Tense

Cathrine Fabricius-Hansen Preprint

1

Tense Tense is a grammatical category serving to locate situations in time; it is the basic grammatical category which, together with lexical and other indications of ordering in time, enables the hearer to reconstruct the temporal relation between the speech situation and the situation described in a sentence, and the relative order of situations described in a text. It is important to distinguish between tense forms, the rather abstract grammatical meanings expressed by such forms (semantics of tense), and the actual use and interpretation of tenses in natural discourse (pragmatics of tense). European languages have more or less elaborate tense systems while many non-European languages lack grammatical tense. In languages without tense temporal ordering is expressed by non-grammaticized means (e.g. temporal adverbials) alone; and the temporal location of the described situation is not necessarily indicated by linguistic means at all.

1. On tense marking Tense is usually expressed overtly on the (finite) verb, by way of morphological markers (cf. happen-s vs. happen-ed, sing vs. sang) or specific auxiliary words (will happen, has happened, had happened), depending on typological properties of the individual language. Auxiliaries in their turn are often functionally specialized tensecarrying verbs (cf. has, had) governing a non-finite (e.g. infinitival or participial) form of the main verb. Combinations of tense auxiliaries and non-finite main verbs are often termed periphrastic (or compound) tenses as opposed to non-periphrastic (one-word) tense forms. Non-finite verb forms by themselves (e.g. to happen, Tense. In: Encyclopedia of Language & Linguistics, Elsevier, p. 566-573

Cathrine Fabricius-Hansen Preprint

2

happening, the past participle happened) do not encode tenses in the stricter sense of finite tense. Tense languages vary with respect to how many tenses they distinguish formally and how their tenses are organized semantically. In general, tense forms that are unmarked from a formal point of view tend to encode the tense that is least marked from a semantic point of view. Thus, the present tense is zero-marked in most European languages; but in many languages outside Europe, zero-marked form express present or past time reference dependent on the type of situation (state or process versus completed event) described by the tensed verb.

2. The semantics of tense: basic principles From a semantic point of view, tenses are abstract meanings locating the time of the situation or eventuality (action, event, activity, process, state) described in a sentence – the event or situation time (E, mnemotechnic for ‘event’) – directly or indirectly with respect to some other time, often termed the time reference or the time of orientation (O). In the default case, it is the time of utterance (speech time, speaker’s now) (U) that serves as time of orientation. That makes tense an inherently deictic category, anchored in the unique utterance situation ('I - here - now'). Often, the term absolute is used about tenses that in their normal use are anchored directly in the time of utterance, i.e. the present (happens, is happening), the (simple) past (happened, was happening), and the future (will happen, will be happening) tenses in English. Tenses that in their normal use locate the situation time with respect to some other time of orientation than the time of utterance are traditionally called relative. Relative tenses are discussed in sect. 4, together with the so-called (present) perfect. This section is concerned with absolute tense and principles underlying the meaning of tense in general. Tense. In: Encyclopedia of Language & Linguistics, Elsevier, p. 566-573

Cathrine Fabricius-Hansen Preprint

3

2.1 Absolute tense (present, past, future) There are three basic ordering relations between E and U (or O) to be distinguished, corresponding to the three deictic notions of present, past, and future: the described situation may overlap with the time of utterance, it may precede the time of utterance, or it may be located after the time of utterance. These relations are encoded crosslinguistically by past, future and present tenses, respectively; cf. (1a-c). (1a)

the pope is visiting South-America

(1b)

the pope visited South-America a couple of years ago

(1c)

the pope will visit South-America next month

These three basic tenses, then, may be represented diagrammatically as in (1’), using the symbols o, < and > for the three basic relations. (1’a)

present:

situation time overlaps utterance time

EoU

(1’b)

past:

situation time precedes utterance time

EU

The definition of present tense above, while readily covering such sentences as (1a), might seem less appropriate to (1d) in its most usual interpretation, that is, indicating that Anna has the habit of reading Donald Duck. (1d)

Anna reads Donald Duck

Sentence (1d) can be true even if it is not the case that Anna is reading Donald Duck at the present moment. However, (1d) does not refer to a particular action of reading Donald Duck but rather to a property of Anna, namely that she habitually or regularly reads Donald Duck, and this property is claimed to be true of Anna even at those moments when she is not actually reading Donald Duck. The same applies in the case of sentences like (1e), with so-called universal time reference.

