation that
different researchers have found in certain syntactic configurations in which the
edge, that is, the specifier and head of a syntactic constituent, is phonologically
empty. Interestingly, An () has proposed a PF explanation for cases such as
the following:
() An (: )
*I believe very strongly [CP Ø [IP John liked linguistics]].
(Cf. I believe very strongly that John liked linguistics.)
An () argues that in examples such as this the adverb intervening between the
verb and the complement CP (strongly) forces the CP to be parsed as an Intonational
Phrase (Selkirk ; Nespor and Vogel ). An (: ), following Nespor and
Vogel () and Schütze (), observes that I-phrasing ‘must occur at the
juncture between two prosodic words’, which is not possible in the above example
because the edge of the Intonational Phrase (i.e., the specifier-head portion of the
CP), being null, is not aligned with the edge of any prosodic word. A counterpart of
this example involving an overt complementizer (that) meets the requirement. See
An () for more relevant examples and extended discussion. My speculation is
that an analogous lack of alignment, at a different prosodic level, could be invoked in
ruling out a prosodic word whose left edge shows a null exponent:
() Path-BIB-v-T
∅-bib-∅-it
In such a sequence the left edge of the morphological word is not aligned with the left
edge of the prosodic word. One way of meeting the alignment requirement is to raise
all the material of the PathP, including the root. Thus, prefixed verbs in Latin would
result from the prefixal character of Path, as encoded in its Vocabulary Items, and, in
the case of complex resultative constructions, from the raising of the material in
PathP to v in order to create a well-formed prosodic word. Finding independent
evidence for the alignment requirement for prosodic words speculatively proposed
here has to be left for future research.
Appendix: Latin telic predicates
with prefixed manner-of-motion
verbs (section ..)
In this appendix I collect the totality of the Latin telic predicates featuring a prefixed manner-
of-motion verb which results from the search referred to in section .. (see footnote ).
I provide just a translation for every example, without glosses. However, I italicise each
prefixed verb and the telicity-signalling expression. See footnote for the criteria established
for the search.
Telic predicates headed by prefixed curro ‘run’
() Ter. Ad.
Nunc ubi me illi non uidebit, iam huc re-curret, sat scio.
‘As soon as he does not see me there, he will run back here at once, I know well.’
() Ter. Phorm.
Vbi in gynaeceum ire occipio, puer ad me ad-currit Mida.
‘As soon as I set off for the women’s apartments, the slave Midas runs up to me.’
() Lucr. ,
Anne [ . . . ] simul ac volumus nobis oc-currit imago [ . . . ]?
‘Is it so, that images come to us as soon as we want?’
() Liv. , ,
Quod ubi videre ipsum Camillum, [ . . . ] vadentem in hostes, pro-currunt pariter omnes.
‘As soon as they see Camillus marching against the enemy, all of them run forth in like
fashion.’
() Liv. , ,
Etruscorum cohortes repente [ . . . ] Romanis oc-currunt.
‘The Etruscan cohorts suddenly run against the Romans.’
() Liv. , ,
Hic vincendum aut moriendum, milites, est, ubi primum hosti oc-curristis.
‘Here it is either win or die, soldiers, as soon as you run against the enemy.’
() Liv. , ,
Pedestris [ . . . ] repente in hostium equites in-currit.
‘The foot soldier suddenly runs against the cavalry of the enemy.’
Appendix
() Liv. , ,
Marcellus [ . . . ] ubi primum in agris pabuli copia fuit, ad Canusium Hannibali
oc-currit.
‘Marcellus, as soon as there was plenty of food on the fields, hurried against Hannibal at
Canusium.’
() Liv. , ,
Repente milites [ . . . ] ex omnibus locis [ . . . ] con-currerunt.
‘Suddenly the soldiers ran together from all places.’
() Liv. , ,
Tanta enim rabies multitudinem inuasit ut [ . . . ] repente omnes ad caedem coniugum
liberorumque dis-currerent.
‘Such a frenzy invaded the crowd that all of a sudden everybody hurried away to kill their
wives and children.’
() Liv. , ,
Subito ad arma dis-currerunt.
‘Suddenly they ran away in different directions for the weapons.’
() Liv. , ,
L. Acilium [ . . . ] iubet [ . . . ] ubi clamorem audisset, de-currere ad castra eorum.
‘He orders L. Acilius to run down to their camp as soon as he hears the call.’
() Liv. , ,
Repente [ . . . ] populus in medium de-currit.
‘Suddenly the people ran down into the middle.’
() Sall. Iug. ,
Repente Maurus [ . . . ] ad Sullam ad-currit.
‘Suddenly Maurus runs up to Sullam.’
() Caes. Civ. , ,
Statim castris exeundum atque oc-currendum putaret.
‘He thought that they had to go out of the camp and run to find them at once.’
() Caes. Gall. , ,
Hostes repente celeriterque pro-currerunt.
‘The enemy ran forth suddenly and rapidly.’
() Caes. Gall. , ,
Matres familiae repente in publicum pro-currerunt.
‘Suddenly the mothers ran forth into the streets.’
() Bell. Afr. ,
Subito [ . . . ] pedites [ . . . ] pro-currunt.
‘All of a sudden the foot soldiers run forth.’
Appendix
() Cic. Verr. Actio secunda, ,
Repente Agrigentini con-currunt.
‘Suddenly the Agrigentines appear in haste.’
() Cic. Verr. Actio secunda, ,
Subito ipse ac-currit.
‘Suddenly he himself appears in haste.’
() Cic. Verr. Actio secunda ,
Statim ac-currunt.
‘They appear in haste at once.’
() Cic. De orat. ,
Habere certos locos, qui [ . . . ] ad causam explicandam statim oc-currant.
‘To have certain topics at hand which come to mind at once to help develop the subject.’
() Cic. De orat. ,
Statim oc-currit naturali quadam prudentia [ . . . ] quid faciat causam.
‘With a little common sense it occurs to us at once what the main point of the subject is.’
() Cic. Div. ,
Istae imagines ita nobis dicto audientes sunt, ut, simul atque velimus, ac-currant?
‘Are those images so docile that come to us as soon as we want them to?’
() Cic. Att. , , ,
Ad me statim iussi re-currere.
‘I told him to come back to me at once.’
() Cic. Att. , , ,
Ita subito ac-currit ut ne Trebatium [ . . . ] possim videre.
‘He suddenly comes over, so I won’t be able to see Trebatius.’
() Cic. Fam. , , ,
Suadeo [ . . . ] ad nos quam primum re-curras.
‘I insist that you come back to us as soon as you can.’
() Cic. Fam. , , ,
Dexippo [ . . . ] imperavi statim ut re-curreret.
‘I ordered Dexippus to come back at once.’
() Cic. Fam. , , ,
In mea ne potestate ut sit spectrum tuum, ut, simul ac mihi collibitum sit de te cogitare,
illud oc-currat?
‘Is it possible for me to conjure up your spectre, so that it comes to me as soon as I think
of you?’
() Cael. Cic. Fam. , , ,
At ego, simul atque audivi, [ . . . ] ad subsellia rei oc-curro.
‘But I, as soon as I heard it, I run to the bench of the culprit.’
Appendix
() Planc. Cic. Fam. , a,
Cum primum poterit, istoc re-currere non dubitabit.
‘As soon as he is able to, he will not hesitate to come back.’
() Galba Cic. Fam. , , ,
Repente Antonius [ . . . ] suas copias de vico produxit et sine mora con-currit.
‘All of a sudden Antonius led forth his troops out of the village and attacked without
delay.’
() Sen. Contr. , ,
Subito fastidiosus raptor oc-currit.
‘Suddenly the loathsome plunderer appears in haste.’
() Curt. , ,
Quod ubi exercitus [ . . . ] conspexit, [ . . . ] con-currit.
‘The army ran up to the place as soon as they spotted this.’
() Curt. , ,
Subito [ . . . ] rex Indus [ . . . ] oc-currit.
‘Suddenly the king Indus runs to meet them.’
() Cels. ,
Ne suc-currere quidem statim sibi possunt.
‘They are not able to assist themselves immediately.’
() Sen. Dial. , , ,
Ne statim cum eo con-curram.
‘I shall not come to fight against it immediately.’
() Sen. Epist. , ,
Non statim, cum haec legeris, hoc tibi oc-curret [ . . . ]?
‘Will it not come to your mind at once, when you have read this?’
() Sen. Frg. ,
Numina vocant, quae [ . . . ] subito oc-currerent.
‘They call up spirits to come to them at once.’
() Homer.
Huic subito [ . . . ] similis Tritonia fratri oc-currens iuuenem [ . . . ] decipit.
‘Tritonia deceived the youth, appearing to him suddenly in the shape of his brother.’
() Petron. ,
Unus ex noviciis servulis subito ac-currit.
‘One of the new serfs suddenly comes up in haste.’
() Stat. Theb. ,
Subito [ . . . ] regina chori de-currit in aequum.
‘Suddenly the queen of the choir runs off into the plain.’
Appendix
() Frontin. Strat. –, , ,
Iussit [ . . . ] eum [ . . . ] statim [ . . . ] pro-currere.
‘He ordered him to run forth at once.’
() Suet. Diuus Augustus, ,
In-currenti repente fero apro [ . . . ] obiectus est.
‘He was thrown in front of a wild boar which suddenly ran against him.’
Telic predicates headed by prefixed equito ‘ride’
() Liv. , ,
Qui ubi ad-equitavit portis, cum duobus equitibus vallum intravit.
‘As soon as he rode up to the gates, he passed through the fence with two riders.’
Telic predicates headed by prefixed fluo ‘flow’
() Ov. Met. ,
Extemplo tristi medicamine tactae de-fluxere comae.
‘Her hair, touched by the horrible venom, suddenly falls off [lit. ‘flows down/away’.].’
() Liv. , ,
Fecit ut intra paucos dies sex milia peditum armatorum, quattuor equitum ad eum con-
fluerent.
‘That made that in a few days six thousand armed foot soldiers and four thousand
cavalry crowded together where he was.’
() Cic. Fam. , , ,
Ubi salutatio de-fluxit, litteris me involvo.
‘As soon as the visits stop, I bury myself in my books.’
() Val. Max. , ,
[Vires atque opes humanae] ad-fluunt subito, repente dilabuntur.
‘The vigour and the wealth of humans come suddenly in a flow, and suddenly slip asunder.’
() Cels. ,
Subito nigra alvus pro-fluxit.
‘Suddenly a black flux flows forth.’
() Colum. ,
Deinde ubi liquatum mel in subiectum alveum de-fluxit, transferetur in vasa fictilia.
‘Then as soon as the liquefied honey has flowed completely into the vessel located
underneath, it should be transferred into earthenware vessels.’
() Colum. ,
Ubi [ . . . ] oliva [ . . . ] in lutum de-fluxit [ . . . ] [aenum] calefieri debet, ut inmundae bacae
eluantur.
‘When the olives have fallen down into the mud a copper pot must be heated up to wash
off the dirty fruits.’
Appendix
Telic predicates headed by prefixed gradior ‘step, walk’
() Ter. Eun.
Sed quid hoc, quod [ . . . ] subito e-greditur Pythias?
‘But why does Pythias walk out all of a sudden?’
() Verg. Aen. ,
Quam simul ac tali persensit peste [ . . . ] Iovis coniunx, [ . . . ] talibus ad-greditur Venerem
Saturnia dictis.
‘As soon as Jupiter’s spouse sensed that she was the prey of this passion, the Saturnian
approached Venus with these words.’
() Liv. , ,
Ubi inluxit, e-greditur castris Romanus.
‘As soon as the sun comes up, the Romans walk out of their camp.’
() Liv. , ,
Extemplo [provinciam] aliam Romanam ad-grederentur.
‘That all of a sudden they attacked another Roman province.’
() Liv. , ,
Ne extemplo castra hostis ad-grederetur.
‘Lest the enemy should suddenly attack the camp.’
() Sall. Iug. ,
Romanos [ . . . ] repente ad-greditur.
‘He suddenly attacks the Romans.’
() Sall. Iug. ,
Statim in collis re-gredi.
‘To step back up to the hills at once’.
() Sall. Iug. ,
Repente magna vi murum ad-greditur.
‘Suddenly he attacks the wall with a great force.’
() Sall. Iug. ,
Statim di-grediens.
‘Suddenly stepping aside.’
() Nep. Themistocles, ,
Quos si statim ag-grederetur, uniuersos oppressurum.
‘If he attacked them suddenly he would crush them all.’
() Cic. Cluent.
Tum repente [ . . . ] L. Clodium [ . . . ] ad-greditur.
‘Then he suddenly walks up to L. Clodius.’
Appendix
() Cic. Mil.
Videte nunc illum [ . . . ] e-gredientem e villa subito.
‘See him now suddenly walking out of the villa.’
() Cic. Att. , , ,
Hanc epistulam dedimus L. Tarquitio simul e portu e-gredienti.
‘This letter we gave to L. Tarquitius as soon as he walked out of the harbour.’
() Tac. Hist. , ,
Statim re-gredi.
‘To come back at once’.
() Tac. Hist. , ,
Quod si statim con-grediantur, nullas esse Ceriali nisi e reliqui<i>s Germanici exercitus
legiones.
‘If they attacked suddenly, Cerialis had no legions except for those remaining from the
army of Germany.’
() Ter. Phorm.
Hisce ostendam me, ubi erunt e-gressi foras.
‘I will appear to them as soon as they walk out into the street.’
() Liv. , ,
Si urbem extemplo ad-gressurus Scipio foret, ita conclamatum ad arma est.
‘If Scipio suddenly attacked the city there would be a call to arms.’
() Liv. , ,
Ipsam Carthaginem repente ad-gressurum credebant.
‘They thought that he would attack Carthago itself all of a sudden.’
() Liv. , ,
Tres quadriremes, [ . . . ] quinqueremem Romanam [ . . . ] repente ad-gressae sunt.
‘Three quadriremes suddenly attacked the Roman quinquereme.’
() Liv. , ,
Tum simul ab omni parte [ . . . ] urbem est ad-gressus.
‘Then suddenly he attacked the city from all sides.’
() Liv. , ,
Extemplo e-gressi sunt.
‘They walked out all of a sudden.’
() Liv. , ,
Repente opera est ad-gressus.
‘He attacked the works all of a sudden.’
() Liv. , ,
Qua triduo ascenderat biduo est de-gressus.
‘He descended in two days, although he had ascended in three.’
Appendix
() Sall. Iug. ,
Is ubi primum magistratum in-gressus est [ . . . ] ad bellum, quod gesturus erat, animum
intendit.
‘As soon as he entered the magistracy, he devoted his attention to the war which he was
to conduct.’
() Sall. Iug. ,
Ubi eum Rutilius praeter-gressus est, paulatim suos in aequom locum deducit.
‘As soon as Rutilius passed by him, he slowly led his men to a flat space.’
() Sall. Iug. ,
Sed ubi dies coepit et Numidae [ . . . ] oppido e-gressi, repente [ . . . ] portas obsidere iubet.
‘As soon as the day began and the Numidians walked out of the citadel, he ordered them
to block the doors immediately.’
() Sall. Hist. Frg. , ,
Exercitum dimisit, ut primum Alpis di-gressus est.
‘He send the army away as soon as he departed from the Alps.’
() Caes. Civ. , ,
Statimque e-gressus et nouissimum agmen consecutus celeriter ex conspectu castrorum
discessit.
‘He suddenly marched out and, quickly joining his rearguard, walked far from the view
of the camp.’
() Bell. Afr. ,
Statim inde di-gressus Rebilo proconsule [ . . . ] Messala Vticam ante praemisso [ . . . ],
ipse eodem iter facere contendit.
‘Suddenly walking away from the proconsule Rebilus and having sent Messala to Utica
in advance he set off for the same place.’
() Cic. Orat.
Itaque hoc sum ag-gressus statim Catone absoluto.
‘So I have addressed this work as soon as I have finished the work on Cato.’
() Vell. , ,
Neque Pompeius, ut primum ad rem publicam ad-gressus est, quemquam <aequo>
animo parem tulit.
‘Nor could Pompeius, as soon as he entered public affairs, brook an equal with
equanimity.’
() Val. Max. , ,
Seruus barbarus Hasdrubalem [ . . . ] subito ad-gressus interemit.
‘A foreign serf killed Hasdrubal after suddenly stepping at him.’
() Tac. Agr. ,
Statim in partes trans-gressus est.
‘Vespasianus suddenly passed over to his cause.’
Appendix
() Plin. Nat. , ,
Statim ad solis occasum trans-gressus.
‘He immediately passed into the west.’
() Suet. Nero, ,
Statimque in gymnasium pro-gressus.
‘And he immediately went to the gymnasium.’
() Gell. , ,
Ille, ubi hoc dixit, di-gressus est.
‘He marched away as soon as he said this.’
() Flor. Epit. , ,
Subito ad-gressus a tergo Camillus adeo cecidit ut omnia incendiorum vestigia Gallici
sanguinis inundatione deleret.
‘Camillus, having suddenly stepped at them from the back, killed so many that the rest of
the fire was extinguished with the flow of Gallic blood.’
() Flor. Epit. , ,
Statimque in Asiam trans-gressus [ . . . ] [regem] obprimit.
‘Having immediately passed over to Asia, he subdued the king.’
Telic predicates headed by prefixed navigo ‘sail’
() Plin. Nat. , ,
Ichthyophagi tam longo tractu, ut XXX dierum spatio prae-navigaverint.
‘The territory of the Ichthyophagi is so long that they sailed past it in thirty days.’
Telic predicates headed by prefixed repo ‘crawl’
() Colum. ,
Ubi deinde pro-repseri<n>t plantae, stramenta colligemus.
‘Then, as soon as the plants have crept forth, we will gather the straw.’
() Suet. Diuus Augustus, ,
Draconem repente ir-repsisse ad eam pauloque post egressum.
‘That suddenly a snake glided up to her and a short time after went away’.
() Apul. Met. ,
Statim latenter in-repit eius hospitium temerarius adulter.
‘At once a bold lover secretly crept into his house.’
Telic predicates headed by prefixed salio ‘jump’
() Lucr. ,
Nam <cum> cita saepe obvia conflixere, fit ut diversa repente dis-siliant.
‘Since, if they happen to collide in their agitation, they immediately jump away in
different directions.’
Appendix
() Lucr. ,
Divolsa repente maxima dis-siluisse capacis moenia mundi
‘That the strongest walls in the world, torn in pieces, explode at once in all directions.’
() Verg. Aen. ,
Ubi clara dedit sonitum tuba, funibus omnes [ . . . ], pro-siluere suis.
‘As soon as the clangor of the trumpet gave the sign, they jumped forth from their
lines.’
() Ov. Fast. ,
Scintillam subito pro-siluisse ferunt.
‘They say that a spark suddenly leapt forth.’
() Ov. Met. ,
Haec ubi disposuit, patria Iove natus ab arce de-silit in terras.
‘When he had settled these things, the son of Jupiter leapt down onto the earth from his
father’s stronghold.’
() Hor. Sat. , ,
Haec ubi dicta agrestem pepulere, domo levis ex-silit.
‘Soon as these speeches had roused the peasant, he leaps nimbly from his cave.’
() Liv. , ,
Repente ex equis de-siliunt.
‘They suddenly leap down from their horses.’
() Liv. , ,
De-silire perniciter ubi datum signum esset.
‘To leap down briskly at a given signal’.
() Liv. , ,
Ubi ad coniectum teli uentum est, signo dato uelites de-siliunt.
‘As soon as they have come within range the signal is given and the light-armed soldiers
spring down to the ground.’
() Cic. Verr. Action secunda, ,
Qui tot dies tacuisset, repente [ . . . ] ex-siluit conscientia sceleris.
‘Who had been silent for so many days suddenly started up out of the awareness of his
crime’.
() Cic. Cael.
Ut eo mitteret amicos qui delitiscerent, dein repente, cum venisset Licinius venenum-
que traderet, pro-silirent hominemque comprehenderent.
