W I L L I A M G. L Y C A N

ETERNAL

SENTENCES

AGAIN

(Received 16 October, 1973)

Quine has claimed that we need not posit language-independent propositions in order to explain how it is that the truth-values of sentences containing pronouns, indexicals, etc. fluctuate from context to context. He argues 1 that 'eternal sentences' will serve our purposes: ... there is no evident reason not to appeal simply to the eternal sentences themselves as truth vehicles. If we undertake to specify the proposition 'expressed' by the utterance of some non-eternal sentence, e.g. 'The door is open', on some particular occasion, we do so by bracketing some eternal sentence that means that proposition; thus we have had to compose an eternal sentence anyway, and we could just as well stop there. (p. 208)

(An 'eternal sentence" is one which is either always true or always false - such as 'It is raining in Columbus, Ohio, at 4:00 EST on October 9, 1973', as opposed to 'R is raining, here, now', which may be true or false depending on the circumstances under which it is tokened. 2) According to Quine, then, every non-eternal sentence may be paraphrased into (and regarded as a conversational surrogate for) the relevant eternal one, on any given occasion of its use. Charles Sayward has recently challenged this view,~ arguing first that "there is a certain sort of case in which it is quite plausible to think that no eternal sentence can be used to express what is asserted" (pp. 537-538), i.e., that not every natural sentence has an eternal sentence which captures what we intuitively take to be the sense or purport of the original. Suppose Smith asks Jones ' W h a t time is it ?', and Jones replies 'It is 4:30 p.m.'. Suppose this episode takes place in Lincoln, Nebraska, and that what Jones said is true. Now it seems unlikely that any eternal sentence could be used to express the proposition Jones expressed. For what could such a sentence be? It could not be 'It is 4:30 p.m., March 1, 1967, in Lincoln, Nebraska, at 4:30 p.m., March 1, 1967, in Lincoln, Nebraska.' For the proposition expressed by uttering this sentence would be a necessary truth, and

Jones' assertion was not a necessary truth. (p. 538) (Sayward goes on to turn aside an unpromising objection to this argument.) Philosophical Studies 26 (1974) 411-418. All Rights Reserved Copyright 9 1974 by D. Reidel Publishing Company, Dordrecht-Holland

412

WILLIAM

G. L Y C A N

I want to maintain that this case is not as troublesome as Sayward supposes. Let us grant that the eternal sentence for which 'It is 4: 30 p.m.' is a surrogate ought to be, in the case in which Jones' answer is correct. "It is 4:30p.m., March 1, 1967, in Lincoln, Nebraska, at 4:30p.m., March 1, 1967, in Lincoln, Nebraska,' or, for short, 'It is 4:30 at 4: 30'. I shall contend that there is a natural way of construing 'It is 4:30 at 4:30' according to which the sentence is contingent and non-trivial. Consider the sentence, 'The chairman is the chairman'. Although it looks like a trivial identity-statement, it bears a contingent and substantive construction: if we read the first occurrence of 'the chairman' more or less as a fused referring expression and the rest of the sentence as a predicate, the whole sentence will function in such a way as to pick out a particular person and assert that that person (by whatever name or however described) is the chairman (has the property of being chairman)4 - and this assertion, an ordinary singular predication, is plainly non-trivial. Thus one can say, 'It's lucky that the chairman is the chairman', without thereby expressing one's gratitude to God for having made the logical truths true. (One can also (coherently) wish that the chairman were not the chairman. And compare 'George IV wished to know whether the author of Waverley was the author of Waverley'; read in the way I am suggesting, this sentence says in effect that George IV wished to know whether Scott wrote Waverley, which assertion betrays no logical blindness on George IV's part.) The same way of interpreting apparent identity-statements can be applied to 'It is 4:30 at 4:30'. Let us construe the second occurrence of '4:30' as a fused referring expression, serving merely to pick out a particular moment (the moment at which the speaker happened to speak, under whatever designation); the whole sentence may then be read as a singular predication having the second occurrence of '4:30' as its subject. Thus, one who utters the sentence would be predicating something of the moment 4:30 (however it is referred to), viz., the property of its-havingbeen-4: 30-at that moment. Analogously, we can say without oddity, 'It's lucky that it was 4:30 at 4:30; if it had been 5:30 already, I'd have been in big trouble', 'I wished that it hadn't been 4:30 at 4:30', 'George IV wished to know whether it was 4:30 at 4: 30', etc. Just what property is expressed by the predicate 'It is 4:30 at... ?' We might be tempted to say it is the property of being (identical with) 4: 30;

