Tense - Oxford Handbooks
Page of

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). (c) Oxford University Press, 2015. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Handbooks Online for personal use (for details see Privacy Policy).
Subscriber: null; date: 25 November 2015

Abstract and Keywords

For Plato and Aristotle, a verb was above all a word that indicates time. The Greeks and Romans made no very strong distinction between a marker and the concept that it marked. Consequently, for some two and a half millennia, the different forms of the verbal paradigm were called times or tenses. It was only in the twentieth century that a consistent distinction began to be made between two kinds of meaning distinctions morphologically marked in synthetic languages, namely “tense” as an indication of different times (such as past, present, and future) and a quite different indication of time, namely aspect, although that distinction was not unknown in the ancient world. This article explores tense and the representation of time, the coherence of verbal systems, tense and aspect, the binary tense systems of Indo-European languages, perfectives and performatives, the allosemes of verbal forms, the vast present, tense and aspect in Swahili, the tense systems of Ruhaya and Kikuyu, and the use of modal forms in tense function.

Keywords: time, tense, aspect, verbal systems, Indo-European languages, perfectives, performatives, allosemes, vast present, modal forms

1. Tense and the Representation of Time

For Plato and Aristotle, a verb was above all a word which indicates time (prosse:máinei khrónon).1 In Greek and Latin, which are synthetic languages, distinctions of meaning are typically indicated through the use of affixes. The Greeks and Romans made no very strong distinction between a marker and the concept that it marked. Consequently, for some two and a half millennia, the different forms of the verbal paradigm were called times or tenses. It was only in the twentieth century that a consistent distinction began to be made between two kinds of meaning distinctions morphologically marked in synthetic languages, namely “tense” as an indication of different times (such as past, present, and future) and a quite different indication of time, namely aspect, though that distinction was not unknown in the ancient world. The Roman grammarian Varro (116–28 BCE), for example, in his treatise De lingua latina “On the Latin Language” (VIII:20, IX:96, 99), described the six forms of the Latin indicative paradigm as comprising three tenses, Past, Present and Future, each divided into complete (Perfectum) and incomplete (Infectum), as in the third person singular indicative paradigm of laudo ‘praise’ in Table 17.1.

Table 17.1








3 sg was praising

3 sg praises

3 sg will praise





3 sg had praised

3 sg has praised

3 sg will have praised

The stem is lauda-, and since final /-t/ marks third person singular, the Present Infectum is unmarked for both tense and aspect.2 The Infectum forms are all unmarked for aspect, whereas the Perfectum forms are all marked by /-v-/ in the typologically expected position immediately after the stem. Present tense is unmarked, Non-Present marked with /-b-/ in the Infectum, and /-er/ in the Perfectum. Finally, Past is marked by /‑a/, and Future by /-i-/ immediately before the (p. 508) personal (pronominal) inflection. Aspectual markers are typically placed closer to the stem than the tense markers. There is probably a justification for this: the verbal lexeme brings its own aspectual features (Aktionsart)3 to the verbal form, and these interact closely with the grammatical aspectual forms.

Twentieth century linguists, even though aware of the categorical differences between tense and aspect, nevertheless frequently continued to refer to the “six tenses” of Latin. Robins, for example, when writing of Varro’s description of tense and aspect in Latin, comments “Formally there are six tenses in Latin, and in their meanings we can recognize, as in Greek, aspect and time reference” (1951, p. 56). Even Gustave Guillaume, who gives us (1929, 1933) the earliest clear definitions of tense and aspect, still refers to the “six tenses” of Greek and Latin, and fails to see that Latin has three tenses with two aspectual contrasts, whereas Greek has two tenses with three aspectual contrasts. As a result, another early writer on tense and aspect, Jens Holt (1943), whose definitions influenced Comrie (see Comrie, 1976, p. 3, fn. 1), comments that he can make no sense of Guillaume’s analysis of the Greek verb.4

This confusion has continued, to a certain extent, right up to the present day. Reichenbach (1947) analyzes all distinctions on his “line of time” as tenses, since it is assumed that the various forms (e.g., the pluperfect) are no more than indicators of times in the real world. Even in Comrie (1985) the term “absolute-relative tense” is used for what is just a simple combination of tense and aspect.

1.1. Tense as a System of Concepts

It must be observed that language is a mental phenomenon, and cannot be adequately described as a physical or behavioral phenomenon, as in the definitions of early to mid twentieth century positivists. Nor is it correct to reduce it to an abstract algebra that supposedly exists independently of the speaker. A language does not exist independently of those who speak it, and the scientific investigation of linguistic phenomena cannot ignore the speaker’s mentally stored knowledge, without which there would be no language. Such stored content is not an inaccessible abstraction; it is accessible to physical intervention (Penfield and Roberts, 1959) and to reconstruction by the classical methods of historical linguistics (Hoenigswald, 1960).

Several important observations can be made about the system in Table 1: (i) Latin has a ternary system of tense, where the three contrastive tenses and two contrastive aspects are clearly marked in the morphology; (2) the base form of the whole system (p. 509) is clearly identifiable (lauda-), since it is unmarked for both tense and aspect; (3) to this base form a variety of morphemes are added to mark the more complex forms of the system (see above), namely the one aspectual marker /-v/ immediately suffixed to the stem, and the more complex tense markers /-ba,-bi,-era,-eri/ immediately preceding the personal inflections (here third person /-t/).

The regularity of this morphology led to a major error in twentieth century linguistics: the assumption that the morphological forms constitute the verbal system. This was reinforced by the behaviorist and positivist bias that science only deals with the directly observable, and that consequently the morphology had to be the verbal system. The reality, as pointed out long ago (1916) by Saussure in his analogy of the game of chess, is that the paradigm is only a set of markers for a contrastive set of grammatical meanings, a “content system” in the terminology of Jakobson (1933, 1936) and Hjelmslev (1935). The game of chess, as Saussure made explicit, is not a set of chess pieces, but the moves that these pieces make, marked by the pieces on the chessboard. A tense, or an aspect, is a position in a mental, a conceptual system.

Grammar is where marker and meaning, form and function, meet. As in the Latin paradigm in Table 1, individual concepts are formally represented by morphosyntactic markers, and one of the central issues where the tense system of a particular language such as Latin or Greek is concerned, and where tense in general in concerned, is what the system of concepts is, what the system of markers is, and how they relate. But it is important to stress that to refer to a form such as laudaverat as the pluperfect tense or the meaning of the pluperfect tense as “a time before a given time in the past” (as Reichenbach effectively does) is derivative, and to some extent an abus de langage.

1.2. The Coherence of Verbal Systems

The verbal system of a language typically shows the same kind of balanced and coherent patterning that we see in the Latin indicative system in Table 1 above. This does not imply that the system of morphosyntactic markers which represent that system need itself be coherent, though the principle of compositionality would suggest a tendency in that direction.5 Systems which are represented by remarkably different systems of markers can not only be equally coherent, but remarkably similar as conceptual systems.

To show how conceptually very similar systems can be nevertheless remarkably different in structure, the six forms of the Ancient Greek Indicative are presented in Table 2 below.6 The verb stem is /graph-/ ‘write’ (the aspiration of the /p/ being lost before sibilants); the forms are again third person singular. Hyphens have been added to show the sequential morphemes. In the left hand column are the traditional names of the morphological forms, and on the right the modern aspectual terminology.

Table 17.2





Present stem




3 sg was writing

3 sg is writing

Aorist stem




3 sg wrote

3 sg will write

Perfect stem



Retrospective, Perfect, Anterior

3 sg had written

3 sg has written

The Present Stem includes the Imperfect (Past) and the Present (Non-Past). The Future grápsei was always considered a separate tense; the first to suggest that it was a Perfective aspect of the Non-Past was Kuryłowicz (1964, p. 115), as noted by Comrie (1976, p. 67). The term Perfect, because it has been continually confused with (p. 510) Perfective, is now normally avoided; it will be replaced by the term Retrospective (abbreviation RTR) in what follows.

Here we notice that the past tense is regularly marked with the augment /e-/, and that the non-past tense carries no augment. The imperfective (abbreviation IPFV) stem is unmarked, as is the norm in Type A systems (see 3.1 below). The perfective (abbreviation PFV) is marked by /s/ immediately after the stem, and this /-s-/ also marks not only a perfective Past that contrasts with the imperfective Past, but also, a perfective Non-Past which represents the future, just as in Eastern and Western Slavic, where the perfective forms of the Non-Past are the normal representation of the future, and the imperfective forms represent the ongoing present of the Non-Past (e.g., Russian perfective ja pročitaju ‘I shall read’, imperfective ja čitaju ‘I am reading’).7

As noted by Jakobson, apropos of the Slavic languages ([1957] 1984, p. 49), “… futurity is the most usual meaning of the perfective present” (i.e., of the Perfective Non-Past). In the binary tense systems of Czech, Slovak, Polish, Ukrainian, Russian, etc., the Non-Past includes both the ongoing present (represented by the Non-Past Imperfective) and the future (represented by the Non-Past Perfective, since what is complete in the Non-Past must necessarily be in the future; it cannot be in either the past—by definition—or the ongoing continuous present). This division of labor is made explicit in the Russian examples in Figure 17.1, where the translations are only approximate because of differences between the systems.

