The Temporal Strategy: Time and Aspect
The Temporal Strategy: Time and Aspect
Abstract and Keywords
This chapter begins by trying to build on the suggestion, adumbrated at the end of Chapter 2, that what is common and peculiar to events, and what distinguishes them from states, is a certain kind of ‘temporal shape’ — to exploit a so-called ‘temporal strategy’. The temporal strategy appears promising as far as the events needed by the philosophy of mind are concerned, for the following reason. There is room for dispute about whether or not, and in what sense, mental phenomena are physical, whether they are spatially located, and whether they have subjects, and if so, what those subjects might be. The chapter tries to shed some light on the differences between events, processes, and states by drawing on the phenomenon of verb aspect. Though some insight is gained into the category of state by this discussion, the main focus of the chapter is the event-process distinction. It shows how the invocation of aspect can permit us to regard the distinction between events and processes as akin, in some respects, to the familiar distinction between substantial objects and the matter of which they are made.
THUS far, I have considered a number of contemporary views about the nature of events and have argued that none really answers to the purpose of the philosopher in search of an ontology for the mind. In so far as such a philosopher needs an ontology of events, I have suggested, what she needs is an ontology of genuine mental particulars and this some of the most prominent contemporary theories of events fail to provide. In Chapter 2, I considered whether it might be more promising to try to develop a theory of events by contrasting events with states, but eventually concluded that the dynamic strategy, which attempts to draw this distinction by adverting to the close connection between events and change, would not succeed unless it could be incorporated into a broader picture which took note of the differences between the ways in which events and states respectively relate to time.
In this chapter, I want to begin by trying to build on the suggestion, adumbrated at the end of Chapter 2, that what is common and peculiar to events, and what distinguishes them from states, is a certain kind of ‘temporal shape’—to exploit what I called at the beginning of Chapter 2 the ‘temporal strategy’. The temporal strategy seems to me promising as far as the events needed by the philosophy of mind are concerned, for the following reason. There is room for dispute about whether or not, and in what sense, mental phenomena are physical, whether they are spatially located, and whether they have subjects, and if so, what those subjects might be. All these are substantive questions in the philosophy of mind, and some of the difficulties we encountered earlier with various theories of events derived from the fact that the theories themselves dictated an answer to one or more of these questions, and so could not coherently be used as a neutral starting-point for their formulation. But there is no controversy about the temporality of mental (p.76) phenomena—about the fact that they take place in, or persist through, time. In so far as the temporal strategy builds on an aspect of mentality that is an undisputed feature of the metaphysics of mind, then, we have grounds for hope that it might deliver a characterization of events and states sufficiently free of controversial metaphysical consequences to provide the ontological framework for which we have been looking.
The temporal strategy might also seem hopeful for a second reason. For besides event and state, there is a third category often mentioned (though seldom discussed) in connection with the ontology of mind—the category of process. Since there has been so little explicit discussion of processes in the philosophy of mind, it is hard to know what philosophers have meant to include in this category, but ‘doing mental arithmetic’, ‘weighing up the pros and cons’, ‘working out what to do next’ could all comfortably fill the frame ‘…is a kind of mental process’, without any violence being done to ordinary usage. On the other hand, these might also seem to be kinds of mental event. Davidson, for example, recognizes no event-process distinction, and it seems clear that doings of mental arithmetic, weighings up of the pros and cons, and workings out of what to do next would all count as events on his view. Davidson's framework provides no semantic reason for treating doings of mental arithmetic any differently from the butterings of toast and the flyings of spaceships which he explicitly treats.1 Consider, for instance, the sentence
Jones did mental arithmetic in the classroom during lunchtime.
According to Davidson's views about the semantics of adverbially qualified sentences, this sentence goes over into logical notation as follows:
(Ǝx)(Did mental arithmetic (Jones, x) and In (the classroom, x) and During (lunchtime, x)).
In other words, the sentence requires an analysis which delivers ontological commitment to events of doing mental arithmetic, just as the sentence ‘Jones buttered the toast in the bathroom, with a knife, at midnight’ commits us to toast-buttering events, and it is (p.77) not clear how processes are to get in on the Davidsonian act at all. We face the question, then, whether there is really any difference between events and processes, and if so, how it is to be explained. Might ‘event’ and ‘process’ be just two words for the same thing? Or could processes perhaps be a subset of the class of events—or might they be composed from events? There is obviously a need to be clearer about what exactly it is, if anything, that distinguishes events from processes; and it is at least conceivable that the temporal strategy might be able to shed some light on the matter.
Any attempt to explain the difference between events and states—and perhaps between both of these and processes—in terms of the different relations which entities of each of these kinds bear to time will have to draw on a rich conception of temporal character. For example, it would be clearly mistaken to suppose that states are merely long events—that simple duration is the key to understanding the temporal differences between the two categories. It is true that, usually, events tend to be relatively short-lived, lasting for minutes or hours, rather than days and weeks, while states are generally rather more permanent and long-lasting (e.g. someone's knowing that Beijing is the capital of China is a state that will probably last for most of his or her life). But these facts about the relative duration of events and states are contingent, not necessary truths. There can be long events—perhaps the Hundred Years War would count as an example—and also short-lived states, e.g. a liquid's being at boiling-point on an occasion where it reaches boiling-point only for a few seconds and then cools rapidly again. It is not duration, but something else, which marks the crucial difference between the two categories.
In this chapter, I shall try to shed some light on the differences between events, processes, and states by drawing on the phenomenon of verb aspect. Though some insight is gained into the category of state by this discussion, the main focus of this chapter will be the event-process distinction. I shall try to show how the invocation of aspect can permit us to regard the distinction between events and processes as akin, in some respects, to the familiar distinction between substantial objects and the matter of which they are made. I shall then move on, in Chapter 4, to discuss in more detail the category of state.