‘That he sent friends there who lay ambush and then, suddenly, when Licinius had
arrived and was delivering the poison, jump forth, and arrest the man’.
() Cic. De orat. ,
Neque ad-siliendum statim est ad genus illud orationis.
‘You must not jump at once into that kind of discourse.’
Appendix
() Vitr. , ,
Tiburtina [ . . . ] simul [ . . . ] sunt ab [igne] [ . . . ] tacta, dis-siliunt et dissipantur.
‘The Tiburtine stones, as soon as they have been touched by fire burst asunder and
scatter.’
() Curt. , ,
Ut primum rex in conspectu fuit, equo ipsa de-siluit.
‘As soon as she saw the king she leapt down from the horse herself.’
() Curt. , ,
Singuli repente de-siliunt.
‘They suddenly leap down one by one.’
() Curt. , ,
Ex lecto repente pro-siluit.
‘He suddenly leapt forth from the bed.’
() Sen. Nat. , ,
Urbes [ . . . ] cito ac repente dis-siliunt.
‘States quickly and suddenly break apart in pieces.’
() Sil. ,
Subitum nemus [ . . . ] ex-siluit.
‘Suddenly a grove of oak trees jumped out.’
Sil. ,
() Subito abruptis fugiens altaria taurus ex-siluit vinclis.
‘Suddenly, tearing off his bonds, the bull jumps off from the altar and escapes.’
() Ps. Quint. Decl. ,
Illos, qui non statim primo timore pro-siliunt, fragor noctis agitavit.
‘The noise during the night agitated those who do not leap forth at once with the first
hint of fear.’
() Ps. Quint. Decl. ,
Non enim praecipiti raptus inpulsu ex-silui repente.
‘I did not, in fact, start up suddenly with a violent shock.’
() Frotin. Strat. –, ,
Repente in dextrum [ . . . ] cornu pro-siluit.
‘He suddenly leapt forth into the right wing.’
() Suet. C. Caligula, ,
Deinde repente [ . . . ] pro-siluit.
‘Then he suddenly leapt forth.’
() Gell. , ,
Statim [ . . . ] de-siliret [ . . . ] in mare.
‘He would jump down immediately into the sea.’
Appendix
() Apul. Met. ,
Ad eum statim pro-silit.
‘She immediately jumped forth at him.’
() Flor. Epit. , ,
Hinc in Aegypton subito tran-siluit.
‘Then he jumped over suddenly into Egypt.’
Telic predicates headed by prefixed volo ‘fly’
() Quadrig. Hist.
Id ubi rescierunt propinqui obsidum [ . . . ] in uiam pro-uolarunt.
‘The moment the parents of the hostages learnt it, they flew forth onto the way.’
() Liv. , ,
Haec ubi inter signa peditum dicta dedit, ad-volat deinde ad equites.
‘As soon as he has given the signs to the foot soldiers, he flies onto the cavalry.’
() Caes. Gall. , ,
Subito omnibus copiis pro-volaverunt impetumque in nostros equites fecerunt.
‘All of a sudden they flew forth with all their troops and made an attack against our cavalry.’
() Caes. Gall. , ,
Subito ex omnibus partibus silvae e-volaverunt et in nostros impetum fecerunt.
‘Suddenly they flew out of everywhere in the woods and made an attack against our men.’
() Caes. Gall. , ,
Repente ex omnibus partibus ad pabulatores ad-volaverunt.
‘Suddenly they flew onto the foragers from all sides.’
() Caes. Gall. , ,
Illi subito ex omnibus partibus e-volaverunt.
‘They suddenly flew out from all sides.’
() Cic. Phil. ,
Alter ad-uolarat subito ad direptionem pestemque sociorum.
‘One of them had suddenly appeared for the ruin and destruction of the allies.’
() Cic. Mur.
In agros suburbanos repente ad-volabit.
‘In little time he will fly over into the suburban fields.’
() Cic. Cael.
Tum repente e-volasse istos praeclaros testes.
‘Then suddenly those respectable witnesses flew away.’
() Cic. Sest.
Statim me perculso ad meum sanguinem hauriendum [ . . . ] ad-volaverunt.
‘Instantly, the moment I was struck, they flew up to me to drink up my blood.’
Appendix
() Cic. Div. ,
Cor subito non potuisse nescio quo a-volare.
‘That the heart could not suddenly have flown away I know not where’.
() Cic. Fam. , ,
Oblitum ne me putas qua celeritate, ut primum audieris, ad me Tarento ad-volaris
[ . . . ]?
‘Do you suppose that I have forgotten with what speed you flew to me from Tarentum,
as soon as you learned it?’
() Rhet. Her. , ,
Qui fortunis alicuius inducti amicitiam eius secuti sunt, hi, simul ac fortuna dilapsa est,
de-uolant omnes.
‘All those who follow after someone’s friendship, drawn by his wealth, fly away as soon
as the fortune has slid apart.’
() Rhet. Her. , ,
Simul atque hiemem fortunae uiderunt, de-uolant omnes.
‘As soon as they have seen the ruin of the fortune, all of them fly away.’
() Homer.
Patroclus subitoque armis munitus Achillis pro-uolat.
‘And Patroclus suddenly flies forth, provided with Achilles’s weapons.’
() Phaedr. , ,
Subito latrones ex insidiis ad-volant.
‘Suddenly the thieves fly onto them out of the ambush.’
() Phaedr. , ,
Illa [ . . . ], simul cognovit vocem laudari suam, cupide ad-volavit.
‘As soon as she heard that her voice was being praised, she flew over there anxiously.’
() Stat. Theb. ,
Extemplo monitu ducis ad-volat.
‘Suddenly, following the leader’s order, he flies over there.’
() Plin. Nat. , ,
Cogitatio etiam [ . . . ] animum subito trans-volans effingere similitudinem aut miscere
existimatur.
‘Even the thought which suddenly flies across the spirit is believed to determine or alter
the resemblance.’
() Suet. Vitellius, ,
A parte dextra repente aquila ad-uolauit.
‘From the right an eagle suddenly flew over.’
() Gell. , ,
Corvus repente inprovisus ad-volat et super galeam tribuni insistit.
‘A crow suddenly flies over there and lands on the tribune’s helmet.’
References
Acedo-Matellán, Víctor (). Preverbs llatins i moviment postsintàctic. Master’s thesis,
Universitat de Barcelona.
Acedo-Matellán, Víctor (a). Una aproximació sintàctica als verbs prefixats en català,
Estudios Catalanes : –.
Acedo-Matellán, Víctor (b). Prefixes in Latin and Romance and the satellite- vs.
verb-framed distinction, in Universitat de Barcelona (ed.), Actes del VII Congrés de
Lingüística General (CD-ROM). Barcelona: Publicacions i Edicions de la Universitat de
Barcelona.
Acedo-Matellán, Víctor (). Adjectival resultatives cross-linguistically: a morphophono-
logical account, ConSOLE : –.
Acedo-Matellán, Víctor (). Merging Roots in Bare Phrase Structure and the Conflation/
Incorporation Distinction, North-Eastern Linguistic Society : –.
Acedo-Matellán, Víctor (). Can we dance without doing a dance? Two opposite views on
the integration of roots in the syntactic structure of the vP, in Asaf Bachrach, Isabelle Roy,
and Linnaea Stockall (eds.), Structuring the Argument. Multidisciplinary Research on Verb
Argument Structure. Amsterdam: John Benjamins, –.
Acedo-Matellán, Víctor and Jaume Mateu (). The path from satellite-framed Indo-
European to verb-framed Romance: A lexical-syntactic account. Paper presented at the
th Diachronic Issues in Generative Syntax (DIGS ), Ithaca, NY, August.
Acedo-Matellán, Víctor and Jaume Mateu (). Satellite-framed Latin vs. verb-framed
Romance: A syntactic approach, Probus : –.
Acedo-Matellán, Víctor and Jaume Mateu (). From syntax to roots: A syntactic approach
to root interpretation, in Artemis Alexiadou, Hagit Borer, and Florian Schäfer (eds.), The
Syntax of Roots and the Roots of Syntax. Oxford: Oxford University Press, –.
Acedo-Matellán, Víctor and Jaume Mateu (). Parameters and argument structure I:
Motion predicates and resultatives, in Antonio Fábregas, Jaume Mateu, and M. Putnam
(eds.), The Handbook of Parameters. London: Bloomsbury, –.
Acedo-Matellán, Víctor and Cristina Real-Puigdollers (). Inserting roots into (functional)
nodes: categories and cross-linguistic variation, Linguistic Analysis (–): –.
Ackerman, Farrell (). Complex predicates and morpholexical relatedness: Locative alter-
nation in Hungarian, in Ivan Sag and Anna Szabolcsi (eds.), Lexical Matters. Stanford, CA:
CSLI, –.
Acquaviva, Paolo (). Roots and Lexicality in Distributed Morphology, York Papers in
Linguistics (YPL) : –.
Åfarli, Tor A. (). Norwegian verb-particle constructions as causative constructions, Nordic
Journal of Linguistics : –.
Åfarli, Tor A. (). Do verbs have argument structure?, in Eric Reuland, Tanmoy Bhatta-
charya, and Giorgos Spathas (eds.), Argument Structure. Amsterdam: John Benjamins, –.
References
Alexiadou, Artemis and Elena Anagnostopoulou (). Parametrizing AGR: Word order,
V-movement and EPP-checking, Natural Language & Linguistic Theory : –.
Allen, Margaret (). Morphological Investigations. PhD thesis, University of Connecticut.
Ambridge, Ben and Silke Brandt (). Lisa filled water into the cup: The roles of entrench-
ment, pre-emption and verb semantics in German speakers’ L acquisition of English
locatives, Zeitschrift für Anglistik und Amerikanistik : –.
Amritavalli, Raghavachari (). Syntactic categories and lexical argument structure, in Eric
Reuland, Tanmoy Bhattacharya, and Giorgos Spathas (eds.), Argument Structure. Amster-
dam: John Benjamins, –.
Amritavalli, Raghavachari and Karattuparambil A. Jayaseelan (). The genesis of syntactic
categories and parametric variation, in Huang-Jin Yoon (ed.), Proceedings of the th GLOW
in Asia . Generative Grammar in a Broader Perspective. Seoul: The Korean Generative
Grammar Circle and Institute for Cognitive Science, Seoul National University, –.
An, Duk-Ho (). Clauses in noncanonical positions at the syntax-phonology interface,
Syntax : –.
Anderson, Stephen (). On the role of deep structure in semantic representation, Founda-
tions of Language : –.
Anderson, Stephen (). Comments on the paper by Wasow, in Peter Culicover, Thomas
Wasow, and Adrian Akmajian (eds.), Formal Syntax. New York: Academic Press, –.
Arad, Maya (). A minimalist view of the syntax-lexical semantics interface, UCL Working
Papers in Linguistics : –.
Arad, Maya (). VP-Structure and the Syntax-Lexicon Interface. PhD thesis, University
College London.
Arad, Maya (). Locality constraints on the interpretation of roots: The case of Hebrew
denominal verbs, Natural Language & Linguistic Theory : –.
Arad, Maya (). Roots and Patterns. Hebrew Morpho-syntax. Dordrecht: Springer.
Arregi, Karlos and Andrew Nevins (). Morphotactics: Basque Auxiliaries and the Structure
of Spellout. Dordrecht: Springer.
Arsenijević, Boban (). Inner Aspect and Telicity: The Decompositional and the Quantifica-
tional Nature of Eventualities at the Syntax-Semantics Interface. Utrecht: LOT Publications.
Aske, Jon (). Path predicates in English and Spanish: A closer look, Berkeley Linguistics
Society : –.
Atkins, Beryl T., Judy Kegl, and Beth Levin (). Anatomy of a Verb Entry: From Linguistic
Theory to Lexicographic Practice, International Journal of Lexicography : –.
Babko-Malaya, Olga (). Zero Morphology: A Study of Aspect, Argument Structure Case.
PhD thesis, State University of New Jersey.
Baker, Mark C. (). Incorporation: A Theory of Grammatical Function Changing. Chicago:
University of Chicago Press.
Baker, Mark C. (). Thematic roles and syntactic structure, in Liliane Haegeman (ed.),
Elements of Grammar. Dordrecht: Kluwer, –.
Baker, Mark C. (). Lexical Categories. Verbs, Nouns and Adjectives. Cambridge: Cam-
bridge University Press.
Barbelenet, Daniel (). De l’aspect verbal en latin ancien et particulièrement dans Terence.
Paris: Honoré Champion.
References
Bauke, Leah S. (). Symmetry Breaking in Syntax and the Lexicon. Amsterdam: John
Benjamins.
Beavers, John (). Argument/Oblique Alternations and the Structure of Lexical Meaning.
PhD thesis, Stanford University.
Beavers, John, Beth Levin, and Shiao Wei Tham (). The typology of motion expressions
revisited, Journal of Linguistics : –.
Beck, Sigrid and William Snyder (a). Complex predicates and goal PP’s: Evidence for a
semantic parameter, Boston University Conference on Language Development : –.
Beck, Sigrid and William Snyder (b). The resultative parameter and restitutive again, in
Caroline Féry and Wolfgang Sternefeld (eds.), Auditur vox sapientiae: a Festschrift for Arnim
von Stechow. Berlin: Akademie Verlag, –.
Bende-Farkas, Ágnes (). Complex predicates and discourse referents: The case of Hun-
garian. Handout of talk given at the Workshop on Predicative Constructions, Berlin,
October.
Bennis, Hans (). Theoretische aspecten van partikelvooropplaatsing, Tabu : –.
Berman, Ruth A. and Yonni Neeman (). Development of linguistic forms: Hebrew, in
Ruth A. Berman and Dan I. Slobin (eds.), Relating Events in Narrative. A Crosslinguistic
Developmental Study. Hillsdale, NJ: Lawrence Erlbaum Associates, –.
Binnick, Robert (). Studies in the Derivation of Predicative Structures. PhD thesis, Uni-
versity of Chicago.
Biskup, Petr and Michael Putnam (). One P with two Spell-Outs: the ent/aus-alternation
in German, Linguistic Analysis : –.
Boas, Hans C. (). Resultative Constructions in English and German. PhD thesis, University
of North Carolina.
Boas, Hans C. (). A Constructional Approach to Resultatives. Stanford, CA: CSLI.
Boeckx, Cedric (). Elementary Syntactic Structures: Prospects of a Feature-Free Syntax.
Cambridge: Cambridge University Press.
Bohnemeyer, Jürgen and Mary Swift (). Event realization and default aspect, Linguistics
and Philosophy : –.
Bolkestein, Machtelt (). The ‘Ab urbe condita’-construction in Latin: A strange type of
raising?, in Saskia Daalder and Marinel Gerritsen (eds.), Linguistics in the Netherlands .
Amsterdam: North-Holland, –.
Bolkestein, Machtelt (). Parameters in the expression of embedded predications in Latin,
in Gualtiero Calboli (ed.), Subordination and other Topics in Latin. Proceedings of the Third
Colloquium on Latin Linguistics, Bologna, – April . Amsterdam: John Benjamins,
–.
Bonet, Eulàlia (). Morphology after Syntax: Pronominal Clitics in Romance. PhD thesis,
Massachusetts Institute of Technology.
Booij, Geert (). The boundary between morphology and syntax: separable complex verbs
in Dutch, Yearbook of Morphology : –.
Booij, Geert (). Periphrastic word formation, in Geert Booij, Janet DeCesaris, Angela Ralli,
and Sergio Scalise (eds.), Topics in Morphology: Selected Papers from the Third Mediterra-
nean Morphology Meeting, Barcelona, September –, . Barcelona: Institut Universi-
tari de Lingüística Aplicada (Universitat Pompeu Fabra), –.
References
Borer, Hagit (). Deriving passives without theta-grids, in Steven G. Lapointe, Diane
K. Brentari, and Patrick M. Farrell (eds.), Morphology and its Relations to Phonology and
Syntax. Stanford, CA: CSLI, –.
Borer, Hagit (). Exo-skeletal vs. endo-skeletal explanations: Syntactic projections and the
lexicon, in John C. Moore and Maria Polinsky (eds.), The Nature of Explanation in Linguistic
Theory. Stanford, CA: CSLI, –.
Borer, Hagit (a). Structuring Sense, vol. : In Name Only. Oxford: Oxford University Press.
Borer, Hagit (b). Structuring Sense, vol. : The Normal Course of Events. Oxford: Oxford
University Press.
Borer, Hagit (). Structuring Sense, vol. : Taking Form. Oxford: Oxford University Press.
Borer, Hagit (). Wherefore roots?, Theoretical Linguistics : –.
Borik, Olga (). Aspect and Reference Time. Oxford: Oxford University Press.
Bortolussi, Bernard (). L’objet interne en latin, in Heinrich Weber and Ryszard Zuber
(eds.), Linguistik Parisette: Akten des . Linguistischen Kolloquiums, Paris . Tübingen:
Niemeyer, –.
Bortolussi, Bernard (). Le double accusatif avec les verbes préfixés en circum- et trans-, in
Claude Moussy (ed.), La composition et la préverbation en latin. Paris: Presses de l’Université
Paris-Sorbonne, –.
Bouchard, Denis (). The Semantics of Syntax. Chicago: University of Chicago Press.
Brachet, Jean-Paul (). Recherches sur les préverbes de- et ex- du latin. Brussels: Latomus.
Broekhuis, Hans (). Adpositions and Adpositional Phrases. Tilburg: Tilburg University.
Burzio, Luigi (). Italian Syntax: A Government and Binding Approach. Dordrecht: Reidel.
Bye, Patrik and Peter Svenonius (). Non-concatenative morphology as epiphenomenon,
in Jochen Trommer (ed.), The Morphology and Phonology of Exponence. Oxford: Oxford
University Press, –.
Caha, Pavel (). The German locative-directional alternation, The Journal of Comparative
Germanic Linguistics : –.
Cappelle, Bert (). Particle Patterns in English: A Comprehensive Coverage. PhD thesis,
Katholieke Universiteit Leuven.
Carlson, Gregory (). Reference to Kinds in English. PhD thesis, University of California.
Carrier, Jill and Janet H. Randall (). The argument structure and syntactic structure of
resultatives, Linguistic Inquiry : –.
Carvalho, Paulo de (). La grammaire de la préverbation, ou: préverbation et ergativité, in
Benjamín García Hernández (ed.), Estudios de lingüística latina. Actas del IX Coloquio
Internacional de Lingüística Latina. Universidad Autónoma de Madrid, – de Abril de
. Madrid: Ediciones Clásicas, –.
Centineo, Giulia (). A lexical theory of auxiliary selection in Italian, Probus : –.
Chen, Liang and Jiansheng Guo (). Motion events in Chinese novels: Evidence for an
equipollently-framed language, Journal of Pragmatics : –.
Chomsky, Noam (). Lectures on Government and Binding: The Pisa Lectures. Dordrecht:
Foris Publications.
Chomsky, Noam (). A Minimalist Program for Linguistic Theory, in Kenneth Hale and
Samuel J. Keyser (eds.), The View from Building : Essays in Linguistics in Honor of Sylvain
Bromberger. Cambridge, MA: MIT Press, –.
References
Chomsky, Noam (). The Minimalist Program. Cambridge, MA: MIT Press.
Chomsky, Noam (). Minimalist inquiries: The framework, in Roger Martin, David
Michaels, and Juan Uriagereka (eds.), Step by Step: Essays on Minimalist Syntax in Honor
of Howard Lasnik. Cambridge, MA: MIT Press, –.
Chomsky, Noam (). Derivation by phase, in Michael Kenstowicz (ed.), Ken Hale: A Life in
Language. Cambridge, MA: MIT Press, –.
Chomsky, Noam (). On phases, in Robert Freidin, Carlos P. Otero, and María Luisa
Zubizarreta (eds.), Foundational Issues in Linguistic Theory. Essays in Honor of Jean-Roger
Vergnaud. Cambridge, MA: MIT Press, –.
