21.4: Relative tense

As noted in the definitions we cited from Comrie and Bybee, tense systems typically specify location in time relative to the time of the current utterance (TU). This type of tense marking is called absolute tense. For certain tense markers, however, some other reference point is used, which must be determined by the context. This type of tense marking is called relative tense. Because absolute tense marking is anchored to the time of the current utterance, absolute tenses are deictic elements; relative tenses might be considered anaphoric rather than deictic. A Brazilian Portuguese example is presented in (16a).⁶

(16)    a. Quando você chegar,           eu  já         saí.
              when     2sg   arrive.fut.sbjv 1sg already leave.past
              ‘When you arrive, I will already have left.’ [Brazilian Portuguese]

  b. *When you arrive, I already left.

The simple past tense form saí ‘left’ would normally have past reference, but in this context it gets a relative tense interpretation, indicating that the event described in the main clause is located in the past relative to the time of the event described in the adverbial clause. So in this context a verb marked for past tense can refer to an event which is actually in the future relative to the time of the speech event (TU). As demonstrated in (16b), the literal English translation of this sentence is ungrammatical, because the simple past tense in English normally does not allow this kind of relative tense interpretation.

We will refer to the contextually determined reference point of a relative tense marker as the perspective time (PT).⁷ Absolute tense constrains the relationship between TT and TU, while relative tense constrains the relationship between TT and PT. In example (16a), the adverbial clause (‘When you arrive’) establishes the perspective time, which is understood to be in the future relative to the time of speaking. The past tense on the main verb saí ‘left’ gets a relative tense interpretation in this context, indicating that the topic time (i.e., the time about which an assertion is being made) is in the past relative to the perspective time.

The most likely interpretation for ex. (16a) is diagrammed in (17). Relative past tense imposes the constraint that TT < PT, but does not specify whether TT is before or after TU.The fact that TT is later than TU is a pragmatic inference; if the speaker had already left before the time of speaking, it would be more natural and informative to simply say ‘I have already left.’ (The relationship between TT and TSit is determined by the perfective aspect of the simple past form, as discussed in Chapter 20.)

(17)    

In Imbabura Quechua, main clause verbs have absolute tense reference.⁸ Most subordinate verbs use a distinct set of tense affixes which get a relative tense interpretation.⁹ In the following examples, the subordinate verb ‘live’ is marked for relative past, present or future tense according to whether it refers to a situation which existed before, during or after the situation named by the main verb, which determines the perspective time. Since the main verb is marked for past tense, the actual time referred to by the subordinate verb may have been before the time of the utterance even when it is marked for ‘future’ tense, as in (18c):

(18)    Imbabura Quechua (Peru; Cole 1982: 143)

a. [Marya Agatu-pi kawsa-j]-ta    kri-rka-ni.
     Mary  Agato-in  live-pres-acc believe-past-1subj
     ‘I believed that Mary was living (at that time) in Agato.’

b. [Marya Agatu-pi kawsa-shka]-ta kri-rka-ni.
     Mary  Agato-in  live-past-acc     believe-past-1subj
     ‘I believed that Mary had lived (at some previous time) in Agato.’

c. [Marya Agatu-pi kawsa-na]-ta kri-rka-ni.
     Mary  Agato-in  live-fut-acc    believe-past-1subj
     ‘I believed that Mary would (some day) live in Agato.’

Relative past tense is sometimes referred to as anterior tense, relative future as posterior tense, and relative present assimultaneoustense. Relative tense is most common in subordinate clauses, but is also found in main clauses in some languages (e.g., classical Arabic). Comrie (1985) points out that participles in many languages, including English and Latin, get a relative tense interpretation. Example (19) illustrates the simultaneous meaning of the English present participle (flying). Example (10) illustrates the posterior meaning of the Latin future participle: the event of crossing the river is described for a topic time which is in the future relative to the perspective time defined by the main clause (the time when he failed to send over the provisions). Example (21) illustrates the anterior meaning of the Latin past participle: the event of delaying is described for a topic time which is in the past relative to the perspective time defined by the main clause (the time when he orders them to give the signal).¹⁰

(19)    a. Last week passengers flying with Qantas were given free tickets.

          b. Next week passengers flying with Qantas will be given free tickets.