I suggested in the last chapter that certain important ontological distinctions might depend upon the way things fill time, rather than (p.78) on the amount of time they fill. I also suggested that one could glean information about these distinctions by looking at the different verbs and prepositions we use to talk about the relations which events and states respectively bear to time (e.g. events happen in time, states persist through time, etc.). But there is also another resource to be tapped. There are important differences between the kinds of verbs which are associated with states and those which seem to imply the occurrence of events, especially where psychological phenomena are concerned. The features of human nature and experience which have usually been classified as mental are, on the whole, associated either with verbs, or else with nouns closely related to verbs (e.g. ‘belief, ‘desire’); and there seem to be major differences between the behaviour of those verbs which indicate the occurrence of events (’recognize’, ‘notice’, ‘spot’, etc.) and of those which rather suggest the presence of more or less permanent psychological states (’know’, ‘believe’, ‘fear’, etc.). This suggests that perhaps there is insight to be gained from an examination of the different behaviour of these verbs. There is a great deal of temporal information packed into verb use, and it seems reasonable to suppose that an investigation of the way in which this information is conveyed ought to play a major part in the temporal strategy. I want to begin this chapter, therefore, by considering some of the attempts which have been made to elucidate these important temporal distinctions between verb types.
1. Vendler and Kenny
In a paper entitled ‘Verbs and Times’,2 Vendler distinguishes between four kinds of verb and verb phrase: verbs of activity, of accomplishment, of achievement, and stative verbs, on the basis of what he refers to as their differing time schemata. Verbs of the first two kinds are alike in possessing continuous tenses, but differ in other ways. Vendler gives ‘run’ and ‘push a cart’ as examples of activity verbs, and ‘run a mile’ and ‘draw a circle’ as examples of accomplishments, and begins by enlarging on the distinction as follows:
(p.79) If I say that someone is running or pushing a cart, my statement does not imply any assumption as to how long that running or pushing will go on; he might stop the next moment or he might keep running or pushing for half an hour. On the other hand, if I say of a person that he is running a mile or of someone else that he is drawing a circle, then I do claim that the first one will keep running till he has covered the mile and that the second will keep drawing till he has drawn the circle.3
It is not altogether clear, though, that this is a very satisfactory way of drawing the distinction. In order for it to be true of someone that she is running a mile, it does not really seem to be true that she needs to complete the mile—whether or not it can be truly said of somebody that she is running a mile seems to have more to do with intention than eventual success. And neither does it seem obviously right to say, as Vendler does, that it does not make sense to talk of finishing running, though one can finish running a mile. Surely one can finish running (e.g. one might pull up exhausted and say that one has finished running for the day).
But other things that Vendler says about the distinction do seem to be true. Vendler points out that if someone stops running a mile then she did not run a mile; though if she stops running, then she did run. Also, Vendler associates the question ‘For how long did he Φ?’ with activity verbs (e.g. ‘For how long did he run?’, but not ‘For how long did he run a mile?’) and the question ‘How long did it take to Φ?’ with accomplishment verbs (e.g. ‘How long did it take to run a mile?’ but not ‘How long did it take to run?’), and this, too, seems to be accurate.
Neither achievement verbs nor stative verbs admit of continuous tenses, according to Vendler. Achievement verbs are distinguished from stative verbs by the fact that achievement verbs can be predicated only for moments of time, while states, broadly speaking, endure for shorter or longer periods. Thus, in ‘I recognized him immediately’ we have a verb of achievement, while ‘I knew it all along’ contains a stative verb. A simple test distinguishes achievements from states: in the case of the former, the question ‘At what time…?’ makes sense; in the case of the latter, ‘For how long…?’ is usually more appropriate. For example, ‘At what time did you realize that you'd left the oven on?’ is a comprehensible question, but ‘For how long did you realize that you'd left the oven on?’ is (p.80) not; and ‘At what time did you love him?’ would need to be construed as asking for a period of time (’Oh, all through my second year’) before we can make sense of it. ‘For how long did you love him?’, on the other hand, is straightforward.Table 3.1 summarizes Vendler's typology.
Table 3.1. A summary of Vendler's typology
With continuous tenses
Without continuous tenses
Run a mile
Eat an apple
Pick a potato
Reach the hilltop
Vendler notes that it is only verbs which admit of continuous tenses that can be used to respond to the question ‘What are you doing?’ One can be running a mile, or picking potatoes, but one cannot be recognizing or knowing or believing anything. In the latter cases, the appropriate questions are of the form ‘Do/Did you Φ?’, rather than ‘Are/Were you Φ-ing?’ Vendler concludes from this that only verbs found in continuous forms designate processes going on in time:
This difference suggests that running, writing, and the like are processes going on in time, i.e., roughly, that they consist of successive phases following one another in time. Indeed, the man who is running lifts up his right leg at one moment, drops it the next, then lifts his other leg, drops it, and so on. But although it can be true of a subject that he knows something at a given moment or for a certain period, knowing and its kin are not processes going on in time.4
Vendler is clearly right to insist that verbs which cannot take continuous tenses do not correspond to processes. But it is important to be clear that the reason why achievement verbs are not usually found in continuous-tense form is quite different from the corresponding reason in the stative case. Roughly, achievement verbs do not take continuous tenses because they relate to things which can happen instantaneously—things that one can have done, (p.81) or will do—but not normally things that one can be in the process of doing, for the simple reason that they are normally over too quickly for it to be possible for anyone to ‘catch one in the act’, as it were, of doing them. When one recognizes someone, or finds something, it happens (at least usually) in a flash—thus these are not things that one can ‘be doing’.5 But states are different. It is not because knowing is instantaneous that one cannot be knowing or believing. It is because states do not take up time at all; they have no temporal parts.
Anthony Kenny arrived independently at a typology of verbs rather similar to that developed by Vendler, in his book Action, Emotion and Will. While Vendler distinguishes four categories of verb, Kenny makes use of only three, dividing verb types into activities, performances, and states. Broadly speaking, the main difference between the two typologies is that Kenny does not distinguish, as Vendler does, between accomplishments and achievements. Kenny labels all verbs that do not admit of continuous tenses ‘static verbs’, and gives as examples ‘knowing’ and ‘being happy’, which would count as stative verbs also under Vendler's system. This criterion for static verbs can make it appear as though Kenny simply fails to recognize the existence of a class of non-static verbs—Vendler's verbs of achievement—within the class of verbs not admitting of continuous tenses, but some of Kenny's examples make it clear that this is not the case. Rather, he includes Vendler's verbs of achievement, ‘recognize’, ‘discover’, ‘notice’, and the like, within his class of performance verbs, thus indicating that the basis of his disagreement with Vendler is rather over the question whether or not these verbs have genuine continuous tenses.