Tense. In: Encyclopedia of Language & Linguistics, Elsevier, p. 566-573

Cathrine Fabricius-Hansen Preprint

(1e)

4

water flows downhill

It should be noted that the past tense is adequately used in (2a) although John may still be in the garden at the time of utterance; thus, (2b) is a possible continuation. (2a)

half an our ago, John was in the garden

(2b)

in fact, he is still there

That is, E in the formulas of (1’) does not necessarily represent the time of the situation in its totality. With (atelic) activities, processes and states (like ‘be in the garden’), the ‘real’ situation may extend beyond E, which in that case represents just one temporal ‘slice’ of a larger situation of the same type; cf. (2a) and fig. 1. But with (telic) events like ‘leave the house’ which are defined by a natural endpoint, E must include that endpoint; in this case, then, E does represent the time of a completed situation; cf. (2c) and fig. 2. (2c)

half an our ago, John left the house

For the reasons mentioned above, what absolute tense locates with respect to the utterance time could be conceived as a temporal frame or ‘time under consideration’ which in its turn overlaps with the time of the described situation, rather than as the situation time itself. In what follows, then, it should be borne in mind that depending on the situation type and aspectual properties of the sentence, E may represent a complete situation (‘perfective viewpoint’) or a situation that is ‘cut out’ of larger situation of the same type (‘imperfective viewpoint’). We are here in an area where tense interacts with aspect (grammaticized ‘viewpoint’) and aktionsart or actionality (the internal temporal structure of situations). Thus, many languages distinguish formally between perfective and imperfective aspect within the past tense.

Tense. In: Encyclopedia of Language & Linguistics, Elsevier, p. 566-573

Cathrine Fabricius-Hansen Preprint

5

Instead of making a grammaticized three-way distinction between present, past and future tense, as in (1), many languages make only a two-way distinction between past (E < U) and non-past; this is the case in, for instance, Finnish. Whether there are languages that have the two-way opposition future versus non-future, is more controversial, and no languages have been cited with a two-way opposition present versus non-present (in which case the non-present would have to be discontinuous in its range). In general, future tense is less stable as a grammatical category than past tense; it is often expressed periphrastically. Many languages make additional grammaticized distinctions within the past or future based on the amount of time separating E and U, that is, the degree of remoteness between them.

For instance, in Cocama, a language of Peru, three

degrees of remoteness are distinguished in the past, as shown in (3). (3a)

ritama-ca

tuts-ui

town-to

go-PAST1

‘I went to town today’ (3b)

ritama-ca

tutsu-icuá

town-to

go-PAST2

‘I went to town yesterday / a few days ago’ (3c)

ritama-ca

tutsu-tsuri

town-to

go-PAST3

‘I went to town a long time ago’ Tenses characterized by different degrees of remoteness are rather rare in European languages but widespread among languages of Sub-Saharan Africa, Australia, New Guinea, and many parts of the Americas. A two-way distinction is the most

Tense. In: Encyclopedia of Language & Linguistics, Elsevier, p. 566-573

Cathrine Fabricius-Hansen Preprint

6

widespread though as many as five-way distinctions have been reliably reported for some languages. Some languages have divisions based on the daily cycle, in particular concerning the past, the most common cut-off point being today versus previous days (hodernial vs. pre-hodernial past); but mostly, the division is less specific (immediate vs. remote past or future). Another relevant parameter concerning the past or future is (in)definiteness. Some languages make a grammaticized distinction between definite and indefinite future, the former referring to a future event the location of which is given by the context while the latter is used when no specific future time is under consideration. And in English, among other languages, the choice between the simple past and the present perfect is partly governed by the same criterion: (4a) differs from (4b) by indicating that the speaker has a specific past occasion or time in mind (s. sect. 4.2). (4a)

Did you go to Paris?

(4b)

Have you been to Paris?

2.2 Focal versus non-focal tense meanings Viewed from a cross-linguistic semantic perspective, tenses can be conceived as prototypes or idealized conceptual categories – sometimes called gram-types – that tend to be grammaticized across languages but whose individual realizations – individual language-specific grams – may differ from one another in detail while sharing the same focal properties. For instance, the present tense normally establishes a relation of overlap between the event time and the utterance time, as witnessed e.g. in English and German alike; cf. (1a) and (5a). (5a)

das Klima

änder-t

sich

the climate

change-PRS.3.S

PRON.REFL

Tense. In: Encyclopedia of Language & Linguistics, Elsevier, p. 566-573

Cathrine Fabricius-Hansen Preprint

7

‘the climate is changing’ But German – and many other languages – may also use the present tense under conditions where the future tense could be used, too; cf. (5b). That is, the German present tense covers part of the semantic space that also falls under the periphrastic future tense and which may demand the future tense in English. (5b)