Citko, Barbara (). Phase Theory. Cambridge: Cambridge University Press.
Coleman, Robert (). The rise and fall of absolute constructions in Latin case history, in
Gualtiero Calboli (ed.), Subordination and other Topics in Latin. Proceedings of the Third
Colloquium on Latin Linguistics, Bologna, – April . Amsterdam: John Benjamins,
–.
Coleman, Robert (). Latin prepositional syntax in Indo-European perspective, in Robert
Coleman (ed.), New Studies in Latin Linguistics. Selected papers from the th International
Colloquium on Latin Linguistics. Cambridge, April . Amsterdam: John Benjamins,
–.
Comrie, Bernard (). Aspect: An Introduction to the Study of Verbal Aspect and Related
Problems. Cambridge: Cambridge University Press.
Crane, Gregory R. (ed.) (). Perseus Digital Library Project. Medford, MA: Tufts University.
Crocco Galèas, Grazia and Claudio Iacobini (). Lo sviluppo del tipo verbale parasintetico
in latino: prefissi AD-, IN-, EX-, Quaderni patavini di linguistica : –.
Csirmaz, Aniko (). Particles and a two component theory of aspect, in Katalin É. Kiss
(ed.), Event Structure and the Left Periphery. Studies on Hungarian. Dordrecht: Springer,
–.
Cuervo, María Cristina (). Datives at Large. PhD thesis, Massachusetts Institute of
Technology.
Damonte, Federico (). Classifier incorporation and the Locative Alternation, in Nicola
Munaro, Giuseppina Turano, and Walter Schweikert (eds.), Proceedings of the XXX Incontro
di Grammatica Generativa. Venice: Cafoscarina, –.
De Belder, Marijke (). Roots and Affixes. Eliminating Lexical Categories from Syntax.
Utrecht: LOT Publications.
De Belder, Marijke and Jeroen van Craenenbroeck (). How to merge a root, Linguistic
Inquiry (): –.
Demirdache, Hamida and Myriam Uribe-Etxebarria (). The syntax of time adverbs, in
Jacqueline Guéron and Jacqueline Lecarme (eds.), The Syntax of Time. Cambridge, MA:
MIT Press, –.
Den Dikken, Marcel (). Particles. On the Syntax of Verb-Particle, Triadic and Causative
Constructions. Oxford: Oxford University Press.
Den Dikken, Marcel (). On the syntax of locative and directional adpositional phrases.
MS, City University of New York.
Den Dikken, Marcel (). Directions from the GET-GO. On the syntax of manner-of-
motion verbs in directional constructions, Catalan Journal of Linguistics : –.
References
Di Sciullo, Anna-Maria (). Prefixed verbs and adjunct-identification, in Anna-Maria Di
Sciullo (ed.), Projections and Interface Conditions. Essays on Modularity. Oxford: Oxford
University Press, –.
Di Sciullo, Anna-Maria (). Asymmetry in Morphology. Cambridge, MA: MIT Press.
Di Sciullo, Anna-Maria and Edwin Williams (). On the Definition of Word. Cambridge,
MA: MIT Press.
Dowty, David (). Word Meaning and Montague Grammar. The Semantics of Verbs and
Times in Generative Semantics and in Montague’s PTQ. Dordrecht: Reidel.
Dowty, David (). Thematic proto-roles and argument selection, Language : –.
Echarte Cossío, María José (). Casos y preposiciones en latín, in Ángel López García and
Evangelina Rodríguez Cuadros (eds.), Miscel·lània. Homenatge a Enrique García Díez.
València: Universitat de València, –.
Echarte Cossío, María José (). Sobre el valor del dativo en latín: El llamado dativo de
dirección y el dativo con verbos prefijados, Minerva : –.
É. Kiss, Katalin (). The Syntax of Hungarian. Cambridge: Cambridge University Press.
É. Kiss, Katalin (). Introduction, in Katalin É. Kiss (ed.), Universal Grammar in the
Reconstruction of Ancient Languages. Berlin: Mouton de Gruyter, –.
É. Kiss, Katalin (a). The function and the syntax of the verbal particle, in Katalin É. Kiss
(ed.), Event Structure and the Left Periphery. Studies on Hungarian. Dordrecht: Springer,
–.
É. Kiss, Katalin (b). From the grammaticalization of viewpoint aspect to the grammat-
icalization of situation aspect, in Katalin É. Kiss (ed.), Event Structure and the Left Periphery.
Studies on Hungarian. Dordrecht: Springer, –.
Embick, David (). Voice and the Interfaces of Syntax. PhD thesis, University of
Pennsylvania.
Embick, David (). Features, syntax, and categories in the Latin perfect, Linguistic Inquiry
: –.
Embick, David (). On the structure of resultative participles in English, Linguistic Inquiry
: –.
Embick, David (). Localism versus Globalism in Morphology and Phonology. Cambridge,
MA: MIT Press.
Embick, David and Rolf Noyer (). Locality in post-syntactic operations, MIT Working
Papers in Linguistics : –.
Embick, David and Rolf Noyer (). Movement operations after syntax, Linguistic Inquiry
: –.
Embick, David and Rolf Noyer (). Distributed Morphology and the syntax/morphology
interface, in Gillian Ramchand and Charles Reiss (eds.), The Oxford Handbook of Linguistic
Interfaces. Oxford: Oxford University Press, –.
Ernout, Alfred and François Thomas (). Syntaxe latine. Paris: Librairie C. Klincksieck.
Espinal, M. Teresa and Jaume Mateu (). Bare nominals and argument structure in Catalan
and Spanish, The Linguistic Review : –.
Eythórsson, Thorhállur (). Particle, prefix, zero: Unidirectionality in the development of
Indo-European preverbs. Paper presented at the Complex Predicates, Particles and Subevents
workshop, Universität Konstanz, September , October –.
References
Fábregas, Antonio (). An Exhaustive Lexicalisation account of directional complements,
Nordlyd : –.
Fan, Sheng-yang (). Argument Structure in Mandarin Chinese: A Lexical-syntactic Per-
spective. PhD thesis, Universidad Autónoma de Madrid.
Farkas, Donka and Jerrold Sadock (). Preverb climbing in Hungarian, Language :
–.
Fay, Edwin W. (). The Latin dative: Nomenclature and classification, The Classical
Quarterly : –.
Filip, Hana (). Integrating telicity, aspect and NP semantics: The role of thematic
structure, in Jindřich Toman (ed.), Formal Approaches to Slavic Linguistics, vol. : The
College Park Meeting . Ann Arbor, MI: Slavic Publications, –.
Filip, Hana (). Aspect, Eventuality Types and Nominal Reference. New York: Routledge.
Filip, Hana (). Prefixes and the delimitation of events, Journal of Slavic Linguistics :
–.
Fillmore, Charles J. (). Scenes-and-frames semantics, in Antonio Zampolli (ed.), Linguistic
Structures Processing. Amsterdam: North-Holland, –.
Fillmore, Charles J. (). Frame semantics, in The Linguistic Society of Korea (ed.),
Linguistics in the Morning Calm. Seoul: Hanshin, –.
Folli, Raffaella and Heidi Harley (). On the licensing of causatives of directed motion:
Waltzing Matilda all over, Studia Linguistica : –.
Folli, Raffaella and Heidi Harley (). Causation, obligation, and argument structure: On the
nature of little v, Linguistic Inquiry : –.
Folli, Raffaella and Gillian Ramchand (). Getting results in Italian and Scottish Gaelic,
West Coast Conference on Formal Linguistics : –.
Folli, Raffaella and Gillian Ramchand (). Prepositions and results in Italian and English:
An analysis from event decomposition, in Angeliek van Hout, Henriette de Swart, and Henk
J. Verkuyl (eds.), Perspectives on Aspect. Dordrecht: Kluwer, –.
Fong, Vivienne (). The Order of Things: What Directional Locatives Denote. PhD thesis,
Stanford University.
Fong, Vivienne (). ‘Into doing something’: where is the path in event predicates?. MS,
National University of Singapore.
Gaffiot, Félix (). Dictionnaire latin-français. Paris: Hachette.
García Hernández, Benjamín (). Semántica estructural y lexemática del verbo. Tarragona:
Avesta.
García Hernández, Benjamín (). Observo = ‘ob oculos servo’? (Thesaurus y Szantyr ),
in Luis Miguel Macía Aparicio, Vicente Picón García, Mª Esperanza Torrego Salcedo, and
Jesús de la Villa Polo (eds.), Quid ultra faciam?: Trabajos de griego, latín e indoeuropeo en
conmemoración de los años de la Universidad Autónoma de Madrid. Madrid: Ediciones de
la Universidad Autónoma de Madrid, –.
García Hernández, Benjamín (). Synonimie et analyse fonctionelle dans le système
préverbal latin, Revue des études latines : –.
García Hernández, Benjamín (). L’antonymie aspectuelle des préverbes alatifs et ablatifs,
in Claude Moussy (ed.), La composition et la préverbation en latin. Paris: Presses de
l'Université Paris-Sorbonne, –.
References
Gehrke, Berit (). Ps in Motion: On the Semantics and Syntax of P Elements and Motion
Events. Utrecht: LOT Publications.
Ghomeshi, Jila and Diane Massam (). Lexical/syntactic relations without projections,
Linguistic Analysis : –.
Gianollo, Chiara (). Il medio in latino e il fenomeno dell’intransitività scissa. Master’s
thesis, Università di Pisa.
Gibert Sotelo, Elisabeth (forthcoming). Source and Negative Prefixes. On the Syntax-lexicon
Interface and the Encoding of Spatial Relations. PhD thesis, Universitat de Girona.
Gleitman, Lila and Barbara Landau (). Every child an isolate: nature’s experiments in
language learning, in Massimo Piattelli-Palmarini and Robert C. Berwick (eds.), Rich
Languages from Poor Inputs. Oxford: Oxford University Press, –.
Goldberg, Adele (). Constructions. A Construction Grammar Approach to Argument
Structure. Chicago: University of Chicago Press.
Goldberg, Adele (). Constructions at Work. The Nature of Generalization in Language.
Oxford: Oxford University Press.
Green, Georgia (). Some observations on the syntax and semantics of instrumental verbs,
Chicago Linguistic Society : –.
Grimshaw, Jane (). Argument Structure. Cambridge, MA: MIT Press.
Haider, Hubert and Martin Prinzhorn (eds.) (). Verb Second Phenomena in Germanic
Languages. Dordrecht: Foris.
Hale, Kenneth (). Notes on world view and semantic categories: Some Warlpiri examples,
in Peter Muysken and Henk van Riemsdijk (eds.), Features and Projections. Dordrecht:
Foris, –.
Hale, Kenneth and Samuel J. Keyser (). The syntactic character of thematic structure, in
Iggy M. Roca (ed.), Thematic Structure: Its Role in Grammar. Berlin: Foris, –.
Hale, Kenneth and Samuel J. Keyser (). On argument structure and the lexical expression
of syntactic relations, in Kenneth Hale and Samuel J. Keyser (eds.), The View from Building
: Essays in Linguistics in Honor of Sylvain Bromberger. Cambridge, MA: MIT Press,
–.
Hale, Kenneth and Samuel J. Keyser (). Adjectives, other stative predicates and the roots of
stativity. MS, Massachusetts Institute of Technology.
Hale, Kenneth and Samuel J. Keyser (). The basic elements of argument structure, MIT
Working Papers in Linguistics : –.
Hale, Kenneth and Samuel J. Keyser (a). Bound features, Merge, and transitivity alterna-
tions, in Liina Pylkkänen, Angeliek van Hout, and Heidi Harley (eds.), Papers from the
UPenn/MIT Roundtable on the Lexicon. Cambridge, MA: MIT Working Papers in Linguis-
tics, –.
Hale, Kenneth and Samuel J. Keyser (b). A response to Fodor and Lepore, ‘Impossible
words?’, Linguistic Inquiry : –.
Hale, Kenneth and Samuel J. Keyser (). Prolegomenon to a Theory of Argument Structure.
Cambridge, MA: MIT Press.
Halle, Morris (). An approach to morphology. North-Eastern Linguistic Society :
–.
Halle, Morris (). Latvian Declension, Morphology Yearbook : –.
References
Halle, Morris (). Distributed morphology: Impoverishment and fission, MIT Working
Papers in Linguistics : –.
Halle, Morris and Alec Marantz (). Distributed Morphology and the pieces of inflection,
in Kenneth Hale and Samuel J. Keyser (eds.), The View from Building : Essays in
Linguistics in Honor of Sylvain Bromberger. Cambridge, MA: MIT Press, –.
Halle, Morris and Alec Marantz (). Some key features of Distributed Morphology, MIT
Working Papers in Linguistics : –.
Halliday, Michael A. K. (). Notes on transitivity and theme in English. Part I, Journal of
Linguistics : –.
Harley, Heidi (). Subjects, Events, and Licensing. PhD thesis, Massachusetts Institute of
Technology.
Harley, Heidi (). How do verbs get their names? Denominal verbs, Manner incorporation
and the ontology of verb roots in English, in Nomi Erteschik-Shir and Tova Rapoport (eds.),
The Syntax of Aspect. Deriving Thematic and Aspectual Interpretation. Oxford: Oxford
University Press, –.
Harley, Heidi (). The bipartite structure of verbs cross-linguistically (or why Mary can’t
exhibit John her paintings). MS, University of Arizona.
Harley, Heidi (). External arguments and the Mirror Principle: On the distinctness of
Voice and v, Lingua: : –.
Harley, Heidi (). On the identity of roots, Theoretical Linguistics : –.
Harley, Heidi and Rolf Noyer (). Licensing in the non-lexicalist lexicon: nominalizations,
Vocabulary Items and the Encyclopedia, MIT Working Papers in Linguistics : –.
Harley, Heidi and Rolf Noyer (). Distributed Morphology, Glot International : –.
Harley, Heidi and Rolf Noyer (). Licensing in the non-lexicalist lexicon, in Bert Peeters
(ed.), The Lexicon/Encyclopaedia Interface. Amsterdam: Elsevier, –.
Haudry, Jean (). L’instrumental et la structure de la phrase simple en indo-européen,
Bulletin de la Société de Linguistique : –.
Haugen, Jason D. (). Hyponymous objects and Late Insertion, Lingua : –.
Haverling, Gerd (). On the sco-suffix, on prefixes, and on the development of the Latin
verbal system, in József Herman (ed.), Linguistic Studies on Latin. Selected Papers from the
th International Colloquium on Latin Linguistics (Budapest, – March ). Amster-
dam: John Benjamins, –.
Haverling, Gerd (a). On sco-verbs, on prefixes, and on semantic functions, in Hannah
Rosén (ed.), Aspects of Latin: Papers from the Seventh International Colloquium on Latin
Linguistics, Jerusalem, April . Innsbruck: Innsbrucker Beiträge zur Sprachwissenschaft,
–.
Haverling, Gerd (b). Some more remarks on sco-verbs, on prefixes, and on semantic
functions, in Alfred Bammesberger and Friedrich Heberlein (eds.), Akten des VIII. Internatio-
nalen Kolloquiums zur Lateinischen Linguistik. Proceedings of the Eighth International Collo-
quium on Latin Linguistics. Heidelberg: Universitätsverlag C. Winter, –.
Haverling, Gerd (). On the development of the perfect and the imperfect tenses in Late
Latin, in Benjamín García Hernández (ed.), Estudios de Lingüística Latina. Actas del IX
Coloquio Internacional de Lingüística Latina. Universidad Autónoma de Madrid, – de
Abril de . Madrid: Ediciones Clásicas, –.
References
Haverling, Gerd (). On Sco-verbs, Prefixes and Semantic Functions. A Study in the
Development of Prefixed and Unprefixed Verbs from Early to Late Latin. Gothenburg: Acta
Universitatis Gothoburgensis.
Haverling, Gerd (). On prefixes and actionality in Classical and Late Latin, Acta Linguis-
tica Hungarica : –.
Hay, Jennifer (). Causes and Consequences of Word Structure. PhD thesis, Northwestern
University.
Hay, Jennifer (). Causes and Consequences of Word Structure. New York: Routledge.
Hegedűs, Veronika (). Hungarian spatial PPs, Nordlyd : –.
Heslin, Thomas P. Jr. (). Prefixation in Latin, Lingua : –.
Hewson, John (). The Latin verbal system, in John Hewson and Vit Bubenik (eds.), Tense
and Aspect in Indo-European Languages. Theory, Typology, Diachrony. Amsterdam: John
Benjamins, –.
Hoekstra, Teun (). Small clause results, Lingua : –.
Hoekstra, Teun (). Auxiliary selection in Dutch, Natural Language & Linguistic Theory :
–.
Hoekstra, Teun and René Mulder (). Unergatives as copular verbs: Locational and
existential predication, The Linguistic Review : –.
Hofmann, Johann B. and Anton Szantyr (). Lateinische Grammatik, vol. : Syntax und
Stilistik. München: C. H. Beck.
Horrocks, Geoffrey (). Aspect and verbs of movement in the history of Greek: Why
Pericles could ‘walk into town’ but Karamanlis could not, in J. H. W Penney, Indo-European
Perspectives. Studies in Honour of Anna Morpurgo Davies. Oxford: Oxford University Press,
–.
Horrocks, Geoffrey and Melita Stavrou (). Actions and their results in Greek and English:
The complementarity of morphologically encoded (viewpoint) aspect and syntactic resulta-
tive predication, Journal of Semantics : –.
Horrocks, Geoffrey and Melita Stavrou (). Grammaticalized aspect and spatio-temporal
culmination, Lingua : –.
Horvath, Julia (). Verbal prefixes: A non-category in Hungarian, Glossa : –.
Hout, Angeliek van (). Event Semantic of Verb Frame Alternations: A Case Study of Dutch
and its Acquisition. PhD thesis, Tilburg University.
Hout, Angeliek van (). Projection based on event structure, in Peter Coopmans, Martin
Everaert, and Jane Grimshaw (eds.), Lexical Specification and Insertion. Amsterdam: John
Benjamins, –.
Inagaki, Shunji (). Motion verbs with locational/directional PPs in English and Japanese,
Canadian Journal of Linguistics : –.
Irube, Kaneharu (). Argument structure and the out-prefixation, English Linguistics :
–.
Jackendoff, Ray (). X-Bar Syntax: A Study of Phrase Structure. Cambridge, MA: MIT
Press.
Jackendoff, Ray (). Semantics and Cognition. Cambridge, MA: MIT Press.
Jensen, Povl J. (). The use of the perfect in Classical Latin, Classica et Mediaevalia :
–.
References
Julien, Marit (). Syntactic Heads and Word Formation. Oxford: Oxford University Press.
Kaplan, Ronald M. and Joan Bresnan (). Lexical-Functional Grammar: A formal system
for grammatical representation, in Joan Bresnan (ed.), The Mental Representation of Gram-
matical Relations. Cambridge, MA: MIT Press, –.
Kayne, Richard (). French Syntax. Cambridge, MA: MIT Press.
Kayne, Richard (). Principles of particle constructions, in Jacqueline Guéron, Hans-Georg
Obenauer, and Jean-Yves Pollock (eds.), Grammatical Representations. Dordrecht: Foris,
–.
Kayne, Richard (). Antisymmetry and the Lexicon, Linguistic Variation Yearbook : –.
Keyser, Samuel J. and James Roeper (). Re: The Abstract Clitic Hypothesis, Linguistic
Inquiry : –.
Kiparsky, Paul (). Lexical morphology and phonology, in Linguistic Society of Korea (ed.),
Linguistics in the Morning Calm: Selected Essays from SICOL-. Seoul: Hanshin, –.
Kiparsky, Paul (). Remarks on denominal verbs, in Àlex Alsina, Joan Bresnan, and Peter
Sells (eds.), Complex Predicates. Stanford, CA: CSLI, –.
Klein, Wolfgang (). A time-relational analysis of Russian aspect, Language : –.