It is undeniable that it is possible, even if not usual, to find most of Vendler's achievement verbs in continuous-tense form; consider, for example, ‘I am discovering for the first time what everyone saw in him’ or ‘He was slowly recognizing that I had been right all along’. Discovering and recognizing are not always instantaneous—and where they are conceived of as unfolding in a period of time, use of a continuous-tense form may become appropriate. This (p.82) is a point made by Alexander Mourelatos, in a paper that is intended as a critique and elaboration of the views of Vendler and Kenny.6 One of Mourelatos's main aims in the paper is to argue that Vendler and Kenny are mistaken in supposing that the distinctions they are seeking to capture and explore are basically distinctions between verbs (and verb phrases) conceived of as lexical types. Rather, he argues, the relevant typology ought to be one which classifies whole predications into kinds. In the next section, I want to discuss, and in some respects to endorse, Mourelatos's view.
2. Types of Verb versus Types of Predication
Mourelatos notes that not just some, but many verbs fall into more than one of Vendler's and Kenny's categories, a phenomenon he calls ‘multivalence’, which he deems sufficiently widespread to ‘make it quite wrong for us to talk in terms of exceptional or catachrestic uses of certain verbs’.7 In certain cases (the verbs ‘know’ and ‘understand’, for example), where one kind of usage is overwhelmingly predominant (in this case, the stative usage), it might be sufficient to remark idiosyncratic contexts and leave it at that, but, as Mourelatos argues convincingly, very many verbs have no special affinity for one or other of the Vendler-Kenny verb categories. ‘Run’, for example, can function either as an activity verb (’I ran for hours’) or as an accomplishment verb (’I ran to the shop’), as indeed Vendler recognizes, but the distinction between these contexts does not hinge, as Vendler implies, on the presence or absence of a stated end. ‘I was running to the shop’, no less than ‘I was running’, describes an activity, not an accomplishment, according to Mourelatos:
The generic activity of running can be further differentiated into a species (one among indefinitely many) of running-a-mile without losing its character as an activity. In other words, regardless of whether a mile is or fails to be run, any substretch of running-a-mile activity divides homogeneously into sub-stretches of the same. There is, after all, a qualitative distinction between the activity of running a mile and the activity of running the hundred-meter dash or the Marathon.8
(p.83) It looks, then, as though ‘run-to-the-shop’, as well as ‘run’ can be an activity verb, and the basis for the distinction between Vendler's first two categories (and between Kenny's activities and performances) collapses.
In the light of these difficulties for Vendler's and Kenny's analyses, Mourelatos argues that what is wanted to deal with the phenomena which interested Kenny and Vendler is a threefold distinction, at the level not of verb types, but of types of predication. Kenny's activity-performance-state trichotomy, adapted to act as a distinction among types of predication rather than types of verb, suggests itself as suitable, but, as Mourelatos points out, this will not do if we are looking for a quite general typology of predication. Activities and performances are normally things people, or perhaps animals do; it is not clear how Kenny's trichotomy would accommodate sentences not involving agency of any kind. What is needed, Mourelatos argues, is a ‘topic-neutral’ typology to correspond to Kenny's trichotomy; and the relevant categories, on his view, are event, process, and state.
There are other refinements in Mourelatos's account, for example, his distinction within the class of event predications between ‘developments’ and ‘punctual occurrences’. For now, though, it will be sufficient to reproduce here his own schematic representation of his preferred typology (see Fig. 3.1), together with a few examples.
’Situation’ is simply the generic term used by Mourelatos to encompass all the kinds of predication with which he is concerned. ‘Occurrences’ include both processes and events. Here are his examples of predications from each of the four categories.
Process: It's snowing.
Development: The sun went down.
Punctual Occurrence: The cable snapped. He blinked. The pebble hit the water.
It is easy to get an intuitive feel for the distinctions between these various categories of predication just by looking at Mourelatos's examples. But it is not quite so easy to provide a set of criteria which will serve as infallible guide-lines in the categorization of predications as predications of event, state, or process. Indeed, as Mourelatos points out, no simple grammatical or syntactic test may suffice. An interplay of factors may be involved, including the verb's meaning, its tense, the nature of the verb's subject and object, and of the qualifying adverbials, if any. But something on which Mourelatos places great emphasis in his account of the event-process-state distinction is the phenomenon of verb aspect, so I shall take some time to explain what is meant by this term.
Verb aspect is a phenomenon more familiar to linguists than to many philosophers, being more clearly marked in other languages than it is in English. For example, corresponding to the two English sentences ‘I was writing a book’ and ‘I wrote a book’, Russian has ‘Ja pisal knigu’ and ‘Ja napisal knigu’ respectively, where pisat’ and napisat’ are distinct verbs.9 This phenomenon is widespread in Slavonic languages, where most verbs occur in two forms, called the imperfective and perfective respectively. It is now recognized by linguists that many of the features of English and other languages which were traditionally assigned by grammarians to tense are aspectual distinctions.
There seem to be two ways of thinking about aspect in English. The first way ties decisions about whether a predication has perfective or imperfective aspect to the presence or absence of an aspectual marker—that is to say, a grammatical or syntactic feature—in (p.85) the case of English, the presence of a continuous-tense form. It will then be easy to decide which aspect a predication has; where a continuous-tense form is present, the predication will have imperfective aspect; otherwise, it will be perfective. But if we make the distinction in this way, it is obvious that we will need to invoke features other than aspect in order to locate predications within the Mourelatos typology. For example, ‘I was running a mile every day at that time’ will be technically imperfective, but it cannot be straightforwardly accounted a process predication, since what the imperfective encodes here is not a single activity going on in time, but rather the repetition of a kind of event. Likewise, ‘He sang for hours’ will be technically perfective, but it is natural to think that it really ought to fall into the ‘process’ category; ‘for hours’ is the kind of adverbial which Vendler associates with his activity predications, and the translation into Russian of the sentence would use a marker for imperfective aspect. Mourelatos suggests that he thinks of aspect in this first way as only one of the features relevant to categorizing a predication; he mentions it as one of six features which might help to determine the place of a predication in his typology.10
The other way of thinking about aspect is to see the aspect of a predication as something which is itself to be determined by a rather complicated interplay of factors. Thus, one might regard ‘He sang for hours’ as having imperfective aspect, even though it lacks the continuous-tense form. This seems to be Galton's view; he writes that ’almost any element of a sentence can contribute to aspectual character’.11
It does not seem to me to matter much which of these ways of looking at aspect is chosen, as long as usage remains consistent. I shall take over Mourelatos's usage, whereby the aspect of an English predication is determined simply by whether or not its main verb is in a continuous tense. The question we now need to ask is what relevance the presence of a continuous-tense form has for a predication, and how it is connected with the event-process-state trichotomy.