(20)    Trāiectūrus Rhēnum commeātum nōn trānsmīsit
          cross-fut.ptcp Rhine provisions neg send.over-past.pfv.3sg
          ‘Being about to cross the Rhine, he did not send over the provisions.’¹¹ 

(21)    Paululum commorātus,   sīgna     canere    iubet.
          little.bit   delay-pst.ptcp signal.pl to.sound order-pres.3sg
          ‘Having delayed a little while, he orders them to give the signal.

The English be going to construction is sometimes identified as marking posterior tense. It can express future time relative to a perspective time in the past, as in (22a), creating a “future in the past” meaning. It can express future time relative to some generic or habitual perspective time, which may be past or present relative to the time of speaking, as illustrated in (22b–c).

(22)    a. I was just going to tell you when you first came in, only you began about Castle Richmond.¹²
          b. John keeps saying that he is going to visit Paris some day.
          c. Dibber always did tell me Pat was going to study to be a doctor.¹³
          d. John is going to visit you very soon.

Comrie (1985) points out that if a relative tense is used in contexts where the perspective time is equivalent to the time of speaking, then its meaning is equivalent to the corresponding absolute tense. For example, the interpretation of the posterior tense in (22d) is equivalent to a simple future tense. English does not have a fully natural way of indicating “future in the future”. Comrie states that the closest equivalent would make use of the about to construction, which marks immediate future: he will be about to X.

21.4.1 Complex (“absolute-relative”) tense marking

The perspective time (PT) for relative tense markers like those discussed above is not grammatically specified but is determined by contextual features. However, Comrie points out that some languages do have tense forms which grammatically specify both the location of PT (relative to TU) and the location of TT (relative to PT). Comrie refers to such cases as “absolute-relative” tense marking; we will use the term complex tense.

The English Pluperfect construction (I had eaten) can be used to express “past in the past”, as illustrated in (23). In example (23a), the event of Sam reaching the base camp is asserted to be true at a topic time which is in the past relative to a perspective time in the past, which is defined by the preceding clause (the time when the speaker arrived there). In example (23b), the event of Einstein publishing a paper (in 1905) is asserted to be true at a topic time which is in the past relative to a perspective time in the past, i.e. the time at which he won the Nobel prize (1922).

(23)    a. I reached the base camp Tuesday afternoon; Sam had arrived the previous evening.

          b. Einstein was awarded the Nobel prize in 1922, for a paper that he had published in 1905.

Similarly, the Future Perfect construction (I will have eaten) can be used to express “past in the future”. In example (24a), the event of Sam reaching the base camp is asserted to be true at a topic time which is in the past relative to a perspective time in the future (the time when the speaker arrives there). Another complex tense, “future in the past”, is illustrated in (24b). This sentence asserts that the event of Einstein winning the Nobel prize (1922) was in the future relative to a perspective time in the past, i.e. the year in which he published four ground-breaking papers (1905).

(24)    a. I expect to reach the base camp on Tuesday afternoon; Sam will have arrived the previous evening.

          b. Einstein published four ground-breaking papers in 1905, including the one for which he would win the Nobel prize in 1922.

The relative positions of TT, PT and TU for the italicized verbs in examples (23b), (24a), and (24b) are shown in the diagrams in (25).

(25)    a. 

  b. 

  c. 

As we will see in Chapter 22, the Pluperfect and Future Perfect forms are ambiguous. In addition to the complex tense readings illustrated in (23–24), they can also be used to indicate perfect aspect. In this chapter we consider only their tense functions.¹⁴

Comrie (1985) points out that cross-linguistically, most forms which express complex tense meanings are morphologically complex; that is, they involve combinations of two or more morphemes, like the English Pluperfect and Future Perfect constructions. However, occasional exceptions to this generalization do exist, e.g. the mono-morphemic pluperfect –ara in literary Portuguese.

