Sei sulla pagina 1di 18

Chapter 2

Aspect: Problem of lexicon and morphology

In this chapter we outline several of the influential linguistic analyses of

aspect in the literature, and examine how these analyses treat grammatical
aspect, lexical aspect, and the interaction between the two. We will begin
by looking at Comrie's analysis of perfective and imperfective aspect,
and then proceed to Vendler's categorization of verbs and times, Smith's
two-component theory of aspect, and Klein's view on aspect in terms of
the relationship between topic time and situation time. We end with a
discussion of how these various linguistic analyses bear on children's
acquisition of aspect.
2.1. Grammatical aspect
2.1.1. Perfective and imperfective
The grammatical encoding of aspectual notions, which we call grammatical aspect, is realized in different languages in different ways, for example, through the use of inflectional morphology, derivational morphology, auxiliary, or periphrastic constructions. This variation does not
mean, however, that grammatical aspect is wholly idiosyncratic and
language-specific in the way it is encoded. Typological studies of how
languages of the world encode aspectual notions have uncovered recurring patterns of aspectual marking (Comrie 1976; Bybee 1985; Dahl
1985; Bybee and Dahl 1989; Bybee, Perkins, and Pagliuca 1994). In this
section, we characterize these grammatical aspectual patterns and show
how they are related to each other.
The most basic aspectual opposition that is often encoded grammatically is that of perfective and imperfective. As noted in Chapter 1,
perfective aspect presents a situation as an unanalyzed whole (external
view), whereas imperfective aspect presents a situation from within
(internal view). In the following example from English,

John built a house.

John was building a house.
Download Date | 4/2/16 6:26 PM

12 Aspect
sentence (1) is aspectually perfective, whereas sentence (2) is aspectually
imperfect!ve. In sentence (1), the perfective aspect of the verb reports the
situation in its entirety, with the speaker presenting the situation including its initial and terminal points. Thus, sentence (1) entails completion:
John finished building a house. In contrast, the imperfective aspect in
sentence (2) communicates the internal structure of the event, without
regard to its beginning and end points. Thus the meaning of the sentence
is non-committal as to whether or not John has finished building the
Comrie (1976: 25) proposed the following hierarchical classification
of aspectual categories.







Figure 2.1. Classification of aspectual oppositions (adapted from Comrie 1976: 25)

Comrie divided imperfective aspect into habitual and continuous, and

further subdivided continuous into progressive and non-progressive. Let
us characterize these categories from the bottom up, following Comrie
Progressive aspect is a category of imperfective that has the properties
of dynamicity and change as its defining features: typically it denotes a
dynamic, continuously changing action in progress, and is generally
incompatible with Stative predicates (*He is knowing the answer).
Habitual aspect denotes a situation that spans an extended period of
time, typically involving repetition of an action over multiple occasions.
English has a habitual aspect marker (used to) which is used in past-time
reference only.
Imperfective aspect, sometimes referred to as general imperfective,
denotes both habitual and continuous (i.e., progressive and nonprogressive) qualities. For example, Romance languages grammatically
encode the perfective-imperfective distinction in the past, and the imperfective past can describe a past action in progress (i.e., progressive), a
Download Date | 4/2/16 6:26 PM

Grammatical aspect 13
past state (i.e., non-progressive),6 and a past habit. For instance, the
Spanish imperfective past form can express the equivalents of he was
dancing, he loved Mary, and he used to dance?
It is interesting to note here that Comrie failed to explicitly characterize the categories of continuous and non-progressive. He defined continuousness in the negative, that is, as imperfectivity that is not habitual,
and he did not define "non-progressive" (Bybee, Perkins, and Pagliuca
1994: 138). This may be related to the fact that the aspectual markers
that specifically denote these categories are extremely rare.8 For this
reason, the two categories appear in parenthesis in Figure 2.1.

