Taalportaal | Semantic classification of main verbs

Dutch
Frisian
Afrikaans

Show full table of contents

1.2.3.Semantic classification of main verbs

quickinfo

This section discusses some of the semantic classifications of main verbs proposed over the last fifty years. The discussion starts with Vendler's (1957) distinction between states, activities, achievements and accomplishments, which has been the starting point for most semantic classifications proposed later. A problem with Vendler's classifications is that it became clear very quickly that it is not a classification of main verbs but of events expressed by larger structures headed by these main verbs. For example, one of the features that Vendler uses in his classification (and which is taken over in one form or another in most classifications of later date) is whether the event denoted by the verb has some logically implied endpoint, and the examples in (52) show that this need not be an inherent property of the verb itself but may be (partly) determined by, e.g., the internal argument of the verb: a singular indefinite object headed by a count noun introduces an inherent endpoint of the event denoted by the verb eten'to eat' (the event ends when the roll in question has been fully consumed), whereas a plural indefinite object does not (the endpoint depends on the number of rolls that Jan will consume).

Example 52

a.	Jan	eet	een broodje	met kaas.
	Jan	eats	a roll	with cheese
	'Jan is eating a role with cheese.'

b.	Jan eet	broodjes	met kaas.
	Jan eats	rolls	with cheese
	'Jan is eating rolls with cheese.'

Another problem with discussing the semantic classifications proposed since Vendler (1957) is that they often involve different dividing lines between the categories so that certain verbs may be categorized differently within the different proposals. Nevertheless, it is useful to discuss some specific proposals, given that the tradition that started with Vendler (1957) is still very much alive and continues to play an important role in present-day linguistics. Furthermore, we will see that a number of more recent proposals are formulated in such terms that make it possible to relate the semantic classification to the syntactic classification proposed in Section 1.2.2.

I. Aktionsart: Vendlerʼs aspectual event classification
II. Participant roles in events
III. Extensions of Vendlerʼs four-way distinction
- A. ±time extension
- B. ±control
IV. Other semantic classifications

[+] I. Aktionsart: Vendlerʼs aspectual event classification

Verbs are often classified according to the Aktionsart (which is sometimes also called inner aspect) they express. The term Aktionsart refers to the internal temporal organization of the event denoted by the verb and thus involves questions like (i) whether the event is construed as occurring at a single point in time (momentaneous aspect) or as evolving over time (durative aspect); (ii) whether the event is inherently bounded in time, and, if so, whether the event is bounded at the beginning (ingressive/inchoative aspect), at the end (terminative aspect) or both; (iii) whether the verb expresses a single event or a series of iterated events, etc; see Lehmann (1999) for further distinctions and more detailed discussion.

Example 53

Momentaneous aspect: exploderen'to explode', botsen'to collide'

Durative aspect: lachen'to laugh', wandelen'to walk/hike', zitten'to sit'

Inchoative aspect: ontbranden'to ignite', ontkiemen'to germinate'

Terminative aspect: doven'to extinguish', smelten'to melt', vullen'to fill'

Iterative aspect: bibberen'to shiver', stuiteren'to bounce repeatedly'

The Aktionsarts in (53) do not, however, necessarily define mutually exclusive verb classes. Bounded events expressed by the inchoative and terminative verbs in (53c&d), for example, also evolve over time and are therefore durative as well. It therefore does not come as a surprise that there have been attempts to develop a more sophisticated semantic classification based on the aspectual properties of verbs.

[+] A. Vendlerʼs Classification

Probably the best-known and most influential classification of main verbs is the one developed by Vendler (1957), who distinguishes the four aspectual classes in (54).

Example 54

Activities: bibberen'to shiver', denken (over)'to think (about)', dragen'to carry', duwen'to push', hopen'to hope', eten (intr.) 'to eat', lachen'to laugh', lezen (intr.) 'to read', luisteren'to listen', praten'to talk', rennen'to run', schrijven (intr.) 'to write', sterven'to die', wachten (op)'to wait (for)', wandelen'to walk', zitten'to zit'

Accomplishments: bouwen'to build', eten (tr.) 'to eat', koken (tr.) 'to cook', lezen (tr.) 'to read', opeten'to eat up', schrijven (tr.) 'to write', oversteken'to cross', verbergen'to hide', verorberen'to consume', zingen (tr.) 'to sing'