The opposition of past/non-past is fundamental to both the Greek and Slavic systems, but is not rendered overt by their quite different morphological systems of marking. Nor does the use of non-past forms to indicate futurity serve to reveal the underlying principles of either system, with the consequence that to this day scholars continue to refer to the present, past, and future tenses of both Greek8 and Slavic languages such as Russian.9

(p. 511) 2. Definitions of Tense and Aspect

In order to define the category of tense, it must be distinguished from its concomitant category of aspect. The first to present clear, contrastive definitions of these linguistic categories was Gustave Guillaume (1933):10 He produces a simple binary contrast (1964, p. 48):

Est de la nature de l’aspect toute différentiation qui a pour lieu le temps impliqué.

(Every differentiation of the time internal to the event involves aspect.)

Est de la nature du temps toute différentiation qui a pour lieu le temps expliqué.

(Every differentiation of the time external to the event involves tense.)

In short, aspect is concerned with the representation of the time contained in the event, and tense with the representation of the time that contains the event. This is a clarification of earlier comments of Guillaume (1929, p. 15) where the lexical meaning of a verb (now normally referred to as Aktionsart) is considered an essential part of aspect.

Guillaume’s original figure representing Event Time was a simple space, running between two vertical bars A and B. To this Hewson and Bubenik (1997, p. 14) added a set of five cardinal positions preceding, following, and inside the time of the event proper. Guilaume’s vertical bars have also been replaced by square brackets in the present work to symbolize the beginning and the end of an event. In Figure 17.2, A marks the Prospective aspect, with the position preceding the event proper; B marks the Inceptive aspect, with the event just beginning; C marks the Imperfective aspect, indicative of the medial portion of the event; D marks the Perfective aspect, with the event completed; and E, the Retrospective aspect, following the event proper.11 These are the typical range of aspectual representations, but, of course, not the only aspects, nor are languages required to have them all: they simply present the five cardinal positions: A—before the event begins; B—at the beginning, C—between the beginning and the end, D—at the end of the event; and E—after the end, in the result phase.

As Hirtle notes (2007, p. 30), Guillaume’s comment from Temps et verbe (1929, p. 21) that “l’aspect … éveille … l’image même du verbe dans son déroulement” (“aspect evokes the image of the verb in its progress”) is echoed by Holt, who had read Guillaume’s Temps et verbe, and who defines aspects (1943, p. 6). as “les manières diverses de concevoir l’écoulement du procès même” “the various ways of conceptualizing the progress of the event.”12 This definition leads in turn to Comrie’s (1976, p. 3), “aspects are different ways of viewing the internal constituency of a situation,” which he acknowledges is based on Holt’s definition.

Distinguishing tense from aspect is essential for a proper description and understanding of tense. There is still much confusion in the literature, because of (p. 512) a tendency to describe linguistic items in terms of their function rather than their category. If I use a kitchen knife as a screwdriver, I still return it to the kitchen drawer with the cutlery; I do not describe it as a screwdriver and put it in the toolkit. It is still a knife; it has not changed category because of a secondary function.13

The use of an auxiliary such as have, for example, creates a change of aspect, not a change of tense. The tense of the auxiliary is the tense of the whole verbal complex: I have eaten and I will eat are both aspectual forms of the English Non-Past; the Past forms are I had eaten and I would eat, where the auxiliary is in the Past tense. Functionally I have eaten (position E in Figure 17.2 above) represents the past, and I will eat (position A) represents the future, but in terms of the categories of the underlying system of the English verb, these are both aspectual forms of the Non-Past tense. Aspect is used, just as much as is tense, to represent events that took place in the past, or events that will take place in the future. It is, consequently, important to base description and analysis primarily on (morphosyntactic) form and (verbal) category.

3. Time and Tense

Guillaume’s distinction of the time that contains the event (which gives us the category of tense) and the time contained in the event (the varying representations of which are provided by the varieties of aspect) was clarified and simplified by Valin (1994, p. 40), who introduced the terms Event Time (the basis of aspectual distinctions) and Universe Time (the basis of tense forms), which can then be represented in diagram form to make the distinction explicit, and to enable the graphic representation of different tenses and aspects, so that no matter what the terminology used, it can be clearly seen whether two different terms represent the same or different entities. This ability to make the meaning of the terminology visible is of fundamental importance because of the terminological confusions that have arisen, especially in the literature of the last half century.

Event Time, as we have already seen, can be represented by square brackets with a space between, and a variety of cardinal positions. When such forms are tensed, they can be represented over or under a line of time shown as an arrow, flowing from an infinity to an infinity, that represents Universe Time. When the arrow points to the left, there is a representation of the flow of time out of the future, through the present, into the past, as in (a) in Figure 17.3, thereby representing the experience of time in the Working Memory,14 where now is always descending into the past. This is the passive view of time, Descending Time, where the past is irretrievable, and death inevitable. But because time is movement, where either background or figure may be perceived as moving, there is an equal and opposite Ascending Time, where an action begun now, such as the reading of a page of a book, will be completed in the foreseeable future, as represented in (b) in Figure 17.3. This is the active view of time, where the cognitive faculties employ past experience to understand and explain the present, and to plan the future.

(p. 513) As noted by Guillaume (1990, pp. 141–142): “This attribution of time to two opposing movements is a new far-reaching fact…. Descending Time is the objective image of time…. Opposed to this strictly objective image of time there is an inverse subjective image: that of time which opens out before us to allow us to carry out our intentions.”15 These two distinct and contrastive representations of the flow of time are the basis for radically distinct tense systems, that otherwise appear to be very similar: there are some classical examples from the Indo-European families.

3.1. The Binary Tense Systems of Indo-European

In a study of the tense-aspect systems of the twelve Indo-European (IE) families (Hewson and Bubenik, 1997), it was discovered that only three of the IE families had three or more tenses: Italic (see the Latin system in Table 17.1 above), Celtic, and Baltic have future tenses, whereas the remaining nine families16 have binary tense systems of Past vs. Non-Past, and represent the future by aspectual means.

The binary systems divide into two distinctive types, one exemplified by the Greek indicative forms in Table 17.2, which we may now show in schematic form in Figure 17.4. Similar systems are also found in Slavic, one of the largest of the IE families, and in other smaller groups, such as Albanian and Armenian, and also Tocharian (extinct since the middle of the first millennium).

  1. (i) For clarity of exposition the forms that are unmarked for aspect are placed above the line, the marked aspectual forms below. The unmarked forms are the base forms of the paradigm; the marked forms are derivations, aspects developed from the base forms.

  2. (p. 514) (ii) Continuous lines represent accompli, the part of the event that is complete; broken lines represent inaccompli, the part of the event that is yet to be completed.

  3. (iii) The PFV represents the completion of the IPFV; the RTR (Retrospective) represents the time after the completed event. Every RTR necessarily contains a hidden PFV (x as opposed to X in the diagram); many languages consequently use the RTR in both functions (e.g., French Passé Composé). The IE aorist, in this way, merged with the perfect in prehistoric Latin; the sigmatic morphology of the aorist is to be found, as a result, in many of the Latin perfects (e.g., scripsi “I have written”). The two different functions of the Latin perfect (PFV and RTR) are easily distinguished in Latin texts by the primary (with RTR function) and secondary (with PFV function) sequence of tenses.17

The second kind of binary tense system found in IE languages is exemplified by English, shown in schematic form in Figure 17.5, where the unmarked forms of the indicative are normally used to represent complete events, whereas a marked form, the English Progressive, is typically used for the representation of events in progress. As will be shown below, the English Simple Forms are not Perfectives, and the term Performative (Hewson and Bubenik, 1997, p. 12 and passim) is used, along with its abbreviation PFM, for this different kind of Completive form, just as Progressive (abbreviation PRG) is used for the different kind of Incompletive form found in English.

From the above diagrams it can be seen that in Figure 17.4 the Incompletive forms of the Greek verb (Imperfective) are unmarked, whereas the Completive forms of the verb are marked (Perfective). The forms in Figure 17.5 present a mirror image to Figure 17.4: the Completive forms are unmarked (Performative), and the Incompletive forms (Progressive) are marked, as in Table 17.3.

Table 17.3




Imperfective (Unmarked)

Progressive (Marked)


Perfective (Marked)

Performative (Unmarked)

(p. 515) The conclusions to be drawn are as follows:

  1. (i) The unmarked forms in Descending Time represent events as incomplete, in progress, whereas the unmarked forms in Ascending Time represent events as complete.

