Literal meaning and fictional utterance

John McDowell

  • Expression and Meaning: Studies in the Theory of Speech Acts by John Searle
    Cambridge, 187 pp, £8.50, December 1979, ISBN 0 521 22901 4

John Searle’s subtitle alludes to the mode of reflection about language which he recommended, and showed in operation, in his earlier book Speech Acts. What the theory of speech acts offers is a view of language which is to allow us to take full account of the fact that linguistic behaviour is behaviour, without lapsing into a behaviouristic obliteration of the inner life.

Asked about an utterance, the question ‘What is that person doing?’ can be answered at different levels; and the questions about language, and more specifically about meaning, which have fascinated philosophers of language can be formulated in terms of relations between the levels. One might ask, for instance: when does an act describable from one point of view (say, that of an uncomprehending alien) only as the production of certain sounds constitute an intelligible communication, since it is correctly describable as such from another point of view, that of someone who understands its language? And how do the properties ascribable to the utterance at the lower level determine the properties ascribable to it at the higher?

Sharpening these and similar questions into tractability, so that they can usefully stand in for suggestive but unclear formulations such as ‘What are the relations between sound and meaning?’, requires a fairly systematic conception of the structure and content of the higher-level descriptions: those which bring the intuitive notion of meaning into play. Searle’s systematisation organises the descriptions into a specification of propositional content, on the one hand, and an identification of illocutionary force (e.g. a characterisation of the speech act undertaken as an assertion or a request), on the other. (The notion of illocutionary force, and the label, are due to J.L. Austin.)

The division marks out two fields for further investigation. First, there are questions about how, by uttering suitable strings of words, we contrive to express propositions: in Searle’s treatment, these are questions about such contributory speech acts as referring and predicating. (These ‘propositional acts’ cannot be performed except in the performance of particular illocutionary acts. Nevertheless, we can theorise about them in abstraction from particular illocutionary forces.) Second, there are questions about what it is to perform the various types of illocutionary act.

An overarching thesis is that all these speech acts are governed by, and make manifest, intentions of a special audience-directed kind: intentions whose content is in part that they themselves should be recognised by an audience. This is what ensures the non-obliteration of the inner life. (Here Searle’s position incorporates, in a modified version of his own, an insight of H.P. Grice.)

Speech Acts contains extended discussions of referring and predicating, and offers a detailed account of the rules which govern one type of illocutionary act – namely, promising – with suggestions as to how the analysis could be extended. There are also some philosophical applications. But the theory of speech acts is a research programme, not a completed product: part of its point is its fruitfulness in raising questions for investigation. Expression and Meaning collects some characteristically forthright and provocative essays on outstanding topics.

A recurring theme is the possibility of divergence between what one means by an utterance and what one would mean if one’s meaning were straightforwardly determined by the literal meaning of the words one utters. In what Searle calls ‘indirect speech acts’ (Chapter 2), one means what one literally says and more besides. (For instance, uttering the words ‘You are standing on my foot’ in suitable circumstances will constitute not only an assertion that the addressee is standing on one’s foot but also a request to him to get off.) In metaphorical utterances (Chapter 4), one means, not more than what one literally says, but something else instead. Fictional utterances (Chapter 3) differ in a different kind of way from straightforward non-fictional discourse. Chapter 6 uses the apparatus of Chapter 2 in a criticism of Keith Donnellan’s distinction between referential and attributive employments of definite descriptions. And half of Chapter 7 argues that the apparatus is superior to a competitor from the literature of linguistics.

Given the thematic prominence of the notion of literal meaning, one might expect a full-dress explication of it at some point. Chapter 5, titled ‘Literal Meaning’, is not that, but at best a preliminary: it aims to demolish the idea that the literal meaning of a sentence is something it has independently of any context or background. (This essay is decorated with charmingly inept drawings by the author.) The remaining essay (Chapter 1) proposes, in place of Austin’s rather bewildering classification of kinds of illocutionary verb, a taxonomy of illocutionary acts, based, as seems evidently sensible, on difference of illocutionary point or purpose.

