CHAPTERS6 4 страница

Imperative sentences (and clauses), it will be recalled, are sen­tences (and clauses) in the imperative mood, which in many languages is in contrast with other moods, such as indicative, subjunctive, optative or dubitative (6.6). English, as we have also noted, has a relatively poor system of moods by compari­son with many, and perhaps most, of the world's languages.

Other kinds of non-declaratives 195

Imperative sentences, in English and other languages, cannot be put into correspondence with declarative (indicative) sen­tences as readily as can interrogative (indicative) sentences of the kind that were discussed in the preceding section, such as (44) and (45), which are here repeated and renumbered as (61) and (62), respectively.

'Is the door open?'

'The door is open'.

The reason for this is that mood is not independent of tense and aspect. Whereas (61) obviously has the same prepositional con­tent as (62) it is not obvious that

(63) 'Open the door!'

has the same prepositional content as the declarative sentence

(64) 'You open the door',

if (a) tense is held to be a part of the propositional content of a sentence and (b) what is traditionally regarded as the tense of (64) is given its most usual interpretation.

As far as condition (b) is concerned, it should be noted that, as the term 'tense' is traditionally used in the description of English, the grammatical category of tense is not clearly distinguished from that of aspect. As we shall see later, in many languages aspect is more important than tense (as tense is nowadays defined by linguists) and, in contrast with tense, what it expresses is definitely part of the propositional content of sen­tences (10.4). The major aspectual distinction grammaticalized in English is progressive (e.g. 'x is/was opening the door') ver­sus non-progressive (e.g., 'x opens/opened the door'). For present purposes, aspect is important in that 'open' belongs to a particular aspectual class of verbs — the majority in English - which do not normally occur in the simple (non-progressive) present tense with straightforward present-time reference. Moreover, from a semantic point of view it might be argued that the time-reference of a request or command made by uttering (63) is made implicitly, rather than explicitly, in the act of requesting or commanding; that (unless it is made explicit

192 Sentence-meaning and prepositional content

in mood) - may also be used for the expression of doubt without the intention of soliciting from the addressee the resolution of that doubt (or any other kind of response). English lexicalizes the expression of doubt in the verb 'wonder' (in one of its senses), which is commonly used either (a) as a verb of report with an indirect-discourse complement or (b) parenthetically with a first-person subject in a clause which is adjoined (paratactically rather than syntactically) to an interrogative sentence. These two possibilities are exemplified by

'x wondered whether the door was open'

'Is the door open, I wonder?',

respectively. An utterance of (55) by x might be subsequently reported to y by uttering (54) as a statement. But so too might be an utterance of the interrogative sentence 'Is the door open?' without the parenthetical clause 'I wonder', if y had reason to believe, in context (and this might be made clear prosodically or paralinguistically), that x was simply expressing doubt and not asking a question.

To be compared with both (54) and (55) is the declarative

(56) 'I wonder whether the door is open'.

This is syntactically parallel with (54) and can of course be used to make a statement. Much more frequently, however, such sen­tences are used, like (55), either directly to express doubt or indirectly to ask a question. According to whether an utterance of (56) is interpreted in one way or the other, it will be reported with (54) or

(57) 'x asked whether the door was open'.

Similarly, if v has reason to believe that x, in uttering (55), is indirectly asking a question rather than sirhply expressing doubt, it will be appropriate for y to report this by saying (57).

The upshot of this discussion - which could be extended by introducing into it direct-discourse constructions for comparison with both (55) and (56) - is that interrogativity and dubitativ-ity are closely related and, in default of any information, in the context of utterance, as to whether the speaker expects a

Other kinds of non-declaratives 193

response or not, may be ultimately indistinguishable. It is not surprising, therefore, to discover, first, that some languages do not grammaticalize the difference between them and, second, that, when they are grammaticalized, grammarians will argue as to whether it is interrogativity or dubitativity that is charac­teristically expressed by the utterance of sentences of a particular type or in a particular mood. It is perhaps only when semantic distinctions are lexicalized, rather than grammaticalized, that what is expressed is explicit enough for such arguments to be settled empirically. This point, as we shall see, applies in the ana­lysis of imperatives and other non-declaratives, as well as in the analysis of interrogative and dubitative sentences (or indeed of non-indicative declaratives).

6.8 OTHER KINDS OF NON-DECLARATIVES: IMPERATIVES, EX C L A M AT I V E S, VOLITIVES, ETC.

In this section we shall be concerned primarily with imperative and exclamative sentences (and clauses), which are the other principal classes of non-declaratives, in addition to inter-rogatives, that are distinguished grammatically in English. We shall also look briefly at volitives and at one or tw6 other classes of non-declaratives which are found in other languages.