It is certain that many interesting connotative differences can be encoded through aspect. R. L. Allen uses the terms ‘inclusive’ and ‘intrusive’ to describe the perfective-imperfective contrast, and in (p.86) many ways, these terms are more suggestive.12 Allen describes aspect as a speaker's ‘way of looking’ at a predication, and argues that a speaker's choice of aspect depends on whether the event in question is to be looked at ‘from the inside’ or ‘from the outside’. The two sentences
(1) Your teacher told me that you weren't doing well in mathematics, and
(2) Your teacher was telling me that you weren't doing well in mathematics,
for example, differ only in the way in which the occasion of the teacher's imparting of the unwelcome information is presented to the listener. In the first case, the aspect is inclusive (perfective); that is to say, the event is presented as a complete whole which is now to be viewed in its entirety from the perspective of the present. The second sentence, by contrast, seems to invite the listener back into the past to a time at which the event in question was still occurring, when it was still, as we might say, perhaps suggestively, ‘in the process’ of happening; the aspect is intrusive (imperfective) and the event is viewed ‘from the inside’. In many ways, intrusive aspect is a more intimate aspect; several linguists have noted the use made by writers of intrusive aspect to present an event as though the reader were living through it (e.g. ‘Suddenly the bullets were flying past his head’).
But is it really true that aspect is really no more than a speaker's ‘way of looking’ at a predication? Might there not be some semantically significant differences between perfective and imperfective aspects—in which case one might think that talk of ‘ways of looking’ at a predication understates the contrast here? Let us take as an example the perfective-imperfective pair
(1) I was writing a novel yesterday.
(2) I wrote a novel yesterday.
Are there any differences between (1) and (2) which might be of semantic importance?
We might note, to begin with, that while (2) implies (1), the reverse is not true. If I was writing a novel yesterday, I did not (p.87) necessarily finish it, so it does not follow that I wrote a novel yesterday. But if I wrote a novel yesterday, it does follow that I must have been writing it yesterday. Second, there are the points about adverbial modification which Vendler notes, in attempting to draw his contrast between activities and accomplishments; I might say, for example, that I was writing a novel for hours yesterday, but not that I wrote a novel for hours. Similarly, if I say that I wrote a novel yesterday, I could add ‘and it took me eight hours’; whereas one would not add any comment about how long the writing took to the sentence ‘I was writing a novel yesterday’, though we might ask how long it went on for. And thirdly, the two sentences behave differently in connection with tense. This is pointed out by Galton. (1) has a present-tense equivalent—as Galton puts it, ‘I am writing a novel’ assigns to the present moment what ‘I was writing a novel’ assigns to the past. But T write a novel’, which is, grammatically speaking, the present-tense form of (2), does not assign to the present moment what (2) assigns to the past. Indeed, in so far as we can understand this sentence in isolation at all (that is to say, unqualified by an adverbial such as ‘every year’), it would seem that we need to understand it as meaning the same as its imperfective counterpart, ‘I am writing a novel’. In Galton's words, ‘Roughly speaking, we may say that perfective aspect is logically incompatible with present tense meaning’.13
If we are satisfied, then, that there is something of semantic significance about the perfective-imperfective contrast, how does it relate to Mourelatos's typology? Clearly, it bears on the event-process distinction. Normally, we might say, an imperfective predication can be assigned to the process category; and normally, too, where it occurs with the right kind of verb, perfective aspect indicates that a predication ought to be classified as an event predication—unless adverbial modification (as for example in, ‘He sang for (p.88) hours’) suggests otherwise. The relevance of aspect for state predications is less clear. Mourelatos does not give a precise criterion for differentiating between states and occurrences; he is more concerned with establishing the legitimacy of the category of event, and hence with the means for distinguishing event from process predications. Clearly, the Kenny-Vendler criterion for stative verbs, which makes the admissibility of continuous tenses the crucial test, is not really suitable where the typology is of predications, rather than of verbs considered as lexical types. A fuller understanding of the nature of stative predications will demand a closer look at factors other than aspect and this will have to await the discussion of the next chapter. The distinction between event and process predications, however, is largely analysable in aspectual terms, and I want to go on now to discuss Mourelatos's suggestions about the relevance of this distinction for the ontological questions with which I am primarily concerned.
4. Nominalization Transcriptions: Events and Processes
One of the most interesting arguments in Mourelatos's paper is designed to show that the distinction between count nouns and mass nouns has a striking parallel in the domain of verbs. Mourelatos begins by referring to arguments by Geoffrey Leech and others that the distinction between count and mass nouns sometimes has a role to play in the determination of the category of certain predications: those where the predication has an object whose classification as count or mass noun affects the character of the predication as a whole.14 Leech argues, for example, that in the sentence ‘He played a Mozart sonata’, where the object is a count noun (’a Mozart sonata’), the predication as a whole is turned thereby into an event predication, whereas in ‘He played a little Mozart’, where the object is a mass term (’a little Mozart’), we have instead a process predication. This claim is supported, I think, despite the absence of imperfective aspect, by the following observations:
(p.89) 1. There is an entailment not only from ‘I played a little Mozart’ to ‘I was playing a little Mozart’ but also from I was playing a little Mozart’ to ‘I played a little Mozart’ (contrast the novel-writing case above).
2. If I played a little Mozart, one would ask for how long I played, not how long it took.
3. ‘I am playing a little Mozart’ seems to assign to the present moment what ‘I played a little Mozart’ assigns to the past.