  2. (ii) In the IE groupings there are languages with two tenses in Descending Time, as in the Greek model in Figure 17.4. We shall call these languages Type A.

  3. (iii) There are also languages with two tenses in Ascending Time, as in the English model in Figure 17.5. We shall call these languages Type B.

  4. (iv) There are also systems that have tenses in both Ascending and Descending Time, where Perfectives and Progressives are still marked forms, but Imperfectives and Performatives may, in certain circumstances, be marked as well. Languages of this mixed typology will be called Type C.18

  5. (v) These are the building blocks for tense systems in the languages of the world.

The notion of system is of fundamental importance. A language is not an atomistic nomenclature, an unordered list, lacking in coherence, for things and events in the real world. The systems of a language are discernible in its phonology and morphology, and in the syntactic distribution of meaningful elements. Verbal systems have observable morphological contrasts that mark the meaningful contrasts of a cognitive system, and the recurrence, over and over again in the languages of the world, of the same building blocks, is an assurance of the universality of certain cognitive contrasts, such as complete versus incomplete, memorial versus non-memorial, marked (i.e., derived) versus unmarked (i.e., basic). Further observation reveals the universality of two different systemic ways of contrasting complete versus incomplete events,19 and consequently the cognitive reality of verbal representations in Descending and Ascending time, i.e., of time that works in the mind in the automatic operation of the Working Memory, and the mind that freely works in time, as it does when composing a sentence such as the one that I am just completing.

3.2. Perfective and Performative

The unmarked forms of the English tense system (3 sg writes, wrote; 3 sg talks, talked, as opposed to Prospective, Retrospective, and Progressive forms), typically called the Simple Present and Simple Past by the grammarians and also by Comrie (1976, p. 25 and passim), are sometimes labeled Perfectives by others (e.g., Smith, 1997, p. 69). Comrie (1976, p. 25) calls them Nonprogressives, presumably because of the difficulty of naming a category that has so many seemingly unusual functions. Bybee et al. criticize this label on the grounds that the Progressive is the evolved form, not the primary form (1994, p. 138). An even more cogent argument stems from the observation that there are similar forms in other Germanic languages, most of which have no corresponding Progressive, and consequently cannot have a Nonprogressive.

(p. 516) In Hewson and Bubenik (1997, p. 12) these forms were labelled Performatives, for a variety of reasons. First, they typically represent the complete performance of all phases of the event; with Activities, Achievements, and Accomplishments they represent a complete event, and in this function Performatives and Perfectives overlap. In the English Pluperfect, Perfective participles can replace the Performative of the Past tense, as in (1).


Past tense He woke up, went to the door, opened it, and saw the parcel on the mat.

Pluperfect He had woken up, gone to the door, opened it, and seen the parcel on the mat.

With stative verbs, however, where English uses Performatives (2a), French requires the use of Imperfectives, as in (2b). In this usage English (Type B) Progressives are not feasible; in Russian (Type A) Perfectives are not feasible, and Imperfectives must be used (2c). The usage of Type A (descending time, DT) is the mirror image of the usage of Type B (ascending time, AT).


a. Past tense I knew what he wanted. *I was knowing what he was wanting.

b. Past tense Je savais ce qu’il voulait. *J’ai su ce qu’il a voulu.

c. Past tense Ya znal shto on khotel. *Ya uznal shto on zakhotel.

The confusions of this situation reveal the importance of understanding how and why different systems operate differently to produce equivalent results: equally valid representations of the same situation, using forms that in other usages are contrastive.

Stative verbs are monophasal:20 it follows that stative verbs are necessarily phasally complete from the very first moment, since by definition a Performative represents “a complete performance of all phases of the event.” This usage consequently shows a categorical difference between the digital Performative and the analog Perfective: stative verbs, that are phasally complete from the very first instant, are normally temporally incomplete because they are stative, and consequently temporally ongoing. In je savais ce qu’il voulait “I knew what he wanted,” the states of knowing and wanting are represented as continuing in time, as temporally incomplete, but phasally complete.

Second, in English, and in other languages,21 Performatives are used expressly in performative function, where the utterance of the verb contitutes the performance of the event. As originally pointed out by Austin (1962) I resign is a resignation, I am resigning is not; I promise is a promise, whereas I am promising is not. There is in fact, a vast quantity of performative usage, of which the following is but a fragment: I give you my word of honor, I insist, With this ring I thee wed, I acquit you from all charges, I admire your steadfastness, I apologize for the error, I applaud your courage, I assure you I will not.

Third, performative usages taper off into the description of the performance, as in Comrie’s example (1985, p. 37): Red Rover crosses the finishing line, which is “taken to coincide with the event of Red Rover’s crossing the finishing line,” or Hirtle’s (2007, p. 76) example from Canadian ice hockey, “He shoots! He scores!” The instantaneous nature of these events may be a challenge for languages that do not have a (p. 517) Performative and have to use an Imperfective instead: it is notable that the Francophone ice-hockey commentators use nominals: “Le tir! Et le but!” (The shot! And the goal!).

Fourth, there are the descriptions that accompany, or even precede the performance, as in cookery demonstrations on television: “We take a bowl, and put in half a pound of flour.” The same usage is found in instructions in reply to enquiries: “You turn left at the next intersection.”

Finally, the only alternative that can be used in performative function is the Imperfective: the Perfective can never be used in this function. The performative function consequently distinguishes the category of Performative from that of Perfective. To translate Performatives such as I thank you, or I apologize, Imperfectives have to be used in Slavic languages (V. Bubenik, p. c.), and in Greek both ancient and modern: use of a Perfective here would mean “I will thank you; I will apologize.” The major differences of function between Perfectives and Performatives are summarized in Table 17.4. The major overlap of these two forms is that both are similarly used, with active verbs (activities and achievements), in narrative function. The contrast between Perfective and Imperfective, and the considerable overlap of Imperfective and Performative is also revealed by comparing the two completive forms with the functions of the two incompletive forms.

Table 17.4





Complete activities





Complete achievements










Performative function





Instant presents





Habitual function










(*) Progressives may be used with stative and habitual reference when the situation described is temporary, not permanent. He lives with his mother / He is living with his mother.

In the Slavic and Greek Type A systems (Descending Time) events are represented as either complete (PFV) or incomplete (IPFV) in time (i.e., an analog representation). In the English Type B system (Ascending Time) it is the phases of the event that are complete (PFM) or incomplete (PRG), regardless of the flow of time (i.e., a digital representation). The proof of this is that stative verbs, which are monophasal by Aktionsart (a state has only one phase, from the first moment onward) are considered complete events in Type B systems.

It is well known that Imperfectives and Progressives, both of which represent incomplete events, are quite different in their usage. It is not as well known, however, that there are two different completive forms, and much Performative usage is (p. 518) consequently described as “Perfective”; this is the status quo in Semitic languages, for example, as noted by Comrie (1976, p. 78):

In written Arabic, there are two sets of forms, traditionally referred to as aspects, tenses, or states, and distinguished either as Perfect and Imperfect, or as Perfective and Imperfective. Here the terms Perfective and Imperfective will be used, although the meanings of the terms are different from those used in Slavonic linguistics and elsewhere in this book, as will become apparent below. (emphasis added)

When we read therefore (Weninger, 2000, p. 91), “Biblical Hebrew, Syriac and Classical Arabic express the performative with verbal forms of the perfective aspect,” we are looking at a traditional terminology for Semitic languages that fails to distinguish between the two quasi universal forms of completive aspect. What is called “Perfective” or “Perfect” in Semitic is in fact unquestionably a Performative, as in Figure 17.6. Perfectives cannot be used in the performative function.

Typologically, Semitic languages have a verbal system of Type C, where there is a morphology of two forms, as in Classical Arabic (Comrie, 1976, p. 94): the incompletive form is an Imperfective (in Descending Time) and the completive is a Performative (in Ascending Time), a typology that is also found frequently in languages of the Niger-Congo phylum of central and southern Africa (Nurse, Rose, and Hewson, 2010). The Arabic system of two contrastive forms, as in Figure 17.6, is typical also of Hebrew, and of Semitic languages in general. Forms of the Type C system will be analyzed in detail in section 4.