States: begrijpen'to understand', bezitten'to own', haten'to hate', hebben'to have', horen'to hear', geloven'to believe', houden van'to love', kennen'to know', leven'to live', verlangen'to desire', weten'to know'

Achievements: aankomen'to arrive', beginnen'to start', bereiken'to reach', botsen'to collide', herkennen'to recognize', ontploffen'to explode', ontvangen'to receive', overlijden'to die', zich realiseren'to realize', stoppen'to stop', opgroeien'to grow up', vinden'to find', winnen'to win', zeggen'to say'

Vendler argues that activities and accomplishments can be grouped together as processes and that states and achievements can be grouped together as non-processes, as depicted in Figure 3.

Figure 3: Vendlerʼs classification

The distinctions shown in Figure 3 are based on a number of semantic properties, which will be discussed in the following subsections.

[+] 1. Processes versus non-processes ±continuous tense

Vendler claims that verbs fall into two supercategories, which he calls processes and non-processes. Process verbs denote events which involve a specific internal dynamism over time and are characterized by the fact that they can be used to provide an answer to interrogative, progressive aan het + infinitive constructions like Wat is Marie aan het doen?'What is Marie doing?'; see constructions like Wat is Marie aan het doen?'What is Mary doing?'; see also Booij (2010:ch.6).

Example 55

a.	Marie	is naar Peter	aan	het	luisteren.	activity
	Marie	is to Peter	aan	het	listen
	'Marie is listening to Peter.'

b.	Marie	is haar boterham	aan	het	opeten.	accomplishment
	Marie	is her sandwich	aan	het	prt.-eat
	'Marie is eating her sandwich.'

c.	*	Marie	is van spinazie	aan	het	houden.	state
		Marie	is of spinach	aan	het	like
		Compare: '*Marie is liking spinach.'

d.	*	Marie	is aan	het	aankomen.	achievement
		Marie	is aan	het	prt.-arrive
		'Marie is arriving.'

[+] 2. Activities versus accomplishments ±bounded

Vendler divides the processes in activities and accomplishments on the basis of whether or not the event has a logically implied endpoint. Activities like luisteren'to listen' are open-ended; the event referred to in (55a) has no natural termination point and can, at least in principle, last for an infinitely long period of time. Accomplishments like opeten'to eat up', on the other hand, involve some inherent endpoint; the event referred to in (55b) is completed when the sandwich referred to by the object has been fully consumed.
This difference can be made more conspicuous by means of considering the validity of the entailments in (56). When we observe at a specific point in time that (56a) is true, we may conclude that (56a') is also true, but the same thing does not hold for the (b)-examples. This shows that in the case of an accomplishment like opeten'to eat up' it is not sufficient for the subject of the clause to be involved in a specific activity, but that reaching the logically implied endpoint is a crucial aspect of the meaning.

Example 56

a.	Marie	is naar Peter	aan	het	luisteren. ⇒	activity
	Marie	is to Peter	aan	het	listen
	'Marie is listening to Peter.'

a'.	Marie heeft	naar Peter	geluisterd.
	Marie has	to Peter	listened
	'Marie has listened to Peter.'

b.	Marie	is haar boterham	aan	het	opeten. ⇏	accomplishment
	Marie	is her sandwich	aan	het	prt.-eat
	'Marie is finishing her sandwich.'

b'.	Marie heeft	haar boterham	opgegeten.
	Marie has	her sandwich	prt.-eaten
	'Marie has finished her sandwich.'

The same point can be illustrated by question-answer pairs like those in (57), which show that accomplishments can be used in interrogatives of the form Hoe lang kostte het ...te Vinfinitive?'How long did it take to V ...?', which question the span of time that was needed to reach the logically implied endpoint, whereas activities cannot. The primed examples provide the corresponding answers to the questions.

Example 57

a.	*	Hoe lang	kostte	het	naar je leraar	te luisteren?	activity
		how long	took	it	to your teacher	to listen
		Compare: '*How long did it take to listen to your teacher?'

a'.	*	Het	kostte	een uur	naar mijn leraar	te luisteren.
		it	cost	an hour	to my teacher	to listen
		Compare: '*It took an hour to listen to my teacher.'

b.	Hoe lang	kostte	het	je maaltijd	op	te eten?	accomplishment
	how long	took	it	your meal	prt.	to eat
	'How long did it take to finish your meal?'

b'.	Het	kostte	10 minuten	mijn maaltijd	op	te eten.
	it	cost	10 minutes	my meal	prt.	to eat
	'It took 10 minutes to finish my meal.'