Klipple, Elisabeth (). Prepositions and variation, in Anna-Maria Di Sciullo (ed.), Projec-
tions and Interface Conditions: Essays on Modularity. Oxford: Oxford University Press,
–.
Kolehmainen, Leena (). Präfix- und Partikelverben im deutsch-finnischen Kontrast.
Helsinki: Helsingin yliopiston verkkojulkaisut.
Koopman, Hilda (). The Syntax of Specifiers and Heads. London: Routledge.
Kopecka, Anetta (). Étude typologique de l’expression de l’espace: localisation et déplace-
ment en français et en polonais. PhD thesis, Université Lumière Lyon .
Kratzer, Angelika (). Stage-Level and Individual-Level Predicates, in Gregory N. Carlson
and Francis Jeffry Pelletier (eds.), The Generic Book. Chicago: University of Chicago Press,
–.
Kratzer, Angelika (). Severing the external argument from its verb, in Johan Rooryck and
Laurie Zaring (eds.), Phrase Structure and the Lexicon. Dordrecht: Kluwer, –.
Kratzer, Angelika (). Building statives, Berkeley Linguistic Society : –.
Kratzer, Angelika (). Building resultatives, in Claudia Maienborn and Angelika Wöllstein-
Leisten (eds.), Event Arguments: Foundations and Applications. Tübingen: Niemeyer, –.
Krifka, Manfred (). Thematic relations as links between nominal reference and temporal
constitution, in Ivan A. Sag and Anna Szabolcsi (eds.), Lexical Matters. Stanford, CA:
Stanford University, –.
Kühner, Raphael and Carl Stegmann (). Ausführliche Grammatik der lateinischen Sprache,
vol. I: Elementar-, Formen- und Wortlehre. Hanover: Hahnsche Buchhandlung.
Kuno, Susumu and Ken-ichi Takami (). Functional Constraints in Grammar: On the
Unergative-Unaccusative Distinction. Amsterdam: John Benjamins.
Landgraf, Gustav von (). Der Dativus commodi und der Dativus finalis mit ihren Abarten,
Archiv für Lateinische Lexikographie : –.
Larson, Richard K. (). On the double object construction, Linguistic Inquiry : –.
Larson, Richard K. (). Double objects revisited: A reply to Jackendoff, Linguistic Inquiry
: –.
References
Lavency, Marius (). À propos de la question des cas en latin classique, Études Classiques
: –.
Lavency, Marius (). Le Paradigme syntaxique de l’ablatif absolu, in Freddy Decreus and
Carl Deroux (eds.), Hommages à Jozef Veremans. Brussels: Latomus, –.
Lavidas, Nikolaos (). Transitivity Alternations in Diachrony. Changes in Argument Struc-
ture and Voice Morphology. Newcastle upon Tyne: Cambridge Scholars Publishing.
Legendre, Géraldine (). Unaccusativity in French, Lingua : –.
Lehmann, Christian (). Latin preverbs and cases, in Harm Pinkster (ed.), Latin Linguistics
and Linguistic Theory. Proceedings of the st International Colloquium on Latin Linguistics.
Amsterdam, April . Amsterdam: John Benjamins, –.
Lehrer, Adrienne (). Prefixes in English word formation, Folia Linguistica : –.
Lemaire, Maryvonne (). Un aspect de l’échange des arguments du verbe: les verbes à
double régime de l’objet en latin, Bulletin de la Société de Linguistique de Paris : –.
Le Roux, Cecile (). On the Interface of Morphology and Syntax. Stellenbosch: University of
Stellenbosch.
Levin, Beth (). Verb Classes and Alternations: A Preliminary Investigation. Chicago:
University of Chicago Press.
Levin, Beth (). Objecthood: An event structure perspective, Chicago Linguistic Society :
–.
Levin, Beth and Tova R. Rapoport (). Lexical Subordination, Chicago Linguistic Society :
–.
Levin, Beth and Malka Rappaport (). Non-event -er nominals: A probe into argument
structure, Linguistics : –.
Levin, Beth and Malka Rappaport Hovav (). Unaccusativity: At the Syntax-Lexical Seman-
tics Interface. Cambridge, MA: MIT Press.
Levin, Beth and Malka Rappaport Hovav (). The semantic determinants of argument
expressions: A view from the English resultative construction, in Jacqueline Guéron and
Jacqueline Lecarme (eds.), The Syntax of Time. Cambridge, MA: MIT Press, –.
Levin, Beth and Malka Rappaport Hovav (). Argument Realization. Cambridge: Cam-
bridge University Press.
Levin, Beth and Peter Sells (). Unpredicated particles, in Linda Uyechi and Lian Hee Wee
(eds.), Reality Exploration and Discovery: Pattern Interaction in Language & Life. Stanford,
CA: CSLI, –.
Levinson, Lisa (). The Roots of Verbs. PhD thesis, New York University.
Levinson, Lisa (). Arguments for pseudo-resultative predicates, Natural Language &
Linguistic Theory : –.
Levinson, Lisa (). The ontology of roots and verbs, in Artemis Alexiadou, Hagit Borer, and
Florian Schäfer (eds.), The Syntax of Roots and the Roots of Syntax. Oxford: Oxford
University Press, –.
Lewandowski, Wojciech (). The locative alternation in verb-framed vs. satellite-framed
languages: A corpus study of Spanish and Polish, Studies in Language : –.
Lewis, Charlton T. and Charles Short (). A Latin Dictionary: Founded on Andrews’ Edition
of Freund’s Latin Dictionary: Revised, Enlarged, and in Great Part Rewritten by Charlton
T. Lewis, Ph.D. and Charles Short, LL. D. Oxford: Clarendon Press.
References
Liddell, Henry G. and Robert Scott (). A Greek-English Lexicon. Revised and Augmented
throughout by Sir Henry Stuart Jones with the Assistance of Roderick McKenzie. Oxford:
Clarendon Press.
Lieber, Rochelle (). Morphology and Lexical Semantics. Cambridge: Cambridge University
Press.
Lindvall, Ann (). Swedish particle verbs in comparison with Polish aspect marking,
Travaux de l'institut de linguistique de Lund : –.
Löfstedt, Bengt (). Zum lateinischen possessiven Dativ. Zeitschrift für vergleichende
Sprachforschung auf dem Gebiete der Indogermanischen Sprachen : –.
Lohndal, Terje (). Phrase Structure and Argument Structure. A Case Study of the Syntax-
Semantic Interface. Oxford: Oxford University Press.
López Moreda, Santiago (). Interferencias semántico-sintácticas entre preverbios y pre-
posiciones, in Benjamín García Hernández (ed.), Estudios de lingüística latina. Actas del IX
Coloquio Internacional de Lingüística Latina. Universidad Autónoma de Madrid, – de
Abril de . Madrid: Ediciones Clásicas, –.
Lüdeling, Anke (). On Particle Verbs and Similar Constructions in German. Stanford, CA: CSLI.
Luraghi, Silvia (). The relationship between prepositions and cases within Latin prepos-
itional phrases, in Gualtiero Calboli (ed.), Subordination and other Topics in Latin. Proceed-
ings of the Third Colloquium on Latin Linguistics, Bologna, – April . Amsterdam: John
Benjamins, –.
Luraghi, Silvia (). Omission of the direct object in Classical Latin, Indogermanische
Forschungen : –.
MacDonald, Jonathan E. (). The Syntactic Nature of Inner Aspect. A Minimalist Perspec-
tive. Amsterdam: John Benjamins.
McIntyre, Andrew (). German Double Particles as Preverbs: Morphology and Conceptual
Semantics. Berlin: Stauffenberg.
McIntyre, Andrew (). Idiosyncrasy in particle verbs, in Nicole Dehé, Ray Jackendoff,
Andrew McIntyre, and Silke Urban (eds.), Verb-Particle Explorations. Berlin: Mouton de
Gruyter, –.
McIntyre, Andrew (). Preverbs, argument linking and verb semantics: Germanic prefixes
and particles, Yearbook of Morphology : –.
McIntyre, Andrew (). Event paths, conflation, argument structure, and VP shells, Lin-
guistics : –.
McIntyre, Andrew (). Particle verbs and argument structure, Language and Linguistics
Compass : –.
Marácz, László (). Asymmetries in Hungarian. Donostia-San Sebastián: Diputación Foral
de Gipuzkoa.
Marantz, Alec (). On the Nature of Grammatical Relations. Cambridge, MA: MIT Press.
Marantz, Alec (). Clitics, morphological merger, and the mapping to phonological
structure, in Michael Hammond, and Michael Noonan (eds.), Theoretical Morphology.
San Diego, CA: Academic Press, –.
Marantz, Alec (). Cat as a phrasal idiom. MS, Massachusetts Institute of Technology.
Marantz, Alec (). No escape from syntax: Don’t try morphological analysis in the privacy
of your own lexicon. Penn Linguistics Colloquium : –.
References
Marantz, Alec (). Words. Paper presented at the West Coast Conference on Formal
Linguistics (WCCFL) , University of Southern California, February –.
Marantz, Alec (). Subjects and objects. MS, Massachusetts Institute of Technology.
Marantz, Alec (). Objects out of the lexicon: Objects as events. MS, Massachusetts
Institute of Technology.
Marantz, Alec (). Words and phases, in Sook-Hee Choe (ed.), Phases in the Theory of
Grammar. Seoul: Dong In, –.
Marantz, Alec (). Roots, re- and affected agents: Can roots pull the agent under little v?.
Paper presented at the workshop Roots: Word Formation from the Perspective of ‘Core
Lexical Elements’, Stuttgart, – June.
Marantz, Alec (). Syntactic approaches to argument structure without incorporation.
Paper presented at the workshop Structuring the Argument: A Multidisciplinary Workshop
on the Mental Representation of Verbal Argument Structure, Paris, – September.
Marantz, Alec (). Locality domains for contextual allomorphy across the interfaces, in Ora
Matushansky and Alec Marantz (eds.), Distributed Morphology Today: Morphemes for
Morris Halle. Cambridge, MA: MIT Press, –.
Marchand, Hans (). The Categories and Types of Present-Day English Word-Formation.
Munich: Beck.
Massam, Diane (). Cognate objects as thematic objects, Canadian Journal of Linguistics
: –.
Mateu, Jaume (a). Small Clause Results Revisited, ZAS Papers in Linguistics :
–.
Mateu, Jaume (b). Unselected Objects, in Nicole Dehé and Anja Wanner (eds.), Structural
Aspects of Semantically Complex Verbs. Berlin: Peter Lang, –.
Mateu, Jaume (c). Lexicalization patterns and the locative alternation. MS, Universitat
Autònoma de Barcelona.
Mateu, Jaume (). Argument Structure. Relational Construal at the Syntax-Semantics
Interface. PhD thesis, Universitat Autònoma de Barcelona.
Mateu, Jaume (). Parametrizing verbal polysemy: The case of bake revisited. MS, Uni-
versitat Autònoma de Barcelona. [Published as Mateu, Jaume (). Parametrizing verbal
polysemy: The case of bake revisited. Chicago Linguistics Society : –.]
Mateu, Jaume (). Arguing our way to the Direct Object Restriction on English resultatives,
Journal of Comparative Germanic Linguistics : –.
Mateu, Jaume (a). Argument structure and denominal verbs. MS, Universitat Autònoma
de Barcelona.
Mateu, Jaume (b). On the l-syntax of directionality/resultativity: The case of Germanic
preverbs, in Anna Asbury, Jakub Dotlačil, Berit Gehrke, and Rick Nouwen (eds.), Syntax
and Semantics of Spatial P. Amsterdam: John Benjamins, –.
Mateu, Jaume (). Conflation and incorporation processes in resultative constructions, in
Violeta Demonte and Louise McNally (eds.), Telicity, Change, and State: A Cross-Categorial
View of Event Structure. Oxford: Oxford University Press, –.
Mateu, Jaume and Víctor Acedo-Matellán (). The manner/result complementarity
revisited: a syntactic approach, in María Cristina Cuervo and Yves Roberge (eds.), The
End of Argument Structure. Bingley: Emerald Group Publishing Limited, –.
References
Mateu, Jaume and Laia Amadas (). Syntactic tools for semantic construal. Paper presented
at the st Conference on Tools in Linguistic Theory. Utrecht Institute of Linguistics OTS,
Utrecht, July –.
Mateu, Jaume and Gemma Rigau (). A minimalist account of conflation processes:
Parametric variation at the lexicon-syntax interface, in Artemis Alexiadou (ed.), Theoretical
Approaches to Universals. Amsterdam: John Benjamins, –.
Mateu, Jaume and Gemma Rigau (). Verb-particle constructions in Romance: A lexical-
syntactic account, Probus : –.
Matushansky, Ora (). Head movement in linguistic theory, Linguistic Inquiry : –.
Meillet, Antoine (). De l’expression de l’aoriste en latin, Revue de philologie, de littérature
et d’histoire ancienne : –.
Meillet, Antoine and Joseph Vendryès (). Traité de grammaire comparée des langues
classiques. Paris: Honoré Champion.
Mendikoetxea, Amaya (). Construcciones con se: medias, pasivas e impersonales, in
Ignacio Bosque and Violeta Demonte (eds.), Gramática descriptiva de la lengua española,
vol. : Las construccions sintácticas fundamentales. Relaciones temporales, aspectuales y
modales. Madrid: Real Academia Española and Espasa Calpe, –.
Merchant, Jason (). How much context is enough? Two cases of span-conditioned stem
allomorphy, Linguistic Inquiry (): –.
Meurant, Alain (). Itinera Electronica. Lovaine: Université catholique de Louvain.
Miller, D. Gary (). Complex Verb Formation. Amsterdam: John Benjamins.
Molinari, Danielle (). Les préverbés de facio, in Claude Moussy (ed.), La composition et la
préverbation en latin. Paris: Presses de l’Université Paris-Sorbonne, –.
Moussy, Claude (). La polysémie du préverbe com-, in Claude Moussy (ed.), La compos-
ition et la préverbation en latin. Paris: Presses de l’Université Paris-Sorbonne, –.
Mulder, Hotze (). Object continuity and elliptical objects in Latin, Romance Linguistics
and Literature Review : –.
Mulder, René (). The Aspectual Nature of Syntactic Complementarion. PhD thesis, Leiden
Universiteit.
Munaro, Nicola (). Alcuni casi di alternanza di struttura argomentale in inglese, in
Gianluigi Borgato (ed.), Teoria del linguaggio e analisi linguistica. XX Incontro di Gramma-
tica Generativa. Padua: Unipress, –.
Napoli, Donna Jo (). Secondary resultative predicates in Italian, Journal of Linguistics :
–.
Narasimhan, Bhuvana, Vittorio di Tomaso, and Cornelia Verspoor (). Unaccusative or
unergative? Verbs of manner of motion, Quaderni del Laboratorio di Linguistica della Scuola
Normale Superiore di Pisa : –.
Nedjalkov, Vladimir P. (ed.) (). Typology of Resultative Constructions. Amsterdam: John
Benjamins.
Neeleman, Ad (). Complex predicates. Utrecht: Onderzoekinstituut voor Taal en Sprak.
Neeleman, Ad and Hans van de Koot (). Bare resultatives, Journal of Comparative
Germanic Linguistics : –.
Neeleman, Ad and Fred Weerman (). The balance between syntax and morphology:
Dutch particles and resultatives, Natural Language & Linguistic Theory : –.
References
Nespor, Marina and Irene Vogel (). Prosodic Phonology. Dordrecht: Foris Publications.
Newell, Heather (). Aspects of the Morphology and Phonology of Phases. PhD thesis,
McGill University.
Nisard, Désiré (ed.) (). Les agronomes latins. Paris: Firmin Didot, Frères, Libraires.
Olsen, Susan (). Der Dativ bei Partikelverben, in Christa Dürscheid, Karl-Heinz Ramers,
and Monika Schwarz (eds.), Festschrift für Heinz Vater. Tübingen, Niemeyer, –.
Oltra-Massuet, Isabel (). On the constituency structure of the Catalan verbs, MIT Working
Papers in Linguistics : –.
Oltra-Massuet, Isabel and Karlos Arregi (). Stress-by-structure in Spanish, Linguistic
Inquiry : –.
Oniga, Renato (). Composition et préverbation en latin: problèmes de typologie, in Claude
Moussy (ed.), La Composition et la préverbation en latin. Paris: Presses de l’Université Paris-
Sorbonne, –.
Oostendorp, Marc van (). Stress as a proclitic in Modern Greek, Lingua : –.
Oya, Toshiaki (). Ground arguments in German particle verbs: A comparison with Dutch
and English, Journal of Germanic Linguistics : –.
Padrosa-Trias, Susanna (). Complex Word-Formation and the Morphology-Syntax Inter-
face. PhD thesis, Universitat Autònoma de Barcelona.
Padrosa-Trias, Susanna and Angelina Markova (). Some remarks on prefixation: Evidence
from Bulgarian, Catalan and English, Interlingüística : –.
Panhius, Dirk (). Gapping in Latin, The Classical Journal : –.
Pantcheva, Marina B. (). Decomposing Path: The Nanosyntax of Directional Expressions.
PhD thesis, Universitetet i Tromsø.
Pantelia, Maria (). Thesaurus Linguae Graecae. Irvine, CA: University of California.
Perrot, Jean (). Préverbes et cas en hongrois, in André Rousseau (ed.), Les Préverbes dans
les langues d’Europe. Introduction à l’étude de la préverbation. Lille: Presses Universitaires du
Septentrion, –.
Pesetsky, David (). Paths and Categories. PhD thesis, Massachusetts Institute of Technology.
Pesetsky, David (). Morphology and logical form, Linguistic Inquiry : –.
Peyraube, Alain (). A diachronic study of directional complements, in Maya Hickmann
and Stéphane Robert (eds.), Space in Languages: Linguistic Systems and Cognitive Categories.
Amsterdam: John Benjamins, –.
Pfau, Roland (). Grammar as Processor: A Distributed Morphology Account of Spontan-
eous Speech Errors. Amsterdam: John Benjamins.
Pfister, Raimund (). Kompetenz in der lateinischen Syntax, in Harm Pinkster (ed.), Latin
Linguistics and Linguistic Theory. Proceedings of the st International Colloquium on Latin
Linguistics. Amsterdam, April . Amsterdam: John Benjamins, –.
Pinault, Georges-Jean (). Le problème du préverbe en indo-européen, in André Rousseau
(ed.), Les Préverbes dans les langues d’Europe. Introduction à l’étude de la préverbation. Lille:
Presses Universitaires du Septentrion, –.
Pinker, Steven (). Learnability and Cognition. Cambridge, MA: MIT Press.
Pinkster, Harm (). On Latin Adverbs. Amsterdam: North-Holland.
Pinkster, Harm (). Tempus, aspect and Aktionsart in Latin, Aufstieg und Niedergang der
Römischen Welt : –.
References
Pinkster, Harm (). Non-accusative second arguments of two-place verbs in Latin, Cua-
dernos de Filologia Clásica : –.
Pinkster, Harm (). Sintaxis y semántica del latín. Madrid: Ediciones Clásicas.
Pottier, Bernard (). Systématique des éléments de relation. Paris: Klincksieck.
Pourcel, Stéphanie and Anetta Kopecka (). Motion events in French: Typological intrica-
cies. MS, University of Sussex and Max Planck Institute for Psycholinguistics.
Puskás, Genoveva (). Word Order in Hungarian: The Syntax of A’-positions. Amsterdam:
John Benjamins.
Pustejovsky, James (). The Generative Lexicon. Cambridge, MA: MIT Press.
Pustejovsky, James (). Generativity and Explanation in Semantics. A Reply to Fodor and
Lepore, Linguistic Inquiry : –.
Pylkkänen, Liina (). Introducing Arguments. Cambridge, MA: MIT Press.
Ramchand, Gillian C. (). First phase syntax. MS, Oxford University. [Later published as
Ramchand ().]