As noted above, Comrie’s term for the Performative is Nonprogressive, which becomes a dysfunctional term when, as is so often the case, the contrasting incompletive is an Imperfective, and there is no Progressive. The only other term proposed is that of Welmers (1973, pp. 345–348), who, writing on African languages, observes that “In many languages there is a single construction which has explicit and exclusive reference to past action” (i.e., a Perfective), but “that some languages use a single construction to refer to past time for active verbs, and present time for stative verbs” (obviously a Performative), for which he suggests the term “factative” on the following ground: “the construction expresses the most obvious fact about the verb in question, which in the case of active verbs is that the action was observed or took place, but for stative verbs is that the action obtains at present.” There are two problems here: (i) there is inevitable confusion with the term factitive,22 and (ii) Progressives and Imperfectives may also simply express “the most obvious fact about the verb in question”; as noted by Comrie (1976, p. 122): “the Imperfective will (p. 519) always add some special nuance, for instance general factual meaning.” The term Factative offers no more descriptive adequacy than the term Nonprogressive. Performative, on the other hand, not only names a function for which a Perfective can never be used, but also represents the complete performance of all phases of an active event. In this way the term Performative covers all the functions that are listed in Table 4 above in the Performative column; the Perfective vs. Performative terminology then parallels the Imperfective vs. Progressive terminology for the two related incompletive forms.

3.3. The Allosemes of Verbal Forms

Phonemes are always represented in discourse by allophones of varying degrees of resemblance to each other. Different aspectual and temporal representations of verbal systems also have their allosemes, surface meanings with varying degrees of resemblance to each other, as exemplified by the varying functional meanings listed for the Performative aspect in Table 17.4 above.

There are also complex forms such as the Retrospective in Figure 17.5 above with dual usage (Perfective with subject [x] and Retrospective with subject [X]), a usage covered by two different aspectual forms (aorist [Perfective] and perfect [Retrospective]) in Greek. We also see, with considerable frequency, aspectual forms, such as Prospective and Retrospective, that are used to represent what in other languages are tense differences of Past and Future. In Section 4 we shall examine other ways of representing tense functions by aspectual forms, and a very rare case of tense distinctions being used in aspectual function in Kikuyu, a Bantu language of East Africa.

Just as the Performative, often an unmarked form, has a considerable array of different surface meanings or functions, the Imperfective, another frequently unmarked form, also has a great variety of usage. In grammars of French, for example, more than 30 different usages of the Imperfect may be found, all of them attributable to the fact that the systemic meaning can be reduced to the formula accompli + inaccompli = 1, as demonstrated by Valin (1964).23

There are also modal usages, not only of future tenses, which lend themselves to such usage because they necessarily represent imaginary time, but also of past tenses, which when used in conditional clauses can represent the impossible, the non-factual, as in English If I knew, I would tell you. The use of the conjunction If establishes the content of the whole clause as imaginary, i.e., non-memorial time. A past tense, representing memorial time, is therefore counter-factual in this context, a deliberate contradiction in terms, a representation of the impossible.24

4. The Vast Present

Just as the lexical development of a word can be seen in the sequencing of derivational forms, as in just > justly > unjustly, or believe > believable > unbelievable > unbelievably, in the same way one can see the grammatical development of a word (p. 520) as in the Latin verb forms in Table 17.1 above, where similar sequencing can be seen in lauda-v-er-a-t. Similar progressions can be observed in Child Language, in the staged development from the one-word, to the two-word, to the three-word sentence. The first verb forms of the child are also aspectual, as in the gone of the English-speaking child: it is normal for tense forms to develop later, often in the third year of life.25 The verbal system of English is consequently a staged system, with three discernible levels as in Figure 17.7.

The forms at Level 1 are simply aspectual, and by themselves have no tense. As pointed out by the writers on primary language acquisition studies (Brown, 1973; Fletcher 1979, 1985; Bloom et al., 1980; Atkinson, 1982) Aktionsart differences are clearly understood by child learners, who use the -ing forms with activity verbs (drinking, eating, laughing, etc.), stative verbs with the bare stem (know, want, see, need, etc.), and the accomplishment verbs somewhat later with a “past” form which is often a past participle rather than a regular past tense (broken, fell, took).

These forms then become the first finite verbs, with nominative subjects: I drawing, Paul want cookie, car broken.26 Bloom et al. (1980, pp. 406–407) see these as aspectual, not tense, forms because of the influence of the verbal Aktionsart, and report on similar findings by Antinucci and Miller (1976) for Italian, and Aksu-Koç (1988) for Turkish. Fletcher (1985, p. 120) raises the same question, noting that the child Sophie, his subject of study, first used the -ed inflection and then changed it to -en, which is only participial. In that case the aspects would be Imperfective (I drawing), Performative (Paul want), and Perfective/Retrospective (car broken).

Let us be clear what is going on here. These three different forms all have subjects, and the two participles are used without auxiliaries. These are finite forms: present time is represented by the Imperfective participle of active verbs, and the Performative stem of stative verbs. The Perfective/Retrospective past participle then becomes used to represent the past, simply because whatever is complete in the present must have taken place in the past. In short, these are aspectual forms, but with an added value: they represent not only Event Time, but also Universe Time: the event is represented in its relation to the experiential present, which is the speaker’s deictic centre. These are tensed forms, but in the simplest system of all, the (p. 521) Vast Present, there are no tense contrasts: no part of Universe Time is contrasted with any other. A significant percentage of the world’s languages have only one tense, the Vast Present.

The evidence of the forms without auxiliaries suggests that in English this is a learning stage, which occurs, in fact, not only in English, and not only in Indo-European languages.27 Semitic languages, as we have seen above (section 3.2) typically have two verbal paradigms, a Performative used for past reference, and an Imperfective for a present (see Comrie, 1976, p. 95). Niger-Congo languages, which in Africa lie to the south of the Sahara desert, also show many languages of Type C, with varying kinds of development of Type C typology, as in Figure 17.8, a diagram of the system of Ejagham, an analysis based on Watters (1981), taken from Nurse, Rose, and Hewson (2010).

The Imperfective of the Vast Present normally has a global sense,28 and this language has developed a Progressive to give a more concrete representation to the here-and-now. The Situative (a term developed by Africanists, see Rose et al., 2002) is an Imperfective with the accompli reduced to zero29. A Retrospective balanced by a Prospective completes a classically balanced system, a Type C system entirely forged from aspectual forms, each one representing a position in the Vast Present.

With the development of the tense contrast between Past and Non-Past, a third layer is added to the English verbal system, and the second level functions as a subjunctive, as in (a) in Table 17.5, with only the one form, as against the indicative in (b) in Table 17.5, which requires sequence of tenses, using the Past vs. Non-Past contrast.

Table 17.5

Quoted form

Reported form


I insist he be at the meeting

I insisted he be at the meeting


I know he is at the meeting

I knew he was at the meeting

One of the curiosities of this development is the evidence of a U-shaped learning curve, in the English child’s acquisition of past tense forms, consisting of three stages: (1) an early period of correct forms, (2) a period of over-generalization of regular forms, and (3) a final period of correct forms, as in adult language. Stemberger (1994, p. 161) correctly presents three developmental stages as follows: (p. 522)

Stage 1: correct fell varies with base-form fall

Stage 2: correct fell varies with regularized falled

Stage 3: correct fell

At Stage 1 there is no contrast between the two forms; they are in free variation, both with the same meaning: complete event (by Aktionsart), necessarily past, as in Man fall down! from a two year old child. The subsequent development of a tense contrast with regular marked forms will produce the hypercorrection in Stage 2, which is rectified once the new tense system has been mastered. This development presents a problem to the proponents of rules.

5. Contrastive Tenses

As noted in section 3.1, the majority of IE tense systems are binary, with just three groups having three or more tenses.30 Semitic languages, which occupy most of North Africa to the north of the Sahara, and run into Asia as far as Mesopotamia (present-day Iraq), typically have a Vast Present with representations of Ascending and Descending Time. Niger-Congo (NC) languages, on the other hand, which occupy most of Africa south of the Sahara, have a remarkable diversity. The West African NC languages rarely have tense contrasts (Nurse et al., 2010), exploiting aspectual forms in a Vast Present, in similar fashion to Semitic. But the Bantu sub-group of NC, which occupies much of Eastern and Southern Africa presents a quite extraordinary range of different tense systems, some of which will be illustrated in 5.1, 5.2, and 5.3 below.

5.1. The Contrast of Tense and Aspect in Swahili

The fundamental paradigms of Swahili verbal forms that are marked for tense and aspect are presented succinctly in Table 17.6, based on Nurse (1989, p. 295) and Nurse and Hinnebusch (1993, p. 706).

Table 17.6










“we run”

“we have run”

“we are running”

“if we run”






“we ran”









“we will run”




The present forms of /-kimbia/ ‘run’ are marked only for aspect; they cannot be marked for tense. The past and future are marked for tense, and cannot be marked for aspect. This fact may explain why the Present Performative, with its ambiguous formative /-a-/ (not used in any other form), is not found in all dialects, and consequently not used by large numbers of speakers.31

(p. 523) For aspectual forms of past and future, compound forms are created: the auxiliary /-kuwa/ ‘be’ carries the tense marker, followed by the appropriate Vast Present form in second position, as in Table 17.7. Both parts of the compound forms are finite, which is the norm in Niger-Congo, and also in Afro-Asiatic languages.