Mourelatos, though, thinks the connection between the count-mass contrast and the event-process distinction runs deeper than this. He introduces the concept of a nominalization transcription to make the point. Most predications can be given a nominalized form, in which the verb is transformed into a kind of noun. For example, corresponding to ‘Jones brushed his teeth’ there is the nominalized sentence ‘There was a brushing of (his) teeth by Jones’. Mourelatos argues that the character of the appropriate nominalization can give an indication of the nature of the original predication. The nominalized transcriptions of event predications, he claims, are normally count-quantified. In some cases, where the number of occurrences is mentioned in the original sentence (e.g. ‘Jones changed his clothes three times’), this will mean that the occurrences are explicitly counted in the nominalization (e.g. ‘There were three changings of (his) clothes by Jones’); more ordinarily, the number of occurrences will be one, and the fact that the nominalization is count-quantified will be signalled by the appearance of an indefinite article (e.g. in the previous example, ‘There was a brushing of his teeth by Jones’).
The nominalizations of process predications, however, do not share this feature, according to Mourelatos. He gives two examples. The first, ‘John pushed the cart for hours’, becomes ‘For hours there was pushing of the cart by John’. The second, ‘Jones was painting the Nativity’, is rendered ‘There was (some) painting of the Nativity by Jones’. In neither case does an indefinite article seem appropriate. As Mourelatos puts it:
The pushing and the painting in these contexts do not have the terminus or closure that would allow us to speak of a pushing or a painting—we are not told that the cart was pushed some place, or that the Nativity did get painted. The parallel with simple nouns for these transcriptions is not in sentences of the form ‘There is at least one K’; it is rather in sentences of the (p.90) same form as ‘There is snow on the roof’, or ‘There is gold in this mountain’.15
The suggestion is, then, that process predications stand to event predications in something like the way that mass nouns stand to count nouns. But can we go further? Just as the differential behaviour of mass nouns and count nouns might lead us to want to make an ontological distinction between masses and individuals, might not the distinction between event and process predications point to an ontological distinction? And if so, how exactly do we move, in this case, from grammar to ontology?
There are at least two very pressing questions here. First, does every event predication demand the occurrence of at least one event for its truth?16 For example, if the average number of children per British household fell last year, was there an event which was its falling?17 There is undoubtedly something odd, I think, about any such supposition. One can give no spatial location to such an event (other than perhaps saying that it occurred ‘in Britain’—but this does not seem to be the location of the fall in the average number of children in the same way as it is the location of the numerous other events which occur in Britain. It does not occur anywhere in particular in Britain—even if the fall is more marked in some regions than others). And perhaps more importantly, there are also problems about providing it with temporal coordinates. We can say that the fall happened last year, but not that it happened at any particular time in the year—nor even, I think, that it lasted the whole year. The idea that such an event could ‘last’ or ‘take time’ seems wrong—and there is something strange also about the idea that such an event has temporal parts. I shall call events which might cause worries of this kind non-paradigmatic events. I shall not attempt to define this concept precisely, since it seems to me probable that there is no sharp dividing-line between respectable events and those whose status as real occurrences we might be inclined to doubt. But as a rough guide-line, perhaps we might say that any event which does not seem to have a definite time of occurrence is likely to cause us unease—and will therefore count, for my purposes, as non-paradigmatic.
(p.91) And second, there is a question about how we are to understand the event-process distinction itself. For if we base the event-process distinction on an account of the difference between event and process predications, it renders untenable a certain kind of tempting view about events and processes. For example, consider the two sentences
(1) Smith pushed the cart to the top of the hill.
(2) Smith pushed the cart for hours.
(1)would be an event predication, on Mourelatos's view; the aspect is perfective and the nominalization transcription would be:
(1*) There was a pushing of the cart to the top of the hill by Smith,
the presence of the indefinite article indicating an event predication.
(2),on the other hand, would be a process predication; the aspect here is imperfective and the nominalization transcription would be:
(2*) There was pushing of the cart by Smith for hours.
But suppose what actually happened was that Smith pushed the cart for hours in order to get to the top of the hill, where he finally arrived. Both (1) and (2) might be used to say what Smith did, to describe, as we might want to say, what happened ‘in the world’. But (1), we might suppose, indicates the occurrence of an event, (2) the occurrence of a process. It is tempting to suppose that Smith's pushing of the cart for hours just was his pushing of the cart to the top of the hill; but how can this be, if ‘Smith's pushing of the cart for hours’ refers to a process and ‘Smith's pushing of the cart to the top of the hill’ refers to an event? Perhaps we might conclude that ‘event’ and ‘process’ are really just two ways of looking at the same thing. But if so, is it possible to continue to conceive of the event-process distinction as a genuinely ontological—rather than a purely grammatical—distinction ?
Despite the appeal of this thought, I believe that a grammatically rooted event-process distinction can be defended against the charge of ontological profligacy. In Section 6, I shall attempt such a defence. But before doing so, I want to say something about the first question I mentioned—the question whether every event predication corresponds to an event—in particular, whether non-paradigmatic events should be allowed to count.
I am inclined to think that whether or not non-paradigmatic events ought to be counted as events rather depends on the purposes for which one's theory of events is required. A theory which postulates events to provide a semantics for adverbially qualified sentences, for example, will probably need the category of event to be grammatically based; for adverbial modification is not confined to predications which correspond to events of a paradigmatic kind. If the theory is to be a general semantic theory, therefore, there might be good reason to make the class of events as inclusive as possible. But we are looking for an account of events which might serve for the formulation of a number of theories in the philosophy of mind. Should we want to countenance non-paradigmatic events, given this purpose?