Chapter 2 credits a hearer who understands an indirect speech act with ‘first, a strategy for establishing the existence of an ulterior illocutionary point beyond the illocutionary point contained in the meaning of the sentence, and second, a device for finding out what the ulterior illocutionary point is’. These strategies involve bringing to bear facts of various kinds, principles from the theory of speech acts, and principles of conversational co-operation: Searle illustrates how they work with chains of inference whereby a hearer could explicitly derive probabilistic conclusions. He is quick to point out that in ordinary conversation one does not consciously engage in such derivations. But this leaves him quite unforthcoming as to how we are to apply the talk of strategies to the phenomena of ordinary conversation; and there are serious problems here. As Searle remarks, one can simply hear an utterance of ‘Can you pass the salt?’ as a request to pass the salt. How can an inference one does not consciously make, with the conclusion, not that the utterance is a request, but that it is probably a request, account for the fact that its being a request can be part of the content of one’s aural experience? Again, if the utterance is a request, we credit a person who hears it as a request with knowing that it is one. Can we account for that fact by ascribing to him an inferential justification for the conclusion that it is probably a request? Similar questions arise about Searle’s similar talk of strategies in his treatment of metaphor. This use of the notion of inference lands Searle in the middle of a familiar group of problems in epistemology and the philosophy of mind, and he seems insufficiently sensitive to their difficulty. (This assessment is compatible with the thought that there must be something right about his treatment of indirect speech acts.)

The leading idea in Searle’s treatment of fiction is the plausible thought that a storyteller (say) goes through the motions of making assertions without really doing so. The activity of making assertions is governed by vertical rules, rules which correlate ‘words (or sentences) to the world’. In making it clear that one’s discourse is fictional, one suspends those rules.

Put as unspecifically as that, the idea is attractive. But there is a serious indeterminacy in ‘words (or sentences)’. (Both? If not, which?) The ‘essential rule’ governing assertion is that the utterer commits himself to the truth of the proposition he expresses; certainly this rule, and a subsidiary rule that he should be able to provide evidence for its truth, are suspended in fiction. But these rules effect sentence-world correlations, and their suspension leaves quite open what, if anything, happens to the correlations between individual words (or morphemes) and the world: those which enable strings of words to be used to express propositions at all. Searle suggests that in fiction ‘the rules correlating “red” with red are not in force’: but if they are not in force, how can one understand what it is that a fictional utterer of ‘Her lips were red’ is going through the motions of saying?

It is a pity that Searle never explicitly addresses himself to this question: does someone who makes a fictional utterance express a proposition, to whose truth he goes through the motions of committing himself, or is the expressing of a proposition a pretence too? If the former is the case, then word-world correlations (as distinct from sentence-world correlations) are not suspended, and problems about reference in fiction must be more complex than Searle allows. If the latter is the case, the reason for saying so will be found, I think, in considerations about reference in particular, not word-world connections in general; and again we need more detailed attention to reference (or pretended reference) than Searle provides.

In a brief discussion of the ontology of fiction, Searle suggests that the author of a piece of fiction only pretends to refer. This implies the second answer to my question. But Searle’s grounds seem thin. He grants that someone other than the author can really refer (to a fictional character, of course) in uttering a sentence like ‘Sherlock Holmes lived in Baker Street’: but he distances this from the author’s fictional utterances, insisting that it is not fictional discourse but discourse about fiction. But why can only the original author produce fictional utterances? Searle’s own talk of a shared pretence, underlying an ability on, say, my part to effect real reference to the fictional detective, makes it attractive to think of the remark about Sherlock Holmes, made by me, as a move in a complex and continuing game of make-believe assertion: it is a game initiated, certainly, by Conan Doyle, and one in which his moves have a specially authoritative status, but that does not show that others cannot join in. (Surely others have joined in?) And now, if my pretended reference to a real person can be a real reference to a fictional person, and if Searle’s reason for claiming that what I do is relevantly different from what Conan Doyle did falls away, why may we not suppose that Conan Doyle’s pretended references are real references too? The answer ‘Because there was no already existing fictional character for him to be referring to’ (cf. page 71) would apply at best to the first sentence, or perhaps the first few sentences, of the first introduction of the character; and does not seem indisputably cogent even there. I applaud Searle’s wish for a no-nonsense view of the ontology of fiction, but I think the problems are harder than he makes out.