2.1.2. Grammaticization of aspect markers

As noted in Chapter 1, there is a tendency for the meanings and functions of tense-aspect markers to change over time. For example, it is
claimed that progressive aspect markers can develop into more general
imperfective aspect markers by generalizing their applicability to habitual and non-progressive situations. In this subsection, we discuss grammaticization of tense-aspect markers to explain how different tenseaspect markers are related to each other, both semantically and historically.
Building on research from historical linguistics and crosslinguistic
survey of typologically diverse languages, Bybee, Perkins, and Pagliuca
(1994) proposed hypotheses concerning the grammaticization of tenseaspect-modality markers. Particularly relevant to our research are two
hypotheses concerning the universal paths of the development of tenseaspect markers. According to these authors, first, there is a universal
tendency for resultative and completive markers to grammaticize into
perfect markers, which in turn become perfective or simple past markers.
Second, there is a tendency for a progressive marker to develop into a
general imperfective marker by expanding its reference to habitual and
stative situations; in terms of Comrie's hierarchy (see Figure 2.1), imperfective aspect markers develop from the bottom up.9
Here, we would like to discuss the perfective path in more detail,
primarily to explicate the nature of the category perfect. Bybee, Perkins,
and Pagliuca (1994) claimed that completive and resultative aspect
markers grammaticize into perfect markers. Completive markers denote
an action performed completely and thoroughly, and resultative denotes
a state that has been brought about by a prior action. These markers then
Download Date | 4/2/16 6:26 PM

14 Aspect
extend their meaning and become perfect markers, which denote "past
action with current relevance". English perfect (have + past participle) is
a typical example of this category (Dahl 1985), which has such uses as
"perfect of result", "perfect of experience", etc. Perfect further grammaticizes and loses its "current relevance" restriction, and becomes a
perfective aspect marker and/or a simple past tense marker.
It is often the case that a past tense marker also has an aspectual value.
For example, many European languages (e.g., French, Spanish, German,
Dutch) have the perfect form grammaticizing into a past tense form
which also has a perfective value, although the degree of grammaticization differs. Thus, in these languages, they have two past tense markers older past forms (often referred to as preterite), and more recent past
forms (auxiliary plus participle). As noted in Chapter 1, what we see here
is a continuous development of tense-aspect markers, which often makes
it very difficult to explicitly determine whether a grammatical form is a
tense marker or an aspect marker.
Interestingly, this historical development charted for tense-aspect
markers has been found to have a parallel in language acquisition. In
Chapters 3 through 6, we will discuss in more detail the claim that children use past tense marking first as a resultative aspect marker and then
as a pure past tense marker, and that progressive markers are initially
restricted to the action-in-progress meaning and only later develop more
varied meanings (e.g., Antinucci and Miller 1976; Bloom, Lifter, and
Hafitz 1980; Shirai and Andersen 1995). This claim about the acquisition of tense-aspect morphology has also been discussed within the
larger issue of how and why there are parallel processes in language
acquisition and historical change (Slobin 1977, 1997; Ziegeler 1997).

2.2. Lexical Aspect: Vendler, Smith, and Comrie

Grammatical aspect provides a certain amount of information for the
interpretation of the aspectual meaning of verbal predicates, but most
linguists accept that we need to consider another type of aspect - lexical
aspect. Lexical aspect (also known as inherent aspect, situation aspect, or
Aktionsart) refers to the semantic characteristics inherent in the lexical
content of words, usually verbs or verb phrases, that are defined in terms
of the temporal properties of given situations that the verbs describe.
Because lexical aspect deals with lexical semantics, it is generally re-

Download Date | 4/2/16 6:26 PM

Lexical aspect 15
garded as a lexical category as opposed to the grammatical category of
"grammatical aspect".
As a starting point for our illustration, Vendler's (1957) four-way
classification represents an early attempt to categorize lexical aspect.10 He
classified verbs or verb phrases into four categories with respect to the
temporal properties that they encode: activities, accomplishments,
achievements, and states. According to Vendler, activity verbs encode
situations as consisting of successive phases over time with no inherent
endpoint, for example, walk, run, and swim. Accomplishment verbs also
characterize situations as having successive phases, but they differ from
activity verbs in that they encode a natural endpoint and often a change
of state, for example, paint a picture, build a house, and run a mile.
Vendler illustrated the contrast between activities and accomplishments
by the difference between run and run a mile. The sentence John was
running entails that John ran, whereas the sentence John was running a
mile does not entail that John ran a mile, that is, John might have stopped
halfway. Like accomplishments, achievement verbs also encode a natural
endpoint, but they differ from accomplishments and activities in that
they encode events as punctual and instantaneous, that is, as having no
duration, such as in fall, win the race, and reach the summit. Strictly
speaking, every event occupies time. But speakers can construe given
verbs as denoting situations having no time duration. Finally, state verbs,
in contrast to the other three categories, encode situations as homogeneous, with no successive phases or endpoints, involving no dynamicity,
such as know and love. Thus, state verbs cannot usually be combined
with progressive aspect that marks change and development from one
phase to the next (e.g., *John is knowing the story is odd).11 These
categories can be schematically represented as follows (Andersen 1990).