The question-answer pairs in (58) show that the opposite holds for interrogatives of the type Hoe lang auxfinite ...V?'For how long did ... V ...?', which simply question the span of time during which the activity took place; such pairs can be used with verbs denoting activities but not with verbs denoting accomplishments.

Example 58

a.	Hoe lang	heb	je	naar je leraar	geluisterd?	activity
	how long	have	you	to your teacher	listened
	'For how long did you listen to your teacher?'

a'.	Ik	heb	een uur	(lang)	naar mijn leraar	geluisterd.
	I	have	an hour	long	to my teacher	listened
	'I have listened to my teacher for an hour.'

b.	*	Hoe lang	heb	je	je maaltijd	opgegeten?	accomplishment
		how long	have	you	your meal	prt.-eaten

b'.	*	Ik	heb	een uur	(lang)	mijn maaltijd	opgegeten.
		I	have	an hour	long	my meal	prt.-eaten

Another, but essentially identical, test that is often used to distinguish activities and accomplishments is the addition of specific types of temporal adverbial phrases: adverbial phrases like gedurende een uur'during an hour' or een uur lang'for an hour', which refer to the span of time during which the event denoted by the verb takes place, are typically used with activities; adverbial phrases like binnen een uur'within an hour', which measure the span of time that is needed to reach a logically implied endpoint, are typically used with accomplishments.

Example 59

a.	Jan luisterde	gedurende/*binnen een uur	naar zijn leraar.	activity
	Jan listened	during/within an hour	to his teacher
	'Jan listened to his teacher for an hour.'

b.	Jan at	zijn maaltijd	binnen/*gedurende	vijf minuten	op.	accomplishment
	Jan ate	his meal	within/during	five minutes	prt.
	'Jan finished his meal in an hour.'

The (in)validity of the inferences in (56) and the selection restrictions on adverbial phrases in (59) are related to the fact that activities can normally be divided into shorter subevents that can again be characterized as activities: if I have been listening to Peter for an hour, I also have been listening to Peter during the first five minutes of that hour, the second five minutes of that hour, etc. This does not hold for accomplishments due to the fact that they crucially refer to the implied endpoint of the event: if I have finished my meal within five minutes, I did not necessarily finish my meal within the first, second, third or fourth minute of that time interval; cf. Dowty (1979:ch.3).

[+] 3. States versus achievements ±time extension

Vendler claims that states differ from achievements in that the former have a temporal extension, whereas the latter do not. This can be made clear by using the questions Hoe lang Vfinite Subject ... al ...?'For how long has Subject already Vpart ...'. The examples in (60) show that states are easily possible in such question-answer pairs, whereas achievements are not.

Example 60

a.	Hoe lang	weet	Jan	al	wie	de dader	is?	state
	how long	knows	Jan	already	who	the perpetrator	is
	'For how long has Jan known who the perpetrator is?'

a'.	Jan weet	al	een paar weken	wie	de dader	is.
	Jan know	already	a couple of weeks	who	the perpetrator	is
	'Jan has known for a couple of weeks who the perpetrator is.'

b.	*?	Hoe lang	herkent	Peter de dader	al?	achievement
		how long	recognizes	Peter the perpetrator	already

b'.	*?	Jan	herkent	de dader	al	een paar weken.
		Jan	recognizes	the perpetrator	already	a couple of weeks

Achievements occur instead in question-answer pairs that involve the actual moment at which the event took place, which is clear from the fact that they can readily be used in questions like Hoe laat Vfinite Subject ...?'At what time did Subject V ...?'.

Example 61

a.	Hoe laat	herkende	Peter de dader?	achievement
	how late	recognized	Peter the perpetrator
	'At what time did Peter recognize the perpetrator?'

a'.		Peter herkende	de dader	om drie uur.
		Peter recognized	the perpetrator	at three oʼclock

b.	Hoe laat	ontplofte de bom?	achievement
	how late	exploded the bomb
	'At what time did the bomb explode?'

b'.		De bom	ontplofte	om middernacht.
		the bomb	exploded	at midnight

States, on the other hand, normally do not readily enter questions of this type, and, if they do, the answer to the question refers to some moment at which something has happened that resulted in the obtainment of the state denoted by the verb.