Table 17.7


tulikuwa tumekimbia

tulikuwa tunakimbia

tulikuwa tukikimbia

“we had run”

“we was running (then)”

“we ran (continuously)”


tutakuwa tumekimbia

tutakuwa tunakimbia

tutakuwa tukikimbia

“we will have run”

“we will be running”

“we will run (continuously)”

The past and future forms are only found in initial position in these compounds, and may only be followed by forms of the Vast Present.

In these paradigms, consequently, we can clearly discern, using the criteria outlined above, distinctions of both tense and aspect. The distribution of forms in the paradigm shows, for example, that /li-/ marks past tense, /ta-/ marks future tense, and present tense is unmarked, with the result that the prefixal position used for the tense marker may instead be occupied by one of the three aspect markers /me, na, ki/ to give three typical aspects, a Retrospective /-me-/, an Imperfective /-na-/, and a Situative /-ki-/. This post-subject position (immediately after the subject marker) may be filled by either a tense marker or an aspect marker, but not both.

These three aspects are also used with Past and Future, as in Table 17.7, using be as an auxiliary verb to carry the tense marker, and the main verb with the aspect marker. Since a verb phrase can only have only one tense,32 the main verb is necessarily marked only for aspect. If tense is marked in any of these compound forms, it is marked in the auxiliary, as in all the forms in Table 17.7. Both auxiliary and main verb in these constructions are finite verbs, each with prefixed subject /tu-/ representing first person plural ‘we’, and final /-a/ representing the positive mode, which contrasts with potential (marked with /-e/) and negative (marked with /-i/).

The fact that there is no sequence of tenses between the two forms in the compounds leads to the conclusion that there is no clash of tenses in any of these Swahili compounds. This stems from the fact that the second form is always an exponent of (p. 524) the Vast Present, from which both Past and Future are derived, and of which they are consequently hyponyms, so that whatever the tense of the first element of the compound, it is always covered by the over-arching tense of the second element, the Vast Present.

There are, consequently two different and successive representations of tense in Swahili, in the first of which the whole of Universe Time is represented as a Vast Present, with aspectual forms that are typical of Descending Time: Retrospective, Imperfective, and Situative. The second level comprises a simple binary contrast between a Performative Past which runs to its Omega (ω) moment representing the last moment of the Past, and a Performative Future which departs from its Alpha (α) moment representing the very first moment of the Future, as in Figure 17.9, the Present (which contains both ω and α moments) being the Vast Present of Level 1.

The typological differences between this system and the English system in Figure 17.7 reflect certain typological differences between Niger-Congo (NC) on the one hand, and Indo-European languages on the other: (i) the general (but not absolute) lack of participles in NC; (ii) the use in NC (and Afro-Asiatic) of two finite forms in compounds, the second one being typically an exponent of the Vast Present; (iii) the use of the Vast Present not as a Subjunctive, but as a vertically contrastive tense, the hypernym (or generic) from which the contrastive tenses are derived. In IE languages the Present is represented by the majority (75%) of the IE families as part of the Non-Past, and by the remainder as a horizontally contrastive tense which spatially separates Past from Future (Celtic, Italic, Baltic).

5.2. The Tense System of Ruhaya

One recognizes many of the same Swahili features in the much more complex system of Ruhaya (Muzale, 1998; Hewson et al., 2001), outlined in Figure 17.10: as in Swahili, Level 1 represents Descending Time, and Level 2 Ascending Time. In this figure the form (tugúra) that is unmarked is placed above the line of Descending Time, and the derived, marked forms are placed beneath it, so that Stage 1 is presented vertically rather than horizontally. It can be seen that /-ire/ is a suffix that marks a Perfective, and this Perfective is turned into a Retrospective by the pre-stem (p. 525) morpheme /áá/ which adds the sense of “already,” indicating that the Perfective is viewed from a later position, the required condition for a Retrospective. This same element /áá/ is also added to /ki/ which in Swahili forms marks the Situative. The combination of /áá/ “already” and /ki/ “continuous” creates a Persistive (PST): “we are still buying.” That this analysis is correct is clear from the data: the /áá/ element is frequently deleted from the negative, where the prior accompli no longer exists: “we are not still buying” indicates a break between the former buying and the present.

  1. (i) The Vast Present of Level 1 is represented as an unbroken line in Descending Time (arrow at the left), open-ended at each limit. The unmarked form tu-gúra, as the key form of the paradigm, is placed above the line and the marked (derived) aspectual forms are placed beneath it.

  2. (ii) The form tu-guz-îre has no segmental tense marker (immediately after the subject marker /tu-/) as do the forms at Level 2. The suffix-ire, from P(roto)-B(antu) /*-ile:/ is a typical Perfective marker also found occasionally in West African languages. It is not uncommon in Bantu to find pre-stem tense markers, and suffixed aspect markers (cf. Kikuyu data in 5.3). It is the Past form of the Vast Present and is often descriptively included, because of its function, with the tenses of Level 2.33 In spite of overlapping function, it is imperative to distinguish aspect from tense. In terms of its form and its distribution -ire is not a tense formative, and the form has no other TA marker.

  3. (iii) The /ki/ in tu-ki-áá-gura is cognate with Swahili /ki/: it represents the subject situated in initial position, ready to initiate the event. It consequently has a variety of functions: if-clauses, actions seen and heard, open-ended possibility of action, continuity of action (glossed “continuously” in Table 17.7 above). Schematically X = subject, x = prior accompli: the combination of Situative /ki/+ prior accompli (x) = Persistive.

  4. (p. 526) (iv) When /áá/ is added to tuguzîre the Perfective is represented as the prior accompli with the subject (X) occupying the result phase, creating the Retrospective tuááguzire.

  5. (v) The contrastive tenses of Level 2 have been represented as contrasting horizontally and dividing the line of Ascending Time into 4 contrastive positions. There is still, as in Swahili, a Past and a Future, but what were Alpha and Omega moments in Swahili have been developed into separate tenses in Ruhaya. Omega (ω), the last moment of the Past has been extended backwards to the last sleep, and Alpha (α), the first moment of the Future has been extended forwards to the next sleep. The Near Past in the diagram covers time earlier today, and the Near Future covers time later today, from now to the next sleep.

Ruhaya consequently has five distinctive tenses, the first of which, the Vast Present at Level 1 represents the whole of Universe Time, which is divided into four separate tense contrasts at Level 2: the four tenses of Level 2 are all hyponyms of the Vast Present. As in Swahili, the contrast between Present and Non-Present tenses is vertical, not horizontal. Compounds can likewise be made by using the tenses of Level 2 as auxiliaries for the aspectual forms of Level 1: tu-raa-ba tu-áá-guzire ‘we will have bought’; tu-ka-ba tu-áá-guzire ‘we had bought’.

5.3. The Tense System of Kikuyu

Kikuyu is similar to Ruhaya in that tense at Level 1 is a Vast Present, and Tense at Level 2 has four temporal domains traditionally described as a Near and Far Past, and a Near and Far Future.34 As elements of the system, these represent, as in Ruhaya, a Past, a Memorial Present, a Non-Memorial Present, and a Future. But unlike Ruhaya, which resembles Swahili in having Descending Time (DT) at Level 1 and Ascending Time (AT) at Level 2, Kikuyu has both AT and DT at both levels, so that Level 1, as in Figure 17.11 is already quite complex (data from Barlow, 1960, with further details—including tones—from Bennett, 1969).

  1. (p. 527) (i) Not all verbs have the /-ko-/ focus marker of the PFM; there is a small remnant subset of half a dozen stative verbs that have null marking. Barlow (1960, pp. 128–129) quotes the following: end-a ‘loves’; um-a ‘comes from’; haan-a ‘resembles’; igan-a ‘is of a quantity or size’; ereg-a ‘lasts’. These are all statives with a degree of permanence (comes from = is from).

  2. (ii) The only form unmarked for aspect is the Performative. The Imperfective that represents the whole of Universe Time and has a generic sense is marked by /aga/, and the Perfective by the same /irε/ element that is used in Ruhaya, and common elsewhere, and the RTR by /ete:/.

  3. (iii) The pre-stem element /ko/in both Imperfective and Performative is a locative-based focus marker, which emphasizes the accompli of the form in DT: ‘we were cooking (earlier today)’, and the completion of the event in the form in AT: ‘we cook, will cook (later today)’.

Only the Performative and the Imperfective have complete paradigms of the four contrastive tense spaces of Level 2, which parallel those found in Ruhaya, as in Figure 17.10 above. In Kikuyu the Perfective and Retrospective are only used in the Present and the two Past tenses, but this still gives a plethora of tensed forms, and there is also extensive compounding. The parallelism of Level 1 and Level 2 is shown in Figure 17.12: there are four aspectual contrasts in the forms of the Vast Present at Level 1, exploiting both AT and DT, and four temporal distinctions at Level 2, exploiting both AT and DT for a total of eight different paradigmatic tense forms at Level 2, given that AT and DT are two different representations of Universe Time.

