Searle on Descriptions

Citation preview

Searle on Descriptions Simon Blackburn Mind, New Series, Vol. 81, No. 323. (Jul., 1972), pp. 409-414. Stable URL: http://links.jstor.org/sici?sici=0026-4423%28197207%292%3A81%3A323%3C409%3ASOD%3E2.0.CO%3B2-O Mind is currently published by Oxford University Press.

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/oup.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is an independent not-for-profit organization dedicated to and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact [email protected].

http://www.jstor.org Sun Jun 17 04:43:37 2007

SEARLE ON DESCRIPTIONS1

INSpeech Acts2 J. R. Searle presents an attack on Russell's theory of definite descriptions. The purpose of this paper is to show that one argument which he uses to support this attack is not as strong as it might appear. The argument is supposed to support the charge that the theory " presents the propositional act of definite reference, when performed with definite descriptions, as equivalent to the illocutionary act of asserting a uniquely existential proposition " (p. 159). Now this charge is not entirely transparent. For if we consider an example of the propositional identity which forms the theory of descriptions: (p) The king of France is bald = (q) There is at least one man who is a king of France and there is at most one man who is a king of France and anyone who is a king of France is bald. it is quite obviously no part of the claim that the sentence on the right hand side can only be used assertively, i.e. to assert (q). We might, for example, say ' If (q) then any Royal barber must have an easy time ', and here no illocubionary act of asserting a uniquely existential proposition has been performed, although were the description satisfied we might well want to hold that in saying 'If the king of France is bald then any Royal barber must have an easy time ' the propositional act of definite reference is accomplished. I t is hard then to see how the identity confuses the two. However, Searle's principal argument may be considered without reference to this supposedly " fundamental objection " to the theory of descriptions. So taken, it is the argument that the theoryis faced with a dilemma. I t must, according to Searle, say one of two things in applying the analysis to the occurrence of definite descriptions in certain contexts, and each of them is false: " For example, either we must construe " Is the king of France bald?" as " There is one and only one thing which is a king of France. Is that thing bald?" or " Is there one and only one thing which is king of France and is that thing bald?" Symbolically, letting " t " be an illocutionary force indicator for assertions and " 2 3> . be an illocutionary force indicator for questions and letting square brackets indicate the scope of the illocutionary force indicator, we have a choice between:

1 I wish to thank E. J. Craig and J. J. Altham for their very helpful comments on this paper. Cambridge University Press, 1969, pp. 157-162.

410

S. BLACKBURN

:

But both interpretations involve us in absurditieq. Consider a general application of the second alternative. Can we plausibly suppose that every questioner who uses a definite description is questioning the existence of the referent of the definite description?" ( p 161). However, the first alternative is no better: " We would regard i t as absurd to greet the command, " Take this to the queen of England ", with " What you say is true, she does exist ". The retort is absurd because the command is not an assertion, nor does i t contain an assertion " (p. 162). Clearly, to rebut this argument, it is necessary either to show that one of these alternatives is not as absurd as Searle claims, or to show that Russell can avail himself of other alternatives which Searle has overlooked. I shall follow Searle's order of discussion by starting with the alternative (2), which construes ' Is the king of France bald?' as asking whether (q), which I shall call the Russellian conjunction, is true. Searle's objection is that in asking whether the king of France is bald we are not always asking whether there is a king of France. To see whether it is really a consequence of the alternative (2) that this is what we would be doing, let us consider an absolutely certain analysis as it transforms in various contexts, such as questions and commands. Let us consider the concept of a l i ~ i d o w : ~ x is a widow

=

(a) x is a woman & (b) x was once married to some nlan & (c) that nlan died while married to x & (d) x has not since remarried