Example 62

a.	*?	Hoe laat	houd	je	van Jan?	state
		how late	love	you	of Jan
		'At what time do you love Jan?'

b.	Hoe laat	weet	je	of	je	geslaagd	bent?	state
	how late	know	you	whether	you	passed	are
	'At what time do you know whether you passed the exam/get the results of the exams?'

[+] B. What did Vendler classify?

Note that we have labeled the top node in Figure 3, repeated below for convenience, not as verbs, but as states of affairs. The reason is that, although Vendler seems to have set out to develop a classification of verbs, he actually came up with a classification of different types of states of affairs; see, e.g., Verkuyl (1972) and Dowty (1979).

Figure 3: Vendlerʼs classification

For example, it seems impossible to classify the verb schrijven'to write' without additional information about its syntactic environment. The judgments on the use of the adverbial phrases of time in example (63) show that schrijven functions as an activity if it is used as an intransitive verb, but as an accomplishment if it is used as a transitive verb.

Example 63

a.	Jan	schreef	gedurende/*binnen	een uur.	activity
	Jan	wrote	during/within	an hour
	'Jan was writing for an hour.'

b.	Jan schreef	het artikel	binnen/*gedurende	een uur.	accomplishment
	Jan wrote	the article	within/during	an hour
	'Jan wrote the article within an hour.'

It is not, however, simply a matter of the adicity of the verb. First, the examples in (64) show that properties of the object may also play a role: the interpretation depends on whether the object refers to an unspecified or a specified quantity of books; cf. Verkuyl (1972/1993), Dowty (1979) and Dik (1997). In the (a)-examples this is illustrated by means of the contrast evoked by a bare plural noun phrase and a plural noun phrase preceded by a cardinal numeral, and in the (b)-examples by means of the contrast evoked by noun phrases headed by, respectively, a non-count and a singular count noun.

Example 64

a.		Jan schreef	gedurende/*binnen	twee jaar	boeken.	activity
		Jan wrote	during/within	two year	books

a'.		Jan schreef	binnen/*gedurende	twee jaar	drie boeken.	accomplishment
		Jan wrote	within/during	two year	three books

b.		Jan at	spaghetti.	activity
		Jan ate	spaghetti

b'.		Jan at	een bord spaghetti.	accomplishment
		Jan ate	a plate [of] spaghetti

A similar effect may arise in the case of verbs like ontploffen'to explode'. If the subject is a singular noun phrase, we are dealing with a momentaneous event, that is, with an achievement. If the subject is a definite plural, however, the adverbial test suggests that we can also be dealing with an activity, and if the subject is an indefinite plural the adverbial test suggests that we can only be dealing with an activity.

Example 65

a.		De bom	ontplofte	om drie uur/*de hele dag.	achievement
		the bomb	exploded	at three oʼclock/the whole day

b.		De bommen	ontploften	om drie uur/de hele dag.	achievement or activity
		the bombs	exploded	at three oʼclock/the whole day

c.	Er	ontploften	de hele dag/??om drie uur	bommen.	activity
	there	exploded	the whole day/at three oʼclock	bombs
	'There were bombs exploding the whole day.'

Second, the addition of elements other than objects may also have an effect on the interpretation; the examples in (66) show, for instance, that adding a complementive like naar huis'to home' or a verbal particle like terug'back' turns an activity into an accomplishment.

Example 66

a.	Jan wandelde	twee uur lang/*binnen twee uur.	activity
	Jan walked	two hours long/within two hours
	'Jan walked for two hours.'

b.	Jan wandelde	binnen twee uur/*twee uur lang	naar huis.	accomplishment
	Jan walked	within two hours/two hours long	to home
	'Jan walked home within two hours.'

b'.	Jan wandelde	in twee uur/*twee uur lang	terug.	accomplishment
	Jan walked	in two hours/two hours long	back
	'Jan walked back within two hours.'

Third, the examples in (67) illustrate that the categorial status of the complement of the verb may also affect the aspectual nature of the event: whereas the nominal complement in (67b) triggers an accomplishment reading, the PP-complement triggers an activity reading.

