In this dissertation I examine a distinction made in Speech Act Theory between normal uses of language and uses of language that are said to be parasitic on them. Fictional, theatrical, comedic and metaphoric uses of language may be said to be parasitic on normal language in so far as their intelligibility requires a prior grasp of the rules or conventions of normal language such as is used in everyday cases of asserting, promising, marrying and ordering, for instance.
Jacques Derrida argued that uses of language could not be determined as exclusively either normal or parasitic and that thus such a distinction could not be made. That is, he argued that it was not possible to make a distinction between fictional promises and real life promises, for instance; or between literal uses of words and metaphorical uses. I show that the distinction can be made and that, although uses of language cannot be determined as exclusively either normal or parasitic in the work of J.L. Austin, they can be in that of John R. Searle.
In arguing for this thesis, I show how Searle, in his attempt to defend Austin and Speech Act Theory against Derrida’s criticisms, failed to appreciate many aspects of Derrida’s work and thus misconstrued his critique and defended Austin and Speech Act Theory against somewhat of a straw man.
I should like to thank my supervisor, Barry Allen, for the considerable effort he put into getting me to turn out a defensible dissertation. His comments were always penetrating and stimulating, and he worked through my drafts with lightening speed. My second reader, Donald Stewart, commented on all my drafts (which is much more than one can expect from a second reader) and his criticisms were always fair and brought many new considerations to my attention. His constant and good-humoured insistence on clarity of presentation helped me turn my chaotic first draft into a considerably better organized final draft. Mark Vorobej, who was kind enough to show an interest in my thesis from the earliest stages, came to the rescue when a former third reader felt unable to continue. I am very grateful to Dr. Vorobej for working so hard on the penultimate draft of this dissertation and under such pressing circumstances.
Besides those on my supervisory committee, others have helped me focus my ideas by discussing various aspect of Speech Act Theory and Deconstruction with me. They include Stephen Liem, Sajahan Miah and Douglas Odegard. I should also like to thank Evan Simpson for his interest in my work and for his valuable help with administrative matters. And I thank Christopher Tindale for sharing some work of his own with me. There are many others I perhaps ought to thank for their challenging scepticism, especially regarding Deconstruction: in this context I must mention my friend, Vincent Ngan.
For bibliographic details, click on the abbreviations.
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
N.B. Clicking on the green endnote numbers (e.g. Note [0]) brings you to the relevant endnote; once there you can click again to get back to where you were. Try this one by clicking on the green nought!
In this dissertation I examine, and defend in a certain form, the drawing of a distinction between normal uses of language and uses that are said to be parasitic on them. This distinction is made in the philosophies of J.L. Austin and John R. Searle. It was attacked by Jacques Derrida. I support Derrida’s criticism of it in Austin’s philosophy but defend its use by Searle. More specifically, I show that, although the distinction is defensible in both philosophies, it is only in Searle’s that one can distinguish between speech acts that are exclusively either normal or parasitic.
The distinction is first introduced by Austin in work of his (most notably his How to Do Things with Words) in which he attempts to see language as a kind of social activity rather than merely as a matter of stating truly or falsely. Austin showed how language may be used to make promises or declarations (as well as assertions), to baptize or to marry, to bet or to express emotion. He also showed how such acts might fail. He called their failure, their being unhappy or infelicitous. Also he felt that these acts could be parasited. By this he meant that they could be used in novels, poems, jokes, on stage and even in quotations. There was something the matter with such parasitic uses which, Austin thought, was related to unhappiness.
Distinctions between uses of language that are happy or unhappy, and normal or parasitic, are also made by Searle in his theory (specifically in Speech Acts, Expression and Meaning and Intentionality). His way of accounting for them is different to Austin’s. I shall show that the Searlean way is better. I do this largely by means of considering a penetrating critique of Austin’s theory by Jacques Derrida in which he questions the possibility of making the distinctions that Austin makes given the theoretical apparatus that Austin adopts. I shall show that Derrida raises problems with Austin’s theory that are insuperable given Austin’s mechanisms (some of these problems however may be due to his theory’s not having been sufficiently developed); but I argue that, although Searle misunderstood Derrida’s critique and defended Speech Act Theory against what was, in effect, a watered-down version of what Derrida had to say, Searle’s theory can be defended against the types of criticism that Derrida brings against Austin and against those criticisms that Derrida explicitly brings against Searle. In other words, I shall argue that a distinction between normal speech acts and their parasites can be defended in the work of Searle but not in the work of Austin.
By way of further introducing the topic of this dissertation I shall present a characterization of the basic issues to be dealt with. The normal/parasitic distinction is sometimes rendered the normal/abnormal distinction and sometimes it is also said to account for two distinctions at a lower level of generality, viz. serious/non-serious and literal/non-literal. To distinguish between normal and parasitic language is to distinguish between uses of language that are primary and other uses that are in various ways dependant on them (or derived from them). This in turn presupposes that language is like an instrument which, although it may have proper functions, may be used for others too. As such, it may be likened to a feather duster made to dust around fragile objects and in awkward corners. The properties of the duster however allow it to be used to tickle. Here dusting is the primary use and tickling is secondary.
If language is characterized in some such way, then this implies that language has proper functions, and that it can be intentionally used according to its functions or in secondary ways. The normal, serious, primary purpose of language will be intentionally to perform certain conventional acts such as asserting or promising, for instance. Other uses will be secondary. Uses that are not (necessarily) primarily concerned with the performance of conventional social activities include joking, writing poems and novels and teaching languages. For instance, in joking the point is to amuse. So if something is asserted or promised that will be a secondary consideration. Thus a primary function of language is made secondary and vice-versa. Similarly in writing poetry, the assertions and declarations involved will be secondary to the aesthetic use of language. And in teaching a language, phrases of the language will not be used (at least not initially) to perform any conventional social activity but for pronunciation exercises, tests of comprehension and translation.
In one sense, a parasitic use of language is not a failed use. One can distinguish between normal uses of a language such as asserting, promising or commanding and the various failed attempts at these uses. For instance ‘commanding’ one’s superior officer in an army will not succeed since a condition of giving a command is being in a position somehow superior to that of the person one is commanding. Such a ‘command’ will be a failure but it will not (or not always) be parasitic. One might, of course, issue such a command as a joke. If it is not a joke (or some other sort of parasite), for instance, then it will simply be a failed, ‘unhappy’ or ‘infelicitous’ command. To distinguish between normal and parasitic uses of language is not per se to distinguish between successful and failed uses. A parasitic use of language is not a mistaken use but rather is quite deliberate (or at least it may be so). So the normal/parasitic distinction is not the successful/failed distinction. We shall see however that, for Austin, they may be related.
It is important to be able to make distinctions between the happy and the unhappy and between the normal and the parasitic because without these distinctions Speech Act Theory would not be possible. If one could not make the normal/parasitic distinction, then one could never give the conditions under which a promise, for instance, would be successfully made. This is because a promise that could not be determined as serious or literal could not be essentially characterized just as one could not define what a duster was unless one could distinguish between its primary function and its secondary functions (which may not be enumerable). If one could not maintain a general difference between promises made by people playing roles on stage and promises made by the same people off stage, then one could not say that the ‘promise’ made on stage entailed commitments which the promise made off stage did not entail. One would not be able to say that in one case certain rules were in operation but that some of them were suspended in the latter case (because of the nature of the context or because certain conventions were being invoked).
For Austin, in order to distinguish normal uses of language from parasitic uses, and similarly in order to distinguish successful from failed uses, one must consider their specific ‘total’ context—i.e. context including ‘internal’ features such as intentions and ‘external’ features such as a certain social situation. Uses of language cannot be unhappy or parasitic independently of their total context. These utterances will have contexts that are proper to them (e.g. one says ‘I declare you man and wife’ at a wedding). The proper total context of a use of language is the social situation in which it is used but it also must include the user’s intentions too since some utterances may not be determinable as normal or parasitic without also considering the utterer’s intention (e.g. an unhappy promise is an insincere one and insincerity is largely a matter of ‘internal’ features).
For Searle, the distinction is made solely in terms of the utterer’s (or writer’s) intentions. Initially this makes his theory look more impoverished than Austin’s since the differences between plays and novels, and real life, seem to be more than a matter of what their authors intended. Indeed such differences seem to be textual and contextual. But, as we shall see, Searle’s intentional criterion (together with his drawing a logical distinction between the intention to represent and the intention to communicate) enables him to escape certain of Derrida’s criticisms of Austin. It does this, as we shall see, by allowing that there may be no textual or contextual mark of the speech act in question which would be observable by any hearer (or reader).
Take some line of poetry and consider whether there is anything about it that tells whether it is a line of poetry or perhaps a mistaken use of language. Sometimes it may appear to be obvious that it is poetry (if the line is evidence of much skill, for instance). But at other times it will not be clear and one will (according to Austin) have to consider context. One could check, for instance, whether it was published in a book of poetry. Those learning English sometimes come up with rather interesting utterances that might in other contexts be taken as poetry. For instance, they might come up with utterances that might seem to be complicated metaphors, rather than simply mistakes, were it not clear that the context in question was a person beginning to learn a language. For Searle, the only way of deciding whether such an utterance was intended metaphorically would be by discovering (in whatever manner) that the speaker (or writer) intended it metaphorically. There might be no textual or contextual mark of it.
In this dissertation I examine how such distinctions and presuppositions are expounded in the speech act theories of Austin and Searle. I examine the reasons for such distinctions, their feasibility and whether they stand up to close critical examination by Derrida. The latter made a very fundamental attack on the normal/parasitic distinction, on the distinction between happy and unhappy utterances, and on the presuppositions of (what he considered to be) any possible system that would embrace such distinctions. His examination of the type of framework in which such distinctions are made forces one to reconsider the basics of Speech Act Theory. The notions of ‘proper’ and ‘total’ contexts, which are fundamental to, and foundational in, Austin’s Speech Act Theory, are shown to be indefensible. And it is because of this that there can be no nice discrimination of the serious from the non-serious, or linguistic hosts from their parasites in his theory. It no longer becomes possible to distinguish the normal from the parasitic except relatively or perspectivally.