love, contain, know

run, walk, swim
paint a picture, build a house
fall, drop, win the race

In this schematization, a solid line is used to represent states, because

states have no apparent beginning point or endpoint and endure indefinitely unless some external force changes them. The wavy lines for
activities and accomplishments indicate the dynamic duration of an
action, while for accomplishments and achievements represents a
Download Date | 4/2/16 6:26 PM

16 Aspect
punctual point of change of state, signaling telicity (i.e., natural endpoint).
Vendler's analysis, now probably the most widely accepted and the
best known, has become an important starting point for subsequent
research on lexical aspect. For example, Smith (1997) recently modified
this system and applied it to an analysis of English, French, Chinese,
Russian, and Navajo. Smith's modification involved mainly the addition
of "semelfactive" verbs, such as cough, tap, and knock in English and
their equivalents in other languages. She argued that semelfactives
resemble achievements in being punctual, but differ from achievements
in that semelfactives encode no endpoint. In particular, semelfactives and
achievements behave differently with progressive aspect marking: when
semelfactives are marked with progressive, they are interpreted as specifying a repeated event (e.g., coughing or knocking several times); when
achievements are marked with progressive, they are interpreted as indicating preliminary, detachable stages of the event rather than a repeated
event (e.g., John was reaching the summit means that John was at a stage
just prior to being at the summit, not that he arrived at the summit several
times). In addition to Smith, other researchers have extended Vendler's
analysis in other ways. For example, instead of using a categorical fourway classification, some linguists have classified verbs as pairs of contrasts, such as stative versus dynamic verbs, telic versus atelic verbs, and
punctual versus durative verbs (Comrie 1976). These pairs of contrast
have the advantage of making the semantic features of lexical aspect
explicit and transparent. These contrasts can be reanalyzed as features,
for example, [dynamic], [telic], and [punctual]. Table 2.1 presents
such an analysis, adapted from Smith (1991), to characterize Vendler's
four categories plus semelfactives.
Table 2.1. Semantic features for the five categories of lexical aspect (adapted from
Smith 1991: 30)

activities accomplishments semelfactives



We should point out a confusion that has been in the literature on

lexical aspect since Vendler (1957). Vendler did not use the terms lexical
aspect, Aktionsart, situation aspect, or inherent aspect. He was speaking
Download Date | 4/2/16 6:26 PM

Lexical aspect 17
of "time schemata". However, it is not clear whether time schemata are
part of the lexical semantic contents of verbs, or part of the temporal
properties of situations to which the verbs refer. Admittedly, it is not
always easy to separate lexical semantics from the objects and situations
to which the lexical items refer. But we would like to make this distinction here, for reasons that will become clear later on. Our intention in this
book is to treat lexical aspect as lexical categories according to the kinds
of temporal properties that the lexical items inherently encode, not as
situational categories according to the temporal properties that the
situations typically display. We contend that situational properties are not
necessarily reflected directly in lexical semantics, and that the confusion
between the two often leads to negative consequences in the study of
language acquisition (see further discussion in Chapter 3).
The abstract concepts of, for example, states and activities can be
viewed as universal semantic notions, independent of particular languages. In other words, every language may have lexical means to
encode states and activities. However, which verbs encode states and
which encode activities is language-specific, and can be determined only
within the context of a given language. Accordingly, children have to
learn to identify which verbs encode which temporal properties of a
situation in the language being acquired. An example from Comrie
(1976) clearly demonstrates the need to distinguish between inherent
verb meanings and temporal properties of situations. According to
Comrie, English and Portuguese treat perception verbs (e.g., see, hear}
differently. English treats them as stative, and these verbs consequently
do not accept progressive marking, while Portuguese treats them as nonstative, so they can naturally accept progressive marking. Take Japanese
for another example. Japanese does not have a stative verb corresponding to the English know. In order to express a notion like / know him,
Japanese speakers use the verb siru which means 'come to know' and
attach to it the durative aspect marker -teiru, to denote the resultative state
after coming to know him. Thus, in Japanese, siru is an achievement
verb. These examples show that verbs in different languages may differ
with respect to their lexical characteristics even though they refer to the
same situations. Such crosslinguistic differences show that there is no
one-to-one mapping between given types of situations and given types of
lexical items.
Having emphasized the importance of this distinction, we must acknowledge that correspondences between lexical contents and situational
Download Date | 4/2/16 6:26 PM

18 Aspect
properties are not totally arbitrary. There are clear cases where languages
do not differ. For example, there is probably no language that treats the
verb referring to killing someone as stative, and the verb "jump" in any
language has to somehow encode the punctuality of the action as part of
its meaning. Thus, while it is important for theoretical and empirical
reasons to distinguish classifications of situations from classifications of
lexical aspect categories, in practice, they often coincide.