Example 67

a.		Jan dronk de wijn.	accomplishment
		Jan drank the wine

b.		Jan dronk van de wijn.	activity
		Jan drank of the wine

The examples in (68) show a somewhat similar alternation between states and activities. The (a)-examples show that if the verb denken'to think' takes a propositional complement like a clause, it cannot occur in the progressive aan het + infinitive + zijn construction, and we may therefore conclude that we are dealing with a state. The (b)-examples show that if the verb denken selects a PP-complement, it can occur in the progressive construction, and that we are thus dealing with an activity. The (c)-examples show that we get a similar meaning shift if we supplement the verb with the verbal particle na.

Example 68

a.	Marie denkt	dat	Jan	een deugniet	is.	state
	Marie thinks	that	Jan	a rascal	is
	'Marie thinks that Jan is a rascal.'

a'.	*	Marie is aan het	denken	dat Jan een deugniet is.
		Marie is aan het	think	that Jan a rascal is

b.	Marie denkt	over het probleem.	activity
	Marie thinks	about the problem
	'Marie is thinking about the problem.'

b'.		Marie is	over het probleem	aan het	denken.
		Marie is	about the problem	aan het	think

c.	Marie denkt	na.	activity
	Marie thinks	prt.
	'Marie is pondering.'

c'.		Marie is	aan het	nadenken.
		Marie is	aan het	prt.-think

[+] C. Alternative approaches to Vendlerʼs classification

The previous subsections have briefly discussed some distinctive semantic properties of verbs and events that Vendler (1957) used to motivate his classification in Figure 3. This discussion leads to the following characterizations of the four subclasses.

Example 69

Activities +continuous tense, -bounded: events that go on for some time in a homogeneous way in the sense that they do not proceed toward a logically necessary endpoint.

Accomplishments +continuous tense, +bounded: events that go on for some time in a non-homogeneous way in the sense that they proceed toward a logically necessary endpoint.

States -continuous tense, +time extension: stable situations that last for some period of time.

Achievements -continuous tense, -time extension: events that are perceived as occurring momentaneously.

One problem with this classification is that the features used are in fact more widely applicable than simply for making the distinctions given in (69). The feature ±bounded, for example, may be just as relevant for states and achievements as for activities and accomplishments. In fact, this feature may group states and activities as unbounded, and accomplishments and achievements as bounded states of affairs. The examples in (70) show that states behave like activities in that they can be used in perfective questions of the form Hoe lang auxfinite ...V?'For how long did ... V ...?', whereas accomplishments and achievements cannot.

Example 70

a.	Hoe lang	heeft	hij	naar zijn leraar	geluisterd?	activity
	how long	has	he	to his teacher	listened
	'For how long did he listen to his teacher?'

b.	*	Hoe lang	heeft	hij	zijn maaltijd	opgegeten?	accomplishment
		how long	has	he	his meal	prt.-eaten

c.	Hoe lang	heeft	hij	van spinazie	gehouden?	state
	how long	has	he	of spinach	liked
	'For how long did he like spinach?'

d.	*	Hoe lang is	de bom	ontploft?	achievement
		how long has	the bomb	exploded

If an interrogative phrase refers to a specific time, on the other hand, the acceptability judgments are reversed. This is shown in (71) by means of the adverbial phrase hoe laat'at what time'.

Example 71

a.	*	Hoe laat	heeft	hij	naar zijn leraar	geluisterd?	activity
		how late	has	he	to his teacher	listened

b.	Hoe laat	heeft	hij	zijn maaltijd	opgegeten?	accomplishment
	how late	has	he	his meal	prt.-eaten
	'At what time did he eat his meal?'

c.	*	Hoe laat	heeft	hij	van spinazie	gehouden?	state
		how late	has	he	of spinach	liked

d.	Hoe laat	is	de bom	ontploft?	achievement
	how late	has	the bomb	exploded
	'At what time did the bomb explode?'

Distribution patterns like these suggest that the four verb classes can be defined by means of a binary feature system of the form in Table 4, in which the features ±bounded and ±continuous tense can be construed as given in Figure 3; cf. Verkuyl (1993).

