A reply to Mr Alston

Nothing that has happened since J. L. Austin’s 1950s lectures ‘Words and Deeds’ or their publication¹ seems to me to have undermined the approach or made obsolete the kind of semantical theory typified by Frege or Russell or, in our own times, Carnap. In the broadest possible terms, however, to look at what a speaker does in saying what he says (a formula vague and inclusive enough to comprehend the very different researches both of Grice and of Austin) may enable us to place the philosophy of language in its principal proper context, the philosophy of mind; it may fill certain gaps and hiatuses left by the more formalistic approach which I follow these authors in espousing²; and it may put to rest metaphysical perplexities which maybe nothing else can relieve.³ It is a very pleasant discovery that, coming on the problem of ambiguity from opposite directions, Mr Alston and I can concur in the conviction ‘that for a certain word to have a certain meaning is for it to make a certain contribution to some appropriate semantic property of sentences in which it occurs’ and that ‘ a word is properly assigned as many senses as is necessary to account for the facts about the meanings of sentences in which it occurs’. I shall call this doctrine A. Our agreement in A helps me to believe that complementarity and mutual consistency, not the annihilation or unconditional surrender of the other side, should be the proper aim of both the factions to this dispute.⁴ There are three or four points in Mr Alston’s reply which I should like to take up, however.

(a) Of the semantical property of sentences to which we agree word-senses contribute Mr Alston says that the present writer wishes to explicate it ‘ in terms of the conditions under which an indicative sentence. . .will be true or factually licensed. [Wiggins] correctly anticipates that some will find this approach intolerably restrictive ’, and Alston says that he counts himself amongst their number, seeking as he does a theory according to which the ‘ saying potential ’ of statements would be merely ‘ a special case ’ of a ‘ more general notion ’ of saying and ‘ saying potential ’.

Evidently the most that Alston could make peace with me on is my ‘minimal contention ’ of page 22 of my paper, but I still believe that the indicative utterance is much more than a special case, and that the autonomy and syntactical and semantical completeness of the indicative mood, as opposed to the subjunctive, optative, and imperative, simply must confer theoretical primacy upon it. (The questioning mood, if mood it is, does match the indicative in completeness, but hardly in autonomy. There could be statements without provision for the possibility of questions perhaps. Could there be questions without provision for the possibility of answers to them?)

Leaving aside the power and theoretical simplicity of an indicative-oriented theory, if Alston still holds that performative utterances are an important obstacle to the construction of one then I must briefly resume the problem. The distinction I appealed to in footnote b on p. 21 between tense and aspect was not forced upon me by any sort of desperation. It is a commonplace of Russian grammar, and is in fact a pervasive feature of both English and Greek. But the self-fulfilling properties of performatives do require a little more comment.

Take ‘ It’s yours ’ said by Y to X to make X a present. Encouraged by the idea that the appropriate dimension of assessment for the utterance has less obviously to do with truth than with the efficacity of the uttering, we may try to spell out conditions for this in terms of Y’s seriousness, Y’s commitment, the thing’s being Y’s to give X in the first place, etc. Note that the prudence or imprudence or motivation or whatever else of the act will have no bearing at all on the particular ‘felicity’ or aptness which is in question. But once we have worked out the conditions for the efficacity which is our only concern here they may surprise the opponent of a contative analysis. The conditions are nothing other than the truth-conditions of the statement that the object is indeed X’s! If one wants to know why saying ‘ It’s yours ’ verifies ‘ it’s yours ’, however, he must look at the social and legal conventions governing the making of gifts. There is no further semantic problem.

I would add one question. If the sentence ‘ It’s yours ’ is not in the indicative mood then what mood is it in? The donatory mood? Are there then as many myriad moods as there are myriads of illocutions? This makes nonsense of the idea that a language is a system of communication with determinate semantics, and as Davidson would point out it completely confuses mood with mode.⁵ It is simply not worth denying that ‘ It’s yours ’ is in the indicative mood.