2.3. Lexical aspect and grammatical aspect

As discussed above, it is clear that interpretation of the aspectual properties of a sentence cannot be determined without considering the contribution of the lexical aspect of the verb. It is also clear that the aspectual
meaning of a sentence can, by and large, be predicted on the basis of
which grammatical aspect is combined with which lexical aspect category. One explicit proposal in this regard is that of Smith (1991, 1997).
In this section, we present our view of how aspectual meaning is determined as a function of the combinatorial properties of lexical aspect and
grammatical aspect on the basis of Smith's theory of aspect and the
grammaticization theory of Bybee, Perkins, and Pagliuca (1994).
Smith (1997) proposed a comprehensive theory of aspect, which she
labeled as a "two-component theory". The two components are what
she calls "situation type" and "viewpoint aspect", which roughly
correspond to what we have described so far as "lexical aspect" and
"grammatical aspect", respectively. Smith's theory attempts to account
for diverse aspectual phenomena related to the interaction between
situation type (achievement, accomplishment, activity, state, and semelfactive) and viewpoint aspect (perfective, imperfective, and neutral)12.
We have been using the term "lexical aspect of verbs", but strictly
speaking, the lexical aspect value is determined by both the verb and its
arguments, which Smith (1997) calls the verb constellation. Examples
include [John love Mary] (state), [John run] (activity), [John run a mile]
(accomplishment), [John reach the summit] (achievement), and [John
jump] (semelfactive). Note that what is inside [ ] is not a linguistic form
but the proposition underlying it, without any verb morphology to signal
the viewpoint aspect. In such instances, we are talking about the semantic
structure of verb-plus-arguments without any value imposed by grammatical aspect.

Download Date | 4/2/16 6:26 PM

Lexical and grammatical aspect 19

Another component of the two-component theory is viewpoint aspect.
Smith used the term "viewpoint" since it is essentially the speaker's
choice as to which aspectual perspective should be used. To describe the
fact that John ran a mile yesterday, the speaker can choose to present it
from either a perfective viewpoint (John ran a mile) or an imperfective
viewpoint (John was running a mile). If the story is about what happened
while John was jogging yesterday, imperfective viewpoint would be used.
If the story is about what John did after he ran a mile, then perfective
viewpoint would be used.13 The speaker chooses to use a certain combination of lexical items (with lexical aspect value) and morphology (with
grammatical aspect value), which results in a particular intended meaning, such as action-in-progress, completion, stative, etc.
2.3.1. Interaction between lexical aspect and grammatical aspect
How do particular combinations of lexical aspect and grammatical aspect
result in particular interpretations of a sentence? What are the principles
behind it? In this subsection, we present an account of how the interaction between the two levels of aspect works. Our account differs from
Smith's in minor details, but the principles are essentially the same.
Imperfective and progressive aspect

As discussed in 2.1, both general imperfective and progressive markers

have the same viewpoint aspect value, except that the progressive viewpoint is [+dynamic]. Therefore, they show very similar distributions in
their interaction with lexical aspect. Since imperfective viewpoint presents
a situation from within, disregarding its beginning or endpoint, it requires duration upon which the internal view can be imposed. If there is
no duration, it is impossible to treat the beginning point or endpoint
separately. This requirement predicts that some achievements are incompatible with imperfective aspect, as seen in the ungrammaticality of *He
is noticing a friend.
To see the properties of lexical aspect clearly, we repeat the following
diagram from section 2.2 here, with the addition of semelfactives. The
punctual point of action for semelfactives is represented by , since it is
not telic, but atelic.

