364 8. Temporal domains and relative tenses: theoretical foundations I. Introduction In chapter 2 and elsewhere a number of concepts and definitions have been introduced. Since these are essential to a good understanding of what follows, we will start by summarizing them (and in some cases elaborating on them). 8.1 Full situation vs predicated situation 8.1.1 We have defined the full situation as the complete situation as it actualized in the past, has actualized (or has been actualizing) in the pre-pres- ent, is actualizing at t 0 or is expected to actualize in the post-present. The predicated situation is that part of the full situation (possibly all of it) about which a claim is made in the clause. As is clear from Ten minutes ago John was in the library (which does not exclude the possibility that John is still there), it is the predicated situation rather than the full situation that is located in time by the use of the tense form Ϫ see 2.12.1. 8.1.2 In connection with the ‘continuative W-interpretation’ (see 5.6) of the present perfect, we have distinguished further between the factual full situa- tion and the potential full situation (see 5.7). The factual full situation is that part of the full situation (possibly all of it) that leads up to and includes t 0 . The potential full situation is the full situation including that part, if any, that follows t 0 . Thus, in John has been sleeping for seven hours now, the factual full situation covers a seven-hour period reaching up to (and including) t 0 . 1 Since John may keep on sleeping after t 0 , there is also an idea of a potential full situation whose length is unpredictable at t 0 . However, the potential part of a full situation (referred to by a ‘continuative present perfect’ Ϫ see 5.6) is not linguistically referred to: it can only be inferred. Thus, the above sentence John has been sleeping for seven hours now only asserts the factual (part of the) full situation, and the duration adverbial indicates the length of this factual part only. That John will probably go on sleeping for some (short or long) time after t 0 is a pragmatic inference based on our knowledge that it would be an exceptional coincidence if John woke up at the precise moment when the sentence was uttered (i. e. at t 0 ). This distinction between a factual and a potential part of the full situation is not only necessary in connection with ‘continuative perfects’ but is also perti- nent to present tense sentences like John is sleeping, because the full situation is durative and has a potential part (since it can extend into the future) while 1. Since, in the present perfect, the situation time can lead up to t 0 but cannot include it (see 5.2.1), the predicated situation differs from the factual full situation in that it does not include t 0 . I. Introduction 365 the predicated situation, which coincides with t 0 , is punctual. The distinction may also apply to sentences in the past tense, provided the ‘implicature of discontinuation’ (ϭ the suggestion that the situation is over at t 0 Ϫ see 4.3) is cancelled, as in Five minutes ago John was in the kitchen Ϫ he is probably still there. In this case the full situation may possibly extend beyond t 0 and thus have a potential part. As far as reference to the post-present is concerned, there is no distinction to be made between ‘factual’ and ‘potential’ full situation because there is no part of the full situation that includes t 0 (and is therefore factual): the entire full situation lies in the post-present. In the present chapter we will not need to refer explicitly to the distinction factual vs potential, because only factual situations are relevant to the discus- sion. We will therefore simply speak of ‘full situation’. This term will always have to be interpreted as ‘factual (part of the) full situation’. (As a matter of fact, in this chapter, which is concerned with temporal domains and relative uses of tenses, the notion of ‘full situation’ will hardly ever need to be referred to. Much more important will be the notion of ‘predicated situation’, since relative tenses express a relation between a situation time (ϭ the time of a predicated situation) and another orientation time Ϫ see 8.1.3.) 8.1.3 The distinction between ‘full situation’ and ‘predicated situation’ leads to a further distinction between the ‘ time of the full situation’ and the ‘ time of the predicated situation’ (or situation time). It is the situation time rather than the time of the full situation that is located in time by the tense form used Ϫ see 2.12.1. (on the telephone) [You can come at three if you like.] I’ll be at home. (It is possible that the full situation of the speaker being at home has already started Ϫ i. e. the speaker is at home and the full situation of his being at home will still be continuing at three. However, the statement I’ll be at home does not represent the time of the predicated situation as coinciding with t 0 . It represents it as coinciding with the time indicated by at three.) 