I proceed as follows. First I outline (in Chapter Two) the main features of Austin’s theory, especially that part of it which he calls ‘Speech Act Theory’. I emphasize what he says about meaning, intentionality and context in order to clarify his understanding of the nature of language. Then I examine some of Searle’s criticisms and modifications of Austinian Speech Act Theory. I shall be especially interested in his distinction between literal sentence meaning and speaker’s utterance meaning.
Then (in Chapter Three) I consider Austin’s and Searle’s ways of making the normal/parasitic distinction. I show how Austin relates parasitism to a certain kind of infelicity associated with misunderstanding or failing to ‘secure uptake’, and I point to initial problems with the distinction in Austin’s speech act theory. For example, some normal speech acts seem to be parasitic on what Austin’s theory would characterize as parasites. Thus the normal speech act would seem sometimes to be a parasite. In this chapter too I show how useful Searle’s distinction is between speaker’s utterance meaning and literal sentence meaning in explaining the relation between normal speech acts and their parasites. I shall show how it enables him to explain how one utterance can function in many different ways. For instance, one utterance can be made as an assertion of a philosophical point, an indirect contradiction of a point made by someone else, and as a joke.
Next (in Chapter Four), by way of introducing Derrida’s criticisms of Austin, I examine Derrida’s treatment of traditional ways of making a distinction between speech and writing. But first I outline a basic concept in his philosophy which he uses in his criticism of Austin and Searle. This is the notion of ‘iterability’ (and ‘iteration’). It is basically the notion of sameness in spite of difference; and it is intended to explain how things that are very different (e.g. the word ‘cat’ as written and as pronounced) can nevertheless be the same thing. Using this notion he argues that something, which he calls the ‘Classical’ theory of writing (supposedly common to all of Western thought and thus a fortiori to Austin), is wrong. He sees his attack on this theory (and the notions of irreducible polysemy and the permanent or structural possibility of failure that go hand in glove with his attack on this theory) as undermining the foundation of Austinian Speech Act Theory and, specifically, the normal/parasitic distinction. The foundation in question is the ‘proper’ context. Derrida sees Austin’s characterization of the proper context as ordinary, normal and serious, and the concomitant exclusion from consideration of non-serious utterances, as enabling conditions of Speech Act Theory. He sees his investigation as showing that this putative foundation and the attempted exclusion of the parasitic, are arbitrary and, in fact, impossible. He affirms the permanent, structural possibility of parasitism and, as a consequence, the impossibility of Austinian and Searlean Speech Act Theory. I show however that his investigation does not compel one to agree that Speech Act Theory, as such, is impossible but only that it needs to be revised. Indeed I show that Speech Act Theory survives in Searle.
Finally (in Chapter Five) I examine Searle’s interpretation of Derrida’s critique of Austinian Speech Act Theory. Here I show that Searle was largely misguided in his attack on Derrida mainly due to his not understanding Derrida’s terminology well. However I show that, despite the failure of his critique of Derrida, his theory of normal and parasitic speech acts is defensible against Derrida’s criticisms. More than this though, his theory is able to account for insights that Derrida has while including them in a theory that accounts for the uses of language in a systematic way. In short, his theory is able successfully to explain the relations of normal and parasitic uses of language in a way that is clearer and more systematic than Derrida’s. It shows that Derrida’s critique, or attempted deconstruction, of Speech Act Theory fails.
In short this dissertation will defend the view that the normal/parasitic distinction is defensible and that, with Searle’s theory of speech acts, it is possible to distinguish between normal speech acts and their parasites. I do this by means of rejecting Derrida’s fundamental attack on the possibility of making the distinction and of distinguishing between normal uses of language and parasitic uses.
In this chapter I shall show what a speech act is and how it may be either happy or unhappy. In order to do this efficiently I must approach the matter systematically. Although it might seem that the first question that I should ask would be ‘What is a speech act?’, I shall first of all investigate the motivation for a theory of speech acts. I do this in order to show later on, after I have introduced the distinction between serious and parasitic speech acts, that, just as the statement (or ‘constative’) and the performative must be (in a sense) synthesized in order best to explain linguistic practice, so also the normal (or ‘serious’) and the parasitic must also be synthesized. I begin therefore by explaining how the speech act emerges from a synthesis of what are known as constatives and performatives. When I say that the two must be synthesized, I am essentially just saying that what was once seen as evidence of two different things is now seen as evidence of one thing with two dimensions (or general characteristics).
As well as showing how the speech act emerges, in this chapter I also show how it may be happy or unhappy (or ‘felicitous’ or ‘infelicitous’—the difference in terminology here will not be significant in what I shall be saying). Later I shall show that, just as a general distinction cannot be made in Austin’s theory (although it can in Searle’s), between speech acts that are exclusively either serious or parasitic, so also a distinction between speech acts that are happy or unhappy cannot in general be made.
Since there are two main theories of speech acts, and since they differ significantly on the matter of serious and parasitic speech acts, in what follows I shall investigate the two theories (viz. the Austinian and the Searlean) separately. Later on it will be necessary to discriminate between criticisms that affect the distinction as made by Austin and those that affect it as made by Searle. First I shall investigate Austin’s theory and then I shall show what Searle changed and added. In this regard I shall pay especial attention to Searle’s greater emphasis on the meaningful component of speech acts. Later I shall show how it plays a significant part in preserving the normal/parasitic distinction.
In line with the above-mentioned articulation of my proposed treatment of matters, I divide this chapter in the following manner: first, I investigate the development of the speech act from a synthesis of the constative and the performative; secondly, I consider Austin’s speech act theory; and thirdly, I show how Searle’s speech act theory is a development of it.
Austin’s theory of speech acts emerges from his consideration, and rejection, of a distinction which he sees as central to philosophy of language up to his own work. This is the distinction between utterances which are meaningful, which are all thought to be statements of what is or is not the case, and utterances which are meaningless. This view holds that only statements are ever meaningful. But Austin rejects this pointing to another class of ordinary utterances which are neither meaningless nor constative (i.e. of the nature of a statement). He calls such non-constative, meaningful utterances ‘performatives’ since they are utterances the production of which, given certain conditions (to be investigated), serves as the performance of some conventional social act. So instead of the traditional constative/nonsense distinction Austin in effect postulates two distinctions: constative/performative and meaningful-utterance/meaningless-utterance. In this section I shall explain why Austin rejects the traditional distinction (as he sees it) and why he ultimately even rejects his own constative/performative distinction. I shall call his view between the rejection of the traditional distinction and the rejection of the constative/performative distinction Austin’s theory of performatives. This I shall contrast with what took its place after the latter distinction was abandoned, viz. Speech Act Theory.
So to investigate the theory of performatives. The utterance of a performative, as already mentioned, is the performance of a conventional act through making a certain utterance in a certain context. For instance, saying ‘I will’ in response to the judge’s or priest’s question ‘Will you take...?’ at a marriage ceremony at which one is being married is eo ipso the act of marrying the person named. Those words do not report the event of one’s marrying, rather they effect it. The utterance here of ‘I will’ does not assert anything which could be discovered to be either true or false. It states neither truly nor falsely that one is marrying, rather it makes it true that one is marrying. Nor, of course, does its being in the future tense mean that it is a prediction about what one will do.
There is an asymmetry between constatives and performatives which Austin expresses by talking about their different directions of fit [see HDTW, 47ff]. To state something is, as it were, to fit words to the world. The statement will be true if it does actually fit (however determined) and false otherwise. To utter a performative though is to fit the world to (one’s) words—to use language to bring about a new state of affairs in the world. That is, a constative reports a state of affairs whereas a performative is a conventional means for bringing one about and often without further ado. For instance, if I say ‘I promise...’, then, without further ado, I have promised. [1]
A difference between the performance of a promise and, for example, a baptism is that in the former case but not in the latter uttering certain words is sufficient to perform the act. Simply saying ‘I promise...’ counts as promising whereas simply saying ‘I baptise you...’ does not count as baptizing without further ado. In order to baptise one must ordinarily (i.e. not in emergency cases) be some kind of religious cleric, e.g. a priest. Also, one must perform certain actions such as pouring water on the baby’s brow or immersing it. Without these concomitant actions the utterance in question would not effect the child’s baptism. Similarly, without the words there would be no baptism.
This draws attention to the importance of the context of a performative utterance. The utterance of certain words in a certain context constitutes a performative. What the words are and what the context is will have been established through convention. [2] Later I shall show that convention cannot ‘establish’ contexts but for now I shall, for the purposes of explication, let this pass. Performatives that are uttered in appropriate contexts are said to be ‘happy’ (or ‘felicitous’) as opposed to ‘unhappy’ (or ‘infelicitous’) [see HDTW, 14ff]. So, for instance, for the words ‘I promise’ to be a promise I cannot be in the situation where I have just been ordered, by someone in command over me, to do that very act I purportedly promise to do. More clearly, in the case of a baptism, the child must not be known to have been baptized already.
Not only this though; performatives may be assessed for happiness or unhappiness in another manner. A performative is defective if it is uttered insincerely. To say ‘I promise...’ while intending not to fulfil what one thereby promises is to promise defectively (as opposed to not promising at all as in the previous example). One’s utterance is unhappy in that it has not been executed with the appropriate intentions, beliefs or attitudes. [3]
Just as a statement is judged, according to Austin, as to whether it corresponds to the state of affairs it purports to represent, so also a performative is judged as to whether it brings about the state of affairs it purports to. One judges a person’s statement about a certain state of affairs by asking whether what he said was a true account of it. One judges a person’s performative utterance purporting to bring about a certain state of affairs by judging whether it conventionally succeeded (if there is a possibility that it could have failed) or whether it was sincere (in certain cases where sincerity is germane). To take the case of promising, one will not have promised if one utters ‘I promise...’ and it is clear that there is no possibility of one’s doing what one promises. In this case, it is conventionally accepted, such an utterance does not count as a promise. So this utterance would be unhappy and the person would not have managed to promise.