Download Date | 4/2/16 6:26 PM

20 Aspect



love, contain, know

run, walk, swim
paint a picture, build a house
fall, drop, win the race
jump, knock, cough

Although achievements do not normally allow for an internal view of

imperfective aspect, it is sometimes possible to impose an internal view
on an achievement verb by focusing on the preliminary stages of an
event (i.e., the process leading up to the endpoint). To elaborate, John
was reaching the summit means that John was at a stage just prior to
being at the summit, not that he was in the middle of reaching the summit. As another example, John is winning the race is appropriate when
John's horse is ahead of all other horses in the race, even though he has
not actually reached the finishing line.
Semelfactive is another punctual, instantaneous category that implies
no duration. In this case, the speaker can only impose an internal view on
the situation when the duration is created by repetition. Thus, John is
jumping normally refers to John's repetitive acts of jumping, that is,
multiple jumps in a single activity, which Smith (1997) called Multipleevent Activity.14
In contrast to achievements and semelfactives, categories that have
duration, i.e., activities, accomplishments, and states, can readily accept an
internal view. Both general imperfective and progressive markers yield
the action-in-progress meaning when attached to activity and accomplishment verbs, focusing on the dynamic duration of the situation
denoted by the verbs.
With respect to stative verbs, the picture is a little different. Stative
verbs do not usually combine with progressive aspect: a large set of
English stative verbs and their translation equivalents in other languages
do not normally accept progressive marking, including psychological
and cognitive verbs such as want, need, like, love, believe, and know,
perception verbs such as see, hear, and feel, and relational verbs or verbs
of existence such as resemble, possess, have, and be. The incompatibility
seems to be that since progressive aspect presents a situation as ongoing,
it requires that the situation have successive phases, i.e., that it be dynamic, whereas the cognitive, perception, and relation or existence verbs
indicate only undifferentiated and homogeneous situations. "Progress"
presupposes the dynamic development of a situation. Thus, progressive
Download Date | 4/2/16 6:26 PM

Lexical and grammatical aspect 21

aspect combines naturally with activity and accomplishment verbs, but

not with stative verbs. However, under certain circumstances some Stative
verbs do appear in the progressive, which we will call "stative progressive". For example, Smith (1983) showed that progressive aspect can be
used with stative verbs to present a state as an event, such as in John is
believing in ghosts these days. Another example is John is being stupid,
in which the stative progressive refers to a particular behavior of John' s
at the time of speech, in contrast to John is stupid which refers to a
general characteristic attributed to John. John is being stupid does not
make any claim about John's intelligence before or after the moment of
speech, but rather that John is behaving stupidly.
General imperfective aspect, in contrast to progressive, is naturally
compatible with stative verbs and does not yield special meanings.15
Note that general imperfective has the features of [+intemal view] and
[-dynamic], and therefore, it can easily find duration in a state, and it
does not trigger the action-in-progress meaning because stative verbs
also have the feature [-dynamic]. The obtained meaning of general
imperfective with stative verbs is thus 'state continuing', as in La mer
etait calme 'the sea was calm' in French (Smith 1997: 197) or Wuli-de
deng kai-zhe 'the light in the room is on' in Chinese.
Perfective aspect
Perfective aspect combines naturally with achievement verbs, because by
definition, perfective aspect presents a situation as a single whole, and
achievement verbs provide ideal instantiation of such a viewpoint in that
they depict punctual situations as single points without internal structure.
Because achievements involve an endpoint, their combination with
perfective aspect denotes the completion of a situation, although in this
case the beginning point of the situation coincides with its point of
Perfective aspect also naturally combines with accomplishments and is
normally interpreted as indicating the completion of a situation, since by
definition accomplishments incorporate an endpoint, and perfective
aspect views the situation as a single whole externally, with both initial
and terminal points. For example, John built a house not only presents
the building process as a single whole, but also indicates that the house
was indeed finished.
The notion of completion is not applicable to uses of the perfective
with activity verbs, since activities encode situations with no inherent
Download Date | 4/2/16 6:26 PM

22 Aspect
endpoints (Comrie 1976; Lyons 1977). For example, John ran cannot
indicate that John "completed" his running. To ask Has John finished
running? is not appropriate unless the speaker has in mind in advance
some delimiter to the distance or time John ran. John ran, then, indicates
simply that John engaged in the activity of running for a while and that
at some arbitrary (unspecified) point this activity terminated. In the same
vein, perfective aspect with semelfactive verbs conveys a termination and
not a completion of the situation (e.g., John coughed).
Earlier we pointed out that the progressive aspect is not compatible
with stative verbs because of the verbs' lack of dynamic meaning.
Perfective aspect is not compatible with stative verbs, either. This is
because stative verbs do not include either a beginning point or an
endpoint in their temporal structure. Therefore, perfective aspect, which
includes both the beginning and end points of a situation in its focus,
cannot normally combine with stative verbs. In rare cases when the
combination is possible, it indicates an entry into a state as in Then
suddenly I knew it! by focusing on the punctual point of entry into that
state. This in effect translates a stative verb into an achievement verb.
The interaction discussed in this subsection is summarized in Table
2.2, which covers the most common grammatical aspect types, although
it is not meant to be exhaustive or without exception. Excluded from our
analyses here are the more marked cases of habitual reference (e.g., He is
walking to school these days) and futurate (e.g., We are eating out
tonight), which can be obtained with any of the dynamic verb classes.
Table 2.2. Interaction of lexical aspect with grammatical aspect