Ramchand, Gillian C. (). Time and the event: The semantics of Russian prefixes, Nordlyd
: –.
Ramchand, Gillian C. (). Verb Meaning and the Lexicon. A First Phase Syntax. Cambridge:
Cambridge University Press.
Randall, Janet, Angeliek van Hout, Jürgen Weissenborn, and Harald Baayen (). Acquiring
unaccusativity: A cross-linguistic look, in Artemis Alexiadou, Elena Anagnostopoulou, and
Martin Everaert (eds.), The Unaccusativity Puzzle. Explorations of the Syntax-Lexicon
Interface. Oxford: Oxford University Press, –.
Rapoport, Tova R. (). Verbs in depictives and resultatives, in James Pustejovsky (ed.),
Semantics and the Lexicon. Dordrecht: Kluwer, –.
Rapoport, Tova R. (). Structure, aspect, and the predicate, Language : –.
Rappaport, Malka and Beth Levin (). What to do with Theta Roles, in Wendy
Wilkins (ed.), Syntax and Semantics : Thematic Relations. San Diego, CA: Academic
Press, –.
Rappaport Hovav, Malka and Beth Levin (). Building verb meanings, in Miriam Butt and
Wilhelm Geuder (eds.), The Projection of Arguments: Lexical and Syntactic Constraints.
Stanford, CA: CSLI, –.
Rappaport Hovav, Malka and Beth Levin (). An event structure account of English
resultatives, Language : –.
Rappaport Hovav, Malka and Beth Levin (). Reflections on Manner/Result Complemen-
tarity, in Edit Doron, Malka Rappaport Hovav, and Ivy Sichel (eds.), Syntax, Lexical
Semantics, and Event Structure. Oxford: Oxford University Press, –.
Real Puigdollers, Cristina (). La classe dels verbs de mesura: Un cas de verbs fronterers.
Master’s thesis, Universitat Autònoma de Barcelona.
Real Puigdollers, Cristina (). A microparametric approach on goal of motion construc-
tions: Properties of adpositional systems in Romance and Germanic, Catalan Journal of
Linguistics : –.
Real Puigdollers, Cristina (). Lexicalization by Phase. The Role of Prepositions in Argument
Structure and its Cross-linguistic Variation. PhD thesis, Universitat Autònoma de Barcelona.
References
Reinhart, Tanya (). The Theta System: Syntactic realization of verbal concepts. Utrecht:
UiL OTS Working Papers.
Reinhart, Tanya (). The theta system: An overview. Theoretical Linguistics : –.
Reinhold, Heinz (). Zum lateinischen Verbalaspekt, Zeitschrift für Vergleichende Sprach-
forschung auf dem Gebiete der Indogermanischen Sprachen : –.
Richards, Marc D. (). On feature inheritance: An argument from the Phase Impenetra-
bility Condition, Linguistic Inquiry : –.
Rigau, Gemma (). Locative sentences and related constructions in Catalan: ésser/haver
alternation, in Amaya Mendikoetxea and Myriam Uribe-Etxebarria (eds.), Theoretical Issues
at the Morphology-Syntax Interface. Bilbao and Donostia-San Sebastián: Universidad del
País Vasco and Diputación Foral de Gipuzkoa, –.
Ritter, Elizabeth and Sarah T. Rosen (). Delimiting events in syntax, in Miriam Butt and
Wilhelm Geuder (eds.), The Projection of Arguments: Lexical and Syntactic Constraints.
Stanford, CA: CSLI, –.
Rivero, María-Luisa (). Adverb incorporation and the syntax of adverbs in modern Greek,
Linguistics and Philosophy : –.
Roberts, Ian G. (). Agreement and Head Movement: Clitics, Incorporation, and Defective
Goals. Cambridge, MA: MIT Press.
Rojina, Nina (). English Particles, Russian Prefixes, and Prepositional Phrases. Master’s
thesis, Universitetet i Tromsø.
Romagno, Domenica (). Azionalità e transitività: il caso dei preverbi latini, Archivio
Glottologico Italiano : –.
Romagno, Domenica (). Ancora su preverbazione e sistemi verbali. Il caso dei preverbi
greci, Archivio Glottologico Italiano : –.
Romagno, Domenica (). Applicative and causative: Some further reflections upon verbal
prefixation in Greek and Latin, Archivio Glottologico Italiano : –.
Roßdeutscher, Antje (). A syntax-semantics interface for verbs with P-elements in
German, Working Papers of the SFB Incremental Specification in Context: –.
Rosselló, Joana (). El SV I: Verb i arguments verbals, in Joan Solà, Maria Rosa Lloret, Joan
Mascaró, and Manuel Pérez Saldanya (eds.), Gramàtica del català contemporani, vol. .
Barcelona: Empúries, –
Rubio Fernández, Lisardo and Tomás González Rolán (). Nueva gramática latina. Madrid:
Coloquio Editorial.
Schäfer, Florian (). The Syntax of (Anti-)Causatives. External Arguments in Change-of-
State Contexts. Amsterdam: John Benjamins.
Schütze, Carson T. (). Serbo-Croatian second position clitic placement and the
phonology-syntax interface, MIT Working Papers in Linguistics : –.
Selkirk, Elisabeth (). Phonology and Syntax: The Relation between Sound and Structure.
Cambridge, MA: MIT Press.
Serbat, Guy (). Essai de définition du datif latin, in Alfred Bammesberger., and Friedrich
Heberlein (eds.), Akten des VIII. Internationalen Kolloquiums zur Lateinischen Linguistik.
Proceedings of the Eighth International Colloquium on Latin Linguistics. Heidelberg:
Universitätsverlag C. Winter, –.
References
Serbat, Guy (a). Le rôle syntaxique du préverbe: Étude à partir de gradior, Studia
Philologica Valentina : –.
Serbat, Guy (b). Autour de l’objet interne, in Benjamín García Hernández (ed.), Estudios
de lingüística latina. Actas del IX Coloquio Internacional de Lingüística Latina. Universidad
Autónoma de Madrid, – de Abril de . Madrid: Ediciones Clásicas, –.
Shim, Ji Y. and Marcel den Dikken (). The tense of resultatives: The case of Korean. MS,
City University of New York. [Later published as Shim, Ji Y. and Marcel den Dikken ().
The tense of resultatives: The case of Korean, North-Eastern Linguistic Society : –.]
Siegel, Dorothy (). Topics in English Morphology. PhD thesis, Massachusetts Institute of
Technology.
Simpson, Jane (). Resultatives, in Lori Levin, Malka Rappaport Hovav, and Annie Zaenen
(eds.), Papers in Lexical-Functional Grammar. Bloomington, IN: Indiana University Lin-
guistics Club, –.
Slobin, Dan I. (). The many ways to search for a frog: Linguistic typology and the
expression of motion events, in Sven Strömqvist and Ludo Verhoeven (eds.), Relating Events
in Narrative: Typological and Contextual Perspectives. Mahwah, NJ: Lawrence Erlbaum
Associates, –.
Slobin, Dan I. (). Relating narrative events in translation, in Dorit Ravid and Hava Bat-
Zeev Shyldkrot (eds.), Perspectives on Language and Language Development: Essays in
Honor of Ruth A. Berman. Dordrecht: Kluwer, –.
Slobin, Dan I. (). What makes manner of motion salient?, in Maya Hickmann and
Stéphane Robert (eds.), Space in Languages: Linguistic Systems and Cognitive Categories.
Amsterdam: John Benjamins, –.
Smith, Carlotta (). The Parameter of Aspect. Dordrecht: Kluwer.
Snyder, William (). Language Acquisition and Language Variation: The Role of Morph-
ology. PhD thesis, Massachusetts Institute of Technology.
Snyder, William (). On the nature of syntactic variation: Evidence from complex predi-
cates and complex word-formation, Language : –.
Snyder, William (). Parameter theory and motion predicates, in Violeta Demonte and
Louise McNally (eds.), Telicity, Change, and State: A Cross-Categorial View of Event
Structure. Oxford: Oxford University Press, –.
Son, Minjeong (). Directionality and resultativity: The cross-linguistic correlation
revisited, Nordlyd : –.
Son, Minjeong (). Resultatives in Korean revisited: Complementation versus adjunction,
Nordlyd : –.
Son, Minjeong and Peter Svenonius (). Microparameters of cross-linguistic variation:
Directed motion and resultatives, West Coast Conference on Formal Linguistics : –.
Sorace, Antonella (). Gradients in auxiliary selection with intransitive verbs, Language :
–.
Spencer, Andrew and Marina Zaretskaya (). Verb prefixation in Russian as lexical
subordination, Linguistics : –.
Starke, Michal (). Nanosyntax. A short primer to a new approach to language, Nordlyd
: –.
References
Starke, Michal (). Course on nanosyntax. Centre de Lingüística Teòrica, Universitat
Autònoma de Barcelona, Bellaterra, January.
Stiebels, Barbara (). Lexikalische Argumente und Adjunkte. Zum semantischen Beitrag von
verbalen Präfixen und Partikeln. Berlin: Akademie Verlag.
Stiebels, Barbara (). Complex denominal verbs in German and the morphology-semantics
Interface, Yearbook of Morphology : –.
Stiebels, Barbara and Dieter Wunderlich (). Morphology feeds syntax: The case of particle
verbs, Linguistics : –.
Stolova, Natalia (). From satellite-framed Latin to verb-framed Romance: Late Latin as an
intermediate stage, in Roger Wright (ed.), Latin vulgaire-latin tardif VIII: Actes du VIIIe
Colloque International sur le Latin Vulgaire et Tardif, Oxford, – Septembre . Hildes-
heim: Olms-Weidmann, –.
Stolz, Friedrich, Joseph H. Schmalz, Manu Leumann, and Johann B. Hofmann ().
Lateinische Grammatik: Laut- und Formenlehre, Syntax und Stilistik. Munich: C. H. Beck.
Stowel, Timothy (). Origins of Phrase Structure. PhD thesis, Massachusetts Institute of
Technology.
Strigin, Anatoli (). Blocking resultative secondary predication in Russian, ZAS Papers in
Linguistics : –.
Strigin, Anatoli and Assinja Demjjanow (). Secondary predication in Russian, ZAS Papers
in Linguistics : –.
Svenonius, Peter (). The verb-particle alternation in the Scandinavian languages. MS,
Universitetet i Tromsø.
Svenonius, Peter (). Limits on P: Filling in holes vs. falling in holes, Nordlyd : –.
Svenonius, Peter (). Slavic prefixes inside and outside VP, Nordlyd : –.
Svenonius, Peter (). Adpositions, particles and the arguments they introduce, in Eric
Reuland, Tanmoy Bhattacharya, and Giorgos Spathas (eds.), Argument Structure. Amster-
dam: John Benjamins, –.
Sznajder, Lyliane (). Verbes transitifs sans objet en latin, in Benjamín García Hernández
(ed.), Estudios de lingüística latina. Actas del IX Coloquio Internacional de Lingüística
Latina. Universidad Autónoma de Madrid, – de Abril de . Madrid: Ediciones
Clásicas, –.
Tai, James H.-Y. (). Cognitive relativism: Resultative construction in Chinese, Language
and Linguistics : –.
Talmy, Leonard (). Semantics and syntax of motion, in John P. Kimball (ed.), Syntax and
Semantics, vol. . New York: Academy Press, –.
Talmy, Leonard (). Lexicalization patterns: Semantic structure in lexical forms, in Tim-
othy Shopen (ed.), Language Typology and Syntactic Description, vol. : Grammatical
Categories and the Lexicon. Cambridge: Cambridge University Press, –.
Talmy, Leonard (). Path to realization: A typology of event conflation, Berkeley Linguistic
Society : –.
Talmy, Leonard (). Fictive motion in language and ‘ception’, in Paul Bloom, Mary
A. Peterson, Lynn Nadel, and Merrill F. Garrett (eds.), Language and Space. Cambridge,
MA: MIT Press, –.
References
Talmy, Leonard (). Toward a Cognitive Semantics, vol. : Typology and Process in Concept
Structuring. Cambridge, MA: MIT Press.
Tenny, Carol (). Grammaticalizing Aspect and Affectedness. PhD thesis, Massachusetts
Institute of Technology.
Tenny, Carol (). Aspectual Roles and the Syntax-Semantics Interface. Dordrecht: Kluwer.
Terzi, Arhonto (). On null spatial Ps and their arguments, in Guglielmo Cinque and Luigi
Rizzi (eds.), The Cartography of Syntactic Structures, vol. : Mapping Spatial PPs. Oxford:
Oxford University Press, –.
Thayer, William P. (). LacusCurtius. University of Chicago.
Tombeur, Paul (ed.) (). Bibliotheca Teubneriana Latina (CD-ROM). Munich: K. G. Saur
and Turnhout: Brepols.
Tomioka, Naoko (). Resultative Constructions: Cross-Linguistic Variation and the Syntax-
Semantics Interface. PhD thesis, McGill University.
Torrego, Esther (). Unergative-unaccusative alternations in Spanish, MIT Working Papers
in Linguistics : –.
Torrego, Ma Esperanza (). Restricciones de la categoría verbal aspecto sobre los elementos
de duración en latín, Minerva : –.
Tortora, Christina (). Verbs of inherently directed motion are compatible with resultative
phrases. Linguistic Inquiry : –.
Travis, Lisa deMena (). Parameters and Effects of Word Order Variation. PhD thesis,
Massachusetts Institute of Technology.
Travis, Lisa deMena (). Event structure in syntax, in Carol Tenny and James Pustejovsky
(eds.). Events as Grammatical Objects. Stanford, CA: CSLI, –.
Tungseth, Mai E. (). Two structural positions for locative and directional PPs in Norwe-
gian motion constructions, Nordlyd : –.
Turcan, Isabelle (). La dépréverbation dans les langues classiques, Bulletin de la Société de
Linguistique de Paris : –.
Untermann, Jürgen (). Sprachwandel, beobachtet an lateinischen Präverbien, in Hannah
Rosén (ed.), Aspects of Latin: Papers from the Seventh International Colloquium on Latin
Linguistics, Jerusalem, April . Innsbruck: Innsbrucker Beiträge zur Sprachwissenschaft,
–.
Vairel, Hélène (). La Valeur de l’opposition infectum/perfectum en latin, Revue des Études
Latines : –.
Van der Heyde, Klaas (). L’aspect verbal en latin. Problèmes et résultats. Le rôle du
préverbe, Revue des études latines : –.
Van Laer, Sophie (). Le Préverbe ob- en latin: Considérations sémantiques, in Benjamín
García Hernández (ed.), Estudios de lingüística latina. Actas del IX Coloquio Internacional de
Lingüística Latina. Universidad Autónoma de Madrid, – de Abril de . Madrid:
Ediciones Clásicas, –.
Van Riemsdijk, Henk C. (). A Case Study in Syntactic Markedness. The Binding Nature of
Prepositional Phrases in Dutch. Lisse: The Peter de Ridder Press.
Van Valin, Robert D. Jr. (). A synopsis of role and reference grammar, in Robert D. Van
Valin Jr. (ed.), Advances in Role and Reference Grammar. Amsterdam: John Benjamins,
–.
References
Vendler, Zeno (). Linguistics in Philosophy. Ithaca, NY: Cornell University Press.
Vendryès, Joseph (). Les Notions de ‘boire’ et ‘manger’ en indo-européen et l’aspect verbal,
Bulletin de la Société de Linguistique : –.
Verkerk, Annemarie (). Diachronic change in Indo-European motion event encoding,
Journal of Historical Linguistics : –.
Verkuyl, Henk J. (). On the Compositional Nature of the Aspects. Reidel: Dordrecht.
Verkuyl, Henk J. (). A Theory of Aspectuality. The Interaction between Temporal and
Atemporal Structure. Cambridge: Cambridge University Press.
Vikner, Sten (). Verb Movement and Expletive Subjects in the Germanic Languages.
Oxford: Oxford University Press
Vincent, Nigel (). The evolution of c-structure: Prepositions and PPs from Indo-European
to Romance, Linguistics : –.
Washio, Ryuichi (). Resultatives, compositionality and language variation, Journal of East
Asian Linguistics : –.
Wechsler, Stephen (). Resultatives under the ‘event-argument homomorphism’ model of
telicity. Paper presented at the workshop The Syntax of Aspect, Ben Gurion University of
Negev, June. [Later published in Nomi Erteschik-Shir and Tova Rapoport (eds.) (), The
Syntax of Aspect. Deriving Thematic and Aspectual Interpretation. Oxford: Oxford Univer-
sity Press, –.]
Wechsler, Stephen (). Resultatives under the ‘event-argument homomorphism’ model of
telicity, in Nomi Erteschik-Shir, and Tova Rapoport (eds.), The Syntax of Aspect. Deriving
Thematic and Aspectual Interpretation. Oxford: Oxford University Press, –.
Weerman, Fred (). The V Conspiracy. A Synchronic and a Diachronic Analysis of Verbal
Positions in Germanic Languages. Dordrecht: Foris.
Whelpton, Matthew (). Resultatives in Icelandic. A preliminary investigation. MS, Uni-
versity of Iceland.
Whelpton, Matthew (). Building resultatives in Icelandic, Western Conference on Linguis-
tics : –.
Williams, Edwin (). Argument structure and morphology, The Linguistic Review : –.
Wunderlich, Dieter (). An investigation of lexical decomposition: The case of German
be-verbs, Linguistics : –.
Wunderlich, Dieter (). Towards a structural typology of verb classes, in Dieter Wunder-
lich (ed.), Advances in the Theory of the Lexicon. The Hague: Mouton de Gruyter, –.
Wurff, Wim van der (). Null objects and learnability: The case of Latin, Working Papers of
Holland Institute for Generative Linguistics : –.
Wurmbrand, Susanne (). The structure(s) of particle verbs. MS, University of
Connecticut.
Xu, Dan (). Typological Change in Chinese Syntax. Oxford: Oxford University Press.
Žaucer, Rok (). The Role of Verbal Prefixes in Slavic: Evidence from Slovenian Locative
Denominal Verbs. Master’s thesis, University of Ottawa.
Žaucer, Rok (). Slavic prefixes as state morphemes. From state to change-of-state and
perfectivity, in Wolfgang U. Dressler, Dieter Kastovsky, Oskar E. Pfeiffer, and Franz Rainer
(eds.), Morphology and its Demarcations. Selected Papers from the th Morphology Meeting,
Vienna, February . Amsterdam: John Benjamins, –.
References
Žaucer, Rok (). A VP-internal/Resultative Analysis of ‘VP-External’ Uses of Slavic Verbal
Prefixes. PhD thesis, University of Ottawa.
Zeller, Jochen (a). How syntax restricts the lexicon: Particle verbs and internal arguments,
Linguistische Berichte : –.
Zeller, Jochen (b). Particle Verbs and Local Domains. Amsterdam: John Benjamins.
Zhang, Niina N. (). Root merger in Chinese compounds, Studia Linguistica : –.
Zubizarreta, María Luisa and Eunjeong Oh (). On the Syntactic Composition of Manner
and Motion. Cambridge, MA: MIT Press.