das Klima

änder-t

sich

bald

the climate

change-PRS.3.SG

PRON.REFL

soon

‘the climate will soon change / be changing’ However, this use does not represent the focal meaning of the present tense in German or any other language. In a case like (5b), it depends crucially on the co-occurrence of a future time adverbial (or other contextual information) indicating the location of E (definite future). In (5c), on the other hand, the future-like interpretation of the present tense hinges on the nature of the described situation. (5c)

mein

best-er Kollege

my

best-NOM.M

verläss-t

colleague

die Uni

leave-PRS.3.S the university

‘my best colleague is going to leave / is leaving the university’ The act of leaving, or the endpoint of that act, lies ahead of the utterance time; but the intention – and probably some preparatory activities – leading up to the final act will be taken to hold at the utterance time. Thus this kind of (immediate) future is situated between the focal areas of present and future tenses, i.e. situations evidently holding at the time utterance, on the one hand, and future situations that are predicted but not yet ‘initiated’ – by intention, decision, schedule – at the time of utterance, on the other hand; cf. (5d). (5d)

das Klima wird sich

änder-n

Tense. In: Encyclopedia of Language & Linguistics, Elsevier, p. 566-573

Cathrine Fabricius-Hansen Preprint

the climate will PRON.REFL

8

change-INF

‘the climate will change’ A comparison of (5a, b) and (5d) shows that in the absence of other indications of future reference, future tense is demanded in order to ensure the intended interpretation – in German as in English. It follows from the discussion above that competition between tenses – across languages and within a single language – is a natural phenomenon, to be expected in domains that, viewed from a cross-linguistic perspective, are not focal to the competing tenses. The sentence pair (6a, b) illustrates the same point with respect to another pair of tenses. (6a)

wir

wohn-en

seit

1990 in Berlin

we

live-PRS.3.P

since 1990 in Berlin

‘we have lived /have been living in Berlin since 1990’ (6b)

vi

ha-r

boet

i

Berlin siden 1990

we

have-PRS

lived

in

Berlin since 1990

‘we have lived / been living in Berlin since 1990’ The German sentence (6a) and the Danish sentence (6b) both state that ‘we’ live in Berlin at the utterance time and that the situation of our living in Berlin goes back into the past until 1990. (In fact, (6b) has a second, so-called existential or indefinite interpretation to the effect that ‘we’ lived in Berlin for a certain period located between 1990 and the utterance time, but not including the latter.) Reference to a situation holding at the utterance time calls for the present tense while reference to a time span extending backward from and including the utterance (or orientation) time – a so-called extended now – falls under the focal domain of ‘anterior’ or ‘perfect’

Tense. In: Encyclopedia of Language & Linguistics, Elsevier, p. 566-573

Cathrine Fabricius-Hansen Preprint

9

tense, i.e. the present perfect in English (see sect. 4.2). In this domain, where both tenses are motivated, German prefers the simple present, as illustrated in (6a) although the (present) perfect does occur, too; in Danish however – as in English – only the perfect is allowed; cf. (6b).

3. Tense in context: pragmatics of tense Tense is only one of several ways in which language can convey information concerning location in time; and the lexical resources in any language for expressing location in time are much richer than the grammatical devices. Thus, sentence-initial adverbials like in June 2004, yesterday, in two months provide more precise temporal frames for the situation time E than can apparently be done by the past or future tense alone. (7a)

in June 2004, Venus passed the sun

(7b)

yesterday, the weather was really bad

(7c)

in two months, my colleague will leave for Australia

In the examples of (7), the time adverbials restrict or specify the abstract past or future time reference established by the tenses; the tense itself may in fact seem quite redundant: as far as the location of E is concerned, it does not add anything to the information conveyed by the adverbial. But (8) – (9) demonstrate that, conversely, tense can help restrict the interpretation of time adverbials: Since on Sunday does not specify its relation to the utterance time, it is the finite tense in (8a, b) that indicates which Sunday is meant; and in (9), with this week referring to the week surrounding the utterance time, the tenses specify the location of E with respect to U within that week. (8a)

on Sunday, my colleague left for Australia

Tense. In: Encyclopedia of Language & Linguistics, Elsevier, p. 566-573

Cathrine Fabricius-Hansen Preprint

(8b)

on Sunday, my colleague will leave for Australia

(9a)

this week, the Beatles gave a concert

(9b)

this week, the Beatles are in town

(9b)