Activity Accomplishment Achievement

prelim, stage
prelim, stage
stative prog
termination completion

iterative prog
iterative prog

prog = progressive
prelim, stage = preliminary stages of event
?? = questionable combination

Our discussion in this subsection, as summarized in Table 2.2, shows

clearly that there are combinatorial constraints or compatibility between
certain grammatical aspect and certain lexical aspect. Comrie (1976)
Download Date | 4/2/16 6:26 PM

Lexical and grammatical aspect 23

discussed this kind of lexical-grammatical relationships with the "naturalness of combination" principle. According to Comrie, perfective
aspect combines naturally with punctual verbs, for example, because
perfective aspect presents a situation as a whole without regard to its
internal structure and punctual verbs encode a situation as a single point
lacking internal structure. Conversely, imperfective aspect is not compatible with punctual verbs (except semelfactives, which can be construed
iteratively), because imperfective aspect presents a situation as having an
internal structure while a punctual verb encodes it as a point lacking
internal structure. Activity verbs lend themselves naturally to the internal
perspective of imperfective aspect because they encode the successive
phases of an event over time. Such interactions between lexical and
grammatical aspect presumably originate in certain natural relationships
between events in reality. For example, when we know that a situation
comes to its end with a clear result, this situation has probably already
become a past event, and we are therefore likely to comment on its
completion by means of perfective aspect (Brown 1973). Many situations with an end result last for such a brief time that they are almost
certain to have ended before one can comment on them, such as situations denoted by verbs like drop, fall, and crash. These situations almost
exclude the internal perspective of imperfective aspect. In sum, there are
likely to be strong associations in the real world between resultativity or
telicity and the use of past and perfective verb forms, and between atelic
activities and the use of present and progressive verb forms.
2.3.2. Variation across languages in grammaticization of aspect
As noted earlier, there is variation across languages in how aspectual
notions are grammaticized. There are two types of variation to consider:
in the pattern of grammaticization and in the degree of grammaticization.
First, different languages develop different types of grammatical aspect, and some languages do not have grammatical aspect (e.g., Hebrew,
Finnish). In terms of combinatorial properties of aspect and tense,
Romance languages have a perfective-imperfective distinction only in the
past tense. Many Slavic languages, although grammatically marking both
tense and aspect, do not have perfective aspect in the present tense (e.g.,
Polish, Russian). In terms of combinatorial properties of grammatical
aspect and lexical aspect, the Japanese imperfective marker -te i- cannot
Download Date | 4/2/16 6:26 PM

24 Aspect
focus on the preliminary stage of an event as in English; instead it
focuses on the duration after the terminal point of a situation, and has the
perfect and resultative meaning. The Chinese perfective -le does not
necessarily yield the meaning of completion with telic verbs (see Chapter
5 for details). These variations are not random. Typological research
(Dahl 1985; Bybee, Perkins, and Pagliuca 1994) has shown that some of
these tendencies reflect common patterns of grammaticization. For
example, it has been suggested that the reason many languages do not
have perfective aspect for present is that the most natural way of talking
about situations existing at speech time is imperfective (state, progressive,
or habitual) and not perfective, because perfective indicates something
has been completed or terminated. This pattern indicates the importance
of functional motivation in shaping systems of grammatical aspect in
natural language.
Second, grammatical aspect markers in different languages also vary
in their degree of grammaticization, which is in turn reflected in their
different combinatorial patterns with lexical aspect. As noted earlier, it
has been suggested that progressive markers grammaticize into general
imperfective markers (Bybee, Perkins, and Pagliuca 1994: 140-149). In
English, the progressive aspect is highly grammaticized. In addition to its
unmarked progressive meaning (i.e., action in progress), it can also be
used for stative progressive, preliminary stages, habitual, and futurate.
Progressive markers in other languages may be less grammaticized, and
practically limited to the typical action-in-progress meaning, for example, in Chinese, Malay, and Thai. Thus, in these languages, progressive
markers cannot be used for other occasions (e.g., with states or achievements). Similarly, perfective aspect is also sensitive to the degree of
grammaticization in its interaction with lexical aspect. As discussed
above, perfective aspect is either incompatible with stative verbs, or has an
inchoative meaning with some stative verbs. As perfective aspect grammaticizes, however, it can be used as past tense marker, that is, to describe
a state that existed in the past (e.g., The book was there yesterday; see
Bybee, Perkins, and Pagliuca 1994: 92). For example, the English past
tense is aspectually perfective (Smith 1997), signaling completion or
termination with dynamic verbs or inception with some stative verbs (e.g.,
Then I knew //!). But as a highly grammaticized form in the function of
simple past tense, it can also be freely applied to stative verbs to locate
states in the past time. The Japanese past tense marker -ta has a similar