The various types of performative unhappiness must be investigated further. Infelicity is a matter of how performative utterances operate in a given context. This is the ‘total speech act’ or ‘total speech situation’ [HDTW, 52 & esp. 76 & 148]. It consists of both ‘internal’ and ‘external’ circumstances (or intentional and worldly circumstances). These two aspects of the total situation or context are approached by systematically outlining the types of things that can go wrong with a performative and thus, by contrast, what can go right. And it is worth noting here that a performative that is happy is one that is not unhappy. It is determined negatively by examining the total context to see whether there is anything wrong. This notion of total context (i.e. context including the speaker’s intentional states) will be important when I come to consider criticisms of Austin in later chapters. I shall show later that, because a context cannot be ‘totalized’ in this manner, the performative (and more especially the speech act) cannot actually be determined.
Where a performative is unhappy due to external circumstances it is said to ‘misfire’. This may be due to ‘misinvocation’ or ‘misexecution’. In the former case conventions either do not exist as appealed to or are incorrectly appealed to. There are thus two types of misinvocation: ‘non-plays’ [see HDTW, 31, 31 n. 1, & 18 n. 1] and ‘misapplications’ [see HDTW, 17f]. The former is where a convention does not exist although one seems to be appealed to (as, for instance, if a man were to stand his wife in front of him in company and utter ‘I divorce you’ [see PP, 238] ); the latter is where a convention is wrongly applied (as, for example, occurs when a married man commits bigamy [see HDTW, 16f] ).
In the case of ‘misexecutions’, the other type of misfire, the conventional procedures are not fully carried out [see HDTW, 17ff]. Here again there are two sorts: ‘flaws’ and ‘hitches’. To take the former; if one party during the marriage ceremony says ‘I will’ and the other says ‘I won’t’, then the marriage ceremony is ‘flawed’ [see PP, 238]. An example of a ‘hitch’ would be where one offers a bet but it is not accepted by anyone. Here, according to Austin, one has not succeeded in betting because the conventional procedure has not been completed [see PP, 238].
To complete the classification that Austin offers one must consider performatives that are unhappy due to intentional (or ‘internal’) circumstances. [4] These are ‘abuses’ rather than misfires and there again are two types: ‘insincerities’ and ‘non-fulfillments’ or breaches of commitment [see HDTW, 16ff; PP, 238f; PC, 14f]. To promise without the intention of keeping to what one promises is to abuse the procedure [see HDTW, 16; PP, 239] or even, as Austin sometimes puts it, to abuse the formula ‘I promise...’ [see PC, 14]. A case of the abuse described as non-fulfilment is where one sincerely promises but does not fulfil one’s promise [see PP, 239]. [5]
Here is a schematic rendering of these distinctions, based on Austin’s [see HDTW, 18] :
This typology of infelicitous performatives does not purport to be a table of categories of infelicity. Austin is quick to point out that it “is not complete, and they are not mutually exclusive; they never are” [PP, 239; cp. HDTW, 23f]. As an instance of the types’ not being mutually exclusive, he gives the example of promising a donkey to give it a carrot. Is this a non-play (there being no convention of promising to donkeys) or a misapplication (the convention of promising not extending to donkeys)? Austin thinks that it is perhaps both. A case of possible ‘overlap’ would be where, at a ship’s launching, the wrong person seizes the champagne bottle and proclaims ‘I name this ship the Generalissimo Stalin’ while smashing it against the ship’s bow and then kicking away the chocks. Again Austin is not worried about how to classify this act (wrong-person-right-act or incomplete procedure, either a misapplication or a hitch).
This typology is not complete either because, emphasizing that performatives are actions, Austin points out that they are “subject to certain whole dimensions of unsatisfactoriness to which all actions are subject but which are distinct—or distinguishable—from what we have chosen to discuss as infelicities” [HDTW, 21]. Here he has in mind such factors as being constrained to act, as when one promises with a knife to one’s throat, or generally acting unintentionally.
Austin is thus not proposing his list of infelicities as either complete or mutually exclusive; nor is he claiming that no one could do any better.
Thus the way we should classify infelicities in different cases will be perhaps rather a difficult matter, and may even in the last resort be a bit arbitrary. But of course lawyers ... have invented all kinds of technical terms and have made numerous rules about different kinds of cases, which enable them to classify fairly rapidly what in particular is wrong in any given case [PP, 240; cp. HDTW, 23].
Here Austin is being pragmatic; he is in effect, as his typical recourse to the habits of lawyers evidences, saying that our classification will largely depend upon our concerns.
Conventions are in fact, according to Austin, inherently vague:
It is inherent in the nature of any procedure that the limits of its applicability, and therewith, of course, the ‘precise’ definition of the procedure, will remain vague. There will always occur difficult or marginal cases where nothing in the previous history of a conventional procedure will decide conclusively whether such a procedure is or is not correctly applied to such a case [HDTW, 31].
Conventions, however they arise, clearly apply to certain activities and not so clearly to others. It may never have been necessary for a society clearly to have used a convention in a certain way and in that case it will not have been important for that society to have been clear about that particular case. Austin’s example here, although it is rather bizarre (it concerns baptizing a dog), does show how difficult it is to decide what type of infelicity would be involved in trying to baptize an animal; the reason, or part of the reason, is the fact that society does not have to consider such cases. It is, so to speak, beyond the scope of (the use of) that convention.
So Austin has put forward a theory of performatives which reflects the vagueness of conventions in that it refuses to categorize nicely the various ways one may make, or fail to make, a performative utterance. At this point he only seems to be certain, on the one hand, that there are meaningful utterances that cannot be either true or false but only happy or unhappy, and, on the other hand, that the only other meaningful utterances are capable of truth or falsity but not of happiness or unhappiness [see PP, 235; HDTW, 5]. Constatives are assessed for truth or falsity and performatives are assessed for happiness or unhappiness.
Eventually Austin finds that performatives can be assessed in a true-or-false dimension and constatives in a happy-or-unhappy dimension. This will render the constative/performative distinction either just fuzzy or useless, and will motivate their synthesis. But before I come to that I shall briefly contrast the logic of performatives with that of constatives. This, together with a demonstration of how constatives can be assessed for felicity and performatives for truth-value, will enable us to consider whether the constative/performative distinction is just fuzzy but still useful or simply not useful.
With regard to the logic of performatives and constatives Austin considers presupposition, implication and entailment. Statements are said to imply other statements (or the truth or falsity of other statements) whereas acts are not said to imply other acts. Similarly, constatives presuppose certain states of affairs and, together with other statements, entail certain conclusions whereas acts are not said to presuppose or entail anything. My act of sneezing, for example, does not logically presuppose, imply or entail any other act or statement. If performatives are acts, then one would likewise expect that they would not presuppose, imply or entail anything.
Consider Austin’s examples of presupposition, implication and entailment. Here are three sentences he examines:
(1) ‘All John’s children are bald, but John has no children.’
(2) ‘The cat is on the mat, but I don’t believe it.’
(3) ‘All the guests are French, but some of them aren’t.’ [PC, 17]
With regard to (1), the first conjunct presupposes that John has children and thus that the statement ‘John has children’ is true. The second conjunct contradicts this. Thus the statement is involved in a contradiction since it denies what it presupposes.
Turning to (2), notice that, although the first part presupposes neither the second part nor its contrary (the cat may be on the mat with or without my believing it), to assert the first conjunct implies that one believes it (if one ignores lying, joking, etc.); it thus implies that it would be true for the utterer of the first conjunct also to assert ‘I believe the cat is on the mat’ and false for him to assert the contrary. Since he does assert the contrary, his assertion contradicts what his asserting it implies that he believes.
Similarly, the utterer of (3) contradicts himself. To take the first part of his claim, ‘All the guests are French’ entails that ‘It is not the case that some of the guests are not French’. But the second part of the assertion, viz. ‘some of them aren’t’, that is ‘Some of the guests are not French’, entails ‘It is not the case that all the guests are French’. So the first part of the sentence entails the contradiction of the second part and the second part of the sentence entails the contradiction of the first part. That is,
(3) "x (Fx ® Gx) & $x (Fx & - Gx)
(i.) "x (Fx ® Gx) ® -$x (Fx & - Gx)
(ii.) $x (Fx & - Gx) ® -"x (Fx ® Gx)
The first conjunct of (3) entails the contradiction of the second; and the second conjunct entails the contradiction of the first. Thus (3) is a contradiction.
Austin shows next that there are factors similar to presupposition, implication and entailment that concern performatives: “these three ways of failing to get by correspond to three of the ways in which a performative utterance may be unhappy” [PC, 18]. Here too there are three examples:
(4) ‘I bequeath you my watch, but I haven’t got a watch.’
orNot owning a watch, saying ‘I bequeath you my watch.’
(5) ‘I promise to be there, but I have no intention of being there.’
orNot intending to be there, saying ‘I promise to be there.’
(6) ‘I welcome you, but get to Hell out of my house.’
orIn the course of abusing a guest, saying ‘I welcome you.’
orSaying ‘I welcome you’, but proceeding to abuse the guest. [6]
From what has been said above about infelicities it should be clear that (4) is a misapplication since the convention of bequeathing is invoked but cannot be applied. Also (5) is an abuse of the institution of promising since it involves insincerity. And (6) is also an abuse but, in this instance, that of non-fulfilment. Austin in this context does not use all of these terms but it may help in what follows to have them fixed as such.