Table 4: Binary feature system for defining Vendlerʼs verb classes I

	–bounded	+bounded
–continuous tense	states	achievements
+continuous tense	activities	accomplishments

Note that the feature ±bounded correlates with other semantic properties of the events. Accomplishments like opeten'to eat up' and achievements like ontploffen'to explode' in (72) both indicate that some participant in the event (here, respectively, the object and the subject) undergoes a change of state and that obtaining the new state marks the end of the event; the only difference is that the transformation requires some time in the former but is perceived as taking place instantaneously in the latter case.

Example 72

a.	Jan	at	de boterham	op.	accomplishment
	Jan	ate	the sandwich	prt.
	'Jan ate the sandwich.'

b.		De bom	ontploft.	achievement
		the bomb	explodes

Activities and states, on the other hand, typically do not involve a change of stage and refer to more or less homogenous states of affairs with the result that the end of these states of affairs is more or less arbitrarily determined. This shows that it is not a priori clear whether the feature ±bounded is the correct feature; it might just as well have been ±change of state, as shown in Table 5. It therefore does not come as a surprise that there are a variety of binary feature systems available; see Rosen (2003: Section 1.3) for a brief discussion of some other proposals.

Table 5: Binary feature system for defining Vendlerʼs verb classes II

	–change of state	+change of state
–continuous tense	states	achievements
+continuous tense	activities	accomplishments

Other alternatives to Vendler's classification readily come to mind. Figure 4, which is based on Smith (1991) and Dik (1997), takes the basic division to be that between states and events: states lack internal dynamism in that they do not require any input of energy as nothing changes while they hold (Lehmann 1999:44), while events do have some form of internal dynamism. Events can be divided further on the basis of their boundedness: activities are not inherently bounded, whereas accomplishments and achievements are. The latter two differ in that only the former evolve over time. This gives rise to the hierarchical or at least more layered classification in Figure 4.

Figure 4: Hierarchical feature system for defining Vendlerʼs verb classes

As this is not the place to discuss the pros and cons of the available feature systems, we will confine ourselves to summarizing some of the conspicuous properties of the verb classes as distinguished by Vendler (1957) by means of the table in (73); we refer the reader to Miller (1999) and Rosen (2003) for more discussion.

Example 73

	state	activity	accomplishment	achievement
dynamic	—	+	+	+
bounded/change-of-state	—	—	+	+
punctual	—	—	—	+
continuous tense	—	+	+	—

[+] II. Participant roles in events

This subsection discusses alternative approaches to Vendler's classification that do not primarily appeal to the internal temporal organization of the events, but instead to specific properties of the participants in the event. One example of this was already discussed in Subsection I, where it was observed that the aspectual feature ±bounded can readily be replaced by the feature ±change of state, which involves a property of one of the participants in the event. This shift in perspective may have been (unintentionally) initiated by Dowty (1979), who suggested (in line with the basic principle of Generative Semantics) that verbs can be semantically decomposed by means of a number of atomic semantic elements like do, become and cause, which combine with a stative n-place predicate πn in (74a) to form the more complex events in (74b-d), and, in fact, a number of more complex subclasses of these event types such as inchoative achievements like ontbranden'to ignite', which would be assigned the structure become [do (α1, [πn(α1, ..., αn)])].

Example 74

State: πn(α1, ..., αn)

Activity: do (α1, [πn(α1, ..., αn)])

Achievement: become [πn(α1, ..., αn)]

Accomplishment: Φ cause (become [πn(α1, ..., αn)])

The status of the three semantic atoms is quite complex. The element do seems to function as a simple two-place predicate taking an argument of the stative predicate πn as well as the stative predicate itself as arguments. The element become, on the other hand, functions as an operator expressing that the truth value of the stative predicate πn(α1, ..., αn) changes from false to true. The element cause, finally, is a connective that expresses that event Φ is a causal factor for the event expressed by the formula following it (here: the achievement become [πn(α1, ..., αn)]); there is some event that causes some other event to come into existence.
The semantic structure attributed to accomplishments in (74d) correctly accounts for our intuition about example (75a) that the referent of the noun phrase het documenten'the documents' undergoes a change of state as the result of some unspecified action performed by the referent of the subject of the sentence, which may be further clarified by adding an instrumental met-PP like met een papierversnipperaar'with a paper shredder'; Jan has destroyed the documents by putting them in a shredder. It should be noted, however, that it is not immediately clear whether the inference that Jan is involved in some action is part of the meaning of the verb or the result of some conversational implicature in the sense of Grice (1975). The answer to this question depends on whether an example such as (75b) likewise expresses that there is some event that involves the referent of the noun phrase de orkaan'the hurricane' that causes a change of state in the referent of the noun phrase de stad'the city'.