Index
∅, see exponent: null/zero exponent in Mateu’s theory –
a instantiated as PlaceP ,
a-/ab(s)- (Latin prefix) –, – prefixed to the verb –, –,
ablative: proleptic
marking a demoted argument – resultative –, , , ,
in directional DPs ,
interpreted as the Ground with prefixed in Unselected Object Constructions
verbs see also passive: adjectival; resultative
and prepositions – construction: adjectival; resultative
abutor (Latin verb) – construction: prefixed adjectival;
accomplishment , resultative construction: complex
accusative: adjectival
directive/directional –, adjunct:
double n., in activity predicates
of extension in change-of-state alternants
interpreted as the Ground with prefixed directional ,
verbs –, further specifying a final location/state ,
in Ground Unselected Object , ,
Constructions expressing manner ,
non-measuring to PathP , ,
and prepositions – to PlaceP , , , , ,
relation to Voice to vP –,
Acedo-Matellán, Víctor n., n. , n. , see also root: adjunction to functional
, , , n., , , n., n. , heads; Figure: demoted
n., , , , n. adposition:
achievement and roots
Ackerman, Farrell in Hale and Keyser’s theory
Acquaviva, Paolo n., in Mateu’s theory
activity , , , , , see also p; Path; Place; PP; preposition
ad- (Latin prefix) , Åfarli, Tor A. n. ,
adjacency: affixation , , , , , , ,
and insertion frames ,
and Morphological Merger –, Afrikaans
adjective: agent noun –
and agreement inflection –, , agentivity ,
–, Agent –
attributive vs. predicative n. see also event: non-agentive
in Hale and Keyser’s theory ago (Latin verb) ,
Index
agreement inflection, see adjective: and Arregi, Karlos , n.
agreement inflection; Place: and Arsenijević, Boban , , , , ,
agreement inflection ,
Aktionsart , , ; see also situation aspect Aske, Jon ,
Alexiadou, Artemis aspect –
alignment Asp ,
Allen, Margaret AspQP (Aspectual Quantity Phrase) –,
allomorphy , , , –
allosemy n. default aspect
Amadas, Laia grammatical, see viewpoint aspect
Ambridge, Ben n. inner, see situation aspect
Amritavalli, Raghavachari outer, see viewpoint aspect
an (German particle) n. see also atelicity; situation aspect; telicity;
An, Duk-Ho viewpoint aspect
Anagnostopoulou, Elena atelicity –
anar (Catalan verb) –, and the central coincidence
Ancient Greek , , –, , – relation ,
andare (Italian verb) with an Effected Object
Anderson, Stephen , n. and motion constructions , –
AP, see adjective with a non-quantity Measurer –,
applicative ,
applicative head , atelic predicate , , , , –
Arad, Maya n. , , and prefixation –
argument , , , and resultativity
added atelicity-signalling expression
external , , , , , , see also event: atelic; FSP
internal , –, Atkins, Beryl T.
interpretation – auxiliary selection:
unselected , with birthing verbs , –
see also Figure: demoted; Ground: in directed motion constructions –,
promoted , –, ,
argument structure in existential constructions
alternations –, and particles
in Borer’s theory – and unaccusativity
configurations –
and deponency Baayen, Harald
in endo-skeletal vs. exo-skeletal Babko-Malaya, Olga
theories – bake –
in Hale and Keyser’s theory – Baker, Mark C. , , ,
in Mateu’s theory – ballare (Italian verb) –
and prefixation , –, –, Barbelenet, Daniel , n., ,
semantics – Bare Phrase Structure ,
see also Causative Alternation; Locative bare plural –, , ,
Alternation; with/against alternation Basque , ,
Index
Bauke, Leah S. n. Carrier, Jill n., ,
be- (Dutch prefix) , – Carvalho, Paulo de , ,
be- (German prefix) , case:
BE-auxiliary, see auxiliary selection alternations –
Beavers, John n., n. and directional DPs and PPs –
Beck, Sigrid –, – Catalan , , , , , , , , , ,
Bende-Farkas, Ágnes , , , , , , , , , , , ,
Bennis, Hans , , , , , , , ,
Berman, Ruth A. , ,
Biblioteca Teubneriana Latina Causative Alternation
bibo (Latin verb) –, –, causativization
Binnick, Robert CAUSE –
Biskup, Petr n. Centineo, Giulia
Boas, Hans C. , – central coincidence relation , ,
Boeckx, Cedric n. change of location
Bohnemeyer, Jürgen n. , change-of-location alternant –,
Bolkestein, Machtelt n. –, , , –
Bonet, Eulàlia change-of-location predicate , ,
Booij, Gert –, n. ,
Borer, Hagit , , , –, , , , , , see also Locative Alternation
, , , , , , , –, change of state
, , alternant –, , , –
Borik, Olga see also Locative Alternation; verb: change-
Bortolussi, Bernard n., of-state verb
Bouchard, Denis n., Chen, Liang n.
boundedness , , –; see also Chomsky, Noam n., , , , , ,
directional PP: bounded; quantity n.,
Brachet, Jean-Paul , , n. Citko, Barbara
Brandt, Silke n. clitic
Bresnan, Joan n. proclitic
Broekhuis, Hans Co-event –, –, –; see also Manner
Bulgarian , , , , co(m)- (Latin prefix) –
Burzio, Luigi n., Coleman, Robert n., n.
Burzio’s generalization complete affection , , –,
Bye, Patrick Complex Directed Motion Construction
in Ancient Greek –, –
c-command , , , –, , in Dutch
C(omplementizer) in German
CP and Hebrew –
Caha, Pavel n. and Javanese –
camminare (Italian verb) – and Korean –
capio (Latin verb) , in Latin –
Cappelle, Bert in Mandarin Chinese
Carlson, Gregory n. and prefixation –, , –
Index
Complex Directed Motion Construction resultative construction; s-framed
(cont.) construction; Unselected Object
situation aspect –, – Construction; v-framed construction
in Slavic Core Schema –
unaccusativity – as an adjective
Complex Effected Object Construction and the goal/source distinction
absence in v-framed languages – and the s-/v-framed distinction , ,
in English , –
in Latin – corpora –, ,
complex predicate n. , , , correre (Italian verb) , , ,
–, courir (French verb)
composition n. ; see also compound Craenenbroeck, Jeroen van n. , n. ,
compositional meaning , ; see also n.
structural semantics Crane, Gregory R.
compositional semantics, see structural Croatian –
semantics Crocco Galèas, Grazia –
compound: cross-linguistic variation:
Compounding Parameter and adjectival resultative
compounding v and a Manner root , constructions –
endocentric root compound – and complex resultative predicates
particle-verb compound and directionality
VV compound as microparametric
computational system , n. and realization of syntactic nodes ,
Comrie, Bernard –, –, –
conceptual content , , , , , , , Csirmaz, Aniko n. ,
, ; see also Encyclopaedia: Cuervo, María Cristina
encyclopaedic semantics cumulative predicate
conceptual semantics, see Encyclopaedia: curro (Latin verb) , ,
encyclopaedic semantics cycle:
conceptual system , , – cyclic head –,
conflation cyclic domain –
in Hale and Keyser’s theory , – non-cyclic head
in Mateu’s theory , of syntactic computation –,
in Talmy’s theory , see also phase; Spell-Out
Conformation –, –, – Czech , , , , , , , ,
construct
construction D-structure –,
causative construction n. Damonte, Federico
constructionist approach – Danish
as a lexically stored unit , dat’ (Russian verb) , –
motion construction dative:
see also Complex Directed Motion benefactive/malefactive ,
Construction; Effected Object directional –
Construction; gerundive construction; with give-verbs
Index
dative: (cont.) durative adverbial , , , , –,
possessive , , ,
with verbs of utterance Dutch –, , –, n., –, ,
dativus commodi/incommodi, see dative: –, –, n.,
benefactive/malefactive dynamicity ,
De Belder, Marijke n. , n. , n. in s- vs. l-syntax
De lingua latina, see Varro
de- (Latin prefix) n., – e(x)- (Latin prefix) , –
deambulo (Latin verb) Early Insertion, see roots: Early Insertion
deictic , , ebibo (Latin verb) –, –, –
Deictic –, Echarte Cossío, María José ,
deixar (Catalan verb) – edge
Demirdache, Hamida Effected Object –, –, –, ,
Demjjanow, Assinja ein (German particle) –,
Den Dikken, Marcel n., n. , n., eixir (Catalan verb) –
, , , n. É. Kiss, Katalin , n. , –
depictive secondary predication , el (Hebrew preposition) –
derivation: Embick, David , , , –, , n.,
in the Locative Alternation , , , ,
morphological n. , n. Empty Category Principle
PF derivation – en-extraction ; see also ne-cliticization
syntactic – Encyclopaedia , , , ,
destroy encyclopaedic entry , ,
Di Sciullo, Anna-Maria n. , n. encyclopaedic semantics , ,
Dictionnaire Latin-Français , see also conceptual content
directed/directional motion , , , end state, see result state
–, , , –, , , , endo-skeletal approach –
; see also Complex Directed Motion endo-skeleton –
Construction see also lexicalist analysis; projectionist
directional PP: theory
bounded , –, , , , entity type
unbounded –, , eo (Latin verb) ,
see also directed/directional motion; EP (Eventive Phrase) –
trajectory equito (Latin verb) –,
dis- (Latin prefix) , -er, see agent noun
Distributed Morphology – Ernout, Alfred , , –, , ,
divisive predicate Espinal, M. Teresa
do (Latin verb) – essive case, see -na
Dowty, David , n., , event , , ,
DP: atelic , , ,
as a cyclic domain change of state/location event , ,
as a non-relational element complex , ,
directional creation/consumption event , , ,
duplication – ,
Index
event (cont.) fictive motion ,
dynamic Figure , ,
externally originated demoted ,
Framing event see also adjunct: in change-of-state
motion event –, alternants; Unselected Object
non-agentive , Contruction: Figure Unselected Object
non-dynamic – Construction
non-externally originated Filip, Hana n., , ,
event participant , , , fill n.
resultative Fillmore, Charles J. n.
event structure , , , final location , , ,
telic , , , –, –, , specified by a DP
transitive –, , further specified by a low adjunct
unaccusative –, , in Ground Unselected Object
Exceptional Case Marking Constructions
Exhaustive Lexicalization Principle , modified by a durative adverbial
existential sentence , encoded by a particle
exo-skeletal approach –, , , encoded by a prefix , , , ,
exo-skeleton – in resultative constructions –
see also generative-constructivist theory; encoded by the verb
neo-constructionist approach see also change of location: change-of-
exponent – location alternant; dative: directional;
default/elsewhere exponent result state
exponent-defectiveness – final state, see result state
null/zero exponent – Finnish –, –,
of roots Finno-Ugric –,
see also PF crash; Vocabulary Insertion; flo (Latin verb)
Vocabulary Item Folli, Raffaella n., , , , n.,
Extension Condition n. n., , ,
Eythórsson, Thorhállur n. Fong, Vivienne , ,
for-adverbial, see durative adverbial
Fábregas, Antonio , , , frame adverbial , , ,
facio (Latin verb) –, – Frame Semantics n.
Fan, Sheng-yang French –, , n., , n., , n.
Farkas, Donka FSP (shell Functional Projection)
Fay, Edwin W. functional architecture, see functional
feature: structure
contextual features in Vocabulary functional head –,
Items , with an adjoined root ,
lexical , – in Borer’s theory –
in Mateu’s theory interpretation –, ,
morphosyntactic , , , phonological realization , –, –
phi-feature , , , , functional structure –, ,
uninterpretable Further Specification Constraint
Index
Gaffiot, Félix , , , , , , , Halliday, Michael A. K. n.
, , , Harley, Heidi , –, , , , , ,
galleggiare (Italian verb) –, , – , , n., , , n.,
García Hernández, Benjamín n., , Haudry, Jean
n., , , Haugen, Jason D. n. , n.
Gehrke, Berit , , , , , , n., HAVE-auxiliary, see auxiliary selection
, , , , , , , , Haverling, Gerd , , n., n.
, , Hay, Jennifer n.
generative-constructivist theory ; see also head:
exo-sketal approach; complex , , , –
neo-constructionist approach cyclic head, see cycle: cyclic head
German , , n., , , –, n. , eventive head –, , , –
, , n., n., , n., , , head movement , , , ,
, –, , –, n., , Head Movement Constraint
Germanic , , , n. , , , , , lexical –, –,
, , , , , , –, and Morphological Merger –, ,
gerundive construction n. see also functional head; specifier: specifier-
Ghomeshi, Jila n. head relation; v
Gianollo, Chiara n. Hebrew , , , –
Gibert Sotelo, Elisabeth n. Hegedűs, Veronika
Gleitman, Lila n. Heslin, Thomas P. Jr. n.
goal , Hewson, John n.
in the dative Hoekstra, Teun , , , , , ,
as opposed to probe , , n., , ,
Goldberg, Adele n., n. Hofmann, Johann B. , , , n.,
González Rolán, Tomás –, ,
grammatical formative holistic effect , –,
grammatically irrelevant meaning n. hornear (Spanish verb)
grammatically relevant meaning n., , Horrocks, Geoffrey –, , –,
Greek , , –, n., ,
Green, Georgia Horvath, Julia
Grimshaw, Jane n. Hout, Angeliek van n. , ,
Ground , , –, Hungarian , , , –
Central Ground
promoted I-phrasing
Terminal Ground Iacobini, Claudio –
see also Unselected Object Contruction: Icelandic n. , , –, ,
Ground Unselected Object Contruction idiomatic interpretation ; see also non-
Guo, Jiansheng n. compositional meaning; special
meaning
Haider, Hubert immisceo (Latin verb) –
Hale, Kenneth , –, , , , , , imperfective, see viewpoint aspect:
, , , , imperfective
Halle, Morris , , , , , , , in- (Latin prefix) –
Index
in-adverbial, see frame adverbial kuleh (Korean verb)
Inagaki, Shunji n. Kuno, Susumu n.
inalienable possession , ,
incorporation , n., –, n. l(exical)-syntax –,
individual-level predicate n. l-syntactic representation –, , ,
inflectional morphology , n. , , LacusCurtius
–, ; see also adjective: and Landau, Barbara n.
agreement inflection; Place: and Landgraf, Gustav von n.
agreement inflection Larson, Richard K. ,
-ing (gerund suffix) Late Insertion , , –
insertion frame – Late Linearization Hypothesis
Intonational Phrase Latin prefix:
Irube, Kaneharu and the directional dative –
Italian –, , –, , , , , , in the expression of bounded
, , , –, – directionality , –
Itinera Electronica in the Locative Alternation –
izdat’ (Russian verb) – in Unselected Object
Constructions –
Jackendoff, Ray , n. , see also prefix
Japanese , , –, , Latin:
Javanese , – Late Latin n.
Jayaseelan, Karattuparambil A. Latin Dictionary
Jensen, Povl J. periodization –
Julien, Marit , Lavency, Marius , n.
Lavidas, Nikolaos , n.
k (Czech/Russian preposition) le (Hebrew preposition) –
Kaplan, Ronald M. n. Legendre, Géraldine n.
Kayne, Richard , , n., Lehmann, Christian , n., n., –,
-key (Korean suffix) , ,
Keyser, Samuel J. , –, , , , , , Lehrer, Adrienne n.
, , , , , Lemaire, Maryvonne , , n.,
Kiparsky, Paul , ,
Klein, Wolfgang Le Roux, Cecile
Klipple, Elisabeth Leumann, Manu
Kolehmainen, Leena – Levin, Beth , , , , , , , , ,
Koopman, Hilda , n. , , , , , , , ,
Koot, Hans van de n. ,
Kopecka, Anetta , Levinson, Lisa , , , , n.
Korean n., – Lewandowski, Wojciech , ,
Kratzer, Angelika n. , , , , n., ,
, , n. , , – Lewis, Charlton T. , , ,
Krifka, Manfred lexical category:
-ksi (Finnish case suffix) in endo-skeletal theories
Kühner, Raphael in Hale and Keyser’s theory
Index
lexical entry , , MacDonald, Jonathan E. , , , ,
lexical gap ,
lexical item –, , – McIntyre, Andrew n. , , , , ,
in Hale and Keyser’s theory , , n. , n. , , ,
lexical marking n., –, n., n., , , , ,
lexical semantics , , malefactive dative, see dative: benefactive/
and deponency malefactive
lexical semantic representation – Mandarin Chinese
lexical subordination , Manner , ; see also Co-event; verb:
lexicalist analysis ; see also endo-skeletal manner-of-motion verb
approach; projectionist theory mapping algorithm ,
lexicalization , , Marácz, László
in Nanosyntax ; see also Exhaustive Marantz, Alec , , , –, n. ,
Lexicalization Principle –, , n., , –, –, ,
lexicon-syntax interface , , , , n., , n.,
LF –, , Marchand, Hans n.
Liddell, Henry G. Markova, Angelina , , , ,
Lieber, Rochelle n. mass , , –, , ; see also
Lindvall, Ann quantity: non-quantity
Linearization , , – Massam, Diane n. ,
linear intervention , –, Mateu, Jaume , n., , , –, , , ,
listeme , , –; see also root , , , , , , , , , , , ,
little a, see a , n., , , , , n.,
little n, see n , n. , , n., n. , ,
little p, see p , n., , , , , –
little v, see v Matushansky, Ora n.
locality , , measure phrase –
Location (theta-role) Measurer –, –, –
Locative Alternation –, – meg (Hungarian particle)
with adjectives n. Meillet, Antoine n., , ,
derivational analysis Mendikoetxea, Amaya n.
non-derivational analysis Merchant, Jason
and prefixation – Merge ,
and the s-/v-framed distinction – metonymy
see also change of location: change-of- Meurant, Alain
location alternant; change of state: microparametric theory, see cross-linguistic
change-of-state alternant; fill variation: as microparametric
locatum verb, see verb: location/locatum verb Miller, D. Gary , , n.
Löfstedt, Bengt Minimalist Program ,
Lohndal, Terje n. minimality
López Moreda, Santiago n. Modern Greek, see Greek
Lowering Molinari, Danielle
Lüdeling, Anke , , morph
Luraghi, Silvia , , portmanteau morph ,
Index
morpheme: non-cyclic head, see cycle: non-cyclic head
dissociated morpheme n. non-eventive relation, see [r]
in Distributed Morphology – non-quantity, see quantity
f(unctional)-morpheme – non-relational elements –, , –, ;
l(exical)-morpheme , see also roots
realization , Norwegian ,
and Talmy’s typology , Noyer, Rolf , , , , –, –,
see also morph; root
Morphological Merger – ob- (Latin prefix) , –
Motion , obduco (Latin verb) –
Moussy, Claude n. object:
Mulder, Hotze n. cognate
Mulder, René , , , , , , , and mapping algorithms ,
, , non-prototypical
Munaro, Nicola n. null –
omission
n , unselected
-na (Finnish case suffix) see also Unselected Object Construction
Nanosyntax n., , oblique case
Napoli, Donna Jo , Occitan
Narasimhan, Bhuvana n. Oh, Eunjeong , , , , , n.,
Narrow Lexicon Olsen, Susan
ne-cliticization –; see also en-extraction Oltra-Massuet, Isabel n.,
neco (Latin verb) , – Oniga, Renato , , , n., n., n.
Nedjalkov, Vladimir P. Oostendorp, Marc van
Neeleman, Ad n., open
Neeman, Yonni open value, see functional head: in Borer’s
negation theory
see also verb: ab-verb of denial Originator
neo-constructionist approach , , , , in Borer’s theory
and Complex Directed Motion out-prefixation
Constructions , Oya, Toshiaki –, n.
and Unselected Object Constructions
see also exo-skeletal approach; generative- p
constructivist theory in Svenonius’s theory
Nespor, Marina , see also adposition; Path; Place; PP;
Nevins, Andrew preposition
Newell, Heather , Padrosa-Trias, Susanna ,
Nisard, Désiré Panhius, Dirk n.
node-pruning Pantcheva, Marina B.
nominative , – Pantelia, Maria
non-compositional meaning , , , , participle:
; see also idiomatic interpretation; in absolute constructions n.
special meaning of deponent verbs –
Index
participle: (cont.) Peyraube, Alain n.
in resultative constructions PF
in Small Clauses PF crash –
particle: operations –,
licensing AspQP n. see also syntax-morphology interface
in complex predicates n. Pfau, Roland
in Dutch – Pfister, Raimund
in English , , –, , , , phase –, ; see also cycle: cyclic
, – domain; special meaning
in German n., –, n., –, , Pinault, Georges-Jean
– Pinker, Steven , , n., n.
in the Locative Alternation Pinkster, Harm , , , , n. ,
particle shift , n., n., , , ,
in Pseudoreversatives – Place –, , –, ,
reversative and agreement inflection ,
topicalization –, Compl-Place, see Ground
unaccusative realization –
as an unergative preposition – spec-Place, see Figure
unpredicated n. see also adjunct: to PlaceP
in Unselected Object po- (Russian prefix) , –
Constructions –, n. , – Polish , –
see also Danish; Finnish; Hungarian; Pottier, Bernard n.