8.1.4 The predicated situation can only form a proper subpart of the full situation if it is representative of the latter, i. e. if the full situation is repre- sented as homogeneous. A situation is represented as homogeneous if the verb phrase representing it can be used not only to refer to the full situation but also to refer to parts of that situation (see 1.36.4). This is the case in each of the following examples: John said he was ill. (The time taken up by John’s utterance is much shorter than the time of the full situation of his being ill. Still, the sentence is fine, even on the reading in which the being ill is taken to be represented as T-simultaneous (i. e. as coinciding Ϫ see 2.17) with the saying, because he was ill represents the full situation as homogeneous, which means that the time of the predicated situation can be a 366 8. Temporal domains and relative tenses: theoretical foundations proper subinterval of the time of the full situation, and can therefore coincide with the time of saying.) John said that at that very moment he {was reading / *read} a book. (The time of saying is much shorter than the time of reading the book. The that-clause is fine in the progressive because be reading a book is a homogeneous VP. The sentence is ungrammatical with he read a book because this clause is nonhomogeneous. 2 In this case the time of the predicated situation (ϭ the situation time of he read a book) cannot coincide with the time of saying because it cannot be a proper subinterval of the time of the full situation of reading a book: this renders read a book unacceptable because reading a book takes up considerably more time than saying something.) The artist whistled while she worked. (Both clauses are homogeneous.) At 5 p.m. I was at home. (homogeneous) At the time I walked to work. (The sentence is interpreted as ‘at that time I was in the habit of walking to work’, which is a homogeneous situation.) At 5 p.m. I was walking. (atelic VP ϩ progressive aspect J homogeneous clause) At 5 p.m. I was walking home. (telic VP ϩ progressive aspect J homogeneous clause) When the full situation is not (represented as) homogeneous, the predicated situation cannot be a subpart of the full situation. (A proper subpart of a nonhomogeneous situation cannot be representative of the whole situation). Instead, when the actualization of a situation is represented as nonhomoge- neous, the predicated situation coincides with the full situation. The following examples represent the situation as heterogeneous (nonhomogeneous): I wrote two letters this morning. (The sentence can only refer to the situation as a whole, not to any portion of it.) John ran around the lake before breakfast. (id.) John was in his office until 8 p.m. (Idem: there is no portion of ‘being in one’s office until 8’ that is itself an instance of ‘being in one’s office until 8’. In this case it is the until-phrase that renders the sentence bounded and hence nonhomogeneous.) 8.1.5 The distinction between homogeneous and nonhomogeneous (hetero- geneous) situations coincides with the distinction between L-nonbounded (i. e. not linguistically represented as bounded) and L-bounded (ϭ linguistically represented as bounded) situations Ϫ see 1.44.2. As explained in 1.44.1, a situation is L-bounded if it is represented as reaching a terminal point. Thus, John ran a mile represents its situation as bounded because it combines a telic VP (implying an inherent terminal point) with nonprogressive (perfective) aspect: the situation is referred to as a whole (from beginning to end), and therefore as reaching the inherent terminal point. The sentence John will run 2. A clause representing a telic situation as bounded (by using a nonprogressive verb form) is automatically heterogeneous (ϭ nonhomogeneous) Ϫ see 1.45. I. Introduction 367 for two hours represents its (atelic) situation as L-bounded in the not-yet-fac- tual-at-t 0 world referred to because the duration adverbial involves reference to a terminal point and the nonprogressive aspect refers to the post-present situation as a whole. The reader should keep in mind that, because L-boundedness is not a char- acteristic of the situation referred to (which is something extralinguistic) but rather of a particular kind of linguistic representation of a situation, we have adopted the habit of applying the label ‘L-bounded’ not only to (actualizations of) situations that are represented as L-bounded by the clauses referring to them but also to clauses which represent (the actualization of) a situation as L-bounded Ϫ see 1.44.1. 