Example 75

a.		Jan vernietigde	de documenten	(met een papierversnipperaar).
		Jan destroyed	the documents	with a paper shredder

b.		De	orkaan	vernietigde	de stad	(*met ....).
		the	hurricane	destroyed	the city	with

The fact that it is not possible to add an instrumental met-PP to example (75b) suggests that the causal relation is more direct in this case and, consequently, that the inference we can draw from (75a) that it is some action of Jan that triggers the change of state is nothing more than a conversational implicature. Given this conclusion, it is tempting to simplify Dowty's semantic structures in (74) by construing all semantic atoms as n-place predicates, as in (76).

Example 76

State: πn(α1, ..., αn)

Activity: do(α1, [πn(α1, ..., αn)])

Achievement: become(β, [πn(α1, ..., αn)]], where β ∈ {α1, ..., αn}

Accomplishment: cause(γ, (become (β, [πn(α1, ..., αn)])), in which β ∈ {α1, ..., αn} and γ ∉ { α1, ..., αn}

The interpretations of states and activities remain the same, but those of achievements and accomplishments change: an achievement is now interpreted as a change of state, such that β becomes an argument of πn, and an accomplishment is now interpreted as a change of state, such that β becomes an argument of πn as the result of some external cause γ. This reinterpretation of Dowty's system in fact seems to come very close to the proposals of the kind proposed in Van Voorst (1988) and Tenny (1994), who claim that Vendler's classes can be defined as in (77) by assuming that the nominal arguments in the clause may function as originator (typically the external argument) or delimiter (typically an internal argument of the verb) of the event; note that states do not fall in this classification since they are characterized by the absence of event structure; see also Levin & Rappaport Hovav (1995), Van Hout (1996), Van der Putten (1997) and many others for proposals in a similar spirit, and Levin & Rappaport Hovav (2005) for a recent review of research in this line of investigation.

Example 77

Activity:
Achievement:
Accomplishment:

An advantage of taking participant roles as the basis of the aspectual classification of events is that this immediately accounts for the fact that the intransitive and transitive uses of verbs like schrijven'to write' and eten'to eat' differ in interpretation in the way they do: only the transitive primed examples have an internal argument that may function as delimiter.

Example 78

a.		Jan schreef	twee uur lang/*binnen twee uur.	activity
		Jan wrote	for two hours/within two hours

a'.		Jan schreef	de brief	binnen twee uur/*twee uur lang.	accomplishment
		Jan wrote	the letter	within two hours/for two hours

b.		Jan at	vijf minuten lang/*binnen vijf minuten.	activity
		Jan ate	for five minutes/within five minutes

b'.		Jan at	zijn lunch	binnen vijf minuten/*vijf minuten lang.	accomplishment
		Jan ate	his lunch	within five minutes/for five minutes

Furthermore, this approach may provide a better understanding of the fact established earlier that properties of the nominal arguments of the verb may effect the aspectual interpretation by postulating additional conditions that the nominal arguments must meet in order to be able to function as delimiters; cf. the discussion of the examples in (64), repeated here as (79), which show that verbs like schrijven'to write' or eten'to eat' are only interpreted as accomplishments if the objects refer to specified quantities. This suggests that bare plurals and noun phrases headed by a mass noun cannot function as delimiters.

Example 79

a.		Jan schreef	gedurende/*binnen	twee jaar	boeken.	activity
		Jan wrote	during/within	two year	books

a'.		Jan schreef	binnen/*gedurende	twee jaar	drie boeken.	accomplishment
		Jan wrote	within/during	two year	three books

b.		Jan at	spaghetti.	activity
		Jan ate	spaghetti

b'.		Jan at	een bord spaghetti.	accomplishment
		Jan ate	a plate [of] spaghetti

In fact, we can now also account for the fact illustrated in (65), repeated here as (80), that the subject may affect that the aspectual interpretation of the sentence by placing a similar restriction on the originator.

Example 80