One might think that it is possible simply to remain indifferent about this matter. For it is natural to suppose that neither mental events, nor the kinds of physical events which are involved in the statement of the relevant class of theories, are in any danger of being non-paradigmatic. Mental events, even if token physicalism is false, it might be argued, can at least be located where the bearers of the relevant mental predicates are located; and they seem to have definite times of occurrence. Sudden rememberings, the occurrences of mental images, the makings of conscious decisions, the thinkings of definite thoughts, and actions, are all events which seem to take place in time in a straightforward sense. But caution is needed here. For there do seem to be event predications involving predicates which would standardly be regarded as ‘mental’, which might give more grounds for doubt. Consider, for example, the question ‘Did you remember to lock the back door?’ and the answer ‘Yes, I remembered’. Now, it is certainly arguable that it may be true of me that I remembered to lock the back door, without its also being true of me that I ever thought consciously about locking it at any stage, without there ever having occurred to me the thought ‘Oh, I must lock the back door’. It may just have been part of my dreary, habitual round of nightly duties—I did it unthinkingly, as I always do it. In such a case as this, there would seem to be no definite answer to the question when I remembered to lock the back door; it was not a dockable event, like my suddenly remembering that I am supposed to be at the dentist's. We might think, then, that such (p.93) remembering events are a bit like fallings in the average number of children per household—mere ontological shadows, as it were, of the nominalized versions of certain predications, not to be taken at face value. What ought the token physicalist to say about such ‘events’ as this?
There seem to be at least four possible views one might take about such cases:
1. The mental attribution is, strictly speaking, false—it is not really true that I remembered to lock the back door, if there was no special time at which I remembered to do so.
2. The mental attribution is true, but does not require the occurrence of a mental event for its truth—some other account of the semantics of a sentence of this kind can be given.
3. The mental attribution is true and does require the occurrence of a mental event for its truth; it is just that the event in question is not a conscious occurrence, but an unconscious one.
4. The mental attribution does require the occurrence of a mental event for its truth—but this event can be non-paradigmatic. There is no need to conceive of it as having occurred at any definite time at all—even unconsciously. Rather, we need to shake ourselves free of an unduly restrictive conception of what constitutes a ‘real’ event.
I do not intend to choose here between these four positions; nothing I have to say in what follows will depend upon making such a choice. But I do want to point out that any choice one might make will have consequences for the scope and perhaps for the plausibility of certain theories about mental events, including token physicalism. If one is inclined to choose position (1), for example, the theory that all mental events are identical with (or constituted by) physical events will extend only to those events which are conscious and dockable—those which afford no room for scepticism about their occurrence. One might think that this is the most plausible version of the theory—that such events are identical with physical events occurring in the brain is a highly attractive view, while one might be less sure about mental events our evidence for the occurrence of which is quasi-theoretical, rather than introspective. At the other end of the scale, position (4) is hard to make cohere with token physicalism at all. For if a mental event can be (p.94) said to have occurred merely in virtue of the fact that some event predication is true, whether or not it can be supposed to have occurred at any particular time, it seems difficult to square this with the view that all mental events are identical with, or constituted by, physical events. For it is natural to think that the neural events with which, on most conceptions of what token physicalism amounts to, mental events are supposed to be identical, or out of which they are constituted, must take place at particular times—that they are paradigmatic events. But it does not seem to make sense to say both that an event has a definite time of occurrence and that it has no such definite time of occurrence.
The token physicalist, therefore, might have good reason to reject (4). But any of the other three views, it seems to me, is consistent with the theory. (3) might be held, for example, by someone who believed that folk-psychological explanation demonstrates a quasi-theoretical commitment to a number of non-introspectible mental states—and who did not see why we should not also be theoretically committed by folk psychology to non-introspectible events. The question whether every event predication corresponds to an event, therefore, seems to remain an open one—a token physicalist can simply define the class as widely or as narrowly as required to give the sense she wishes to the claim that all mental events are identical with, or constituted by, physical events. If a narrower class is chosen, an account of event predication alone will not be enough to explicate the category of event—something will need to be said about what distinguishes ‘real’ events from those which are merely the ‘ontological shadows’ of nominalizations. But even if one chooses to opt for some narrow definition which excludes non-paradigmatic events, like falls in the average number of children per household, it still seems to me that at least the necessary conditions of eventhood should be sought in the temporal features of event predication. I shall try to make good this claim in Section 7. I turn, now, to the event-process distinction.
6. Event and Process as Ontological Categories
An example may be the best way of making persuasive the case for an ontological distinction between events and processes. Consider (p.95) the present humming of my computer. I might think or speak about this humming; I might say, for instance, that the humming of my computer is distracting me. Now, this humming is undoubtedly something which is taking place in time—it is an entity which has temporal parts, and thus is somewhat event-like in nature. But there are good reasons for thinking that it is nevertheless not an event. I might describe the humming, for example, as persistent, or continuous—but it does not really seem to make sense to describe an event as persistent or continuous. And the humming might stop, but do events really stop? They come to an end, but that is different. These are small and subtle differences, but they are real enough. And they seem to me to show that when I speak of the present humming of my computer, I am not referring to any individual event. Rather, I am referring to a continuous activity, something which is going on through time, a process.
Against this argument for a distinction, though, one might bring the following, forceful reasoning. There seems to be nothing in the nature of time itself, as it were, or in the nature of change, which dictates that happenings should divide themselves into two categories, events and processes. Looking back over any given period of time, for example, it is not instantly obvious that happenings of two entirely distinct varieties were occurring throughout. All we seem to have, when we view the world from this kind of temporal standpoint, is a succession of events, longer or shorter, more or less spatially dispersed chunks of which can be (perhaps only inexactly) picked out by event names and event descriptions. All we really have, it might be said, are events succeeding other events in time—and different ways of looking at those events—from the ‘intrusive’ or from the ‘inclusive’ point of view, to use Allen's terminology. The event-process distinction arises as a result of the existence of these different viewpoints, and to this extent, is just a by-product of grammar.
But this way of looking at things is, I think, unduly dismissive of the importance of grammatical distinctions for ontology. Just because we can conceive of a point of view from which ‘intrusive’ aspect is redundant—a kind of ‘God's-eye view’ of the history of the world, say—does not mean that the referring terms which are associated with our use of intrusive aspect in the time-bound circumstances in which we actually live our lives can only refer to the very same entities as are available from the ‘God's-eye’ perspective. (p.96) We do possess referring expressions deriving from imperfective contexts, as well as ones deriving from perfective predications; and there are reasons for wanting to insist that the referents of these expressions must be genuinely distinct, even where the entities they single out appear to coincide both spatially and temporally, as did Smith's pushing of the cart for hours and Smith's pushing of the cart to the top of the hill. For processes have properties which it would be inappropriate to ascribe to events, and vice versa. For example, as I have already noted, the humming of my computer in the process sense can be persistent; but it does not really make sense to think of an event as persistent. And events, it is natural to say, take time, while the same does not seem to be true of processes. Smith's pushing of the cart to the top of the hill, for example, took four hours; but his pushing of the cart for hours did not take four hours, though it lasted for four hours. Arguments from Leibniz's Law, then, can be straightforwardly brought to bear against any proposals for the identification of processes with events.