Imperative and exclamative sentences are different from declaratives and interrogatives, and from one another, in several respects. But the same general point can be made about them as was made, in the preceding section, about declaratives and interrogatives: in addition to their propositional content, they also encode and grammaticalize (in those languages in which the relevant distinctions are indeed grammaticalized) some kind of non-propositional component of sentence-meaning. As declarative sentences grammaticalize their characteristic use for making statements and interrogative sentences grammatica­lize their characteristic use for asking (or posing) questions, so imperative sentences grammaticalize their characteristic use for issuing commands, requests, entreaties, etc., and exclamative sentences their characteristic use for uttering what are

196 Sentence-meaning and propositional content

by means of a temporal adverb or adverbial) its reference is to the future, immediate or less immediate as the case may be; and that the sentence itself is tenseless. In support of this view is the fact that in many languages in which tense is encoded inflectionally the imperative is clearly unmarked for tense. As to the inherently future reference of commands and requests (in normal circumstances), it is to be noted that, even if their temporal reference is made explicit by means of the word 'now' or the phrase 'at this very moment', it must be to a point or period of time that is later, if only infinitesimally, than the time of utterance. From this point of view it is interesting to consider a structurally ambiguous utterance such as

(65) I am telling you to open the door now,

in contrast with the non-ambiguous utterances,

I am now telling you to open the door
and

I am now telling you to open the door now.

Two points may be made in relation to this example. First, (65) can have the meaning of either (66) or (67). Second, in (67) the reference of 'now' differs according to whether it locates the act of telling or the anticipated act of opening the door in time.

There is the further point that the grammatical categories of mood and tense are undoubtedly interdependent in all languages that have both. And mood, whose function is usually if not always non-propositional, is far more common throughout the languages of the world than tense. Only a minority of the world's languages have tense as a grammati­cal category; and many of the functions of tense in those languages that have it are quite definitely non-propositional. I will come back to this point in Part 4.

Condition (a) is even more important, and more controver­sial. From the point of view of classical logic, propositions are eternally true or false, and therefore of their very nature tense-less. It is when propositions are treated as objects of mental acts or attitudes, on the one hand, or of such communicative acts as

Other kinds of non-declaratives 197

assertion and denial, on the other, that one is tempted to introduce tense into propositions themselves, anchoring them to the moment at which the mental or communicative act is performed. We shall not be able to deal with the problem of reconciling these two different views of propositions in the present book. It should be noted, however, that it is a problem that is all too often ignored in general treatments of tense, not only by linguists, but also by logicians. Since natural languages differ considerably as to how they grammaticalize and lexicalize indirect discourse, it is possible that different analyses are appropriate for different kinds of languages.

In fact, standard tense-logic, so called, is demonstrably inad­equate for the analysis of tense as it actually operates in tliose natural languages that have it. But richer and more powerful systems of tense-logic are now being developed by formal seman-ticists; and it may well be that these will prove to be more suit­able for the semantic analysis of tense in natural languages than currently available systems are. Whether they can successfully integrate the propositional (and purely temporal) and the non-propositional (modal and subjective) functions of tense is as yet uncertain.

But let us now return to imperative sentences without considering any further the question of tense. Imperative sentences constitute a subclass of sentences that are used, characteristically, to issue what are nowadays commonly called directives (commands, requests, prohibitions, etc.). For example, (63)

might be used by x to order or request y (or in the appropriate context to grant y permission) to perform a particular action. The effect of y 's compliance with this order or request would be

to bring about a state of affairs, or situation, in which the door, having been closed, is now open: i.e., to bring about a change in the world, in consequence of which the truth-conditions, not just of (62), but, more specifically, of

(68) ' y has opened the door'

and the truth-conditionally equivalent passive sentence

(69) 'The door has been opened by y,

 

 

198 Sentence-meaning and propositional content

are satisfied. It follows that, although imperative sentences, as such, may not have truth-conditions, they can be put into sys­tematic correspondence with declarative sentences that do. This being so, it is clearly possible in principle to bring impera­tive sentences within the scope of truth-conditional semantics; and various attempts have been made to do this. The question remains, however, as to what exactly is the propositional content of an imperative sentence.

If we adopt the methodological principle of saving the appearances for those languages in which there is a systematic and morphologically transparent relation between imperative and indicative sentences, we can say that, not only the impera­tive, but also the indicative, operates semantically upon the propositional content. This means that we can then say of (63) that it does indeed have the same propositional content as the declarative sentence (64) - but only when (64) is used to refer to a point rather than a period of time. Such uses of present-tense, non-progressive, sentences with verbs of the same aspec­tual class as 'open', though unusual in making straightforward descriptive statements, are quite normal in English, in the appropriate contexts, as we shall see when we look at so-called performative utterances in Part 4.

What has just been said about tense holds true of many natural-language phenomena. It is not difficult to demon­strate the inadequacy of current treatments of natural languages within the framework of standard propositional logic. Much of this chapter has been devoted to just that task. But my purpose throughout has been constructive. We learn more from a demon-strably inadequate, but precisely formulated, theory than we do from one that is so vaguely expressed that we do not even see its inadequacy. Let us bear this point in mind as we move on to consider some of the recent work in formal semantics.