21.4.2 Sequence of tenses in indirect speech

The difference between direct vs. indirect speech is that direct speech purports to be an exact quotation of the speaker’s words, as in (26a), whereas indirect speech does not (26b).¹⁵

(26)    a. Yesterday Arthur told me, “I will meet you here again tomorrow.”                        [direct]

          b. Yesterday Arthur told me that he would meet me there again today.                    [indirect]

One of the most important differences between the two forms is seen in the use of the deictic elements. Deictics within the direct quote (26a) are anchored to the perspective of the original speaker (Arthur) and the time and place of the original speech event: I = Arthur; you = the addressee in the original speech event, who is also the speaker in the current, reporting event; here = place of the original speech event; tomorrow = the day following the original speech event; etc. Deictics within the indirect quote (26b) are anchored to the perspective of the speaker in the current, reporting event (= the addressee in the original speech event), and the time and place of the current speech event. So I shifts to he; you shifts to me; here shifts to there; tomorrow shifts to today, etc.

Notice that the tense of the verb also shifts: will meet in the direct quote (26a) becomes would meet in the indirect quote (26b). Since (absolute) tense is a deictic category, anchored to the time of speaking, this is hardly surprising. It would be natural to assume that this shift in tenses follows automatically from the shift in deictic reference point. This may in fact be the case in some languages, but in English and a number of other languages, the behavior of tense in indirect speech is more complex. (The same issues often arise in other types of finite complements, e.g. complements of verbs of thinking and knowing, in addition to verbs of saying.)

Comrie (1985) presents an interesting contrast between the use of tense in indirect speech in English vs. Russian. In Russian, the tense of the verb in indirect speech is identical to the tense in direct speech, i.e., the tense that was used by the original speaker in the original speech act. However, all of the other deictic elements shift to the perspective of the current speaker, just as they do in English. An example is presented in (27), reporting a speech act by John at some unspecified time in the past:¹⁶

(27)    a. Džon skazal: “Ja   ujdu        zavtra.”
              John said      1sg will.leave tomorrow
              John said, “I will leave tomorrow.”                                                        [direct]

  b. Džon skazal, čto     on  ujdet      na  sledujuščij den. 
      John said     comp 3sg will.leave on next          day
      John said that he would leave (lit: will leave) on the following day.           [indirect]

In other words, verbs in Russian indirect speech complements (and other finite complements) get relative tense marking: the reference point is not the current time of speaking, but the time of the reported speech event (or, more generally, the topic time of the main clause). English verbs behave differently in this regard. For example, in (26b) and the English translation of (27b), where the original speaker used a simple future tense (will leave), the form used in indirect speech is the complex “future in the past” tense (would leave). As noted above, this is what we would expect to happen due to the shift in the deictic reference point, from the time of the original speech event to the time of the current, reporting speech event. However, there are other contexts where this shift by itself cannot account for the English tense forms.

The examples in (28) suggest that the form of the complement verb depends on the tense of the matrix (main clause) verb. Assume that John’s actual words in both (28a) and (28b) use the present progressive form (I am studying). When the matrix verb occurs in the future tense, as in (28b), English seems to follow the same pattern as Russian: the tense of the complement verb in indirect speech is identical to the tense that would have been used by the original speaker. However, when the matrix verb occurs in the past tense, this is not always true: in (28a), for example, we see the past progressive form (was studying) instead of the present progressive (is studying).

(28)    a. Yesterday I asked John what he was doing, and he said that he was/*is studying.

          b. If I ask him the same thing tomorrow, he will say that he is/*will be studying

Some additional examples illustrating this contrast are presented below. One general pattern that emerges is that, when the complement clause contains an auxiliary verb, that auxiliary retains its original tense form if the matrix verb occurs in the future (b, b, b). However, if the matrix verb occurs in the past, the auxiliary is normally “back-shifted”, i.e., replaced by the corresponding past tense form, as seen in (a, a, b).¹⁷

(29)    a. Yesterday I invited John to go out for pizza, but he said that he had/*has just eaten.

          b. If you invite him for pizza tomorrow, he will say that he has/will have just eaten.  

(30)    (spoken in 1998):

            a. In 2008 Ebenezer will say, “I will get tenure in 2011.”

            b. In 2008 Ebenezer will say that he will get tenure in 2011.

(31)    (spoken in 1998):

            a. In 1987 Ebenezer said, “I will get tenure in 1992.”

            b. In 1987 Ebenezer said that he would/*will get tenure in 1992.