this week, the Beatles will give a concert

10

In other words, tense and frame-setting time adverbials cooperate in narrowing down the temporal location of E. Consequently, they should not point in different directions: (10), where a future time adverbial co-occurs with past tense, does not allow a coherent interpretation – at least not without further context (see sect. 5). (10)

tomorrow, Peter left for Glasgow

Often, however, adverbials or other lexical means are not needed in order to identify the time under consideration in sentences concerning past or future time; it may be deduced by pragmatic reasoning, drawing on the linguistic and non-linguistic context of the sentence. For example, a question like (11a) would normally be interpreted with respect to the same day or evening and not, for instance, the whole preceding week or the entire lifetime of the hearer; in (11b, c) E of the second sentence is understood as overlapping the situation time of the preceding sentence; and in narrative discourse we preferably take the relative order of sentences to mirror the order of the events described; cf. (11d). (11a) have you had dinner? (11b) (we spent the summer in Italy -) did you visit your sister? (11c) (we shall spend the summer in Italy -) will you visit your sister? (11d) he went to the phone and dialed zero; a woman’s voice answered

Tense. In: Encyclopedia of Language & Linguistics, Elsevier, p. 566-573

Cathrine Fabricius-Hansen Preprint

11

4. Relative tense and the (present) perfect Relative tenses (s. sect. 2.1) differ from absolute tenses by locating the situation time E not directly with respect to the utterance time (U) but as preceding or following a secondary time of orientation (O) which, in its turn, may precede or follow – or overlap with – the utterance time. Since the utterance time does play a role, albeit indirectly, ‘absolute-relative’ might be a more adequate term, as suggested by Comrie (1985). Compared to anterior – or perfect – tenses expressing the relation ‘situation time precedes time of orientation’ (cf. had happened, will have happened), grammaticized posterior tenses conversely locating the situation time after the time of orientation (would happen) are quite rare. Corresponding to their meaning, relative tenses are often periphrastic, or have developed from periphrastic forms, combining auxiliary verbs marked for absolute tense (cf. had, will have, would) with a (nonfinite) participial or infinitival of the main verb. 4.1 Anterior (‘relative past’) and posterior (‘relative future’) tense The so-called pluperfect, or past perfect, (had happened) and the future perfect (will have happened) belongs to the category of anterior tense tenses mentioned above, locating the described situation (E) prior to the time of orientation (O), which in its turn is located before or after the time of utterance, respectively; cf. the formulas in (12) and the corresponding examples (13a) and (13b); compare (1’) in sect. 2.1. (12a) past perfect/anterior: situation time precedes orientation time, and orientation time precedes utterance time

EU

In English, past posterior is encoded by would combined with the naked infinitive; it is used predominantly in indirect or reported speech (or thought), expressing future time relative to the (past) time of the reported utterance, as witnessed in (15a). (15a) (they said) they would start at noon Past posterior and anterior tense may even combine into tenses that depend on multiple orientation times. Thus, would have finished is the anterior corresponding to the past posterior would finish; cf. (15b). (15b) (they said that if they started at noon,) they would have finished the task by five o’clock in the afternoon Here, the event described as ‘start (at noon)’, which precedes U, serves as orientation time (O1) for the past posterior would + infinitive; the time denoted by the adverbial by five o’clock in the afternoon follows O1, – as demanded by the posterior – and serves as time of orientation time (O2) for the anterior (infinitive) have finished. In the end, then, E – the act of finishing the task – is located prior to O2, i.e., somewhere between noon and five o’clock. That is, the location of E is computed in three steps, as illustrated in fig. 5.

4.2 The (present) perfect From a formal point of view, the so-called (present) perfect in English (e.g. has happened) differs from the past perfect (had happened) merely by the present tense auxiliary (has vs. had). Correspondingly, as far as its semantics is concerned, the perfect can most adequately be understood as a relative (anterior) tense, differing from

Tense. In: Encyclopedia of Language & Linguistics, Elsevier, p. 566-573

Cathrine Fabricius-Hansen Preprint

14

the past and the future perfect with respect to the relation between O and U: in the default case, its time of orientation may neither precede nor follow the time of utterance, that is, the time of orientation has to overlap – or be identified with – the time of utterance; cf. (16a, b). Thus, the (present) perfect can be represented as in (17); compare (12) in sect. 4.1. (16a) my colleague has left for Australia (so I am alone on the job now) (16b) I have been waiting for too hours (so I leave now) (17) (present) perfect/ anterior: situation time precedes orientation time, and orientation time overlaps utterance time

E