It is in virtue, then, of the conventions of ownership and transfer that to say ‘ It’s yours ’ makes it yours. But consider explicit performatives. Can there be a convention which decrees that to say that one ψs is to ψ ? How could a mere convention make it true that you ψ if you don’t in fact ψ . Consider the two following conversations:⁶

1.

Y: ‘ I assert that the persistence of the Conservative party is a monument to the impossibility of intellectual progress.’

X: ‘Go on then, assert it.’

Y: ‘ I just have.’

X: ‘No, all you’ve done so far is assert that you assert it.’

2.

Y: ‘ I promise I will be there.’

X: ‘Go on then. Promise!’

Y: ‘ I already have.’

X: ‘You haven’t. All you’ve done is assert that you do.’

Y: ‘All right. I will be there.’

X: ‘ But I want you to promise to be there, not just say that you will be! ’

Mad though all this is it will steer the second conversation back to a haven of sanity to extend it.

Y: ‘All right. I will be there. That’s a promise. Is that enough? I promise this. I will be there.’

for which Y might have substituted:

Y: ‘ All right. I promise that I will be there. Which is what I said in the first place!’

The one missing piece has now come to light in the shape of the paratactic analysis advanced by Davidson, and once anticipated by Meinong,⁷ of ‘ Y ψs that.’ We analyze it into two utterances of which the first, ‘ Y ψs this . . . . . .:’, makes a reference to the second which is an autonomous but subordinate utterance of the sentence‘ . . . . . . . .’.⁸ It is not convention but the logical form of ‘I promise that I will be there ’ or ‘ I state that the persistence of. . . ’ which makes each of them do the complex of distinct things it does do.

(b) I do not dissent from Alston’s general insistence on the ‘ ineradicably systemic character ’ of semantic hypotheses in general. And I am gratified that we can concur in many of the consequences which Alston ends his reply by deducing from A. But Alston sometimes talks and he often proceeds as if A were equivalent to what really is a quite different doctrine, which I shall call B. ‘ My justification for supposing that e.g. “run” has at least two meanings, operate and force, can be no greater than my justification for supposing that an adequate lexicon for English would include two such entries for “ run Now the view I took and still take is that this is at once too much and too little to ask. It is the wrong test in other words. It is too little to ask because even this justification would not show that no other adequate lexicon and grammar could make do with a single and more abstract entry for ‘ run ’. It is too much to ask because one may sometimes be able to show directly that no correct lexicon could get by without treating some word as ambiguous. (Consider ‘ bank ’ again and my sentence ‘ All banks in the U.S.A. are now guaranteed by the Central Reserve Bank ’, p. 28. Andl suggested other methods of bringing these matters to decision.) I think it is because Alston confuses B and A that, suspicious though he is, he is less suspicious than I should counsel of the bad old traditional method for suggesting differences of sense in a word w in contexts ‘____w____’ and ‘. . . . .w. . . . .’. To find a reading of w which makes the first sentence true and the second false if the reading is substituted for w in each is not enough. Nor is Alston correct in suggesting - if he does suggest it -that the validity of the old substitution method is anywhere involved in my demonstration that all unitary analyses of ‘good’ so far offered are defective. The method of substitution was used there to disprove an analysis, not to prove an ambiguity. My claim that we can understand or ‘hear’ the sentence ‘she has good legs’ in three different ways is itself independent of any substitution test.

(c) Alston identifies what I called directives with what Katz and Chomsky have called selection restrictions. This identification I very strongly deprecate. Katz and Chomsky’s notion of a selection restriction comes with a notion of a semantic theory, and a whole apparatus of ‘semantic markers’ and what not, which is expressly designed to do all sorts of things I am inclined to believe that no semantic theory should attempt to do: for instance rule out certain weird readings of sentences which are impossible, but which must in my opinion count as theoretically possible so far as semantics are concerned. The notion of a directive, on the other hand, was introduced in connexion with a phenomenon which is perfectly incontrovertible from whatever point of view. It was then generalized by me from the case of snub, quite vaguely and probably quite recklessly. Aristotle’s insight, for whatever it is worth, is not the creature of any particular linguistic methodology. For the time being I should wish to leave matters like that - in the air. And I should leave the notion of ‘directive’ tentative and imperfectly defined.