When the original, reported utterance contains a verb in the simple past tense, the original tense form is again retained if the matrix verb occurs in the future (32). This can result in a past tense form being used to describe an event which is in the future relative to the current time of speaking, as in (32b). Back-shifting of a simple past form is often optional when the matrix verb occurs in the past, as in (33).

(32)    (spoken in 1998):

            a. In 2008 Ebenezer will say, “I got tenure in 2004.”

            b. In 2008 Ebenezer will say that he got/*will get tenure in 2004.

(33)    (spoken in 1998):

            a. In 1987 Ebenezer said, “I got tenure in 1982.”

            b. In 1987 Ebenezer said that he got/had gotten tenure in 1982

There are certain other contexts where back-shifting appears to be optional as well. For example, if the matrix verb occurs in the past and the complement clause describes a situation which is still true at the current time of speaking, either past or present can often be used for the complement verb in place of the present tense used by the original speaker (34). However, even in this context back-shifting is sometimes obligatory, as illustrated in (35).

(34)    a. Yesterday the mayor revealed that he is/was terminally ill.

          b. Last week John told me that he likes/liked you.

          c. The ancient Babylonians did not know that the earth circles/circled the sun.

(35)    a. I knew you liked/*like her.

          b. This is John’s wife.
              — Yes, I thought he was/*is married.

The set of rules which determine the tense forms in indirect speech complements is traditionally referred to as the “sequence of tenses.” A full discussion of the sequence of tenses in English is beyond the scope of this chapter. Scholars disagree as to whether the sequence of tenses in English can be explained on semantic grounds. Some (e.g. Comrie 1985) argue that the rules are purely grammatical, and cannot be predicted from the semantic content of the tense forms. Others (e.g. Declerck 1991) argue that a semantic analysis is possible, though the rules would need to be fairly complex.

Our purpose in this section has been to show that verb forms in indirect speech complements may require special treatment: these verbs may not exhibit the same kind of relative tense marking found in other kinds of subordinate clauses within the same language, and the normal shift in deictic reference point may not explain the usage of the tenses. Finally, this is an area where even closely related languages can exhibit significant differences from each other.

---

⁶ Comrie (1985: 31). A reviewer notes that the Persian translation for this example would use exactly the same tenses as the Portuguese. Note that the Portuguese future subjunctive is homophonous with the infinitive paradigm for most verbs, including chegar; but the paradigms are distinct for certain irregular verbs, including ter ‘have’, haver ‘have’, ser ‘be’, estar ‘be’, querer ‘want’, trazer ‘bring’, ver ‘see’, vir ‘come’ (Jeff Shrum, p.c.).

⁷ This terminology follows Kiparsky (2002) and Bohnemeyer (2014).

⁸ Cole (1982), Comrie (1985: 61).

⁹ Verbs in relative clauses use the main-clause tense markers with absolute tense reference.

¹⁰ Example (21) comes from Sallust, Catilina 59; cited in Allen & Greenough (1903: §496). The past participle in Latin, as in English, normally has a passive meaning; but the verb meaning ‘delay’ in Latin is a deponent verb, meaning that passive morphology does not create a passive meaning.

¹¹ Suetonius; cited in Comrie (1985: 61).

¹² Anthony Trollope (1860), Castle Richmond; cited at: http://grammar.about.com/od/fh/g/Fut...n-The-Past.htm

¹³ John Fante, “Horselaugh on Dibber Lannon”; cited at: http://grammar.about.com/od/fh/g/Fut...n-The-Past.htm

¹⁴ As discussed in Chapter 22, the temporal adverbs used here ensure that only the complex tense readings are available.

¹⁵ Most languages probably make a distinction between direct vs. indirect speech, but in some languages the difference is quite subtle. A number of languages are reported to have an intermediate form, “semi-direct speech”, in which some but not all of the deictic elements (especially pronouns and/or agreement markers) shift their reference point.

¹⁶ Data from Comrie (1985: 109). The non-past tense used in these examples would be interpreted with future reference in this context.

¹⁷ Many of the examples in the remainder of this section are adapted from Declerck (1991)