These three examples of performative infelicities show similarities with the invalid forms of reasoning associated with the constatives in the first three examples. Compare (1) and (4): just as ‘All John’s children are bald’ presupposes (given an existential interpretation of ‘all’) that John has children, so ‘I bequeath you my watch’ can be said to presuppose that I own a watch. [7] So the claim is stronger than merely pointing out that bequeathing a watch involves having one to bequeath; the claim is that there is an interesting similarity between this involvement and presupposition in so far as the prerequisite for genuinely bequeathing a watch is that one should own one, just as the prerequisite for making genuine assertions about John’s children is that John should have children (or that ‘John has children’ should be true or, at least, be believed to be true). With regard to (2) and (5): Austin says that
Just as my saying that the cat is on the mat implies that I believe it is, so my saying I promise to be there implies that I intend to be there. ...If we don’t hold the belief, or again don’t have the intention, appropriate to the context of our utterance, then in each case there is a lack of sincerity and abuse of the procedure [PC, 18f].
Just as making statements implies that one has certain beliefs, so similarly uttering certain performatives implies that one has certain intentions. If one in either case does not, then one is abusing either the convention of stating truly (although it is not clear that there is such a convention) or of promising. In the one case one will be generally expected to be telling the truth as one sees it and, in the other, promising something that one actually intends to perform.
Regarding (3) and (6): here Austin’s attempt to find a parallel is a little more strained. Entailment has to do with the compatibility of statements as to their truth-value: ‘All the guests are French’ and ‘Some of the guests are not French’ cannot both be true in the same universe of discourse. Making the first claim commits me to holding claims that are consistent with it. Saying ‘I welcome you’ also involves commitments, specifically to behave welcomingly and not abusively. Just as accepting ‘All the guests are French’ requires one also to accept ‘Some of the guests are French’, so, somewhat similarly, ‘I welcome you’ requires treating you welcomingly (especially in what one goes on to say) on pain of being incoherent or capricious. In greeting, just as in making assertions, one will be generally expected to be consistent (construing this word broadly enough to avoid equivocation).
So far it has been shown how (4) looks like (1); (5), like (2); and (6), like (3). Austin however also wants to see things the other way around where instead of (1), (2) and (3) being the paradigms, (4), (5) and (6) are. He wants to show that just as (4) seems to involve presupposition, (5) implication and (6) entailment, so, similarly, (1) seems to involve a misfire, (2) to involve an abuse of procedure (viz. insincerity) and (3) also to involve an abuse of procedure (viz. non-fulfilment).
Consider first (1) as a type of misfire. A performative that misfires (by means of non-play, misapplication, flaw or hitch) is said to be ‘void’ and so similarly may a statement that only purports to refer:
we can take over for [the] doctrine [of the constative] the term ‘void’ as employed in the doctrine of the unhappiness of the performative. The statement on the subject of John’s children is, we may say, ‘void for lack of reference’, which is exactly what lawyers would say about the purported bequest of a watch. So here is a first instance in which a trouble that afflicts statements turns out to be identical with one of the unhappinesses typical of the performative utterance [PC, 18].
Just as the performative is only purported, so also the statement is only a purported statement (on the ultimately perhaps unacceptable assumption that a statement must be exclusively either true or false); in both cases there is lacking some factor which could make the utterance happy. Alternatively one can say that in both cases there is presupposed some fact or event which does not actually obtain.
The case regarding (2) is clearer. Saying that the cat is on the mat implies that one believes it is because it is nonsense to say ‘The cat is on the mat, but I don’t believe it’. This is not to deny that both conjuncts can be true together, but only that they can be truly asserted together. Austin wants to say that there is an abuse of procedure here too. The convention of stating something is for stating truly to the best of one’s knowledge. Not to do so is to be insincere. Even if (2), as it stands, is convincingly characterized as insincere (and it is not), it is such nonsense that, rather than believing that someone who makes it is being insincere, one would believe that they were perhaps mentally ill. However if one did not believe that the cat were on the mat but asserted that the cat was on the mat, an alternative version of (2) [see PC, 17] , then one would clearly be being insincere and would be universally regarded as being so. Asserting what one does not believe might, in this case, be considered an abuse of the convention of assertion. [8]
Austin also shows that (3) may be construed as a non-fulfilment or ‘breach of commitment’. He rhetorically asks whether, having claimed that ‘All the guests are French’, I do not “commit myself in a more of less rigorous fashion to behaving in future in such-and-such a way, in particular with respect to the statements I will make?” [PC, 19]. Clearly like (6) there is abuse involved in (3). One will expect someone to derive the statement ‘Some of the guests are French’ rather than ‘Some of the guests are not French’ in a way similar to the way one expects someone who says ‘Welcome!’ to behave welcomingly.
In his comparison of the logic of constatives and performatives, Austin is successful in showing interesting similarities between the compatibility of conventional acts in society and the compatibility of utterances in rational speech. There is the similarity between denying what one presupposes and bequeathing what one does not own. Also, saying what one does not believe, or asserting that one does not believe what one says, is similar to promising something that one does not intend to fulfil, or promising something and saying that one does not intend to fulfil one’s promise. Likewise denying what one’s assertions entail is like committing oneself to a certain course of action and then acting in a contrary manner. Such similarities show at least that constatives and performatives are similar in the ways they are bound up with the conventions of social life. They both either fit or fail to fit happily into the total speech situation, or context, and they both commit the utterer to accepting other statements or to behaving in certain ways.
In HDTW Austin concludes this investigation of the logic of constatives and performatives by pointing out that “there is a danger of our initial and tentative distinction between constative and performative utterances breaking down” [HDTW, 54; cp. PP, 251]. Considering the happy-or-unhappy dimension of assessment he points out that it “may infect statements (or some statements)” just as consideration of truth-or-falsity “may infect performatives (or some performatives)” [HDTW, 55]. The language of danger, breakdown and infection is noticeable and I shall examine it more closely later. [9]
What should one conclude from all of this? That there is no constative/performative distinction? That there is no exact constative/performative distinction but only a fuzzy one? Should the concepts of constative and performative be given up? It does not follow that, because a distinction is not without exception, it is not a valid distinction. [10] Failure to find a criterion (or set of criteria) to distinguish constatives from performatives does not mean that there is no such distinction. The fact that one recognizes that the criteria do not always work shows that one is able to make the distinction but is simply not able to state a rule whereby one can distinguish them. However, it shows that there is a danger of the “initial and tentative distinction between constative and performative utterances breaking down” [HDTW, 54]. This appears to be Austin’s attitude, or fear, in the first five lectures of HDTW. However in Lecture VI and VII, before the introduction of his Speech Act Theory, Austin abandons his apparent worry about not finding some criterion that would be foolproof.
In fact Austin abandons the constative/performative distinction because his investigation of possible criteria leads him to appreciate a better way of dealing with the issues involved.
It is time then to make a fresh start on the problem. We want to reconsider more generally the senses in which to say something may be to do something, or in saying something we do something (and also perhaps to consider the different case in which by saying something we do something) [HDTW, 91f].
His investigation of the constative/performative distinction can be seen to serve as a warrant for his proceeding to treat all utterances as in a sense performative. [11] Now he will, instead of speaking of performatives and constatives, speak of speech acts. This need not be taken to imply that Austin thinks that a constative/performative distinction cannot be made. Obviously it can be, since it is, but it cannot be done precisely and even where there seem to be clear cases of constatives, the fact that they can be preceded by ‘I state...’ or ‘I assert...’ shows that they could conceivably be performative.
Austin’s investigation of the constative/performative distinction may be viewed (whether or not it was intended to be) as a dialectical investigation. It starts out with two apparently separate classes: utterances that are true or false but not happy or unhappy (statements or constatives) and utterances that are happy or unhappy but not true or false (performatives). But then with claims, as it were, from the performative side that some so-called constatives look somewhat performative, and claims from the constative side that some performatives have a constative dimension, the distinction begins to dissolve, as it were. One sees how much like performatives constatives are and vice-versa. The upshot of this dialectic is a synthesis of the performative and the constative as the speech act. I shall show in the next section how speech acts may be said to have performative and constative dimensions.
The speech act thus arises from a dissatisfaction with constative/performative terminology which in turn arises from dissatisfaction with the statement/nonsense distinction. Each step marks a better model for interpreting linguistic phenomena and is warranted by its plausibility and its strengths as compared to the previous model. One could say that Austin rejects constative/performative as an opposition between two types of acts but not as a distinction between dimensions of one act.
Ultimately therefore the constative/performative distinction is not abandoned because it is useless or fuzzy but because it seemed that it might break down on further investigation and, more importantly, a consideration of it leads to a better approach, one which can deal with linguistic phenomena, especially those that cannot be determined as exclusively either constative or performative, without making that distinction. Later I shall show how similar points may be made about the other distinctions, especially the normal/parasitic distinction.
Having examined what he calls the ‘sea-change’ [HDTW, 150] that his theory undergoes during the first seven lectures of HDTW, as a result of which all language use comes to be viewed as having a performative and a constative dimension, Austin proceeds in the remaining five lectures to investigate speech acts, i.e. those units of speech that have both a performative and a constative dimension.
The speech act can be investigated under three different headings: (1) as meaningful speech, (2) as speech with a certain conventional force, and (3) as speech with a certain non-conventional effect. Here (1) can be regarded as the speech act’s constative dimension while (2) and (3) can be regarded as together constituting its performative dimension. The first of these in turn can be investigated under three subheadings: (a) the production of the actual noises that are, so to speak, the ‘vehicles’ of meaning, (b) the production of certain words in certain syntactical order and in a certain language by means of the production of those noises, and (c) the production of the latter to communicate a specific message, usually but not necessarily about a concrete situation.
To introduce Austin’s terminology: the speech act as meaningful utterance is the locutionary act; as meaningful utterance with a certain conventional (performative) force, it is an illocutionary act; as meaningful utterance with a certain conventional force non-conventionally bringing about a certain effect, it is a perlocutionary act. [12] The locutionary act is at one level the production of certain noises and as such it is dubbed the phonetic act; through the production of those noises the speaker intentionally produces words in syntactic arrangements and, in this respect, the act is called a phatic act; finally through the production of words in syntactic arrangements, with certain intentions and in certain contexts, it conveys certain messages and is in this respect dubbed a rhetic act.
