22.3.1 The present perfect puzzle
As illustrated in (1c–d), the English present perfect cannot normally co-occur with adverbial phrases which name the time in the past when the event occurred; further evidence is provided in (8b).
(8) a. George left for Paris yesterday/last week.
b. George has left for Paris (*yesterday/*last week).
This constraint may seem puzzling, since the use of the present perfect clearly indicates that the described event took place in the past. Klein’s definition of perfect aspect as indicating that Time of Situation precedes Topic Time may offer at least a partial explanation.3
In English, the perfect can be combined with the various tenses to create the present perfect, past perfect, and future perfect forms: the present perfect combines present tense with perfect aspect, and so forth. Recall that tense indicates the position of Topic Time relative to the Time of Utterance; so in the present perfect, the Topic Time equals or includes the Time of Utterance. This helps to explain the “current relevance” constraint on the use of the present perfect: if the Topic Time is now, then in using the perfect to describe an event or situation in the past, we are actually “talking about” or making an assertion about the present moment.
Time adverbials like those in (1) and (8) generally modify the Topic Time. In the present perfect, the Topic Time is “now”; so time adverbials which locate the Topic Time in the past will be incompatible with the present perfect. The present perfect is, however, compatible with time adverbials which include the present moment, as illustrated in (9).
(9) a. I have now built hospitals on five continents.
b. I have interviewed ten students today/*yesterday.
c. I have built five hospitals this year/*last year.
The use of the perfect aspect constrains the Time of Situation by indicating that it precedes Topic Time, but it does not provide a precise location in time for the Time of Situation. The result is an “indefinite past” interpretation, which stands in contrast to the simple past form of the verb. The simple past tense indicates that Topic Time precedes the Time of Utterance (past tense) and contains the Time of Situation (perfective aspect). Topic Time must be identifiable by the hearer, and so will generally be specified, with whatever degree of precision is required, by some combination of adverbial phrases, contextual clues, etc.
Comrie (1976: 55) points out that past time adverbials actually can be used with the present perfect form of the verb in certain contexts, such as in non-finite clauses (10a–c), or in the presence of a modal auxiliary (10d–f).
(10) a. Having eaten a whole pizza last night, I skipped breakfast this morning.
b. Einstein’s having visited Princeton in 1921 eventually led to his permanent appointment there.
c. Charlie Chaplin was believed to have been born on April 16, 1889.
d. I should not have eaten a whole pizza last night.
e. Einstein must have visited Princeton in 1921.
f. Charlie Chaplin may have been born on April 16, 1889.
McCawley (1971: 101) observes that these environments have something in common: in these contexts past tense cannot be morphologically expressed by the normal past tense suffix -ed. This suggests that the perfect form (have + V-en) may have a different function in such contexts, namely as a marker of past time, i.e., an “allomorph” of past tense.
The acceptability of the time adverbials in (10) shows that the italicized verbs in these sentences do not have the interpretation normally associated with the present perfect form. But of course, it is also possible for a true perfect to occur with modals or in non-finite clauses, as illustrated in (11). So in these contexts the perfect form is ambiguous: it may either mark past tense, as in (10), or perfect aspect, as in (11). The two uses are distinguished by the interpretation of the time adverbials: if the time adverbs specify the time of the situation itself, as in (10), we are dealing with past tense.
(11) a. Having lived in Tokyo since 1965, I know the city fairly well.
b. Arthur was believed to have climbed Mt. Fuji four times.
c. Einstein must have visited Princeton several times before he emigrated to America.
The same ambiguity can be observed in the past perfect and future perfect as well. The examples in (12) involve true perfect aspect. The time adverbials shown in boldface in these examples refer to Topic Time, which precedes the time of speaking in the past perfect (12a) and follows the time of speaking in the future perfect (12b). In both cases, perfect aspect indicates that the Situation Time (the time when Mt. Fuji is climbed) occurs before Topic Time.
(12) a. In 1987, when I first met Arthur, he had (already) climbed Mt. Fuji four times.
b. Next Christmas, when you come to see me, I will have climbed Mt. Fuji four times.
The examples in (13) illustrate the use of the perfect form as a tense marker. In these examples the time adverbials shown in boldface refer to the time when the event actually took place. The perfect form is used to locate the situation prior to some perspective time which is different from the time of speaking. The result is a compound tense, as discussed in Chapter 21: “past in the past” in (13a), “past in the future” in (13b).
(13) a. Einstein was awarded the Nobel prize in 1922, for a paper that he had published in 1905.
b. I will reach Tokyo at 6:00 pm, but George will have arrived at noon.
22.3.2 Distinguishing perfect aspect vs. relative tense
There is a long tradition of regarding the aspectual vs. complex tense uses of the past perfect and future perfect forms as instances of polysemy.4 However, some authors disagree with this view. Klein (1994) for example argues that both the “perfect in the past” (12a) and the “past in the past” (13a) interpretations of the pluperfect (= past perfect) form can be assigned to a single basic sense: TSit<TT<TU. He states, “The notion of relative tense is not necessary to account for the Pluperfect nor for the Future Perfect” (1994: 131).
Bohnemeyer (2014) argues that perfect aspect does need to be distinguished from anterior (relative past) tense. The empirical basis for this claim is that some languages (e.g. Kalaallisut (=West Greenlandic) and Yucatec Maya) have a perfect aspect that cannot be used to express anterior tense, while other languages (e.g. Japanese, Kituba, and Korean) have anterior tenses that cannot be used to express perfect aspect. The critical diagnostic that Bohnemeyer uses is the interpretation of time adverbials. Time adverbials can be used with the perfect aspect in Kalaallisut and Yucatec Maya to specify a topic time before which the event had occurred, as illustrated in (14), but not to specify the time of the event itself, as in (15). The opposite pattern holds for the anterior tense forms in Japanese, Kituba, and Korean: these are compatible with time adverbials that specify the time of the event itself, as in (15), but not with time adverbials that specify a topic time before which the event had occurred, as in (14).
(14) Perfect aspect:
a. In 1912, when Theodore Roosevelt challenged William Howard Taft for the Republican nomination, both men had been elected President of the United States. Taft was now an unpopular incumbent, Roosevelt his beloved predecessor.
b. When you see me again next Christmas, I will have graduated from law school.
(15) Anterior tense:
a. Arthur’s theft of government documents was discovered on May 21st, but he had left the country on April 16th.
b. I expect to reach the base camp on Tuesday afternoon; Sam will have arrived the previous evening.
The crucial difference between perfect aspect vs. anterior (= relative past) tense is this: With relative past tense the time of the described situation can be specified precisely, as seen in (15), because TSit must overlap with Topic Time. With perfect aspect, however, the time of the described situation is generally not specified precisely; all we know is that TSit must be sometime prior to Topic Time, as illustrated in (14).
3 See Klein (1992) for a detailed discussion of this topic.
4 Jespersen (1924); Comrie (1976).