Searle takes the problem of metaphor to be this: how can a speaker, by making an utterance whose literal interpretation would determine one set of truth-conditions, contrive to commit himself to the obtaining of a different set of truth-conditions? He says very little about this ‘truth-conditions’ way of talking. Probably he regards it as innocent and unnoticeable. But I suspect a piece of high metaphysics is presupposed in the very formulation of the issue. According to Searle, it is trivially obvious that there could in principle be a literal expression of what a metaphorical utterance conveys. His argument for this surprising claim is surprisingly sloppy: he appeals to the principle (which is indeed plausible) that whatever can be meant can be exactly expressed, but of course that leaves open the possibility that some meanings can be exactly expressed only by a metaphorical utterance. (Metaphor is not a species of vagueness.) One can see, however, how some metaphysical presuppositions might conceal the weakness of the argument. A person who makes an assertion commits himself as to how things are. If he is right, then things are as he says: in the world, one is inclined to say (as if that added something). Metaphorical assertions can be right. And surely there cannot be an aspect of how things are (in the world) which can be captured only by a metaphor.

Philosophers are prone to conceive the world in a way that makes this last thought seem obviously compulsory. Nothing else will do, they tend to think, as a correlate for anything recognisable as a notion of truth. But the prevalence of such ideas in philosophy is not necessarily a point in their favour. (Some might think the reverse.) And their correctness ought not to be allowed to pass as an innocent truism, something we must at all costs respect in thinking out the relations between the notions of assertion, propositional content, truth-conditions, and (if you like) world. If we can win through to seeing this conception of the world, and the associated conception of truth-conditions, as a discardable metaphysical prejudice, we make room for the idea that a metaphorical utterance may be the only conceivable way of conveying precisely what it conveys – rather than at best an effective means for expressing something which could also be literally expressed, by a coinage if necessary. This would, arguably, permit a more lifelike picture of the peculiar delight we take in an apt metaphor than any available to Searle.

Chapter 6 strives valiantly to engage with Donnellan while remaining neutral as between theories of reference. In fact, there is enough here, even without Speech Acts, to reveal the general shape of Searle’s view of reference. All reference is under an aspect; and there is no place for a conception of a certain kind of content of a belief or act, such that there would be no content of that kind if there were no object for the belief or act to be about. These two theses determine a version of the position standardly ascribed to Frege: a referring expression has a sense (Searle’s ‘aspect’), available, even if no appropriately related object exists, to serve as a partial determinant of propositional contents expressible by uttering sentences containing the expression. This view, which Searle finds obvious, will strike some as failing to recognise genuine reference at all. Now it may seem unfair to dredge up this iceberg and complain about its nature, when Searle claims that his objection to Donnellan rests on no particular theory of reference. One may well suspect, however, that it is only because of his general view that Searle can be happy to say, against Donnellan, that reference (scilicet: reference in the only sense we need be concerned with) takes place in both sorts of case Donnellan distinguishes. I am no particular friend of Donnellan’s distinction: but I find it most implausible that Donnellan can be satisfyingly answered except in the context of an explicit and adequately defended general account of reference; and, moreover, highly unlikely that a general view on the lines of Searle’s would equip one so much as to appreciate what Donnellan is driving at.

All the essays have been, or will be, published elsewhere, as journal articles or anthology chapters. Do they add up to a book? I think if we make sense of the distinction between a book and a collection of essays, the answer must be ‘no’. The point of an article is to initiate or continue discussion: one can make undefended assumptions, and have loose ends, and not worry unduly if one’s ideas are a little half-baked around the edges, so long as one is confident of having a central suggestion that is worth discussing. A collection of such pieces does not make a book, even if there are thematic connections between them, as there undoubtedly are between these. A book is not just a lot of words on one subject: it should aspire to a certain self-containedness. These pieces do not take up one another’s loose ends: even juxtaposed, they retain the characteristic feel of articles.

Of course this question of genre need not matter much, even if we do make sense of it. Good journal articles are probably still (just) worth collecting. Speech Acts gave Searle a deserved prominence in the field, and there are plenty of people, both inside and outside philosophy, who will be interested in what he has been up to since. Moreover, some of the essays are adequately defended as they stand. All the same, it seems to me to be cause for disappointment, on the whole, that Searle has not undertaken the sort of further thinking and writing on some of these topics that would have been necessary if he had persisted with the discarded project he tells us about in his introduction. His next book – which is to be a discussion of the relation between philosophy of language and philosophy of mind – will, I believe, merit a different kind of respect.