  1. (p. 528) (i) In the tense contrasts the Past begins with yesterday, and the Future begins with tomorrow. The Memorial Present (traditionally Near Present) represents what is in today’s memory, and the Non-Memorial Present (traditionally Near Future) represents what is not yet in today’s memory. This terminology is used simply to emphasize the cognitive underpinnings of the system; it is not intended to replace the traditional terminology.

  2. (ii) It is normal with such systems to have usages that are stylistically creative or innovative. The two Past forms of the Kikuyu Performative, tw-ă-rúg-a and to-ráa-rúg-á, being largely redundant because the representation of past events is covered by use of the Perfective and Retrospective, are, in fact, employed functionally in a way that is quite extraordinary. These are tense forms that are used in aspectual functions, the Memorial Present to-ráa-rúg-á as a “Short Imperfect” and the so-called Far Past tw-ă-rúg-a as a “Short Perfect” (the terms are from Johnson, 1980). This curious phenomenon involves the natural boundaries between the tenses.35

  3. (iii) Systems that have contrastive tenses in both AT and DT frequently bring into focus the problems of overlapping categories, which may be compared with the problems of phonemic overlapping. This is a question that is beyond the scope of this presentation.

6. The Use of Modal Forms in Tense Function

Modal auxiliaries are typically used in Germanic languages to represent the future, and are found elsewhere with similar function because future time is imaginary, and exists only in the realm of the possible. For binary tense systems it is normal that the Non-Past covers both Present and Future, with the tense form covering the Present, and some aspectual element making possible the representation of the Future.

But there are verbal systems, in Niger-Congo, Tibeto-Burman, Amerindian, and elsewhere where a difference between Future and Non-Future is marked by modal contrasts of various kinds.36 In Doyayo, a Niger-Congo language spoken in North Cameroon, which has a simple Vast Present with Imperfective and Performative aspects, the future is represented by a High tone on the Subject marker (Wiering and Wiering, 1994), giving an Imperfective (immediate) and a Performative (remote) future, as in Figure 17.13.

Burmese, for its part, has markers [tε] “realis” and [mε] “irrealis” that are obligatory at the end of clauses to mark the status of the clause, with [mε] typically being used with the future (Romero, 2008, pp. 67–68), the obligatory marking suggesting the status of a tense, but the contrast is evidential rather than temporal. Smith et al. (p. 529) (2007, p. 47) also report a Future for Navaho: “The Future Mode is mainly a temporal location indicator, although it can also have a strong modal meaning.” The question is also discussed by Comrie (1985, pp. 39–40), apropos of Dyirbal, with the following conclusion: “… despite the terminology adopted for Dyirbal, which identifies the two tenses as present-past and future respectively, the distinction between them is more accurately described as one of mood, namely realis versus irrealis respectively.”

Evidentials37 are also found with temporal force in Amerindian: events marked as Attestive (witnessed by the speaker) and Suppositive (hearsay, or probable) in Algonkian languages (Amerindian) are necessarily past events, and such forms were typically treated as past tenses in the grammars of the early missionaries. In Mi’kmaq (formerly spelled Micmac), an Algonkian language of Eastern Canada, if the word is unmarked for evidentiality it will be understood to be a Present (data from Hewson and Francis, 1990): ewi’kiket ‘he writes, is writing’. Marked as Attestive (ewi’kikep ‘I witnessed his writing’) or Suppositive (ewi’kikes ‘I believe he wrote’), it will be understood to be a Past. And many verbs have a reduced stem, representing imaginary time, which is used with the Imperative (wi’kike ‘write!’), the If-Conjunct (wi’kikej ‘if he writes’), the Conditional (wi’kikes ‘he would write’), and also with the Future (wi’kiketew ‘he will write’), which has its own personal inflections, which include evidential elements, as shown by Inglis and Johnson (2002), who consequently conclude (2007, p. 256) that “modal suffixes (existential and evidential) are used on an irrealis stem creating a Future form, but not a future tense,” a conclusion stemming from an uneasiness about a Future tense in a language without any other tense contrasts.

(p. 530) This question of the possibility of a binary Future vs. Non-Future tense contrast deserves an extended study of its own so that the unanswered questions in its regard can be more fully examined.

7. Conclusion

The study of verbal forms in the twentieth century was marked by a variety of important developments, among which the following may be noted: (i) the discovery of the importance of aspect, and aspectual contrasts; (ii) the development of definitions that clearly distinguish tense from aspect; (iii) the development of the discipline of Linguistics as a separate body of knowledge, resulting in the foundation of departments of Linguistics, with students and specialists, and (iv) the development of travel and communications, which gave access to languages and dialects that had never before been recorded.

This essay has attempted to show that the basis of verbal forms is aspect, which involves the representation of Event Time, limited to the representation of the event itself. Tense, which involves the wider view of the unlimited extent of Universe Time, the time that contains the event, is a later development, an extrapolation from Event Time, that occupies a later or secondary level in verbal systems, and may involve a tertiary level in which a variety of systemic tense contrasts are developed.

The tensed forms of the Vast Present typically appear in Child Language when a verb is predicated of the sentence subject, at which point verbs that are incompletive (lexically and grammatically), are automatically understood as representing the experiential present.38 Verbs that are completive (lexically and grammatically), on the other hand, correspondingly represent Memorial Time, because whatever is complete in the Vast Present must necessarily have taken place in the past, in time that is coeval with memory. The development of tense, before there are any tense contrasts, is achieved in this way by predicating aspectual forms of the verb to the subject of the sentence.

Not all verbal systems develop a further, third level with tense contrasts that are either binary (typically Past versus Non-Past), ternary (Past, Present, Future), or even more complex. Examples at this level may be found of three variant types of contrastive tense systems: Type A, Descending Time only; Type B, Ascending Time only; and Type C, with representation of both Ascending and Descending Time. Types A (e.g., the Greek indicatives in Table 17.2) and B (e.g., the English indicatives in Figure 17.5) are common in the Indo-European phylum, whereas Type C systems in IE languages tend to be tertiary (e.g., the Latin indicatives in Table 17.1) rather than binary. Even more complex systems are found elsewhere, especially in the Bantu group of the Niger-Congo phylum, where the system of the Kikuyu group is so complex that two tense forms, being largely redundant for purposes of tense, are actually used in aspectual function.


Aksu-Koç, A. (1988). The acquisition of aspect and modality: The case of past reference in Turkish. Cambridge: Cambridge University Press.Find this resource:

    Antinucci, F., and Miller, R. (1976). How children talk about what happened. Journal of Child Language, 3, 69–89.Find this resource:

      Aristotle. (1932/1995). Poetics. Loeb Classical Library. Cambridge, MA: Harvard University Press.Find this resource:

        Aristotle. (1938). The categories. On interpretation. Loeb Classical Library. Cambridge, MA: Harvard University Press.Find this resource:

          Ashton, E. O. (1944/1993). Swahili grammar. Harlow: Longmans Group.Find this resource:

            Atkinson, M. (1982). Explanation in the study of child language development. Cambridge: Cambridge University Press.Find this resource:

              Austin, J. L. (1962). How to do things with words. Oxford: Clarendon Press.Find this resource:

                Barlow, A. R. (1960). Studies in Kikuyu grammar and idiom. Edinburgh: Blackwood.Find this resource:

                  Bennett, P. R. (1969). A comparative study of four Thagicu verbal systems. Ph.D. dissertation, SOAS, London.Find this resource:

                    Bloom, L. M., Lifter, K., and Hafitz, J. (1980). Semantics of verbs and the development of verb inflections in child language. Language, 56, 368–412.Find this resource:

                      Boadi, L. A. (2008). Akan as an aspectual language. In F. K. Ameka, M. E. Kropp Dakubu (eds.), Aspect and modality in Kwa languages. Amsterdam: John Benjamins.Find this resource:

                        Brown, R. (1973). A first language: the early stages. Cambridge, MA: Harvard University Press.Find this resource:

                          Bybee, J., Perkins, R., and Pagliuca, W. (1994.) The evolution of grammar. tense, aspect, and modality in the languages of the world. Chicago: University of Chicago Press.Find this resource:

                            Cohen, M. (1924). Le système du verbe sémitique et l’expression du temps. Paris: Ernest Leroux.Find this resource:

                              Comrie, B. (1976). Aspect. Cambridge: Cambridge University Press.Find this resource:

                                Comrie, B. (1985). Tense. Cambridge: Cambridge University Press.Find this resource:

                                  Contini-Morava, E. (1989). Discourse pragmatics and semantic categorization. The case of negation and tense-aspect with special reference to Swahili. Berlin: de Gruyter.Find this resource:

                                    (p. 534) Fletcher, P. (1979). The development of the verb phrase. In P. Fletcher and M. Garman (eds.), Language acquisition (pp. 260–284). Cambridge: Cambridge University Press.Find this resource:

                                      Fletcher, P. (1985). A child’s learning of English. Oxford: Blackwell.Find this resource:

                                        Guillaume, G. (1929). Temps et verbe. Paris: Champion. (Reprinted, 1965).Find this resource:

                                          Guillaume, G. (1933). Immanence et transcendance dans la catégorie du verbe: Esquisse d’une théorie psychologique de l’aspect. Journal de Psychologie, 355–368. Reprinted, 1964, in G. Guillaume, Langage et science du langage (pp. 46–58), Paris: Nizet; and 1969, in J.-C. Pariente (ed.), Essais sur le langage (pp. 207–225), Paris: Édition de Minuit.Find this resource:

                                            Guillaume, G. (1945). Architectonique du temps dans les langues classiques. Copenhagen: Munksgard. (Reprint, Paris: Champion, 1965.)Find this resource:

                                              Guillaume, G. (1990). Leçons de linguistique. Vol. 10. Edited by R. Valin, W. Hirtle, and A. Joly. Quebec: Presses de l’Université Laval; Lille: Presses Universitaires.Find this resource:

                                                Hewson, J. (2006). Le système verbal du grec ancien: Trois distinctions de temps, ou deux? Glotta, 82, 96–107.Find this resource:

                                                  Hewson, J. (2007). L’aspect situatif. In J. Bres and T. Ponchon (eds.), Actes du XIe Colloque de l’Association Internationale de la Psychomécanique du Langage. Limoges: Lambert-Lucas.Find this resource:

                                                    Hewson, J., and Bubenik, V. (1997). Tense and aspect in Indo-European: Theory, typology, and diachrony. Amsterdam: Benjamins.Find this resource:

                                                      Hewson, J., and Francis, B. (1990). The Micmac grammar of Father Pacifique. Memoir 7. Algonquian and Iroquoian Linguistics. Winnipeg: University of Manitoba.Find this resource:

                                                        Hewson, J., and Nurse, D. (2005). The relationship of tense and aspect in the Kikuyu verb. In J. Maniacky and K. Bostoen (eds.), Studies in African linguistics, with special focus on Bantu and Mande: Festschrift for Yvonne Bastin and Claire Grégoire (pp. 281–311). Brussels: Royal Museum for Central Africa.Find this resource:

                                                          Hewson, J., Nurse, D., and Muzale, H. (2001). Chronogenetic staging of tense in Ruhaya. Studies in African Linguistics, 29, 33–56.Find this resource:

                                                            Hirtle, W. H. (2007). Lessons on the English verb: No expression without representation. Montreal: McGill-Queen’s University Press.Find this resource:

                                                              Hjelmslev, L. (1935). La catégorie des cas. Aarhus: Universitetsforlaget.Find this resource:

                                                                Hoenigswald, H. (1960). Language change and linguistic reconstruction. Chicago: University of Chicago Press.Find this resource:

                                                                  Holt, J. (1943). Études d’aspect. Acta Jutlandica, 15(2).Find this resource:

                                                                    Inglis, S., and Johnson, E. (2002). The Mi’kmaq future: An analysis. In J. D. Nichols (ed.), Actes du 32e Congrès des Algonquinistes (pp. 249–257). Winnipeg: University of Manitoba.Find this resource:

                                                                      Jakobson, R. (1932). “Zur Struktur des russischen Verbums.” Charisteria V. Mathesio oblata. (Reprinted in Jakobson, 1984, pp. 1–14).Find this resource:

                                                                        Jakobson, R. (1936). “Beitrag zur allgemeinen Kasuslehre.” Travaux du Cercle Linguistique de Prague, 6, 240–288. (English translation in Jakobson, 1990).Find this resource:

                                                                          Jakobson, R. (1957/1984). Shifters, verbal categories, and the Russian verb. In R. Jakobson, Selected Writings II (pp. 130–147). The Hague: Mouton. Reprinted, 1984, in L. R. Waugh and M. Halle (eds.), Russian and Slavic grammar studies 1931–1981 (pp. 145–149), Berlin: Mouton.Find this resource:

                                                                            Jakobson, R. (1984). Russian and Slavic grammar studies 1931-1981. Berlin: de Gruyter.Find this resource:

                                                                              Jakobson, R. (1990). On language. Cambridge, MA: Harvard University Press.Find this resource:

                                                                                Jespersen, O. (1924). The philosophy of grammar. New York: Norton. (Reprinted, 1965, 1992.)Find this resource:

                                                                                  (p. 535) Johnson, M. R. (1980). A semantic description of temporal reference in the Kikuyu Verb. Studies in African Languages. 11, 269–320.Find this resource:

                                                                                    Kuryłowicz, J. (1964). The inflectional categories of Indo-European. Heidelberg: Carl Winter.Find this resource:

                                                                                      Muzale, H. R. T. (1998). A reconstruction of the Proto-Rutura tense/aspect system. Ph.D. dissertation, Memorial University.Find this resource:

                                                                                        Muzale, H. R. T. (Forthcoming.) A reconstruction of the Proto-Rutura tense/aspect system. Cologne: Rüdiger Köppe. Translated into Russian by V. I. Vasilyev as Vid I vremya v yazykax tara (Moscow: RAN Linguistic Institute, 2005).Find this resource:

                                                                                          Nurse, D. (1989). Change in tense and aspect: evidence from northeast coast Bantu languages. In I. Haïk and L. Tuller (eds.), Current approaches to African linguistics (pp. 277–279). Dordrecht: Foris.Find this resource:

                                                                                            Nurse, D., and Hinnebusch, T. (1993). Swahili and Sabaki: A linguistic history. Berkeley: University of California Press.Find this resource:

                                                                                              Nurse, D., Rose, S., and Hewson, J. (2010). Verbal categories in Niger-Congo languages. Retrieved August 30, 2010, from NICO website: www.mun.ca/linguistics/NICO/.

                                                                                              Penfield, W., and Roberts, L. (1959). Speech and brain mechanisms. Princeton: Princeton University Press.Find this resource:

                                                                                                Potapova, N. (1951). Le russe. Paris: Editions Sociales.Find this resource:

                                                                                                  Reichenbach, H. (1947). Elements of symbolic logic. New York: Free Press.Find this resource:

                                                                                                    Robins, R. H. (1951). Ancient and medieval grammatical theory in Europe. London: Bell.Find this resource:

                                                                                                      Romero, N. (2008). Aspect in Burmese: meaning and function. Amsterdam: Benjamins.Find this resource:

                                                                                                        Rose, S., Beaudoin-Lietz, C., and Nurse, D. (2002). A glossary of terms for Bantu verbal categories: With special emphasis on tense and aspect. Munich: LINCOM Europa.Find this resource:

                                                                                                          Saussure, F. de. (1916). Cours de linguistique générale. 5th ed., 1955, edited by Ch. Bally and A. Sechehaye. Paris: Payot.Find this resource:

                                                                                                            Smith, C. S. (1997). The parameter of aspect. 2nd ed. Dordrecht: Kluwer.Find this resource:

                                                                                                              Smith, C. S., Perkins, E. T., and Fernald, T. B. (2007). Time in Navaho: Direct and indirect interpretation. International Journal of American Linguistics, 73(1), 40–71.Find this resource:

                                                                                                                Stemberger, J. P. (1994). Rule-less morphology at the phonology-lexicon interface. In S. D. Lima, R. L. Corrigan, and G. K. Iverson (eds.), The reality of linguistic rules (pp. 147–170). (Studies in Language Companion Series, 26). Amsterdam: Benjamins.Find this resource:

                                                                                                                  Tulving, E., and Craik, F. I. M. (eds.). (2000). The Oxford handbook of memory. Oxford: Oxford University Press.Find this resource:

                                                                                                                    Valin, R. (1964). La méthode comparative en linguistique historique et en psychomécanique du langage. Cahiers de psychomécanique du langage (no. 6). Quebec: Presses de l’Université Laval.Find this resource:

                                                                                                                      Valin, R. (1994). L’envers des mots: Analyse psychomécanique du langage. Quebec: Presses de l’Université Laval.Find this resource:

                                                                                                                        Watters, J. (1981). A phonology and morphology of Ejagham with notes on dialect variation. Ph.D. dissertation, University of California at Los Angeles.Find this resource:

                                                                                                                          Weist, R. M., Pawlak, A., and Carapella, J. (2004). Syntactic-semantic interface in the acquisition of verb morphology. Journal of Child Language, 31, 31–60.Find this resource:

                                                                                                                            Welmers, W. E. (1973). African language structures. Berkeley: University of California Press.Find this resource:

                                                                                                                              Weninger, S. (2000). On performatives in Classical Ethiopic. Journal of Semitic Studies, 45, 91ff.Find this resource:

                                                                                                                                Wiering, E., and Wiering, M. (1994). The Doyayo language. University of Texas at Arlington: SIL.Find this resource:


                                                                                                                                  (1) . Viz., Aristotle, Poetics 20.9 (1932/1995, p. 77), On Interpretation 3 (1938, p. 119). To this day, one of the words for ‘verb’ in German is Zeitwort ‘time-word’.