With regard to the scope of this section: various of Austin’s critics suggest alternative ways of sectioning the speech act. But I shall be concerned only with what Austin said about the various aspects of speech acts and whether that is acceptable. I propose first of all to investigate the locutionary act under its three headings and then the illocutionary act. The perlocutionary act will be mentioned briefly and mainly to indicate the limits of the illocutionary act. Then I shall investigate the problematic distinction between the meaning and force of an utterance, the question being why Austin does not assimilate the force of an utterance to its meaning. [13]
With regard to the locutionary act, Austin claims that in order for there to be a speech act certain noises must be produced by the human voice: “to say anything is ... always to perform the act of uttering certain noises..., and the utterance is a phone” [HDTW, 92]. This is obviously untrue, since one can say something by means of writing, the production of graphemes. There are also many other ‘vehicles’ (so to speak) of speech, other sign-systems such as semaphore, Morse code, smoke signals, etc. At one point however Austin allows that utterances can be in the form of writing. This is when he speaks of “the utterance (in writing) of the sentence” [HDTW, 57]. It is clear however that he considers spoken language to be the paradigm of utterance and writing to be its “rather crude” reproduction [see HDTW, 74]. I shall consider this matter in Chapter Four where I shall show how it can be criticized in such a manner as to unsettle the foundations of Austin’s speech act theory.
Before considering the phatic act I should remark that, whereas phones are just noises, phonemes are the sound-units of a particular language. So we must not take Austin to be distinguishing between phonemic and non-phonemic noises at the level of the phonetic act. His ‘phone’ is not yet a phoneme. Although Austin does not say this, what he goes on to say, as we shall see, calls for this. It is at the phatic level then that actual languages are first considered. Here one utters
certain vocables or words, i.e. noises of certain types belonging to and as belonging to a certain vocabulary, in a certain construction, i.e. conforming to and as conforming to a certain grammar, with a certain intonation, &c [HDTW, 92].
Here the phones become phonemes, which intentionally express words from the lexicon of a certain language, and are intentionally produced in an order prescribed by the syntactic rules of that language. The phones are produced as conforming to the phonemic, lexical and syntactic conventions of a certain language. I take it that this does not mean that the phemes (as the results of phatic acts are called) are always well pronounced or well formed sentences. One does not cease to speak a language if one mispronounces words within certain limits (for instance, native English speakers do not fail to speak Russian merely because they cannot roll, or trill, their r’s). [14] Also, one does not cease to speak a language if one makes certain syntactic errors, again within certain limits (such as, for instance, ‘If I would have been there, I would have seen it’). These limits would probably be determined by the ability of another speaker of the language either mentally to correct the mistake or to get the intended sense in spite of the mistake. [15]
To pass from the phonetic act to the phatic act one must have certain intentions conforming to certain conventions: one must intend one’s phones to express utterances that conform to the conventions of a certain language. The monkey that produces phones indistinguishable from those that the English speaker produces when he says ‘go’ does not say the word ‘go’ because he did not intend his phonetic act to conform to the conventions of English. His act is not an intentional act in accordance with conventions [HDTW, 96].
To show that merely uttering phones is not the same as uttering phonemes, words and phrases, consider the following example of Austin’s. One is asked the following trick question: ‘If cold water is iced water, what is cold ink?’ One responds: ‘Iced ink’. [16] Here one intentionally produces the phonemes /ist’ink/ but the phones one produced could also be interpreted as the phonemes /i’stink/ although they were not intended as such. Or, since Austin does not speak in terms of phonemes, one would have uttered the phones that go to make up the utterance of ‘I stink’ but one would not have uttered those words since one had not that intention as the context makes clear, the relevant context here being the fact that one was asked about iced liquids. This shows the importance of context of utterance: it is context, including the speaker’s intentions (i.e. ‘total’ context), that determines which phatic act the phonetic act gives rise to.
Intentionally conforming to linguistic conventions in specific contexts gives rise to rhetic acts which Austin describes as being “generally to perform the act of using [a] pheme or its constituents with a certain more or less definite ‘sense’ and a more or less definite ‘reference’ (which together are equivalent to ‘meaning’)” [HDTW, 93]. [17] It is clear, although Austin does not actually say so, that it is the total context that determines what rhetic act, if any, is performed in the performance of a phatic act. One can utter a pheme as an example of a piece of English, for instance, in which case it will not be a rheme (as the product of a rhetic act is called) since it will not be used to convey anything. Such a production of the pheme is a mere mention (although Austin does not use this term here). The context generally makes it clear how or whether the speaker intended to use the pheme. [18] We shall see later that Austin excludes mentions as not being serious speech acts but rather parasites.
The relation between phemes and rhemes is somewhat complex. To state the matter briefly first, change of context can affect the same pheme to produce different rhemes but context cannot affect different phemes to produce the same rheme. That is, the same pheme, or different tokens of the same type, can be used to express different rhemes in different contexts but different phemes, or tokens of different types, can never be used to express the same rheme. So different phemes cannot express the same rheme but the same pheme can express different rhemes. The rheme, in other words, is not the same as a proposition in that it is tied to a specific pheme. All Austin will allow is that two different phemes can be ‘rhetically equivalent’ [see HDTW, 97f].
The trouble with this distinction is that there seems prima facie to be no point to it. To say that two phemes are rhetically equivalent but do not express the same rheme seems to be gratuitous hair-splitting. However Austin claims that it is important to keep the distinction in mind. Further, he seems to think that, properly speaking, different phemes do not express the same statement since he says that rhetically equivalent acts express ‘the same statement’, putting the latter three words between inverted commas, but in another sense not the same statement where the identity of the rheme is in question, and here those three words are not put between inverted commas. I take it that these factors show that Austin may be at least suspicious of such entities as propositions. He may not want to allow for different expressions’ being of the same proposition.
I interpret Austin’s distinction as follows: the rheme is the product of the rhetic act; it is what is stated in a statement, promised in a promise, ordered in an order, etc. Various of these products may be equivalent in that they have meanings that would generally be regarded as in practice substitutable for one another. They are thus rhetically equivalent. If we interpret Austin in this manner we can see how it allows him to avoid idealist, or idealist-sounding, entities such as propositions. He thinks he would be committed (as we shall see) to some such metaphysical or language-transcendent entity were he to allow that two different phemes could express the same rheme. [19]
Austin’s move away from his theory of the performative to Speech Act Theory is a move away from such (possibly) ideal entities towards the concrete. He contrasts the two theories respectively as the ‘special’ and the ‘general’ theory, [20] and advocates the need for the general theory as one which would avoid the traditional theory’s problems:
the need for the general theory arises simply because the traditional ‘statement’ is an abstraction, an ideal, and so is its traditional truth or falsity....
...The total speech act in the total speech situation is the only actual phenomenon which, in the last resort, we are engaged in elucidating [HDTW, 148].
This clearly expresses Austin’s dislike of abstract entities such as the ‘statement’. Since the proposition is (or, at least, may be) such an abstract, ideal entity, I take it that Austin’s remarks similarly apply to it. [21] He thus stresses the rhetic act by contrast as a dimension of an actual phenomenon. Rather than saying that some information was communicated or conveyed, Austin prefers to see information as an effect. [22] We shall see that Searle does not think that his commitment to the proposition commits him to language-transcendent entities.
Next, I shall contrast the locutionary act with the illocutionary act. The question to be asked here is whether the force of an utterance is not part of its meaning (and thus whether Austin’s notion of locution is not too narrow). [23] Is it not true to say that ‘The cat is black’ means that the cat is black and that similarly ‘I promise’ means that I promise? To clarify the question, it is true that ‘The cat is black’ means that the cat is black and that ‘I promise’ means that I promise; but is the way that the first utterance means what it means the same as, or similar to, the way that the second utterance means what it means? For Austin, I suggest, the two senses of ‘means’ are to be distinguished and rightly so. Austin marks this type of distinction by calling the first the meaning or sense-cum-reference of an utterance and the second, the force of an utterance. Thus, ‘The cat is black’ said of a specific cat in a specific context means that that cat is black, in other words it refers to that cat and attributes blackness to it. On the other hand, ‘I promise’ means that I promise in that it has the force of bringing it about that I have undertaken to do something. For instance, if I have just said ‘I will be there’ (which might either be a prediction or a promise), the ‘I promise’ indicates that my utterance has the force of a promise and not of a prediction. It is in this sense of ‘means’ that one can say that ‘I will be there’ means that I promise to be there. To conflate these two senses of ‘means’ would instigate confusion: meaning, on one construal, is sense and reference and, on another, is force. In what follows I will use ‘means’ (and its cognates) only in its first sense.
In order to clarify the nature of illocution and to explain why Austin says that the illocutionary force of an utterance is not to be construed as a consequence of the locutionary act of uttering it [see HDTW, 114] , I shall now consider the perlocutionary act, which is said to be a consequence of the locutionary act, and to distinguish it from the illocutionary act.
The perlocutionary act, as already mentioned, is the bringing about of a certain effect by means of the use of language, that effect being non-conventionally brought about. A man who says to his wife, for instance, ‘I promise you a diamond ring’ may please her. There is no convention though whereby uttering ‘I promise you a diamond ring’, or promising something, or even promising specifically good things (even diamond rings), pleases its audience. The effect was purely ‘natural’, we may say. There is however a convention, as already indicated, whereby one who utters ‘I promise’ thereby promises. The utterance in question conventionally brings it about that a diamond ring was promised but non-conventionally brings it about that a woman was pleased. The conventional effect, to use provisionally the language of cause and effect, is the illocutionary effect and the non-conventional effect is the perlocutionary effect.
The act was an illocutionary act of promising and a perlocutionary act of pleasing. However, Austin warns that “we must avoid the idea ... that the illocutionary act is a consequence of the locutionary act” [HDTW, 114].