ETERNAL SENTENCES AGAIN

413

but this would not help, since (I am inclined to say) the m o m e n t 4: 30 has the property of being identical with 4: 30 (itself) essentially and trivially. Rather, I think, the property of its-being-4: 30-at m o m e n t m is defined in terms of its relations to other times. E.g., we might say, to be an x such that it is 4: 30 at x is to be located (in time) four and one half hours after that m o m e n t at which the sun reaches its highest point in the sky - or whatever. I do not insist, o f course, on this particular analysis of a predicative reading of 'It is 4:30 at ...' But it seems clear that that phrase has a predicative reading in English, and (further) that the subject-predicate reading 1 a m proposing is the natural reading of Jones' utterance (it is 'what he intended to convey'), not merely a fanciful interpretation of the sentence imported just for the ad hoe purpose of rescuing the partisan of eternal sentences f r o m an objection. 5 Sayward himself agrees that his first example is less than conclusive, but he goes on to describe a second case which he finds more convincing: Suppose that Smith, after dozing on a bus, wakes up and sees that it is raining outside. A blind companion asks him what the weather is, and Smith replies 'It is raining'. Suppose further that Smith knows neither the time nor the place of the episode. (pp. 538-539) Let t and j be the actual time and place of the episode, respectively. Sayward argues against the Quinean view that Smith's utterance can accurately be paraphrased into the eternal 'It is raining at t and j ' . ... we cannot say.., that Smith asserted that it is raining at time t [and] place j. For, by hypothesis, Smith had no idea as to where it was raining nor what the time was when it was raining. This being so, he could not be said to have believed that it was raining at [time] t [and] place j. And if this is the case, then, since believing p is a necessary condition for honestly assertingp, and since Smith made an honest assertion, Smith did not assert that it was raining at time t [and] place j. (p. 539) As Sayward recognizes, the view of opaque contexts presupposed by this argument is open to dispute. But what I want to object here is that Sayward's conclusion ('Smith did not assert that...') has at best an oblique effect on anything Quine actually says on the subject, and certainly fails to show that 'It is raining at t and j ' is not an accurate paraphrase, in Quine's sense, of Smith's utterance. What, after all, is supposed to be the relation between a non-eternal sentence and the eternal sentence for which it allegedly goes proxy? Quine says that the relation is much the same as in the case of paraphrasing sentences into canonical notation... : the

414

W I L L I A M G. L Y C A N

eternal sentence will be one that the original speaker could have uttered in place of his original utterance in those original circumstances without detriment, so far as he could foresee, to the project he was bent on. (p. 208)

The technical notion of'paraphrase' is crucial to Quine's philosophy taken as a whole, and, although Quine never to my knowledge manages to make it precise, he says enough in various passages (see especially pp. 159ff., 182, 250, and 257-266) to indicate that the claim Sayward apparently attributes to him is a stronger one than he actually accepts. To say that a sentence S may be paraphrased by another sentence T is not to allege any equivalence-relation between the two; it is only to say that T will serve at least as well as S in the context or situation in question and relative to our theoretical aims at that time. (By way of illustration, consider Quine's remarks in the passages mentioned above on the paraphrasing of natural sentences into logical symbolism generally, on the elimination of singular terms in particular, and on the paraphrasing of number theory into set theory; and see w40 ('Propositions and Eternal Sentences') itself.) What are the 'projects' and 'aims' in service of which eternal sentences are to be provided? In Word and Object, Quine is primarily making suggestions about how to regiment ordinary factual sentences into a 'canonical notation for unified science', 6 an ideal language in which to state our serious philosophical]scientific account of everything. This canonical notation will have many technical advantages over present ordinary usage, and will be of great usefulness in advancing theory. But these advantages come only at a price: we cannot hope to preserve, in paraphrasing S into its counterpart, either the 'meaning' of the original in any philosophically interesting sense or the conversational nuances which S may ordinarily carry; we preserve the scientifically important roles that S fills, and forget about other of S's functions which may be more important to the man on the street. (Hence, the phrase 'without detriment" in the above quotation from Quine is used in a very special way.) The quotation itself, it might be objected, 7 is importantly at odds with the account I have given here of Quine's intent. For what matters, according to the quoted passage, is what "the original speaker could have uttered,., without detriment, so far as he could foresee, to the project he was bent on." There are two problems here: (1) Whatever Smith's project may have been (presumably, just the provision of a local weather report for the blind man), it had nothing to do with contributions to unified