Icelandic; Mandarin Chinese; up Pourcel, Stéphanie
passive: PP
adjectival – inducing telicity –
and deponency – in Unselected Object Constructions
and prefixed verbs , , with verbs of utterance
Path –, –, , –, – see also adposition; directional PP; durative
diachronic development – adverbial; frame adverbial; goal;
PathP in Latin – Locative Alternation; Path; preposition;
Spec-Path, see Measurer resultative construction: PP resultative
in strong s-framed languages –, , construction; source
, prefix:
in Talmy’s theory external , –, , –
in v-framed languages –, –, in adjectival resultative
in weak s-framed languages –, –, constructions –,
–, in Ancient Greek –
see also adjunct: to PathP in atelic predicates –
Patient – in Catalan n.
perfect tense , , , , –, in Dutch –, –, –
perfective, see viewpoint aspect: perfective in German n., , , –
Perrot, Jean internal –
Perseus Digital Library Project lexical n.; see also prefix: internal
Pesetsky, David n. , in non-dynamic predicates , n.
Index
prefix: (cont.) Rappaport Hovav, Malka , , , , , ,
stacking , , , , , , , ,
scopal relations with the verb, see verb: ab- re- (English prefix) n.,
verb of denial re- (Latin prefix) n., –,
superlexical n.; see alsoprefix: external readjustament rules
see also Latin prefix; Slavic prefix Real Puigdollers, Cristina , , n. , ,
preposition: n., , , n., , n.,
in activity predicates – realization, see accomplishment
ambiguous , reddo (Latin verb) ,
in Latin – Reinhart, Tanya n.
expressing unbounded direction Reinhold, Heinz
see also adposition; p; Path; Place; PP relational elements , , –
preverb , , , , –, relational semantics
–, n.; see also prefix relational syntax
Prinzhorn, Martin relinquo (Latin verb) ,
probe , result state , , ,
Projection Principle n. in change-of-state alternants
projectionist theory ; see also endo-skeletal further specified by a low adjunct ,
approach; lexicalist analysis , ,
proper government with instrumental verbs –
Pseudoreversative – modified by a durative adverbial ,
Puskás, Genoveva –,
Pustejovsky, James and out-prefixation
Putnam, Michael n. encoded by a particle
Pylkkänen, Liina , , , , encoded by a prefix , , ,
in Pseudoreversatives –
quantity – in resultative constructions –, ,
non-quantity , –,
subject-of-quantity – encoded in the verb , , –, ,
and (a)telicity –
see also aspect: AspQP; Measurer see also final location
result:
r (non-eventive relation) R(esult) ,
R (source relation) result predicate , –, , , ,
-ra/-re (Hungarian case suffix) , , , , –, , ,
Raising –, –, –; see also see also final location
Morphological Merger resultative construction –
Ramchand, Gillian C. , , , , adjectival –, , , , , ,
, n., , , n., , , –, , , , –, ,
, , –, ,
Randall, Janet H. n., –, , expressing change of colour
range assignment complex –, , –, , , ,
Rapoport, Tova R. , , n., , –, , , , , , ,
, –, –, –, , –
Index
resultative construction (cont.) raising to functional heads
complex adjectival , , , , , inducing telicity –
–, –, , , –, , see also listeme
, –, – Rosen, Sarah T. n.
(complex) PP resultative Roßdeutscher, Antje n.
construction –, n., , , Rosselló, Joana
, –, Rubio Fernández, Lisardo –,
with a prefixed verb –,
simple –, , , , s(atellite)-framed construction , , ,
strong –, n. , ,
weak –, , , –, see also Complex Directed Motion
resultative predicate, see result: result Construction; Locative Alternation;
predicate Pseudoreversative; Unselected Object
resultativity: Construction
and prefixation , –, , , , s(atellite)-framed language
, strong , –
see also final location; result: result weak , –, –
predicate; resultative construction; see also Locative Alternation: and the s-/v-
resultative predicate; result state; telicity: framed distinction
and resultativity s(entential)-syntax –,
resulting location, see final location Sadock, Jerrold
resulting state, see result state salio (Latin verb)
Richards, Marc D. saltare (Italian verb)
Rigau, Gemma , , , , , satellite
Ritter, Elizabeth n. Schäfer, Florian n.
Rivero, María-Luisa Schmalz, Joseph H.
Roberts, Ian G. n. Schütze, Carson T.
Rojina, Nina n., , scope, see prefix: scopal relations with the verb
Romagno, Domenica , n., n., Scott, Robert
, Secondary Imperfective, see viewpoint aspect:
Romance –, , , , –, , , Secondary Imperfective
n., , , , n. , n., secondary predicate:
, , , , –, , –, depictive , ,
, – resultative, see result predicate
root Selkirk, Elisabeth
adjunction to functional heads – Sells, Peter n.,
complex semantic construal
Early Insertion , – Sequence of Identical Events
naming an instrument interpretation
interpretation , Sequence of Similar Events interpretation
licensing theory of roots Serbat, Guy , n., ,
naming a result Serbo-Croatian , , ,
root ontologies shell Functional Projection, see FSP
position Shim, Ji Y. , n.
Index
Short, Charles , , , Sorace, Antonella ,
Siegel, Dorothy source ,
Simpson, Jane n. source relation, see [R]
Single Delimiting Constraint span
situation aspect – Spanish , , , , n., , ,
and case alternations – , , ,
in Complex Directed Motion spargo (Latin verb) , –
Constructions –, – special meaning , –, ; see also
and Ground Unselected Object idiomatic interpretation; non-
Constructions – compositional meaning
in complex resultative specifier:
constructions – and roots
see also Aktionsart; aspect; atelicity; telicity; specifier-head relation
viewpoint aspect see also Figure; Measurer
skočit (Czech verb) , Spell-Out
Slavic prefix: cyclic Spell-Out –
and bounded directed motion – see also cycle: cyclic domain
and perfectivity Spencer, Andrew n., ,
in Unselected Object Constructions – Split S-framedness Hypothesis
see also prefix: external; prefix: internal stage-level predicate n.
Slavic: Starke, Michal n. , n.,
absence of complex adjectival stativity
resultatives – state
absence of complex PP resultatives – stative event ,
complex resultative constructions – see also verb: stative
incremental verbs Stavrou, Melita –, , –, n.,
telic unprefixed verbs , – ,
viewpoint aspect – Stegmann, Carl
see also prefix: external; prefix: internal; Stiebels, Barbara , n., n. , ,
Unselected Object Constructions: in ,
Slavic Stolova, Natalia n.
Slobin, Dan I. n., Stolz, Friedrich
Small Clause , Stowell, Timothy
and Germanic prefixed verbs , Strigin, Anatoli ,
in Latin facio-resultative structural semantics , , , ,
constructions – subcategorization frame
in Unselected Object Constructions subject:
and verb-particle combinations n. , and mapping algorithms ,
passive
smile , postverbal
Smith, Carlotta unaccusative ,
Snyder, William , , –, –, see also quantity: subject-of-quantity
–, – sublative, see -ra/-re
Son, Minjeong , n., –, – Subset Principle ,
Index
sum (Latin verb) –, , n. Terzi, Arhonto
suppletion –, tha (Greek particle)
Svenonius, Peter , , , , , –, Tham, Shiao Wei n.
, , , , , , , n., Thayer, William P.
–, thematic role, see theta role
Swift, Mary n. , thematic vowel n.
Syntactic Hierarchical Structure All the Way Theme ,
Down Incremental Theme ,
syntax-morphology interface , – theta-grid
mismatches theta-role ,
Szantyr, Anton , , , n., Thomas, François , , –, , ,
Sznajder, Lyliane n. to
Tomaso, Vittorio di n.
T (transitional relation) Tombeur, Paul
T(ense) , , –, , , Tomioka, Naoko n.
Tai, James H.-Y. n. topicalization ,
Takami, Ken-ichi n. Torrego, Esther
Talmy, Leonard –, , –, , , Torrego, Mª Esperanza n.
, , , , , –, , Tortora, Christina
, trajectory:
telic adverbial, see frame adverbial bounded
telicity – unbounded
determining aspectual reference see also directional PP: bounded;
in Borer’s theory – directional PP: unbounded
in Complex Directed Motion transfer of possession predicate
Constructions – transition , ,
with an Effected Object transitional relation, see [T]
in Ground Unselected Object transitive predicate , ,
Constructions in Mateu’s theory
and internal prefixes , – transitivization , n.
and the Locative Alternation in Hale and Keyser’s theory ,
in Mateu’s theory –, , translative case, see -ksi
induced by particles , Travis, Lisa deMena n. , , n.
and resultativity –, , Tungseth, Mai E.
telic predicate , , , , , Turcan, Isabelle n.
telicity-signalling expression , n.
telos , unaccusativity:
and unaccusativity – in Borer’s theory –
depending on the verbal root –, in Complex Directed Motion
see also aspect: AspQP; event: telic; Constructions –
unergativity: telic unergative predicate diagnostic , , –
Tenny, Carol , , , in existential constructions
terminal coincidence relation , in Mateu’s theory
terminal node –, , unaccusative predicate
Index
unaccusativity: (cont.) and Unselected Object Constructions
in Unselected Object see also existential sentence; Locative
Constructions – Alternation: and the s-/v-framed
see also particle: unaccusative; subject: distinction; Path: in v-framed languages
unaccusative; telicity: and V
unaccusativity; verb: unaccusative Vairel, Hélène n.
unboundedness, see directional PP: Van der Heyde, Klaas , –, ,
unbounded; mass; trajectory: Van Laer, Sophie –
unbounded Van Riemsdijk, Henk C.
unergativity: Van Valin, Robert D. Jr. n.
in Mateu’s theory Varro
telic unergative predicate Vector –, –
see also verb: unergative Vendler, Zeno
univerbation , , –, –, Vendryès, Joseph ,
Universal Grammar , venio (Latin verb) , ,
Unselected Object Construction – ver- (German prefix)
in Ancient Greek verb:
Figure Unselected Object ab-verb of denial –
Construction – birthing verb n., –, –
Ground Unselected Object causative verb , ,
Construction –, , change-of-state verb , –, , , ,
in Slavic – –, , n., , , , –
strong/weak verb class
unspoken languages – creation verb , , n.
Untermann, Jürgen n. deadjectival , –,
up –, , deponent , n.
Uribe-Etxebarria, Myriam determined verb
utor (Latin verb) explicit creation verb n.
give-verb –
v in Hale and Keyser’s theory
and domains for idiomatic implicit creation verb n.
interpretation incremental
flavours , , light , , ,
interpretation location/locatum verb –
relation with Path manner-of-motion verb , –,
see also head: eventive head –, , –, , ,
v(erb)-framed construction , , , , ,
, naming an instrument
v(erb)-framed language –, run-verb –
directed motion constructions , –, sound emission verb
– stative
and Effected Object Constructions surface contact verb ,
and resultative constructions , –, unaccusative ,
unergative ,
Index
verb: (cont.) Washio, Ryuichi –,
of utterance Wechsler, Stephen , ,
see also directed/directional motion; prefix; Weerman, Fred ,
transitive predicate; unaccusativity; Weissenborn, Jürgen
unergativity Whelpton, Matthew n. , , ,
vergärtnen (German verb) Williams, Edwin n.
Verkerk, Annemarie n. with/against alternation n.
Verkuyl, Henk J. word:
Verspoor, Cornelia n. phonological/prosodic –, , ,
viewpoint aspect – and Spell-Out domains –
in Ancient and Modern Greek – relation with syntactic atoms
imperfective –, –, – world knowledge
in Latin vs. Slavic – Wunderlich, Dieter n. , , ,
perfective –, –, – Wurff, Wim van der n.
Secondary Imperfective Wurmbrand, Susanne n.
and situation aspect –, –,
– Xu, Dan n.
see also aspect: default aspect
Vikner, Sten Y-model of grammar
Vincent, Nigel n.
Vocabulary Zaretskaya, Marina n., ,
Vocabulary Insertion –, – Žaucer, Rok ,
Vocabulary Item –, – Zeller, Jochen n. , n., n., ,
Vogel, Irene , , ,
Voice , , –, , , , Zhang, Niina N. n. ,
n. , ; see also Originator zu (German particle) –
volo (Latin verb) Zubizarreta, María Luisa , , , ,
vůči (Czech preposition) , n.,
O X F O R D S T U D I E S I N T H E O R E T I C A L L I N GU I S T I C S
Published InterPhases
The Syntax of Silence Phase-Theoretic Investigations of Linguistic Interfaces
Sluicing, Islands, and the Theory of Ellipsis edited by Kleanthes Grohmann
by Jason Merchant Negation in Gapping
Questions and Answers in Embedded Contexts by Sophie Repp
by Utpal Lahiri A Derivational Syntax for Information Structure
Phonetics, Phonology, and Cognition by Luis López
edited by Jacques Durand and Bernard Laks edited by Anastasia Giannakidou and Monika Rathert
At the Syntax-Pragmatics Interface The Syntax of Sentential Stress
Concept Formation and Verbal Underspecification in by Arsalan Kahnemuyipour
Dynamic Syntax Tense, Aspect, and Indexicality
by Lutz Marten by James Higginbotham
The Unaccusativity Puzzle Lexical Semantics, Syntax, and Event Structure
Explorations of the Syntax-Lexicon Interface edited by Malka Rappaport Hovav, Edit Doron, and Ivy
edited by Artemis Alexiadou, Elena Anagnostopoulou, Sichel
and Martin Everaert
About the Speaker
Beyond Morphology Towards a Syntax of Indexicality
Interface Conditions on Word Formation by Alessandra Giorgi
by Peter Ackema and Ad Neeleman
The Sound Patterns of Syntax
The Logic of Conventional Implicatures edited by Nomi Erteschik-Shir and Lisa Rochman
by Christopher Potts edited by Phoevos Panagiotidis
Paradigms of Phonological Theory Interfaces in Linguistics
edited by Laura Downing, T. Alan Hall, and Renate New Research Perspectives
Raffelsiefen edited by Raffaella Folli and Christiane Ulbrich
The Verbal Complex in Romance Negative Indefinites
by Paola Monachesi by Doris Penka
The Syntax of Aspect Events, Phrases, and Questions
Deriving Thematic and Aspectual Interpretation by Robert Truswell
Edited by Nomi Erteschik-Shir and Tova Rapoport
Dissolving Binding Theory
Aspects of the Theory of Clitics by Johan Rooryck and Guido Vanden Wyngaerd
by Stephen Anderson The Logic of Pronominal Resumption
Canonical Forms in Prosodic Morphology by Ash Asudeh
by Laura J. Downing Modals and Conditionals
Aspect and Reference Time by Angelika Kratzer
by Olga Borik The Theta System
Direct Compositionality Argument Structure at the Interface
edited by Chris Barker and Pauline Jacobson edited by Martin Everaert, Marijana Marelj, and Tal
A Natural History of Infixation Siloni
by Alan C. L. Yu Sluicing
Phi-Theory Cross-Linguistic Perspectives
Phi-Features Across Interfaces and Modules edited by Jason Merchant and Andrew Simpson
edited by Daniel Harbour, David Adger, and Susana Béjar Telicity, Change, and State
French Dislocation A Cross-Categorial View of Event Structure
Interpretation, Syntax, Acquisition edited by Violeta Demonte and Louise McNally
by Cécile De Cat Ways of Structure Building
Inflectional Identity edited by Myriam Uribe-Etxebarria and Vidal Valmala
edited by Asaf Bachrach and Andrew Nevins The Morphology and Phonology of Exponence
Lexical Plurals edited by Jochen Trommer
by Paolo Acquaviva Count and Mass Across Languages
Adjectives and Adverbs edited by Diane Massam
Syntax, Semantics, and Discourse Genericity
Edited by Louise McNally and Christopher Kennedy edited by Alda Mari, Claire Beyssade, and Fabio Del Prete
Strategies of Quantification by Daniela Isac
edited by Kook-Hee Gil, Steve Harlow, and George Sentence and Discourse
Tsoulas edited by Jacqueline Guéron
Nonverbal Predication Optimality-Theoretic Syntax, Semantics, and
Copular Sentences at the Syntax-Semantics Interface Pragmatics
by Isabelle Roy From Uni- to Bidirectional Optimization
Diagnosing Syntax edited by Géraldine Legendre, Michael T. Putnam, Henri-
edited by Lisa Lai-Shen Cheng and Norbert Corver ëtte de Swart, and Erin Zaroukian
Pseudogapping and Ellipsis The Morphosyntax of Transitions
by Kirsten Gengel A Case Study in Latin and Other Languages
Syntax and its Limits Víctor Acedo-Matellán
edited by Raffaella Folli, Christina Sevdali, and Robert
Truswell Published in association with the series
The Oxford Handbook of Linguistic Interfaces
Phrase Structure and Argument Structure
edited by Gillian Ramchand and Charles Reiss
A Case Study of the Syntax-Semantics Interface
by Terje Lohndal In preparation
Edges in Syntax Modality Across Syntactic Categories
Scrambling and Cyclic Linearization edited by Ana Arregui, Mari Luisa Rivero, and Andrés
by Heejeong Ko Pablo Salanova
The Syntax of Roots and the Roots of Syntax Phi Syntax
edited by Artemis Alexiadou, Hagit Borer, and Florian A Theory of Agreement
Schäfer by Susana Béjar
Causation in Grammatical Structures Stratal Optimality Theory
edited by Bridget Copley and Fabienne Martin by Ricardo Bermúdez Otero
Continuations and Natural Language Phonology in Phonetics
by Chris Barker and Chung-chieh Shan by Abigail Cohn
The Semantics of Evaluativity Concealed Questions
by Jessica Rett by Ilaria Frana
External Arguments in Transitivity Alternations Lexical Semantics and Morphosyntactic Variation
by Artemis Alexiadou, Elena Anagnostopoulou, and Flor- by Itamar Francez and Andrew Koontz-Garboden
ian Schäfer Generality and Exception
Control and Restructuring by Ivan Garcia-Alvarez
by Thomas Grano Computing Optimality
The Interaction of Focus, Givenness, and Prosody by Jason Riggle
A Study of Italian Clause Structure Pragmatic Aspects of Scalar Modifiers
The Morphosyntax of Gender by Osamu Sawada
by Ruth Kramer Gradience in Split Intransitivity
The Morphosyntax of Imperatives by Antonella Sorace
Potrebbero piacerti anche
- A Theory of Struktur ParadigmDocumento325 pagineA Theory of Struktur ParadigmSory SukmantaraNessuna valutazione finora
- Parameter Hierarchies and Universal Grammar Ian Roberts Full ChapterDocumento67 pagineParameter Hierarchies and Universal Grammar Ian Roberts Full Chapterrose.bachman558100% (2)
- Pragmatic Aspects of Scalar Modifiers The Semantics Pragmatics Interface 1St Edition Osamu Sawada All ChapterDocumento67 paginePragmatic Aspects of Scalar Modifiers The Semantics Pragmatics Interface 1St Edition Osamu Sawada All Chapterjeanette.healy840100% (2)
- Inflectional Morphology A Theory of Paradigm StrucDocumento26 pagineInflectional Morphology A Theory of Paradigm StrucGabriel BrantNessuna valutazione finora