8.2 Orientation time & situation time 8.2.1 We call any time functioning as the origin of a temporal relation ex- pressed by a tense form a time of orientation (or orientation time). As noted in 2.14, there are five types of orientation time: t 0 , situation times, (other- wise unspecified) orientation times ‘contained’ in an Adv-time, orientation times that are ‘implicit’ in the meaning of a temporal conjunction, and ‘unspec- ified’ orientation times. (a) In sentences like I left yesterday, the orientation time to which the time of the predicated situation is T-related is t 0 . Tenses that T-relate the time of a predicated situation directly to t 0 are called absolute tenses (see 2.44.) (b) In John said that Bill had left, the past tense form said locates the time of its predicated situation anterior to t 0 , whereas the past perfect form had left locates its situation time anterior to the time when John spoke. This shows how a situation time can itself be the orientation time to which another situation time is T-related. In fact, since the time of a predicated situation can always serve as the origin of a temporal relation, any situa- tion time is at least a potential orientation time. 3 John stayed in the kitchen after he had finished his dinner. (t 0 functions as the orientation time to which the situation time of John staying in the kitchen is repre- sented as anterior, and this situation time is the orientation time to which the situation time of John finishing his dinner is represented as anterior.) (c) A third type of orientation time is an (otherwise unspecified) time ‘con- tained in’ an Adv-time (ϭ a time indicated by a time adverbial), as in At 3. This is why in earlier work we referred to a situation time as a ‘situation-time of orienta- tion’ (abbreviated to ‘TO sit ’ in Declerck (1991) and to ‘STO’ in Declerck (1997) and Declerck & Reed (2001)). 368 8. Temporal domains and relative tenses: theoretical foundations that time Bill had (already) left the room. (In section 2.23.1 we went into the special way in which a time adverbial indicates an orientation time by ‘containing’ it in terms of either inclusion or coincidence.) (d) An orientation time may also be implicit in the semantic structure of a temporal conjunction. For example: I will be exhausted {before / when / by the time that} this job is finished. Here the situation time of the job being finished is represented as T-simulta- neous with the implicit orientation time in the temporal structure of before (which means ‘before the time that’ Ϫ see 14.2.1) or when (which means ‘at the time that’ Ϫ see 13.1.2) or by the time that (where the implicit orientation time is overtly lexicalized as the time). (e) When we use Bill had left in isolation, we interpret the time of the predi- cated situation as anterior to an ‘unspecified orientation time’ which is understood as being anterior to t 0 . An orientation time is unspecified if it is neither t 0 nor the time of a predicated situation nor a time contained in an Adv-time nor an orientation time that is ‘implicit’ in the semantic structure of a temporal conjunction. For a good understanding of Bill had left, the unspecified orientation time to which Bill’s leaving is T-anterior will have to be recoverable from the context. 8.2.2 As noted in 8.1.4, the situation time (ϭ the time of the predicated situa- tion) can only be a subinterval of the time of the full situation if the situation is represented as homogeneous. Thus, in At 7 p.m. John was still in the library the situation time is a punctual time coinciding with the punctual Adv-time (see 2.23) specified by at 7 p.m., but still implies that the full situation of John’s being in the library started before 7 p.m. It is therefore important to distinguish carefully between ‘the time of the predicated situation’ (ϭ situation time), which is a T-concept (i. e. a concept relevant to the description of the tense system), and ‘the time of the full situation’, which is a W-concept (i. e. a concept which has to do with the interpretation of situations and temporal relations in the actual world or in whatever alternative world is being referred to). A tense temporally locates a situation time and does not say anything about whether that situation time is the time of the full situation or not. 8.3 Temporal relations expressed by tenses: T-relations A tense form can relate a situation time to an orientation time in three dif- ferent ways: (a) T-simultaneity: the tense represents the situation time as simultaneous with the orientation time. It is in keeping with the distinction between ‘situation I. Introduction 369 time’ and ‘time of the full situation’ that T-simultaneity is defined as a relation of strict coincidence. (b) T-anteriority: the tense represents the situation time that is temporally ‘bound’ (see 2.49) as preceding the ‘binding’ orientation time in either of two ways. Either the situation time lies completely before the orientation time, as in [He claimed that] I had locked the door, or it begins before the orientation time and leads up to it (without including it), as in [By the time Gertie and I went to Iceland,] we had been friends for years. (c) T-posteriority: the tense represents the bound situation time as following the binding orientation time in either of two ways. Either the situation time lies completely after the binding orientation time, as in [I promised] I would do it the next day, or it begins immediately after the binding orientation time, as in [He said that] from then onwards he would call me Jim. We refer to a temporal relation expressed by a tense form as a tense relation (or T-relation). 