ETERNAL SENTENCES AGAIN

415

science and the reduction of all human knowledge to physics. Besides, (2) even if such a contribution had been Smith's project, he still could not have used 'It is raining at t and j ' to this end, because he did not know that he was in fact located at t and j. Let us begin with the latter difficulty. Of course Smith, in the exact physical and epistemic circumstances hypothesized, could not have made an honest assertion by uttering 'It is raining at t a n d j ' . But Quine surely does not mean to suggest that he could have. In virtually no case of Quinean regimentation can a casual, non-technically-minded speaker of average intelligence make an honest assertion by uttering a promising canonical paraphrase of his original sentence, if for no other reason than that all canonical sentences are technical, being couched in the vernacular of some formal logical theory. E.g., a speaker who tells someone that 184 + 372 = 556 (normally) lacks the ability to assert that sentence's canonical paraphrase into set theory. The point of regimentation is that an ordinary speaker could substitute a canonical paraphrase for his utterance if he had the relevant (technical and topical) knowledge (the knowledge that is required for him to see that his paraphrase is an appropriate paraphrase) and if he were concerned to say something of use to unified science. In Sayward's ease, it is easy to see, Smith could honestly have asserted 'It is raining at t and j ' if he had had the knowledge that is relevant to shifting into Quine's technical idiom (viz., that he was located at t and j). This disposes of problem (2); but what about (1)? Again, we must not take Quine's words, in the passage at issue, too literally. That is, we must keep the 'projects' we are talking about within a fairly narrow range; to allow them to vary widely would be to trivialize the point of regimentation, much as it was trivialized by the excessively strict reading of'could' in problem (2) above. Consider a man whose purpose in uttering 'It is snowing here now' on a particular occasion is to activate a voice-triggered electronic door-opening device which responds only to the phonetic sequence exemplified by that remark. The relevant eternal paraphrase would of course not serve his purpose; nor would the set-theoretic translation of '2 + 2 = 4', if instead '2 + 2 = 4" were the sequence that activated the door-opener. It seems clear that, when Quine says 'purpose' or 'project', he means theoretical purpose or project. 8 This interpretation is borne out by other passages from Word and Object which

416

W I L L I A M G. L Y C A N

make the selection quoted above look as carelessly phrased as I have here suggested that it is: To implement an efficient algorithm of deduction is no more our concern, in these pages [w 33, 'Aims and Claims of Regimentation'] than was the implementation of communication... Each reduction that we make in the variety of constituent constructions needed in building the sentences of science is a simplification in the structure of the inclusive conceptual scheme of science. Each elimination of obscure constructions or notions that we manage to achieve, by paraphrase into more lucid elements, is a clarification of the conceptual scheme of science.., the broader framework shared by all the sciences. (10. 161) Writing is essential to serious science, as rendering it cumulative; and the longer the preservation, the dimmer the circumstances of utterance. Furthermore, the spirit of theoretical science encourages fixity of truth values also apart from the demands of writing. What is true here-now tends the more to be true also there-then, the more it is of the sort that scientists aspire to discover. Though scientific data go back to observation sentences, which are true only utterance by utterance, the sentences of the theory that is projected from those data tend to be eternal. (p. 227)

Once we understand 'project' in this narrowed fashion, it is easier to see Quine's point. His claim concerning eternal sentences is, by and large, just that non-eternal sentences can be and ought to be replaced by eternal ones in the ideal language of science. It does not follow from this that (as Sayward says) 'It is raining at t a n d j ' 'expresses the same proposition as', or even serves all the casual purposes of, 'It is raining' said by Smith; nor does Quine accept either of these latter claims. And it certainly does not follow that, if 'It is raining at t and j ' is an acceptable paraphrase (in Quine's sense) of Smith's utterance and if Smith's utterance was an honest assertion, then Smith must have honestly asserted that it was raining at t and j. Quine would certainly deny the validity of the schema T is an acceptable paraphrase of S Smith honestly asserted S Smith asserted T according to which anyone who (honestly) asserts a sentence of number theory thereby asserts the corresponding sentence in Frege's set theory, as well as the corresponding sentence in von Neumann's system, etc. (assuming that both of these corresponding sentences are acceptable paraphrases of the original number-theoretic sentence). 9 So Sayward's second argument, while its conclusion may be true, misses the point, if it is taken