What we do import by the use of the nomenclature of illocution is a reference, not to the consequences (at least in any ordinary sense) of the locution, but to the conventions of illocutionary force as bearing on the special circumstances of the occasion of the issuing of the utterance [HDTW, 115].
Phatic acts and rhetic acts are not the consequences of phonetic acts. We have already seen why this is so: the noises must be made with the intention of conforming to the lexical and syntactic conventions of a particular language. Austin wants to point out now that illocutionary acts, similarly, are not the consequences of performing locutionary acts. They are not consequences “in any ordinary sense” because bringing something about by intending to conform to convention and by being understood to do so, is not a matter of effecting the state of affairs in question so much as holding it to obtain. Thus, saying ‘I promise’ does not causally bring it about that I promise; rather it is (or is constitutive of) the fact that I promise. [24]
However, in some cases it may not be possible to decide whether an act is illocutionary or perlocutionary. As an example of an act which could be construed as either illocutionary or perlocutionary, Austin mentions a man swinging his stick. This act may be equivalent to his saying ‘I warn you’ in which case it is illocutionary or it may be equivalent to his speaking with an (unintentional) ‘edge’ to his voice which serves as a warning to his audience (that he is not to be trifled with, for instance) in which case the fact that the audience is warned is a perlocutionary effect of his swinging his stick. The issue here is whether swinging one’s stick is conventional and, as Austin perceptively remarks, “it is difficult to say where conventions begin and end” [HDTW, 119]. The act could be classified either way. [25]
P.F. Strawson claims, against Austin, that there are illocutions that are not conventional. He gives three examples which will force us either to modify the above criterion, reject the illocution/perlocution distinction or account for the examples. Here is his first example:
Surely there may be cases in which to utter the words ‘The ice over there is very thin’ to a skater is to issue a warning (is to say something with the force of a warning) without its being the case that there is any statable convention at all (other than those which bear on the nature of the locutionary act) such that the speaker’s act can be said to be an act done as conforming to that convention. [26]
The other two examples are similar enough. With regard to this example, it is not clear that there is no such convention. On the one hand, one could say that there is an ethical principle that skaters tend to follow which requires telling other skaters about dangerous situations one has encountered. This is surely not just skaterly bonhomie! It is not obviously wrong to say that skaters do follow such a code of ethics or etiquette and expect others to follow it. [27] Strawson’s examples show at most that we must be cautious about this criterion of illocution.
On the other hand however, it could be said that Strawson is operating with too narrow a conception of convention as is apparent from his comment on a remark Austin makes about convention. Here first is what Austin says:
Speaking of the ‘use of “language” for arguing or warning’ looks just like speaking of ‘the use of “language” for persuading, rousing, alarming’; yet the former may, for rough contrast, be said to be conventional, in the sense that at least it could be made explicit by the performative formula; but the latter could not [HDTW, 103].
Referring to the ‘oddly qualifying remark’ here, Strawson comments that there seems to be no such sense of ‘conventional’ but merely of ‘being capable of being conventional’. [28] Austin’s point though, which may be being missed here, is that the fact that one could say ‘I warn you that...’ or ‘I argue that...’, which are conventional performative formulae, instead of simply saying, for instance, ‘The ice over there is very thin’, shows that making such statements as the latter is conventional in some circumstances. That one need not use the formula in question indicates that there is a convention to be appealed to implicitly by asserting that the ice is thin since otherwise one would have used the formula. [29]
Perhaps though it could also be argued that warning was a perlocutionary act in this instance. The speaker, this argument goes, did not need to appeal to any convention because he knew that mentioning that a certain part of the pond was covered with thin ice would be sufficient to get the skater to keep clear of it (by making him apprehensive or whatever). As I have already pointed out, Austin says that “it is difficult to say where conventions begin and end”. It is thus difficult sometimes to see where illocutions end and perlocutions begin. Strawson’s examples are thus valuable in showing that the illocution/perlocution distinction is somewhat fuzzy, a point which Austin would certainly accept (as I shall now try to show).
In the previous section I showed how Austin rejected the constative/performative distinction as a distinction between two exclusively different uses of language and continued to accept it only as a distinction between two different dimensions of acts of speech in general. Austin then, in the final five lectures (plus the end of Lecture VII), examines the speech act, which is the name given to units of speech with constative and performative dimensions. It is not an act made up of various components which fit together to constitute the total speech act but rather an act of many dimensions which cannot ultimately be sharply divided from one another.
The distinctions are largely matters of empirical research of the total speech situation and thus are not likely to be hard and fast. Austin admits that
typically we distinguish different abstract ‘acts’ by means of the possible slips between cup and lip, that is, in this case, the different types of nonsense which may be engendered in performing them [HDTW, 147].
Here Austin is presumably using the word ‘nonsense’ loosely to cover infelicity in general. In Lecture X he investigates the use of ‘in’ and ‘by’ to distinguish illocutions from perlocutions but ultimately finds that as tests they break down [see HDTW, 123]. Similarly in Lecture XII he attempts “a list of illocutionary forces of an utterance”, of “families of related and overlapping speech acts” [HDTW, 150]. Again he draws attention to the “wide possibilities of marginal or awkward cases” and emphasizes that what he has to say is not to be construed as definitive [see HDTW, 152]. Thus Austin eschews any kind of metaphysic of speech acts; he does not analyze them into various fixed categories, nor does he divide them into various fixed components, or purport to discover non-empirical entities such as propositions in them. [30] In a later paper, in fact, Austin mentions that what he does could be described as ‘linguistic phenomenology’ [see PP, 182]. It is investigating the way language is ordinarily used which demands describing linguistic practices in their everyday situation and in the situations where they for some reason or another are defective. Throughout HDTW Austin shows himself ready to redescribe the total speech situation as soon as one way becomes implausible. Even his final suggestions are offered as suggestions requiring much further work. Searle may be considered to have attempted some such further work. I move on to consider that now.
In this section I shall examine three basic issues. First, I shall consider Searle’s replacing the locutionary act with the propositional act (whose components are the reference act and the act of predication) and show why such a complication is warranted. Secondly, I shall consider Searle’s reasons for rejecting Austin’s locution/illocution distinction—basically he shows that the locutionary act, which he conceives of somewhat differently and calls the propositional act, is a dimension of the illocutionary act. And thirdly, I shall consider Searle’s additions to Speech Act Theory; I shall consider the development of his theory of utterance meaning away from its Gricean origins. I shall show, in the next chapter, how Searle’s theory of meaning enables him to formulate a more sophisticated account, than Austin’s, of the relations between normal speech acts and parasites. What I shall be mainly interested in, in this section, is Searle’s account of reference and predication, his development of Grice’s theory of speaker’s utterance meaning (what Grice calls non-natural meaning, or ‘meaningNN’), and finally his abandonment of such a theory by postulating a logical distinction between the intention to represent and the intention to communicate (the former intention being said to be independent of the latter and prior to it).
Searle accepts Austin’s rejection of the constative/performative distinction as a distinction between two different types of acts. He accepts that the speech act is the basic unit of meaning and force, or the most basic linguistic entity with both a constative and a performative dimension. He also accepts that there are illocutionary acts and perlocutionary acts. His understanding of the latter is similar to Austin’s but his understanding of the former is quite different. Searle does not distinguish between the illocutionary act and the locutionary act but rather between the illocutionary act and both an utterance act and a propositional act. In this section I shall examine why Searle rejects the locution/illocution distinction. As pointed out in the previous section, locution and illocution cover language as meaningful and language as having conventional force. The same phonetic act under one description was meaningful, which means that it had sense and reference, and under another description had a certain conventional force, which means that it counted as a conventional social act of a certain sort (such as ordering or promising).
Although Searle accepts that the speech act is both meaningful and of some conventional force, he analyzes the dimensions of the speech act differently. The major difference is Searle’s postulating a propositional act which is subdivided into a reference act and an act of predication. Searle thus accepts the proposition which, as we have seen, Austin’s scruples prevented him from embracing. He also speaks of the (incomplete) speech act of predication which Austin did not mention. Here is an outline of the two systems:
| AUSTIN | SEARLE |
(a) Locutionary Act:
(i) Phonetic Act, |
(a) Utterance Act. |
(b) Propositional Act:
(i) Reference Act, |
|
| (b) Illocutionary Act. | (c) Illocutionary Act. |
| (c) Perlocutionary Act. | (d) Perlocutionary Act. |
With this outline in mind I shall now investigate Searle’s analysis of the speech act with reference to Austin’s.
The most basic act in Searle’s system is the uttering of morphemes, words and sentences [see SA, 24]. A morpheme is an element of word-form which is functional in a linguistic system. It is thus very different to Austin’s phone. It is phones combined into certain types of units that have a function in a language. Thus the utterance act does not correspond to Austin’s phonetic act and, in fact, there is nothing in Searle’s system which does. This is not to say that he rejects the idea of a phonetic act though. He recognizes it but does not include it [see ALIA, 424].
The utterance act is a speech act without a determinate meaning. [31] To perform an utterance act without performing a propositional act would be to “utter words without saying anything” [SA, 24]. It would seem then that the utterance act corresponds roughly to Austin’s phatic act which was the act of uttering the vocables, words and syntactic units of a specific language. In short, since the utterance act is the producing of morphemes, words and sentences (without regard to whether they are being used or merely mentioned) and the phatic act is the production of vocables, words and grammatical units in a specific language (again without regard to whether they are being used to say anything or are merely being mentioned), the similarity here is close enough to warrant my proceeding with the provisional understanding that Searle’s utterance act is the same as Austin’s phatic act.