- The Lexical Semantics of The Arabic VerbDocumento216 pagineThe Lexical Semantics of The Arabic VerbAhmad Al-Halabi100% (1)
- Syntax and Its Limits PDFDocumento479 pagineSyntax and Its Limits PDFcristian david urueña uribeNessuna valutazione finora
- The Lexical Semantics of The Arabic Verb Peter John Glanville Full ChapterDocumento50 pagineThe Lexical Semantics of The Arabic Verb Peter John Glanville Full Chapterjason.johnson763100% (16)
- Course Title Syntax Course Code Credit Hours Pre-Requisite ObjectivesDocumento2 pagineCourse Title Syntax Course Code Credit Hours Pre-Requisite ObjectivesAmir HamzaNessuna valutazione finora
- Article About Linguistic TheoryDocumento10 pagineArticle About Linguistic TheoryDoorHingeNessuna valutazione finora
- DWORKIN - A Guide To Old SpanishDocumento163 pagineDWORKIN - A Guide To Old SpanishCamilla RastelyNessuna valutazione finora
- Inquisitive Semantics 1St Edition Ivano Ciardelli Full ChapterDocumento67 pagineInquisitive Semantics 1St Edition Ivano Ciardelli Full Chapterrobert.johnson900100% (5)
- Grammar and SyntaxDocumento6 pagineGrammar and SyntaxBel SterlingNessuna valutazione finora
- Gender and Noun Classification Eric Mathieu Editor Full ChapterDocumento67 pagineGender and Noun Classification Eric Mathieu Editor Full Chapterbruce.glinski869100% (6)
- Dirk Geeraerts - Lexical SemanticsDocumento18 pagineDirk Geeraerts - Lexical SemanticsFederico RodriguezNessuna valutazione finora
- 026213375x.mit Press - Dynamic Asymmetry - Moro, Andrea - Sep.2000Documento154 pagine026213375x.mit Press - Dynamic Asymmetry - Moro, Andrea - Sep.2000Branko NikolicNessuna valutazione finora
- R. MereDocumento6 pagineR. MerebreeNessuna valutazione finora
- Object (Grammar) : Jump To Navigation Jump To SearchDocumento10 pagineObject (Grammar) : Jump To Navigation Jump To SearchbreeNessuna valutazione finora
- A Guide To Old Spanish First Edition Edition Steven N Dworkin Full ChapterDocumento67 pagineA Guide To Old Spanish First Edition Edition Steven N Dworkin Full Chapterthomas.frazier143100% (5)
- Abralin2021 1Documento91 pagineAbralin2021 1Sergio Sánchez PadillaNessuna valutazione finora
- Banaat Ain Shams November 2018 (Prof. Bahaa Mazeed)Documento99 pagineBanaat Ain Shams November 2018 (Prof. Bahaa Mazeed)Omnia N. Elkholy100% (4)
- 6-Lecture Six (Chapter Four-Semantic Analysis)Documento25 pagine6-Lecture Six (Chapter Four-Semantic Analysis)Getnete degemuNessuna valutazione finora
- Construction GrammarDocumento22 pagineConstruction GrammarIsraelEscMarNessuna valutazione finora
- (Claudia Claridge) Hyperbole in English A Corpus PDFDocumento316 pagine(Claudia Claridge) Hyperbole in English A Corpus PDFR FNessuna valutazione finora
- Teaching Discourse (II) : Unit 9Documento27 pagineTeaching Discourse (II) : Unit 9Inés Amaya DíazNessuna valutazione finora
- Archi Minimalist ChapterDocumento63 pagineArchi Minimalist ChapterAlNessuna valutazione finora
- The Notion of Syntactic Valence: "Lexical Information Is Syntactically Represented."Documento7 pagineThe Notion of Syntactic Valence: "Lexical Information Is Syntactically Represented."Alexandra FicicarNessuna valutazione finora
- Comprehending Conceptiual MetaphorsDocumento11 pagineComprehending Conceptiual Metaphorsscribd1comNessuna valutazione finora
- NLP Unit 5Documento15 pagineNLP Unit 5Tejaswini PotabattiNessuna valutazione finora
- Universal Dependencies: A Cross-Linguistic TypologyDocumento8 pagineUniversal Dependencies: A Cross-Linguistic TypologyprogisNessuna valutazione finora
- A Study of Structure of Korean Knowledge DatabaseDocumento6 pagineA Study of Structure of Korean Knowledge DatabaseNur HameliaNessuna valutazione finora
- Language Functions and FormsDocumento11 pagineLanguage Functions and FormsyatzirigNessuna valutazione finora
- Bromberger & Halle 89 - Why Phonology Is DifferentDocumento20 pagineBromberger & Halle 89 - Why Phonology Is DifferentamirzetzNessuna valutazione finora
- UntitledDocumento505 pagineUntitledavelino corral esteban100% (1)
- (Luigi Rizzi) The Structure of CP and IP The Cart (B-Ok - Xyz) PDFDocumento376 pagine(Luigi Rizzi) The Structure of CP and IP The Cart (B-Ok - Xyz) PDFMelisa BenelliNessuna valutazione finora
- Space in Language and Cognition PDFDocumento415 pagineSpace in Language and Cognition PDFGigi Costelus100% (1)
- Semantic Competence and TS LatestDocumento23 pagineSemantic Competence and TS LatestAisha FaridNessuna valutazione finora
- Cognition: A B A C A D A eDocumento24 pagineCognition: A B A C A D A eparisa baseremaneshNessuna valutazione finora
- A Survey of Applications of Finite Automata in Natural Language ProcessingDocumento3 pagineA Survey of Applications of Finite Automata in Natural Language ProcessingHarsh ChaudharyNessuna valutazione finora
- SEMINAR 1 New грамDocumento5 pagineSEMINAR 1 New грамValter ValterNessuna valutazione finora
- MorphologyDocumento19 pagineMorphologyNonameNessuna valutazione finora
- Palmer Propbank CLDocumento33 paginePalmer Propbank CLAtif RaufNessuna valutazione finora
- Weiland Et AlDocumento17 pagineWeiland Et AlLương Nguyễn Vân ThanhNessuna valutazione finora
- Logic, Representation and Inference: Introduction To SemanticsDocumento22 pagineLogic, Representation and Inference: Introduction To SemanticsnabeelajavedNessuna valutazione finora
- The Baboonandthe BeeDocumento27 pagineThe Baboonandthe Beeakindele02muchNessuna valutazione finora
- George Lakoff Instrumental Adverbs and The Concept of Deep StructureDocumento28 pagineGeorge Lakoff Instrumental Adverbs and The Concept of Deep Structureكمال گاطعNessuna valutazione finora
- Seminar 1 Fundamentals of Theoretical Grammar 1. Theoretical CommentDocumento7 pagineSeminar 1 Fundamentals of Theoretical Grammar 1. Theoretical CommentАлександра ГильдунинаNessuna valutazione finora
- Siddiqi-2010-Language and Linguistics CompassDocumento19 pagineSiddiqi-2010-Language and Linguistics CompassKrishna KumarNessuna valutazione finora
- Semantics-Some AreaDocumento9 pagineSemantics-Some AreaZainur RohmanNessuna valutazione finora
- Uts RLT BDocumento2 pagineUts RLT BHonesty TrilaNessuna valutazione finora
- Semantic Pragmatic-Speech ActDocumento26 pagineSemantic Pragmatic-Speech ActDiah Retno WidowatiNessuna valutazione finora
- Asher DISCOURSE ANALYSISDocumento103 pagineAsher DISCOURSE ANALYSISHarry TapiaNessuna valutazione finora
- Lec CS563 Lexical Knowledge NNDocumento208 pagineLec CS563 Lexical Knowledge NNVaraprasad DNessuna valutazione finora
- Ellipsis: Jason Merchant University of ChicagoDocumento43 pagineEllipsis: Jason Merchant University of ChicagoWin certoNessuna valutazione finora
- THE Relationship Between Case AND AgreementDocumento25 pagineTHE Relationship Between Case AND AgreementAnonymous R99uDjYNessuna valutazione finora
- Lecture 04Documento35 pagineLecture 04tafadadi272727_40816Nessuna valutazione finora
- Aj Semantics Simple Summary of First 7 Chapters of Saeed Simple With Examples ClearlyDocumento12 pagineAj Semantics Simple Summary of First 7 Chapters of Saeed Simple With Examples ClearlyJaya govinda raoNessuna valutazione finora
- Curriculum Content - August, 2017 Grade 5B: Subject Topic Sub-Topic Resources English Grammar: Connectives BooksDocumento20 pagineCurriculum Content - August, 2017 Grade 5B: Subject Topic Sub-Topic Resources English Grammar: Connectives BooksanduNessuna valutazione finora
- Greenough, J. B. Et Al. (1903), New Latin GrammarDocumento500 pagineGreenough, J. B. Et Al. (1903), New Latin GrammarBeavermaster35Nessuna valutazione finora
- Karlos Arregi, Andrew Nevins Auth. Morphotactics Basque Auxiliaries and The Structure of SpelloutDocumento433 pagineKarlos Arregi, Andrew Nevins Auth. Morphotactics Basque Auxiliaries and The Structure of SpelloutensinodelingamaternaNessuna valutazione finora
- Noam Chomsky - Failed StatesDocumento164 pagineNoam Chomsky - Failed Statesensinodelingamaterna100% (1)
- Manuel Garcia-Carpintero, Max Kolbel-The Bloomsbury Companion To The Philosophy of LanguageDocumento302 pagineManuel Garcia-Carpintero, Max Kolbel-The Bloomsbury Companion To The Philosophy of Languageensinodelingamaterna100% (10)
- A Tentative Syntax of Modern IndoeuropeanDocumento241 pagineA Tentative Syntax of Modern Indoeuropeanvj03eNessuna valutazione finora
- Q2WEEK OneDocumento11 pagineQ2WEEK Onevivien villaberNessuna valutazione finora
- 0510 s08 Ms 11+12Documento13 pagine0510 s08 Ms 11+12Hubbak Khan100% (1)
- ExamDocumento10 pagineExamhell oNessuna valutazione finora
- African Polyphony and Polyrhythm PDFDocumento709 pagineAfrican Polyphony and Polyrhythm PDFEstefanía Gutiérrez Galvis100% (5)
- Poetry PPT Sir EaDocumento52 paginePoetry PPT Sir EaCHRISTIAN EANessuna valutazione finora
- Test 3Documento3 pagineTest 3Zara EdwardsNessuna valutazione finora
- Latin Study Guide 1 Declension Singula R Feminine Nom - A Gen - Ae Dat - Ae Acc - Am Abl - Ā Plural Nom - Ae Gen - Arum Dat - Is Acc - As Abl - IsDocumento7 pagineLatin Study Guide 1 Declension Singula R Feminine Nom - A Gen - Ae Dat - Ae Acc - Am Abl - Ā Plural Nom - Ae Gen - Arum Dat - Is Acc - As Abl - IsChris JagoeNessuna valutazione finora
- 3shape Grammars1Documento94 pagine3shape Grammars1OSHIN BALASINGH100% (1)
- Lecture 5Documento22 pagineLecture 5Даша МусурівськаNessuna valutazione finora
- Nonverbal CommunicationDocumento16 pagineNonverbal CommunicationAra0% (1)
- Unit 2 VocabularyDocumento1 paginaUnit 2 Vocabularysoufian chabliNessuna valutazione finora
- Reading Diagnostic Test n2Documento5 pagineReading Diagnostic Test n2Cleo TokongNessuna valutazione finora
- Teaching SpeakingDocumento46 pagineTeaching SpeakingParadise PNessuna valutazione finora
- Raz Laa02 Farmanim PrevDocumento7 pagineRaz Laa02 Farmanim Prevmadalina.giorgescu3515Nessuna valutazione finora
- Chapter 2 - Writing Business MessagesDocumento60 pagineChapter 2 - Writing Business MessagesrenNessuna valutazione finora
- edited-Q1-Module 6-ADM-Engl7 PDFDocumento24 pagineedited-Q1-Module 6-ADM-Engl7 PDFRassel Jay GumbocNessuna valutazione finora
- Re-Visits The Grand Theory of Geoffrey Leech: Seven Types of MeaningDocumento7 pagineRe-Visits The Grand Theory of Geoffrey Leech: Seven Types of MeaningENG331C F21Nessuna valutazione finora
- 周次 / Minggu: 11 2021: Rancangan Pembelajaran Dan Pengajaran Harian 星期 / Hari: Khamis 日期 / Tarikh: 08-04Documento2 pagine周次 / Minggu: 11 2021: Rancangan Pembelajaran Dan Pengajaran Harian 星期 / Hari: Khamis 日期 / Tarikh: 08-04Hui YingNessuna valutazione finora
- 1A GRAMMAR Verb:, Subject PronounsDocumento8 pagine1A GRAMMAR Verb:, Subject PronounsThefi BlancoNessuna valutazione finora
- Relative Clauses of Time - EnglishDocumento17 pagineRelative Clauses of Time - EnglishGiordano ArgentaNessuna valutazione finora
- Elevator Pitch Rubrics (ICT For Social Chnage)Documento2 pagineElevator Pitch Rubrics (ICT For Social Chnage)Dran OteroNessuna valutazione finora
- (10) أحمد الكبيسي مكانة الإسناد وأثرهDocumento27 pagine(10) أحمد الكبيسي مكانة الإسناد وأثرهzulikhwan alhumawyNessuna valutazione finora
- Heading Forward Starter TBDocumento130 pagineHeading Forward Starter TBSamiCaroNessuna valutazione finora
- Burmese CompleteDocumento313 pagineBurmese CompleteIsaack Neme Huapaya100% (1)
- Test Initial Clasa A X-A ADocumento3 pagineTest Initial Clasa A X-A AJoanneSoutherlandNessuna valutazione finora
- Toastmaster FullDocumento39 pagineToastmaster Fullmachinel1983Nessuna valutazione finora
- Grammar Summary Unit 5Documento2 pagineGrammar Summary Unit 5memm519Nessuna valutazione finora
- Language Skills of Children With Early Cochlear ImplantationDocumento13 pagineLanguage Skills of Children With Early Cochlear Implantationismail39 orthoNessuna valutazione finora
- Intergration of Cultural and Multicultural Literacy in English Lesson PlanDocumento3 pagineIntergration of Cultural and Multicultural Literacy in English Lesson PlanLovely GracieNessuna valutazione finora
- Body Language: Decode Human Behaviour and How to Analyze People with Persuasion Skills, NLP, Active Listening, Manipulation, and Mind Control Techniques to Read People Like a Book.Da EverandBody Language: Decode Human Behaviour and How to Analyze People with Persuasion Skills, NLP, Active Listening, Manipulation, and Mind Control Techniques to Read People Like a Book.Valutazione: 5 su 5 stelle5/5 (81)
- Writing to Learn: How to Write - and Think - Clearly About Any Subject at AllDa EverandWriting to Learn: How to Write - and Think - Clearly About Any Subject at AllValutazione: 4 su 5 stelle4/5 (83)
- Surrounded by Idiots: The Four Types of Human Behavior and How to Effectively Communicate with Each in Business (and in Life) (The Surrounded by Idiots Series) by Thomas Erikson: Key Takeaways, Summary & AnalysisDa EverandSurrounded by Idiots: The Four Types of Human Behavior and How to Effectively Communicate with Each in Business (and in Life) (The Surrounded by Idiots Series) by Thomas Erikson: Key Takeaways, Summary & AnalysisValutazione: 4.5 su 5 stelle4.5/5 (2)
- Stonewalled: My Fight for Truth Against the Forces of Obstruction, Intimidation, and Harassment in Obama's WashingtonDa EverandStonewalled: My Fight for Truth Against the Forces of Obstruction, Intimidation, and Harassment in Obama's WashingtonValutazione: 4.5 su 5 stelle4.5/5 (21)
- The Language Instinct: How the Mind Creates LanguageDa EverandThe Language Instinct: How the Mind Creates LanguageValutazione: 4 su 5 stelle4/5 (916)
- Summary: How to Know a Person: The Art of Seeing Others Deeply and Being Deeply Seen By David Brooks: Key Takeaways, Summary and AnalysisDa EverandSummary: How to Know a Person: The Art of Seeing Others Deeply and Being Deeply Seen By David Brooks: Key Takeaways, Summary and AnalysisValutazione: 4 su 5 stelle4/5 (4)
- Learn French with Paul Noble for Beginners – Complete Course: French Made Easy with Your 1 million-best-selling Personal Language CoachDa EverandLearn French with Paul Noble for Beginners – Complete Course: French Made Easy with Your 1 million-best-selling Personal Language CoachValutazione: 5 su 5 stelle5/5 (81)
- How Not to Write a Novel: 200 Classic Mistakes and How to Avoid Them—A Misstep-by-Misstep GuideDa EverandHow Not to Write a Novel: 200 Classic Mistakes and How to Avoid Them—A Misstep-by-Misstep GuideNessuna valutazione finora
- The Art of Writing: Four Principles for Great Writing that Everyone Needs to KnowDa EverandThe Art of Writing: Four Principles for Great Writing that Everyone Needs to KnowValutazione: 4.5 su 5 stelle4.5/5 (21)
- Wordslut: A Feminist Guide to Taking Back the English LanguageDa EverandWordslut: A Feminist Guide to Taking Back the English LanguageValutazione: 4.5 su 5 stelle4.5/5 (429)
- Learn Spanish with Paul Noble for Beginners – Complete Course: Spanish Made Easy with Your 1 million-best-selling Personal Language CoachDa EverandLearn Spanish with Paul Noble for Beginners – Complete Course: Spanish Made Easy with Your 1 million-best-selling Personal Language CoachValutazione: 5 su 5 stelle5/5 (136)
- The Practice of Poetry: Writing Exercises From Poets Who TeachDa EverandThe Practice of Poetry: Writing Exercises From Poets Who TeachNessuna valutazione finora
- Idioms in the Bible Explained and a Key to the Original GospelsDa EverandIdioms in the Bible Explained and a Key to the Original GospelsValutazione: 4 su 5 stelle4/5 (7)
- Everything You'll Ever Need: You Can Find Within YourselfDa EverandEverything You'll Ever Need: You Can Find Within YourselfValutazione: 4.5 su 5 stelle4.5/5 (41)
- Win Every Argument: The Art of Debating, Persuading, and Public SpeakingDa EverandWin Every Argument: The Art of Debating, Persuading, and Public SpeakingValutazione: 4.5 su 5 stelle4.5/5 (78)
- Spanish Short Stories: Immerse Yourself in Language and Culture through Short and Easy-to-Understand TalesDa EverandSpanish Short Stories: Immerse Yourself in Language and Culture through Short and Easy-to-Understand TalesValutazione: 5 su 5 stelle5/5 (1)
- Messages: The Communication Skills BookDa EverandMessages: The Communication Skills BookValutazione: 4.5 su 5 stelle4.5/5 (22)
- Learn German with Paul Noble for Beginners – Complete Course: German Made Easy with Your 1 million-best-selling Personal Language CoachDa EverandLearn German with Paul Noble for Beginners – Complete Course: German Made Easy with Your 1 million-best-selling Personal Language CoachValutazione: 5 su 5 stelle5/5 (151)
- Learn Mandarin Chinese with Paul Noble for Beginners – Complete Course: Mandarin Chinese Made Easy with Your 1 million-best-selling Personal Language CoachDa EverandLearn Mandarin Chinese with Paul Noble for Beginners – Complete Course: Mandarin Chinese Made Easy with Your 1 million-best-selling Personal Language CoachValutazione: 5 su 5 stelle5/5 (15)
- Writing Screenplays That Sell: The Complete Guide to Turning Story Concepts into Movie and Television DealsDa EverandWriting Screenplays That Sell: The Complete Guide to Turning Story Concepts into Movie and Television DealsNessuna valutazione finora
- 1000 Words: A Guide to Staying Creative, Focused, and Productive All-Year RoundDa Everand1000 Words: A Guide to Staying Creative, Focused, and Productive All-Year RoundValutazione: 3.5 su 5 stelle3.5/5 (13)
- NLP The Ultimate Guide to Learn the Art of Persuasion, Emotional Influence, NLP Secrets, Hypnosis, Body Language, and Mind ControlDa EverandNLP The Ultimate Guide to Learn the Art of Persuasion, Emotional Influence, NLP Secrets, Hypnosis, Body Language, and Mind ControlValutazione: 5 su 5 stelle5/5 (21)