8.4 Temporal relations not expressed by tenses: W-relations 8.4.1 A situation may be interpreted as simultaneous, anterior or posterior to another situation without this being expressed by a tense form. In that case the temporal relation in question is not a tense relation but simply a relation that is inferred (for reasons referred to in 8.4.2) to exist in the real world (or in whatever alternative world is being referred to). We refer to such temporal relations as world-relations (or W-relations). When they had first visited the house, they hadn’t noticed the wet patch on the ceiling of the kitchen. (The past perfect tenses each locate their situation time ante- rior to an (unspecified) past orientation time. The relation of W-simultaneity that is understood to exist between the two situations is not expressed by the tense forms.) John told me he wrote Jim a letter. (The past tense forms merely locate their situa- tion times in the past. The use of the nonprogressive form wrote means that the durative situation of John writing Jim a letter is referred to in its entirety. Since the VP is telic, this means that the writing is represented as finished Ϫ see 1.39.1Ϫ2. Pragmatic knowledge tells us that this interpretation requires that the writing should be understood as W-anterior to the telling.) The distinction between ‘T-’ and ‘W-’ will not only be applied to relations but also to temporal interpretations: a ‘ T-interpretation’ is a temporal reading which is based on the semantics of the chosen tense only; a ‘ W-interpretation’ also takes into account such elements as temporal adverbials, context, situation 370 8. Temporal domains and relative tenses: theoretical foundations of speaking, knowledge of the world and pragmatic principles such as the Gricean Maxims of conversation. 8.4.2 Unlike simultaneity expressed by a tense form (T-simultaneity), W-si- multaneity is not a semantic relation but a pragmatic one. In other words, while T-simultaneity forms part of the semantics of certain tenses, W-simulta- neity is to be inferred from various factors, including the context and/or prag- matic knowledge. It follows (a) that W-simultaneity is a question of pragmatic interpretation rather than linguistic representation only, and (b) that W-simul- taneity is a less rigid temporal relation than T-simultaneity: whereas the latter is equivalent to coincidence, what counts as W-simultaneity can, according to context, be coincidence, overlap or inclusion. In other words, two time in- tervals are W-simultaneous the moment they have one point of time in common (see 2.18.2). This is illustrated by the following: Jim was in London yesterday. I saw him there. The two sentences here use an absolute past tense form, each creating a domain of its own. The two past domains are interpreted as W-simultaneous with each other. In this case the most natural interpretation of the W-simultaneity is that of inclusion: the situation of Jim being in London temporally included the situation of my seeing him there. 8.5 The two ‘time-spheres’ In English, the use of tenses implies a division of time into two (rather than three) distinct ‘time-spheres’: the present time-sphere and the past time-sphere (see 2.33 for arguments supporting this claim). The present time-sphere is conceived of as a time span of indefinite length which includes the temporal zero-point (t 0 ). The past time-sphere is conceived of as a time span of indefi- nite length which lies wholly before t 0 and is disconnected from t 0 . The abso- lute tense locating a situation time in the past time-sphere is the past tense (preterite). (As explained in 2.41, an absolute tense is a tense that relates a situation time directly to t 0 and not to another orientation time.) I left the house around midnight. (Left locates its situation time in the past time- sphere, which is defined in direct relation to t 0 .) 8.6 Present time-sphere zones The zero-time (t 0 ) is taken to divide the present time-sphere into three time- zones (or zones). The portion of the present time-sphere that precedes t 0 (and leads up to it) is the ‘pre-present zone’; the portion that coincides with t 0 is the . interpretation of the W-simultaneity is that of inclusion: the situation of Jim being in London temporally included the situation of my seeing him there. 8.5 The two ‘time-spheres’ In English, the use of. this case the time of the predicated situation (ϭ the situation time of he read a book) cannot coincide with the time of saying because it cannot be a proper subinterval of the time of the full. char- acteristic of the situation referred to (which is something extralinguistic) but rather of a particular kind of linguistic representation of a situation, we have adopted the habit of applying the label