If Russell gave a conjunction which stands to ' The king of France is bald ' as the conjunction of (a) to (d) stands to ' x is a widow ', then he clearly gave an analysis of that proposition and of one use of definite descriptions. Now consider the question: ' Is x a \\idolir?' The analysis makes this the question: ' Does x satisfy the conjunction ((a) & (b) & (c) & (d))?' But then someone might argue that in asking this we must be asking inter alin whether x satisfies (a), i.e. whether x is a woman. But is it really plausible to suppose that in asking whether someone is a widow we are always asking whether that person is a woman? This might then appear to be an argument against the analysis. Again, consider the command: ' Ensure that she is a widow in the morning.' This, according to the analysis, must be the conlmand: ' Ensure that the person satisfies ((a)& (b) & (c) & (d))in the morning.' But is the wicked baron then commanding his henchmen to ensure that the person is a woman in the morning? I t appears that in 1 An example suggested in G. E. Moore, I'hilosopl~iccclP c ~ p e ~George s, Allen & Unwin, 1959, p. 183.

SEARLE O N DESCRIPTIOSS

411

issuing a sensible if deplorable command, he must also have issued a completely idiotic one, and this again might seem to be an argument against the analysis. What is wrong is quite evident if me consider finallythe performance of discussing whether x is a widow. For here i t is obvious that we can discuss whether sonieone satisfies a joint set of conditions without discussing whether they satisfy each member of that set. I n discussing whether x is a widow we need not discuss whether x is a woman precisely because we may assume it for the discussion. I n discussing whether x is a widow we are discussing whether ((a) & (b) & (c) & (d)) are together satisfied by x, but we may, in doing this, be assuming that some of them are satisfied by the person, and be discussing whether the others are. Similarly we can ask whether x satisfies a joint set of conditions, command that x be made to do so, state our intention of making it do so, and so on, while assuming and not therefore commanding, asking or intending the truth of some members of the set. Searle's stated argument against Russell taking the second alternative is therefore invalid. For i t is not a consequence of the theory of descriptions that in asking whether the king of France is bald I am always asking whether there is a king of France. It is a consequence only of the theory and the false assumption that in asking whether a set of propositions is true I must be asking of each member of that set whether i t is true.

I n asking ~ i ~ h e t h exr is a widow I may be assunling that x is a woman. This is quite consistent with holding that being a woman is part of the artalysis of being a widow; it must not be thought that an ' assumption ' is here another name for a presupposition in Strawson's sense. Similarly I am suggesting that in asking whether the king of France is bald we may be assunling that there is one, and that this is quite consistent with the proposition that there is one being part of the analysis of the affirmative answer to the question. Now we may make assumptions without indicating which assumptions we are making, or on the other hand we niay indicate which assumptions we are making. So suppose we admitted what seems very likely to be true, that in asking ' Is the king of France bald?' we are normally not only assuming that there is just one king of France, but also ilzdicatirtg our assumnption that there is just one king of France. Then if we use ' A ' to symbolize the illocutionary act of indicating an assumption, and ' ?[PI 3 d[Q] ' to symbolize that in asking whether P is true we are indicating our assumption that Q is true, then our view is a conjunction of Searle's second alternative : (2) ?[the f is g] = ?[(Ex)(fx. (y)(fy -t y = x) . gx)]

412

S. BLACKBURN :

and the view that normally Now the formal similarity of my (3) to the first alternative which Searle presents to Russell is obvious: indeed we have sinlply replaced his assertion-indicating sign with an assumption-indicating sign. Of course, I am further saying that this is not an alternative to (2)' but rather something that is normally true when we ask questions like ' I s the king of France bald?' even if in so doing we are asking the truth of the Russellian conjunction. Indeed, I should say in passing that a defender of Russell's theory would be quite wrong to accept any alternative to (2), such as Searle's (1). For Russell the question ' Is the king of France bald?' is the question ' Is it true that the king of France is bald?' (the denial of this identity being one of the less plausible consequences of rival theories) and Russell must suppose that this is the question ' I s i t true that there is a t least and a t most . . .?',on pain of having the sentence ' the king of France is bald ' ambiguous over its occurrence in the question and its occurrence in the affirmative answer: ' Yes, the king of France is bald.' Now Searle's argument against alternative (1)is that it is absurd to reply to questions and commands as though an assertion had been made. But this argument has no force a t all against (3), for i t is not a t all absurd to reply to questions and commands as though an assumption had been made. I n fact, if me thought the assumption false, that is exactly how we would reply. ' Is the king of France bald?'-' What do you mean?-You're assuming that there is one.' How is it then that Searle thought that ' t ' had to be the operator occurring in an alternative of the form of (I), and failed to consider the possibility which I have called (3)? I suggest that there is one good reason, namely the thought that if (3) is true, this in itself gives an argument against Russell's analysis.