                                                                                                                                  (2) . On markedness, see Andrews, this volume.

                                                                                                                                  (3) . On Aktionsart, see Filip, this volume.

                                                                                                                                  (4) . For a binary tense analysis (with three aspectual distinctions) of the Greek verb, and a discussion of the issues involved, see Hewson and Bubenik (1997, pp. 24ff.), and Hewson (2006).

                                                                                                                                  (5) . On compositionality, see Verkuyl, this volume.

                                                                                                                                  (6) . Hewson and Bubenik, 1997, p. 28; Hewson, 2006, p. 97.

                                                                                                                                  (7) . Potapova, 1951, pp. 158ff.

                                                                                                                                  (8) . Guillaume, 1945, p. 53.

                                                                                                                                  (9) . In Russian, although it is recognized that the future is represented aspectually, as in on pisat ‘he is writing’, on napisat ‘he will write’, it is still traditional to consider the compound (Imperfective) future on budet pisat’ ‘he will be writing’ as a Future tense. It is normal to teach the compound future first (Potapova, 1951, pp. 98ff.), as a regular future tense, before introducing the simple, Perfective future (pp. 163ff.). The situation is complicated by the fact that the lack of a copula has left the verb to be in Russian defective, lacking a full array of Imperfective forms, so that there is no corresponding Imperfective to the Perfective auxiliary budet above.

                                                                                                                                  (10) . Guillaume’s article, published in a special linguistic issue of the French Journal de Psychologie, is reprinted in Langage et science du language (1964, pp. 46–58).

                                                                                                                                  (11) . The choice of the term Retrospective to replace Perfect (Position E) is based on the balance with Prospective (Position A). Comrie comments (1976, p. 54), under the heading of Prospective aspect: “the perfect is retrospective, in that it establishes a relation between a state at one time and a situation at an earlier time,” and notes the symmetry of Prospective and Retrospective. Jespersen (1924, p. 269) also comments: “… it represents the present state as the outcome of past events, and may therefore be called a retrospective variety of the present.”

                                                                                                                                  (12) . My translation.

                                                                                                                                  (13) . See, for example Comrie’s insistence on the importance of distinguishing between a form and its implicature (1985, pp. 23ff.).

                                                                                                                                  (14) . Alan Baddeley’s Chapter “Short-Term and Working Memory” in The Oxford Handbook of Memory (Tulving and Craik, 2000, pp. 77–92) gives an admirable overview of a topic which is of prime importance to an understanding of the cognitive functions that are often represented in Tense-Aspect systems.

                                                                                                                                  (15) . My translation.

                                                                                                                                  (16) . Albanian, Armenian, Germanic, Hellenic, Hittite, Indic, Iranian, Slavic, Tocharian.

                                                                                                                                  (17) . On sequence of tenses, see Ogihara, this volume.

                                                                                                                                  (18) . Binary versions of Type C will be described in section 4 (The Vast Present). Otherwise Type C versions are typically complex.

                                                                                                                                  (19) . This contrast is, in fact not just verbal, but appears in nominal systems in the contrast between mass nouns and count nouns. English mass nouns such as hair, paper, may be turned into count nouns by the use of an article: a hair, a paper (marked forms), and count nouns often become mass nouns after prepositions: they went to school by bus. A mass noun provides an analog interior view (as does Descending Time), whereas a count noun presents an exterior digital view (as does Ascending Time).

                                                                                                                                  (20) . The term is from Hirtle (2007, pp. 87–89), and contrasts with metaphasal (Activities, Accomplishments, Achievements). In a stative verb each moment is identical to the previous moment, and to the next: there are no different phases, as there are in Activities, Accomplishments, and Achievements.

                                                                                                                                  (21) . Boadi’s description of tense and aspect in Akan, a Niger-Congo language of West Africa, describes an Habitual that is the unmarked form of the paradigm, and is also used with verbs meaning “suppose, beg, request, apologize,” etc., all “members of the subclass of predicates which, following some writers, we refer to here as performative” (2008, p. 20). This Habitual is clearly a Performative, a category which has not been fully and properly described or defined in the literature on tense and aspect, leaving uncertainties in the minds of researchers.

                                                                                                                                  (22) . Which is in most dictionaries, whereas factative is not.

                                                                                                                                  (23) . Which means that both accompli and inaccompli can have values varying between 0 and 1 (since each balances the other), allowing for a wide range of representations.

                                                                                                                                  (24) . A usage well beyond the scope of Reichenbach’s line of time.

                                                                                                                                  (25) . On primary language acquisition, see Wagner, this volume.

                                                                                                                                  (26) . In the active sense of car has broken, “The car broke down.”

                                                                                                                                  (27) . There is also evidence that the IE protolanguage had no past Imperfective, so that the only past form was the aorist (the Perfective), and the Imperfective was used to represent the present: tense functions were represented aspectually. See, for example, Kuryłowicz, 1964, p. 134; Comrie, 1976, pp. 83–84; Hewson and Bubenik, 1997, pp. 351ff.

                                                                                                                                  (28) . As pointed out long ago by Marcel Cohen (1924, p. 56), these Imperfectives were described by some European grammarians as “aorists,” a term that, when correctly used, means “Perfective.” The Greek Imperfective is horistos ‘having a horizon’. [<——X---] between accompli and inaccompli; the Greek Perfective, being totally accompli, [<———-X] has no internal boundary, is a-(h)oristos ‘having no horizon’.

                                                                                                                                  (29) . For the occurrence and function of the Situative in Indo-European languages, see Hewson (2007).

                                                                                                                                  (30) . The five forms of Gaelic and Romance (two Pasts, one Present, two Futures) may be dealt with as three tenses and two aspects (with the second aspect missing in the Present), or as five tenses, since the difference in the Present and the Past is based on different representations of Universe Time (DT vs. AT). It is argued in Hewson and Bubenik (1997, p. 321) that the patterning and distribution of forms suggest five tenses rather than three.

                                                                                                                                  (31) . Ashton (1944/1993, pp. 37–38) gives examples that show that this “Present Indefinite” has usages very similar to the English Simple Non-Past, as in “The cook says he wants some sugar.” Contini-Morava (1989) also has an extended exposition of the usages of the form with examples that confirm its status as a Performative.

                                                                                                                                  (32) . A single event cannot occupy two different spaces in Universe Time.

                                                                                                                                  (33) . Ernest R. Byarushengo’s original unpublished work on Ruhaya listing three past tenses (P1 tu-áá-gura (hodiernal), P2 tu-guz-îre (hesternal), P3 tú-ka-gura (pre-hesternal)—with emphasis added) was used by Comrie (1985, p. 28) to show how adverb usage (i.e., context) clearly demonstrates that P1 is used for situations earlier today, P2 for yesterday, and P3 for any time before yesterday. P2, however, is technically not a past tense: it has no tense marker: tuguzîre is a Vast Present Perfective, contrasting with tugúra, Vast Present Imperfective. What is at issue here is a small complexity of the interface of tense and aspect systems, for which the rigor of scientific method quite properly requires an explanation: here we have Reichenbach’s Line of Time clashing with the linguistic data, which informs us that tu-guz-îre is not a Past tense; it represents the recent past because it is a Perfective form of the Vast Present. It is markedness that reveals the system; function may entail the table-knife screwdriver (see section 2).

                                                                                                                                  (34) . For a different treatment of remoteness distinctions (“metric tense systems”), see Botne, this volume.

                                                                                                                                  (35) . For an extended explanation of the anomaly, see Hewson and Nurse (2005, pp. 302ff.). Performatives often have an inchoative sense: he sat down and ate indicates that he began to eat, and it is this sense of “he just began to cook” (in the omega moments of the past) that leads to to-ráa-rúg-á being translated “he is cooking,” the Short Imperfective. In similar fashion a genuine Past Performative may often function as a Retrospective, as in English I came [= have come] to get your signature (said by a person entering an office) so that forms such as tw-ă-rúg-a can be used for “he has just cooked, just finished cooking,” the Short Perfect (i.e., Retrospective).

                                                                                                                                  (36) . On the temporal values of verbal complexes containing modal verbs, see Depraetere, this volume.

                                                                                                                                  (37) . On evidentials, see de Haan, this volume.

                                                                                                                                  (38) . As in I drawing (Weist et al., 2004, p. 41).