Searle’s propositional act does not correspond to Austin’s rhetic act though. Both of these acts concern language use as meaningful in the sense of having definite sense and reference. Searle however allows that different utterance acts can involve the same propositional act [see SA, 24] , whereas Austin, as we have seen, denies that the different phatic acts can produce the same rhetic act. Also, whereas Austin holds that there can be a rhetic act that is not illocutionary, Searle denies that there can be a propositional act without there being an illocutionary act. [32]
As outlined, one can investigate the propositional act under two headings: the reference act and the act of predication. The former is a complete speech act because one can refer to some object without saying anything about it. One cannot say something though without that something being putatively, if not actually, about some object; so the act of predication is an incomplete speech act. With regard to the reference act, Searle accounts for the use of referring expressions in reference acts as follows:
Any expression which serves to identify any thing, process, event, action, or any other kind of ‘individual’ or ‘particular’ I shall call a referring expression. ...It is by their function, not always by their surface grammatical form or their manner of performing their function, that referring expressions are to be known [SA, 26f].
So ‘a man’ in ‘A man came’ refers; but it does not refer in ‘John is a man’. This is clear from the fact that the expression only serves to identify a man in the first example. In the second it predicates the property manness of John. Clearly then ‘is a man’, which is a predicate, cannot stand on its own; it must accompany some referring expression. This is why Searle says that the act of predication “is not a separate speech act at all” [SA, 122].
Searle says that we must distinguish between the sense of a referring expression and the proposition communicated by its utterance. The sense is conveyed by the descriptive general terms given or implied by the referring expression “but in many cases the sense of the expression is not by itself sufficient to communicate a proposition, rather the utterance of the expression in a certain context communicates a proposition” [SA, 92]. So ‘the man’, for instance, has a sense independently of any particular context but only in some particular context will it be used to refer to an individual.
This means, I think, that an utterance of ‘The man is drunk’ construed as a mention, and not as a use in a specific context, is not a propositional act even though it has a sense. There is an act of predication involved but no reference act since the putative referring expression is not functioning, i.e. is not identifying anything in some context. It would seem then that Searle’s utterance act covers cases of meaningful utterances that do not refer to anything and do not express any proposition. Searle however claims, as we have seen, that utterance acts are not acts of ‘saying anything’. Now since ‘The man is drunk’ is not a propositional act it could only be a mere utterance act. So, while they have a sense, utterance acts do not refer to anything and thus express no proposition. To be clear about the matter then, ‘The man is drunk’ although it is literally meaningful (it means the man is drunk), may not express any proposition (it may not assert that any man is drunk).
The utterance act corresponds to Austin’s phatic act as already mentioned. The phatic act did not have a determinate sense or reference but, to use Forguson’s term (already mentioned), a determinable sense. The same, we now see, can be said of Searle’s utterance act. It requires a specific context in order for its referring expression actually to refer. The propositional act then, like the rhetic act, has by contrast a determinate sense and reference. The main difference between the latter two acts is thus the one already mentioned, viz. that the same proposition can be expressed in two utterances of different types whereas the same rheme can not. The other significant difference, already mentioned, is Searle’s discovery (for Speech Act Theory) of the act of predication.
The question to ask now is whether this different analysis (and I have shown above that it is only slightly different) adds anything to Speech Act Theory or, more specifically, anything that could not be arrived at with Austin’s concepts. Thus far we can see that Searle’s postulation of an act of prediction may be an advance on Austin. However the cost of this ‘advance’ is the adoption of the propositional act. The propositional act makes Searle’s view seem, if not actually be, more metaphysical than Austin’s. It is time therefore to consider whether there is in Searle such a reason for the propositional act that is not a reason for the rhetic act. Is the change worth the cost? The reason can be found in Searle’s postulation of deep structures in language.
I quoted above a passage from Searle that says that it is by their function that referring expressions are recognized and not, or not always, by their surface grammar. Thus ‘That man is drunk’ said here in this text does not refer even though the expression ‘that man’ may look like a referring expression. It is only a referring expression when it is used in a specific context to say of a specific man that he is drunk. It is therefore the function of the expression not its surface form that defines it.
The surface grammar of a sentence is what one thinks of as grammar simpliciter. One uses the word ‘surface’ to qualify it only when one has depth grammar in mind as a contrast, i.e., in this instance, when one has in mind a grammar of sentences concerned solely with their function. If one examines utterances in terms of how they function (or what acts they perform), then one can get a grammar that isolates such features as propositional indicators, referring expressions, predicates and illocutionary force indicators. The illocutionary force indicator, for instance, marks an utterance as either an assertion, an order, a promise or some such performative and may or may not be represented at the surface level by an illocutionary force indicating device such as an Austinian explicit performative. [33] Searle emphasizes that there need be no device at the surface level to indicate what sort of illocutionary act is being performed. Often context will make it clear in ‘actual speech situations’ what the force of an utterance is. Thus context rather than any device may determine the illocutionary force.
Context can also, although Searle does not seem to say so explicitly, determine the propositional indicator. For instance, ‘I’ll be there’ said by a speaker in a particular tone of voice may constitute a promise. Here the proposition expressed is that the speaker will be there and the illocutionary force indicating device is his intonation. [34] The two are not separate at the surface level. And ‘I promise’, which is an explicit performative, has a propositional content when said in response to ‘You’ll be there, won’t you?’. It has the propositional content that the speaker will be there.
It would seem that Searle’s innovations do pay the cost of his introducing the proposition because now it is possible to formulate a deep grammar of speech acts through considering how they operate in their contexts. And the proposition here need only be construed as a logical device for analyzing speech acts in their situation. There is no hint of metaphysics in any of this.
Having presented the differences between Searle’s theory and Austin’s I shall now address Searle’s specific criticisms of Austin’s theory, especially of his locution/illocution distinction, which lead him to his differently structured speech act. Basically Searle’s reason for rejecting the distinction is that, since meaning sometimes determines force, the distinction is not completely general. For instance, the meaning of ‘I promise’ determines the force of that utterance act as an illocutionary act of promising. It is by virtue of its meaning that ‘I promise’ counts as a promise.
To examine the details of this criticism: Searle rightly characterizes Austin as committed to the view that “Utterances which [are] different tokens of the same locutionary type [can] be tokens of different illocutionary types” [ALIA, 407]. Thus ‘I am going to do it’ may sometimes be mere prediction and at other times be a promise without its meaning changing. Searle formulates the criticism as follows:
it seems that [this distinction] cannot be completely general, in the sense of marking off two mutually exclusive classes of acts, because for some sentences at least, meaning, in Austin’s sense, determines (at least one) illocutionary force of the utterance of the sentence [Thus ‘I hereby promise that I am going to do it’] may on occasion be other illocutionary acts as well, but it must at least be a promise [ALIA, 407].
The example here is an explicit performative, an explicit promise. Austin of course would not deny that it will always be used with the force of a promise. Now Searle’s critical point here is that there is no locutionary act here.
To investigate this criticism: one could not convincingly deny that a typical utterance of ‘I promise...’ is an illocutionary act but, as we have seen in §2 above, it is not wrong to say that under one description it is a meaningful act and, under another description, an act with a certain force. Searle however is denying that one can abstract from the illocutionary ‘nature’ of the utterance to consider it solely in terms of locutionary meaning. In other words, he is saying that it can be described as an illocution but not as a locution. Conventionally, ‘I promise [something] ’ determines my promising that something by virtue of its meaning.
Although meaning here determines force, the force of the utterance is not its meaning in the same sense as the cat is on the mat is the meaning of ‘The cat is on the mat’. That would be to equivocate over two senses of meaning already outlined. Searle however shows that force can be assimilated to meaning to the extent that meaning determines force. In so far as he shows this, he shows that not all speech acts can be analyzed into illocutionary and locutionary acts since sometimes the illocution cannot be abstracted from. This therefore justifies his leaving the locutionary act out of his analysis of the speech act. And so his analysis of the speech act is in this respect an improvement on Austin’s. [35]
I now turn to Searle’s theory of meaning and communication in order to examine his further criticism of Austin which is as follows:
Austin sometimes talks as if in addition to the meaning of sentences there were a further set of conventions of illocutionary force; but in precisely those cases where there is a distinction between force and meaning, the force is not carried by a convention but by other features of the context, including the intentions of the speaker; and as soon as force is tied down by an explicit convention it becomes, or in general tends to become, part of meaning [ALIA, 414].
In other words it is either meaning, or intention and other features of context, rather than convention, that determines an illocutionary act which has no illocutionary force indicating device, as the illocutionary act that it is. So ‘How do you do?’ is an idiom whose meaning, according to Searle, cannot be paraphrased ‘In what manner or condition do you perform?’. It has the force of a greeting. But how does it have this force? Is it solely the speaker’s intention that determines it in a certain context as such?
Meaning may sometimes, it would seem, be a matter of intending a certain effect in a certain context. This is what H.P. Grice means by meaningNN (or non-natural meaning). [36] For my purposes the formulation of it given by Strawson will suffice.
S non-naturally means something by an utterance x if S intends (i1) to produce by uttering x a certain response (r) in an audience A and intends (i2) that A shall recognize S’s intention (i1) and intends (i3) that this recognition on the part of A of S’s intention (i1) shall function as A’s reason, or a part of his reason, for his response r. [37]
Searle points out that this makes communication, or saying something and meaning it, a perlocutionary act. Understanding the meaning becomes merely a consequence of the act (viz. a response to it) [see SA, 44]. But sometimes the only ‘effect’ of an act of communication is understanding (and not necessarily some response such as belief). In Austin’s terminology this is ‘securing uptake’ and in Searle’s terminology it is bringing about an ‘illocutionary effect’. Here is how Searle formulates it: “the speaker S intends [i-2] to produce an illocutionary effect IE in the hearer H by means of getting H to recognize S’s intention [i-1] to produce IE” [SA, 47]. Here we have the intention to be understood to be intending to bring about an illocutionary effect.
But sometimes, Searle shows against Grice, meaning is more than that. “Meaning is more than a matter of intention, it is also at least sometimes a matter of convention” [SA, 45]. Sometimes what one means is a matter of convention and sometimes all three factors (viz. intention, context, convention) need to be taken into account. A sentence may have a certain meaning independently of what I intend to effect by uttering it in a certain context. Meanings may thus be conveyed intentionally or conventionally.