An asymmetry between the analysis of ' widow ' and the Russellian theory could be urged as follows. It is true that in asking whether someone is a widow we are normally assuming that that person is a woman, and more often than not we would also be assuming that she was once married. Nevertheless we are not indicating which assumption we are making simply by posing the question ' Is x a widow?': i t is rather the context, in particular shared background knowledge, which will indicate what is being assumed. On the other hand, simply by asking the question ' Is the king of France bald?' in those words we indicate our assumption that there is just one king of France. I n short, in asking ' Is x a widow?' we may be assuming any or none of (a) . . . (d), but in asking ' Is the king of France bald?' we must be assuming that there is just one king of France. Then surely the fact that this must be assumed for

SEARLE ON DESCRIPTIONS

413

the question to be properly asked supports an account more like Strawson's than Russell's. However it is just not true that in asking a question using the definite description we must be assuming that there exists a referent and indicating that assumption. Imagine an historian researching into the possibility of Queen Elizabeth I having had a child, who ~ o u l dthen have been direct heir to the throne. He establishes that if she had a child she could only have had one and, on the internal evidence of the reign, it must have been Montmorency. Montmorency died in a tavern brawl. The historian sets this out, and then asks: ' Did the direct heir to the throne of England end his life in this squalid way? Perhaps we shall never know.' Here the historian has cancelled the normal assumption of existence, for in asking this question he is regarding as established the second and third of the Russellian conjuncts, and questioning the first. He is assuming that anyone who was direct heir to the throne died in this way, and that there was not more than one, and asking whether there was one. In asking this question at the end of his argument, he is not indicating an assumption of existence, for it is precisely that which he is questioning. So the ' assumption-indicating ' force of using a definite description in a question is something which can be cancelled, but perhaps only by an explicit indication that the other two members of the Russellian conjunction are regarded as established. I suggest that this is exactly the result we would expect given that we normally talk about things we know to exist. For, to take the vridow analogy again, we also have to have an explicit indication that (b), (c) and (d)are not in question before we can naturally use ' Is x a widow?' to question the sex of x ; and even then it would be a rather affected way to ask the question, because, of course, there is a much better way of asking it. But this is explained by the rarity of discussion of the widowhood of someone whose sex is unknown. Similarly, once the supposition that we must be indicating an assumption of existence by the use of a definite description is removed, (3), which states that we are normally doing so, is explained on a Russellian theory by pointing out that we normally discuss the properties of things known to exist. And in discourse we like an explicit indication of something abnormal, in the absence of which we suppose what is commanded, questioned, assumed, to be what is normally commanded, questioned, assumed. We are left then with no reason why a joint assumptioiz of (2) and (3) should not provide a perfectly satisfactory way of incorporating the theory of descriptions into the general theory of illocutionary acts. Of course, this is not supposed to be a comprehensive defence of the Russellian theory. One asymnletry with the analysis of ' widow ' which could be suggested is this. I t is clearly correct, even if misleading, to reply just ' No ' when asked if x is a widow when in fact x is a man, in a way in which it is not clearly correct to

414

6 . BLACKBURN : SEARLE O N DESCRIPTIONS

reply ' No ' when asked if the f is g when in fact there is 110 f. But discussion of this would obviously be very close to discussion of whether it's false that the f is g when there is no f. And this is a discussion which Searle explicitly allows us to settle as we like (p. 169). All I hope to have shown is that his argument adds nothing to its results.

Pembroke Collegr, Uwiversity of Oxford

SII~ON BLACKI~URN