It can be harder to grasp the distinction between processes and events than it is to understand the distinction between masses and individuals, which can be regarded as parallel in some respects. For one has to overcome, in this case, not only the natural inclination to assume that the dimensions along which language may carve the world are simple, linear, spatial and temporal axes, but also confusions generated by the fact that the same gerundial nominal is used to refer both to the event and to the corresponding process. Thus, looking at our difficulty with Smith, for example, one is inclined to think that after all, there can only have been one pushing. But this inclination stems from an accident of our language, which has in this case no lexical difference to correspond with the grammatical distinction between event and process. To help with this, I want to introduce the following notation: ‘Φ-ingsE’ are events and ‘Φ-ingSp’ are processes. Φ-ingsE are countable; when Smith pushed the cart to the top of the hill, there was exactly one pushing, of the cart to the top of the hill, and if he does it again the next day, there will have been two. PushingsP, on the other hand, are not countable. ‘Pushing’, in this sense, is an activity of which there can be more or less, but not one or two. The complicated truth about Smith and his pushing is therefore this: that while it is true that, in a sense, there was only one pushing (one pushingE), there was also, alongside this pushingE, some pushingp. And the pushingE and the pushingp are (p.97) items of different ontological types and for the reasons given above cannot be identified with one another. It is a familiar (though still not sufficiently uncontroversial) point that in refusing to identify continuants, like statues, horses, and human beings, with the lumps of matter of which they are composed, we do not ‘double count’. In insisting on the distinctness of spatiotemporally coinciding events and processes, I am merely making a parallel claim in respect of entities with temporal parts.
I now want to go on to say a little more about the concept of temporal shape, to see how it might be used to talk about the differences between events, states, and processes, and to discuss, finally, whether an account of events based on temporal shape might be a plausible rival to the accounts of events I considered in Chapters 1 and 2—in particular, to the view that events are changes.
7. Temporal Shape
How do the distinctions between event and process predications link up with the concept of ‘temporal shape’ which I introduced at the end of Chapter 2? There, I suggested that the question whether or not an entity has temporal parts counts as a question about its temporal shape. But if the difference between events and processes is itself to count as a difference of temporal shape, we need to add to these observations. For both events and processes, on the account I have offered, have temporal parts, and so both can be said to ‘happen’, to ‘occur’, to ‘take place’. How might we extend our conception of the features of an entity which contribute to its temporal shape, so as to account for the event-process distinction?
One thing which helps a little is the difference we have already noted between entities which take time and those which can be said to go on for a time. The humming of my computer, for example, does not take two hours—it only goes on for two hours. But these features of events and processes do not seem to be completely reliable as guides to the distinction. Though many events certainly take time—buildings of houses and runnings of miles, for example—there seem to be other events where this form of words is not appropriate—not only instantaneous (or very short-lived) events, (p.98) but also events which do have temporal duration. Of a picnic, for example, I think one would ask how long it lasted, not how long it took; and the nominalization transcriptions of certain predications which otherwise seem to be perfective also cause problems. Consider, for instance, the sentence ‘I went to the fair’. This would seem to be perfective: one would, I think, produce ‘There was a going to the fair by me’ and not ‘There was going to the fair by me’ as its nominalization transcription; if I went to the fair twice, then there were two goings to the fair by me, etc. But in normal circumstances, I think it would be inappropriate to ask how long it took to go to the fair—though one might ask how long it took to get there. ‘How long did you go for?’ seems more natural—but this was the form we associated above with processes.
Though it is generally true that events take time, while processes last for a time or go on for a time, then, I do not think we can rest entirely content with this account. It may be more hopeful, I think, to look to differences in certain temporally sensitive adjectives which it is appropriate to apply to events and processes respectively. A process like the humming of my computer, for example, can be persistent, continuous, ongoing, constant, incessant, perpetual, unremitting, sporadic, intermittent, irregular, steady. But none of these adjectives can be comfortably applied to an event—at least, where such an adjective is so applied, it must be given a different sense. This is related to the fact that the aspectual features of event predications make certain kinds of adverbial modification inappropriate; for example, if Smith pushed the cart to the top of the hill, we cannot say that he pushed it to the top of the hill constantly (unless we mean that he did it many times over); though we can say that he pushed it for hours constantly. Processes are things which, as it were, go on throughout periods of time—and so we can sensibly ask how they went on through that time—whether they went on constantly, or intermittently, etc. But events simply happen—there is a ‘when’, and a ‘how long’ to be asked, but it does not make sense to ask a certain kind of ‘how’ question, the kind which asks for the distribution of the happening in time.
I suggest, then, that the receptivity of an entity to such temporally sensitive adjectival modification ought to be accounted a feature of its temporal shape. Putting all that we have said so far together, then, we can say, roughly, that those features of a temporal entity (p.99) which determine its temporal shape are those which determine which of the following may be said of it:
1. Whether it persists, occurs, goes on, continues, happens, obtains.
2. Whether it takes time, lasts for a time, goes on for a time, persists for a time, occurs at a time.
3. Whether certain temporally sensitive adjectives may be applied to it—these include ‘intermittent’, ‘continuous’, ‘persistent’, etc.
Of course, this is only rough; in particular, much more might be said about other kinds of adjectives which might serve to indicate differences of temporal shape—those to do, for example, with change, which might be thought to distinguish persisting physical objects from events and processes. But since the considerations alluded to in (1) will suffice to make this distinction, I will not pursue the matter here.
On the basis of this rough outline, then, we can now say this:
1. Physical objects, their parts, and the masses of matter which constitute them all share a temporal shape—they persist through time, last for a time, and may change; none occurs or happens. This is related to the point, raised at the end of Chapter 2, that the part-whole relation depends on identity of temporal shape between part and whole.