(d) Similarly, Alston uses the notion of a projection rule in the exposition of the theoretical framework within which our common doctrine A is to be applied. This again for my part I deprecate, since it would seem to commit both Alston and myself to a particular view, and one I myself reject, of what a semantic interpretation or reading is -, viz. a transcription into a universal conceptual notation or characteristica universalis. So far as I can see, projection rules have been intended at once

(i) to give the dictionary expansions by which some words need to be supplanted in order to display the deep structures, or (as I should prefer to have it) the logical form of the sentences where they occur,

(ii) to mark out and characterize homonymous words, and

(iii) to make translations into the characteristica universalis.

(iii) is an independent enterprise, whatever we think of it. And (i) and (ii) are separate again. My original paper would have done better to distinguish them, (i) is an important and special task. I give here one example of it.

Consider a term occurring in the manner which the scholastics called in sensu diviso, e.g. ‘ a sloop ’ as it occurs in ‘ I want a sloop ’, the sentence being so read that it does not imply ‘there is a sloop such that I want it’. If you see no hope, as I see no hope, of explaining the difference between the cases where such sentences do and do not imply their existential generalizations in terms of a difference of style of oc¬ currence of ‘ a sloop ’ - any old sloop versus a particular one, and the rest of that bad old business - then you must unfold the want we have here into want that one have⁹ and distinguish

(ꓱx)  (x is a sloop and I want that I have x)

from

I want (ꓱx) (x is a sloop and I have x).

To persons as puritanical as myself about referential matters the possibility of a verb’s generating the in sensu divisojin sensu composito ambiguity must always spell the necessity for lexical intervention. The thrust of this suggestion will escape nobody who has tried to characterize the logical form of intentional verbs which appear to take an accusative or direct object. (Nor will the difficulties of finding accurate two-part definitions which will bring out these referential ambiguities.)

This and like business under (i) may serve to divorce some homonyms, but not of course all. The ambiguity of bank, for instance, does not impinge on logical form. Such matters must wait on (ii). But (ii) again, like (i), though it is a new use for the dictionary, need never involve more than assigning English sentences English sentences as their readings. For the enterprise to which my paper was devoted everything that is special and technically distinctive about the notion of a projection rule is superfluous.

¹ Under the title How to do Things with Words (Oxford, 1962).

² See my paper ‘On sentence-sense’ in this volume, pp. 18, 19, 23, 24.

³ Consider negation. If anything can do what the traditional otherness explanations were meant to do to make us comfortable with not, I suppose it is a theory of what a speaker does when he negates a sentence (proposition?). See my ‘Negation, Falsity and Plato’s Problem of Non-Being’ in Plato I (ed. Vlastos, New York, 1970), vii-viii. I still think that is more important, however, to formulate what the problem of negation precisely is than to rush to answer it.

⁴ A thought implicit in note a, p. 19 and now argued in another way by P. F. Strawson in Meaning and Truth (An Inaugural Lecture, Oxford, 1970).

⁵ See page 171 of Donald Davidson ‘On Saying That’ in Words and Objections, ed. D. Davidson and J. Hintikka (Dordrecht, 1969).

⁶ Which I have had occasion to discuss with my student, Mr Dudley Knowles of Bedford College.

⁷ See Uber Annahmen III, 9. I owe the reference to another student, Miss Glen Crowther.

⁸ See again Davidson, p. 171.

⁹ Cp. W. V. Quine, ‘Quantifiers and Propositional Attitudes’ in Ways of Paradox (New York, 1966).