ETERNAL SENTENCES AGAIN

417

to count against the relatively minimal view I have ascribed to Quine. One m o r e objection might still be made. Even given everything I have said so far, we are still hard put to say what fact Smith has c o m m u n i c a t e d to the blind m a n in uttering his non-eternal truth. The simple paraphrastic p r o g r a m we are considering suggests that Smith has c o m m u n i c a t e d (albeit in vulgar style) the fact that it is raining at t and j, and there is something unsatisfactory a b o u t this; the blind m a n would not be particularly interested in that fact. W h a t fact is it that the blind man, in the exact physical and epistemic circumstances hypothesized, comes to k n o w on the basis o f Smith's utterance ? The answer to this is fairly straightforward. W h a t the blind m a n comes to know, superficially speaking, is that it is raining in his vicinity; that is w h a t interests him. ' I n his vicinity' is admittedly an indexical term. But it is easily paraphrased away; we can replace the sentence that the blind m a n might say to h i m s e l f - 'It is raining in m y vicinity' - by ' I t is raining in the vicinity o f the blind m a n sitting next to Smith [or better, substitute the blind m a n ' s name here], at the time at which he wakes up on the bus on [insert the date]'. This is a sentence that Smith could have uttered without detriment to his communicative function, even t h o u g h it is t o o casual and loquacious for the purposes o f physics. 1~

The Ohio State University and Tufts University NOTES 1 W. V. O. Quine, Word and Object, Massachusetts Institute of Technology Press, 1960. 2 "Reports and predictions of specific single events are eternal.., when times, places, or persons concerned are objectively indicated rather than left to vary with the references of first names, incomplete descriptions, and indicator words" (19. 193-194). 3 Charles Sayward, 'Propositions and Eternal Sentences', Mind LXXVII (1968), 537-542. 4 What I am hinting at here is, of course, a general distinction between two different ways of reading definite descriptions; my suggestion is reminiscent of Donnellan's distinction between referential and attributive uses of descriptions ('Reference and Definite Descriptions', PhilosophicalReview LXXV (1966)), and of Kripke's distinction between rigid and non-rigid designators ('Naming and Necessity', in Davidson and Harman (eds.), Semantics of Natural Language, D. Reidel Publ. Co., Dordrecht-Holland, 1972). The subtleties of characterization that commentators have found to attend both these distinctions need not trouble us here - they do not affect the simple point I wish to make.

418

W I L L I A M G. L Y C A N

5 Richard Garner and Charles Kielkopf have pointed out to me that a referentially used singular term need not characterize its referent correctly; hence, someone might object, the same eternalized sentence could still be used to make different statements, and accordingly still might fluctuate in truth-value. But '4" 30', taken referentially, in fact denotes 4:30 in our language, and we know this; so, in saying 'It is 4:30 at 4:30' in our language, on the intended parsing, we know that we are talking about the moment 4:30, and the statement therein made will not fluctuate from speaker to speaker in our language. n See Gilbert Harman, 'Quine on Meaning and Existence, II', Review of Metaphysics XXI (1967-1968), 354. This was initially called to my attention by Sayward, in correspondence. s He cannot mean anything even so general as communicative purpose; what it takes to fulfill our communicative purposes varies too erratically with the psychological and epistemic makeup of one's audience. 9 In fact, we may raise doubts even about a similar schema involving the stronger relation of logical equivalence: T and S are logically equivalent Smith asserted S Smith asserted T. It is not obvious that, were Smith to assert 'Grass is green', he would thereby assert a much longer sentence which is truth-functionally equivalent to 'Grass is green', but which mentions snow, coal, and the sky. 10 Strictly speaking, this is not yet quite right, if the blind man did not know that Smith was named 'Smith' and Smith did not know the blind man's name. In such a case, Smith could substitute a n identifying description of the blind man and achieve his communicative purpose. N.b., if there were no sentence of this sort (an eternal sentence which communicates what Smith wants to communicate to the blind man), we would not be able to say, in an occasion-neutral way, what fact the blind m a n comes to know. We therefore could not specify a proposition that was putatively expressed by Smith's original words; and so it seems that (as our first quotation from Quine on p. 1 hints) the defender of propositions is no better off in a case of Sayward's type than the champion of eternal sentences.