2. Events and processes have many features of their temporal shapes in common; both occur, both have temporal parts. But subtle differences of temporal shape which are related to the aspectual distinctions discussed above also exist; processes, but not events, can be persistent, intermittent, etc. Processes, then, if we are sticking to the principle that only entities which share a temporal shape can be related by the part-whole relation, ought not to be regarded as parts of events (and vice versa).
3. States seem to have many temporal features in common with physical objects. They persist through time (at least usually) and have no temporal parts. That states, but not physical objects, can be said to obtain seems to me to be related not to a temporal difference, but rather to the special relation between the obtaining of a state and the holding of certain truths; e.g. if the state of my believing that p ‘obtains’, that is because I believe that p. I conclude, therefore, that states share a temporal shape with physical objects. (p.100) Having given a rough indication of how the concept of temporal shape might be elaborated, I now want to argue that the best hope for a ‘metaphysical theory of events’ may lie with the concept of temporal shape.
It is, I think, a hopeful feature of the concept of temporal shape that it can be used to explain what unites those events which are described by means of nominals derived from event predications (capsizings, runnings, pushings, etc.) with those which are referred to rather by means of nouns independent of verbs (funerals, picnics, vigils, etc.). For despite the differences between them, all these kinds of events occur. The idea of occurrence, of something which happens in time, which consists of temporal parts succeeding one another, is common to all. Compared to the accounts offered by Kim, Bennett, and Lombard, an account of events based on the concept of temporal shape would, admittedly, be a ‘thin’ account; it will perhaps be felt that to say that what events have in common is that they occur is not to say very much. Certainly, it does not say anything very rich about the ‘metaphysical nature’ of events. But perhaps it may be that no such rich account is really possible. The only characteristics of events which seem both sufficiently general to enable us to include all those things we might want incorporated into the category, and yet sufficiently specific to enable us to distinguish events from other entities, like physical objects, processes, and states, seem to be temporal characteristics. These features are common to vigils and weddings and pushings, of carts and settings, of the sun; and either jointly or individually seem to enable us to distinguish between these entities and things which are not events, like the sky's being blue at t and the humming, of my computer. To say this is not, admittedly, to say a great deal, but perhaps this may seem less unsatisfactory and surprising if one reflects on what a general account of continuants might look like; would we be able to say anything more than that they persist through time and have spatial, but no temporal, parts?18 I suggest that when dealing with such large categories as ‘event’ and ‘continuant’ we might be foolish to expect any very rich account to be forthcoming.
An account of events based on the idea of a common temporal shape might also enable us better to understand the close but non-necessary connection between events and change. Most things (p.101) which can be said to occur either are changes or involve change of one kind or another—for it is quite natural that we should only have reason to single out entities which consist of a succession of temporal parts where the succession is in some way pertinent, noticeable, or causally significant—and, by and large, it is successions of temporal parts such that successive parts differ from one another in various respects that matter to us in these ways. But this is not always true. Sometimes, a succession of resembling temporal ‘slices’ of a thing, person, or area can have event-like significance—as, for example, when a vigil is held, or a person responds by pointedly saying nothing at all. These can merit the status of occurrences because we can interpret the succession of resembling ‘slices’ in such cases as an action or a demonstration—something which deserves to be focused on as a succession, as something which occurs. But we need to beware of conflating such changeless events with states. States differ from events most fundamentally not in being ‘unchanges’, not in being constituted by successions of resembling time-slices, but rather in failing to have temporal parts at all.
I would like to suggest, then, that temporal shape is the key to understanding the category of event and also to making sense of the event-process and event-state distinctions. In the next part of this book, I want to move on to discuss in more detail the concept of a state—a concept which, I believe, has been grossly misused in philosophy of mind. Aspect alone, as I mentioned earlier, seems unlikely to provide any general account of states, although there are some generalizations one can make (’stative’ verbs are rarely found with imperfective aspect, for example). But clearly, more needs to be said. I shall begin the next chapter, therefore, with a discussion of states and the expressions by means of which we refer to them and otherwise demonstrate our commitment to their existence. (p.102)
(1) See ‘The Logical Form of Action Sentences’, in Nicholas Rescher (ed.), The Logic of Decision and Action (Pittsburgh, 1967); repr. in Davidson, Essays on Actions and Events.
(3) ‘Verbs and Times’, 145.
(4) ‘Verbs and Times’, 144.
(5) ‘Usually’ because there are some exceptions—see below. These verbs also have a stative usage. The question ‘Do you recognize…?’ is asking not whether some rather short-lived event has occurred but rather about a present state of familiarity in the respondent. ‘Do you find him attractive?’ exhibits ‘find’ in a similarly stative usage.
(6) Mourelatos, ‘Events, Processes and States’.
(9) This example is taken from A. Galton, The Logic of Aspect (Oxford, 1984), 1.
(10) ‘Events, Processes and States’, 421.
(11) The Logic of Aspect, 70.
(12) In R. L. Allen, The Verb System of Present-Day American English (The Hague, 1966), 218–19.
(13) The Logic of Aspect, 3. ‘Roughly speaking’, presumably, because there seem to be special cases which provide exceptions to the general rule, e.g. performatives. For example, ‘I name this ship’ has present-tense meaning, though its aspectual character seems to be perfective. However, Galton is at pains to point out that it remains true that sentences such as ‘I name this ship’ do not assign to the present what sentences like ‘I named this ship’ assign to the past—the reason being that the present-tense sentence is not a report at all—and so does not, in the relevant sense, assign anything to the present. An utterance of ‘I name this ship’ just is an act of naming; it does not report it. Thus ‘I named this ship’ is not a past-tensed version of T name this ship’, but rather a report of my utterance, at some past time, of the sentence ‘I name this ship’ (The Logic of Aspect, 13–14).
(14) G. N. Leech, Towards a Semantic Description of English (Bloomington, Ind., 1969).
(15) ‘Events, Processes and States’, 427.
(16) A parallel question also arises for processes—but I consider only events here, since the same considerations apply to both.
(17) I owe this example to Paul Snowdon.
(18) It is not obviously true that all continuants are made of matter (for one might want to class, for example, holes, rainbows, shadows, as continuants of a kind).