Download Date | 4/2/16 6:26 PM

Lexical and grammatical aspect 25

property, and is considered to be in a grammaticization transition from
perfect to perfective and further to simple past (Horie 1997).
Finally, a theoretically important variation concerns that of lexical
aspect. As we have discussed, some languages (e.g., English) accept
imperfective/progressive with achievement verbs to focus on preliminary
stages of an event (e.g., He is reaching the summit) whereas others (e.g.,
Chinese and Japanese) do not. On the basis of this observation, Smith
(1997) suggested that achievements in Chinese do not include preliminary stages of an event, whereas those in English do. An alternative
interpretation of this phenomenon is that the Chinese progressive markers zai or the Japanese marker -te i- simply cannot focus on preliminary
stages of an event. In other words, it is not the lexical aspect of the verb
that does not include preliminary stages, but the grammatical aspect
marker that cannot denote such stages (Shirai 1998a). This interpretation
attributes the difference between English and Chinese to variation in
grammatical aspect, not in lexical aspect. Of course, this issue - whether
the differential behavior of progressives in different languages is due to
grammatical aspect or lexical aspect - needs further empirical investigation.
2.4. A time-relational analysis of aspect
Most linguistic analyses, including ours, have adopted the definition of
grammatical aspect given in Comrie (1976). That is, grammatical aspect
involves different ways of "viewing" the temporal contour of a situation: perfective aspect presents an external view of the situation as a
single whole in its entirety without reference to its internal structure, and
imperfective aspect an internal view of the inner constituency of the
situation without regard for the situation's initial or final boundaries.
Smith (1983) treats this definition of aspect as "viewpoint aspect", in
contrast to "situation aspect" (see 2.3). Much the same definition of
aspect has been used in research by other linguists (e.g., Dahl 1985;
Bybee 1985; Bybee, Perkins, and Pagliuca 1994).
This approach to aspect has been challenged recently by Klein (1994,
1995, in press). According to Klein, the definition of aspect as different
ways of viewing a situation is entirely metaphorical, and thus vague: what
does it mean exactly to "view a situation" in its entirety, or as a single
whole, or with or without reference to its internal constituency? For
example, the difference between John stood on his toes and John was
Download Date | 4/2/16 6:26 PM

26 Aspect
standing on his toes is characterized as that between the perfective and
the imperfective, yet it is difficult to grasp the metaphorical differences
of entirety, wholeness, or reference to internal structure in these instances. In view of this problem, Klein proposed that aspect should be
examined on a par with tense, strictly in terms of temporal relations, such
as "prior to" (>), "contained in"(c), or "posterior to" (<), between the
time intervals that are described by aspectual markers. In particular,
Klein argued that three types of time intervals are involved in this timerelational theory of analysis and aspect:
time of utterance, i.e., the time at which an utterance is produced;
T-SIT: time of situation, i.e., the time at which the situation described by
the utterance obtains;
topic time, i.e., the time span about which something is said or for
which an assertion is made.
For example, for the sentence John was writing a letter, TU is the time
at which the speaker produces the sentence, T-SIT is the time at which
the situation of John's writing a letter obtains, and TT is the time span
about which the speaker makes an assertion. In this sentence, all that is
claimed is that there was some time span within which John was performing the action of writing a letter; whether John completed the letterwriting action is left open. Thus, the tense function of the sentence is
represented by TT < TU (topic time preceding time of utterance), while
the aspect function is represented by TT c T-SIT (topic time contained
in time of situation).16 In other words, tense is concerned with the temporal relations between TT and TU, while aspect is concerned with the
temporal relations between TT and T-SIT. In this analysis, the same
kinds of temporal relations (<, >, c) operate on both tense and aspect.
This analysis thus eliminates the vagueness associated with traditional
metaphorical interpretations of aspect in terms of ways of viewing a
situation. The following diagram shows how the major tense and aspect
categories are represented in this framework (see Klein 1994, 1995 for a
detailed analysis):