Take Searle’s example of an American soldier who wants his Italian fascist captors to think him a German so that they will let him go. He can only recite a phrase of a German song to them to make them think him a German-speaker and thus a German soldier. What he intends to bring about in that context by means of reciting ‘Kennst du das Land wo die Zitronen blühen?’ is that they should believe that he is German. But this sentence does not mean what Grice is committed to holding that it means, viz. ‘I am a German soldier’. Its conventional meaning is still what would be translated into English as ‘Do you know the country where lemon trees blossom?’ even though the American intends that it should make the Italians believe that he is a German soldier. So the conventional meaning is different to the meaningNN à la Grice.
Searle derives the following scheme to explain conventional meaning:
S utters sentence T and means it (i.e., means literally what he says)=S utters T and(a) S intends (i–1) the utterance U of T to produce in H the knowledge (recognition, awareness) that the states of affairs specified by (certain of) the rules of T obtain. (Call this the illocutionary effect, IE)
(b) S intends U to produce IE by means of the recognition of i–1
(c) S intends i–1 will be recognized in virtue of (by means of) H’s knowledge of (certain of) the rules governing (the elements of) T [SA, 49f].
Here to understand T is to know its meaning and to know its meaning is to understand the rules for the use of T. So T has a conventional meaning which is expressible in semantic rules for the use of T. Now to use T with its conventional meaning, in other words to say T and mean it with its conventional meaning, is to intend that H understand that T' (the conventional interpretation of T, let us say); to intend him to understand it by virtue of his recognition of one’s intention that he should understand that T'; and to get him to understand T' by means of his knowledge of the rules governing the conventional use of sentence T. In this situation T is “a conventional means of achieving the intention to produce a certain illocutionary effect in the hearer” [SA, 48].
In SA Searle rejects the Gricean account of meaning in favour of the above account. The Gricean account cannot even be modified to take into account features of the American Prisoner example. One cannot
amend Grice’s account so that meaning is analysed in terms of understanding. That would be too circular for one feels that meaning and understanding are too closely tied for the latter to be the basis for an analysis of the former [SA, 47].
In later works, Searle comes back (before, in still later works, again abandoning) to what is basically the Gricean view (although as a supplement to his SA view which then becomes recognized, by contrast, as a notion of literal or sentence meaning). In E&M he distinguishes between literal, sentence meaning and speaker’s utterance meaning [see E&M, 77]. For instance, in introducing metaphor (which I shall investigate in the next chapter) he speaks of “utterances in which the speaker means metaphorically something different from what the sentence means literally” [E&M, 76]. Other instances where sentence meaning and utterance meaning ‘come apart’ are in irony and indirect speech acts. [38] These are
occasions where one succeeds in communicating what one means even though both the speaker and the hearer know that the meanings of the words uttered by the speaker do not exactly and literally express what the speaker meant [E&M, 77].
Utterance meaning here would seem to be Grice’s meaningNN modified to be illocutionary rather than perlocutionary. In Lewis’s words, in such cases “coordination [i.e., in effect, communication] will be achieved not by force of precedent but by force of salience”. [39] The context, in the case of utterance meaning, will make it clear what the intention was. That is, if the audience manages correctly to interpret what S intended, then he will have communicated with him (his meaning will have been communicated to him) by virtue of his having managed to make plain his intention to communicate what he intended. He will not have communicated by virtue of convention.
Before going on to show how Searle modified his view of utterance meaning (moving away again from the Gricean type theory), I shall account for the advantages of this development. It shows how Searle’s Speech Act Theory can explain how the force of an utterance may be determined by context when it is not determined by the meaning of a sentence. In order for someone to appreciate that one is making him a promise, even though one does not use the terms ‘I promise’, it is sufficient to tell him what one will do, in such a context as to make it clear that one is promising. For instance, if he is sick and one says to him ‘I’ll be around tomorrow to get you some groceries’, the situation will make it clear that one is promising. There is no need to postulate (Austinian) conventions of illocutionary force here.
Now I shall investigate how Searle modified his theory of meaning (i.e. as expounded in SA and E&M). In his later work (e.g. Int and MCR) Searle changes his mind concerning the analysis of (sentence and utterance) meaning. He now rejects “the idea that the intentions that matter for meaning are the intentions to produce effects on audiences” [Int, 161]. In the language of SA, uttering sentence T and meaning it will no longer be accounted for as the using of a conventional means to produce an illocutionary effect on some hearer. Likewise utterance meaning will not be accounted for in terms of intentions to produce an illocutionary effect in some hearer. In Austinian language, meaning will no longer be a matter of securing uptake. Searle thus separates the intention to represent from the intention to communicate one’s representation. He points out now that what is represented in not what is communicated: what is represented is a state of affairs, but what is communicated (if there is any communication) is the representation of the state of affairs (not the state of affairs itself) [see MCR, 213]. So Searle says that
representation is prior to communication and representing intentions are prior to communication intentions. Part of what one communicates is the content of one’s representations, but one can intend to represent something without intending to communicate [Int, 166].
One can intend to mean something without caring whether one will be understood or not. That is, one can intend to represent some state of affairs without intending to communicate that representation.
The means whereby utterances (or other sorts of acts) take on meanings is as follows: the act is performed with the intention that it should have the same conditions of satisfaction as the corresponding belief. Searle gives the example of a soldier signalling, by raising his arm, that the enemy is retreating. The soldier intentionally imposes on the act of raising his arm the conditions of satisfaction which his belief that the enemy is retreating would have. The conditions of satisfaction that his belief would have would be the enemy’s retreating; that is, his belief that the enemy was retreating would be true if the enemy were retreating and false otherwise. In general,
an intention to represent is an intention that the physical events which constitute part of the conditions of satisfaction (in the sense of things required) of the intention should themselves have conditions of satisfaction (in the sense of requirement) [Int, 167f]. [40]
Searle gives a schematic outline of the argument in which he distinguishes between the intention to represent and the intention to communicate:
Here points 1 to 3 concern the intention to represent and points 4 and 5 concern the intention to communicate that intention. The intention to represent is not tied to an intention to produce an illocutionary effect such as understanding.
Note that in this theory it is sufficient to intend that U of X should represent A for it to represent A, and thus to mean A. And we have seen that what applies to U of X would also apply to any other type of act (e.g. raising one’s hand). By intending that raising one’s hand should represent that the enemy is retreating, one makes one’s raising one’s hand mean that the enemy is retreating. The American Prisoner then, on this account, only had to intend that ‘Kennst du das Land wo die Zitronen blühen?’ should represent the state of affairs of his being a German soldier (which of course did not obtain) in order for it to represent that state of affairs. He also intended to communicate this representation of the state of affairs. This he did by intending2 that his Italian captors recognize his intending1 that that German sentence represented his being a German soldier.
We shall see later that this distinction between the intention to represent and the intention to communicate, together with the distinction between literal word or sentence meaning and speaker’s utterance meaning, enables Searle’s distinction between normal and parasitic speech acts to be defended from Derrida’s criticisms.
On this later view of Searle’s one can make any utterance or act mean what one wants it to mean by intentional fiat. This is a view widely rejected in contemporary philosophy. It is known as Humpty Dumpty’s view of meaning. In Lewis Carroll’s Through the Looking-Glass, Alice meets Humpty Dumpty with whom she has an argument. At one stage Humpty Dumpty says ‘There’s glory for you!’. Alice responds: ‘I don’t know what you mean by “glory” ’. Humpty Dumpty retorts: ‘Of course you don’t—till I tell you’. Alice couldn’t have known that Humpty Dumpty meant by ‘glory’ ‘a nice knock-down argument’ which is what he intended it should mean. [41] One would think that, in order for ‘glory’ to mean ‘a nice knock-down argument’ in this context, Alice would have had to have had some way of figuring out that that is what it meant. Donald Davidson, for instance, points out that
you cannot change what words mean ... merely by intending to; ... but you can change the meaning provided you believe (and perhaps are justified in believing) that the interpreter has adequate clues for the new interpretation. [42]
In other words, there must be some connection between meaning and communication. There must be some shared understanding in virtue of which Alice would be able to interpret the utterance as meant by Humpty Dumpty. The speaker and the hearer, or interpreter, must share some common understanding that will account for the meanings of their utterances. Davidson points out that this need not be a prior understanding. [43] There must be some feature of what Austin calls the ‘total speech situation’, to which the hearer, or interpreter, has access, which would make it possible for the hearer to interpret the utterance correctly.
However Davidson gives no reason why the intention to represent must be accompanied by an intention to communicate. One can admit that Humpty Dumpty was behaving very strangely by having a conversation with Alice without communicating to her that he was using some words with special meanings and without communicating those meanings to her. What was queer about what Humpty Dumpty did was not that he somehow acted illogically but that he said things to Alice without intending to communicate to her what he meant. In conversation one normally seeks to communicate. Humpty Dumpty conversed without seeking to communicate.
We shall see that Searle, in his account of parasitic discourse (e.g. in his theories of fictional discourse and of metaphor), uses such a pure intentional criterion to distinguish between normal illocutions and parasites. In a later chapter, after I have examined his theory of parasitic utterances, I shall come back to this difference between intentions (or what Austin, as we have seen, calls ‘internal’ circumstances) and ‘total’ context (which comprises both internal and external circumstances).
To conclude this section. I have shown how Searle analyzes the speech act differently. The major differences are (1) his introduction of the propositional act with it component acts of reference and predication, and (2) his introduction of a distinction between literal word or sentence meaning and speaker’s utterance meaning. In the next chapter I shall show how Searle supplements his theory of reference with a theory of parasitic reference which is important to this theory of fictional discourse. I shall also show how he utilizes the distinction between literal or conventional meaning and utterance meaning in his theory of metaphor. Both fictional discourse and metaphors are parasitic. Later too (in Chapter Five) we shall see how useful the distinction between the intention to represent and the intention to communicate can be in defending Speech Act Theory from criticisms brought by Derrida (which