Download Date | 4/2/16 6:26 PM

Time-relational analysis 21



What does this time-relational analysis have to say about the acquisition of aspect? Although currently no studies have used this framework
to directly examine the acquisition of tense and aspect, one can imagine
that a whole new set of interesting empirical questions would arise from
the standpoint of this perspective. For example, which temporal relations
are entertained by children at which time? When and how do children
separate TU from T-SIT, TU from TT, and T-SIT from TT? What
underlies children's development in their understanding of the various
relations between the three temporal intervals, for tense and aspect,
respectively? What language-universal and language-specific patterns do
children display in the acquisition of these temporal relations? 17
While there is a great deal of theoretical motivation and empirical
evidence for the time-relational analysis of tense and aspect, and answers
to the above questions are important in addressing developmental and
crosslinguistic issues, our analysis of aspect in this book remains traditional, for several reasons. First, our original language acquisition studies
were conducted within the framework of a "viewpoint aspect" type of
linguistic analysis. Since the kinds of research questions in the timerelational framework are very different, it would be difficult to test these
questions with our existing data. Second, important to our study is the
relationship between grammatical aspect and lexical aspect in adult as
well as child language, as discussed throughout this chapter. It is not yet
entirely clear how the time-relational analysis of aspect captures the
intricate relationships between lexical and grammatical aspect (see Klein,
Li, and Hendriks, in press, for an attempt to study this issue in Chinese).
Finally, since almost all previous studies have examined the acquisition
of aspect on the basis of the traditional definition of aspect, it would be
important as well as convenient for us to evaluate and compare our
studies against previous findings on the same theoretical and empirical
bases, using the same terminologies. Nevertheless, researchers in language acquisition will find the time-relational analysis of aspect relevant
and useful, as such new linguistic analyses could have considerable
significance for reconceptualizing issues of acquisition.

Download Date | 4/2/16 6:26 PM

28 Aspect

2.5. Summary
In this chapter, we have provided an introduction to theories of aspect by
discussing two kinds of aspect, grammatical and lexical, with respect
particularly to linguistic and typological considerations. We examined
Comrie's classic definition of perfective and imperfective aspect,
Vendler's four categories of lexical aspect (activities, states, accomplishments, and achievements), and Smith's two-component theory of aspect
(the situation type and the viewpoint aspect). We also presented Klein's
approach to aspect in terms of temporal relations, which challenges
classic definitions of aspect.
In this chapter, we have placed a strong emphasis on the interaction
between grammatical aspect and lexical aspect in the interpretation of the
sentence's aspectual meaning. We looked at the combinatorial properties
of given aspectual markers with given types of verbs, and the different
implications that a given grammatical aspect conveys when it combines
with verbs of different lexical aspect. As is the case with most studies in
language acquisition, a clear linguistic analysis of a particular domain
serves as a good starting point for understanding the acquisition of that
domain. Thus, we have laid the basis here for the study of the acquisition
of aspect.
Grammatical aspect, lexical aspect, and the interaction between the two
have important theoretical implications for language acquisition. As we
will see in later chapters, in both child language and adult second language learning, there is a clear interaction between grammatical aspect
and lexical aspect in the acquisition of tense-aspect morphology, and this
interaction occurs across diverse languages. Moreover, as we have
stressed in this chapter, aspect stands at the interface between lexicon and
morphology, and thus aspect not only emerges as a grammatical task for
learners, but also serves as a catalyst for their acquisition of the lexicon.
For example, the morphological marking of aspect is a distinct feature of
verbs and not nouns or other grammatical categories of the sentence.
Children may use aspect markers as a means to help them distinguish
verbs from nouns, in which case aspect serves as a kind of syntactic
bootstrapping device.18 To fully understand the relationships between
grammatical aspect and lexical aspect, not only descriptively for the
language, but also developmentally for the language learner, we now turn
to the acquisition of aspect.

Download Date | 4/2/16 6:26 PM