‘”mumnnnmmmwmmm L LIBRARY 3 1293 00989 7616 Michigan 5mm University "4988 This is to certify that the thesis entitled CONVENTION AND INTENTION IN SPEECH ACTS: A RECONCILIATION OF AUSTIN AND GRICE presented by William S. Dibrell has been accepted towards fulfillment of the requirements for Ph. D. degree in Philosophy fii‘YII/l‘ C (U) (LE-1." Major professor 6/ 27 /7j 0-7 639 l Pa-kajgzfi‘, 2 NR 14":333 OVERDUE FINES ARE 25¢ PER DAY PER ITEM Return to book drop to remove this checkout from your record. CONVENTION AND INTENTION IN SPEECH ACTS: A RECONCILIATION OF AUSTIN AND GRICE BY William S. Dibrell A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Philosophy 1979 ABSTRACT CONVENTION AND INTENTION IN SPEECH ACTS: A RECONCILIATION OF AUSTIN AND GRICE BY William S. Dibrell Illocutionary acts, Austin argued, are always per- formed in conformity with conventions. Strawson and others have suggested that an intentional mechanism of the sort described by Grice in his theory of speaker meaning is basic in the performance of illocutionary acts. In this dissertation an attempt is made to show that these two views are compatible and further, that an intentional mechanism is involved in the performance of illocutionary acts whenever they are convention-governed. In the first of four chapters, Austin's approach to the study of language is examined. Austin's discussion of the performative-constative distinction gives rise to his View that all utterances are in some sense performa- tive. This View is defended against recent criticisms made by Urmson. In the second chapter Austin's view that illocu- tionary acts are conventional is examined in light of both Austin's discussion of performative utterances and subse- quent discussions of his view by Strawson and Searle. Because the counter examples brought against the view that illocutionary acts are always performed in accordance with William S. Dibrell conventions do not seem to be decisive, the issue is left open here. \4 acts involve an intentional mechanism of the sort described n the third chapter the thesis that illocutionary in Grice's theory of speaker meaning is examined. First it is argued that the intentionalist View is compatible with Austin's views on illocutionary acts. Second, the adaptation of Grice's analysis of meaning to an analysis of illocutionary act performance is defended on the grounds that counter examples to Grice's original theory fail when brought against the present analysis. In the final chapter attempts by Strawson and Searle to show a relationship between convention and intention in illocutionary acts is presented first. Second, Lewis's analysis of convention is examined so that it may be utilized to show an essential connection between the thesis that illocutionary acts are conventional and the thesis that they are intentional. The main argument here is derived from those made by Lewis and Bennett to the effect that communicative regularities maintained as conventions involve a Gricean intentional mechanism. Finally, it is argued that if Austin is correct in saying that illocutionary acts are conventional and if Lewis's analysis of convention is correct, then Austin's theory and Grice's theory are reconciled at a basic level. To My Father ii ACKNOWLEDGMENTS I am most grateful to my advisor, George Kerner, for his enthusiastic and consistent support. His comments were most helpful in keeping the project focused and in facilitating its completion. I am also indebted to the other members of my committee for their prompt and detailed criticisms and for their availability and willing- ness to discuss the issues. For their encouragement and tolerance, thanks to my colleagues, Martha Logsdon, Bill Parle, Joyce Pillote, Nolan Kaiser, and Paul Yu. Further thanks are due to Dr. Yu for his suggestions and for listening to my argu- ments. Many thanks to Marge Farmer for all her help, and to the students who typed the first drafts, Renée Patterson, Marsha Luhman, Terri Teal, Marilyn Sanchez, and Jan Hughes. Finally, a very special thanks to Jeanette, who typed the final draft and provided the needed emotional support during this project. iii TABLE OF CONTENTS Page INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . 1 Chapter 1. THE GENESIS OF AUSTIN'S THEORY OF SPEECH ACTS O O O O C O O O O O O O O O O O 6 INTRODUCTION 0 O O O O O C O O O C O C O O O O 6 LANGUAGE AS ACTION . . . . . . . . . . . . . . 8 PERFORMATIVE UTTERANCES . . . . . . . . . . . 10 PERFORMATIVE COMPARED WITH DECLARATIVE OR CONSTATIVE UTTERANCES . . . . . . . . . . 12 THE SUBSTANTIAL COMPARISON OF PERFORMATIVE AND CONSTATIVE UTTERANCES . . . 18 URMSON'S ATTEMPT TO RESIST THE LIBERALIZATION OF THE CONCEPT OF A PERFORMATIVE . O O C C O O O C O O O O O O 23 S UMMARY O O O O O O O O O O O O O O O O O O O 3 3 2. ILLOCUTIONARY ACTS AND CONVENTION: AUSTIN, STRAWSON, AND SEARLE . . . . . . . . . 35 INTRODUCTION . . . . . . . . . . . . . . . . . 35 ILLOCUTIONARY ACTS: PRELIMINARY IDENTIFICATION . . . . . . . . . . . . . . . 39 AUSTIN ON ILLOCUTIONARY ACTS AND CONVENTION . . . . . . . . . . . . . . . 43 AN OBJECTION TO THE VIEW THAT ILLOCUTIONARY ACTS MUST BE CONVENTIONAL I O O O O O O O O O C O O O O O 49 iv Chapter 3. 4. STRAWSON'S CHALLENGE TO AUSTIN'S DOCTRINE THAT ILLOCUTIONARY ACTS ARE CONVENTIONAL . . . . . . . . . STRAWSON'S MODIFICATION OF AUSTIN AND PROBLEMS WITH STRAWSON'S VIEW SEARLE ON ILLOCUTIONARY ACTS AND CONVENTION . . . . . . . . . . . CONCLUSIONS . . . . . . . . . . . . ILLOCUTIONARY ACTS AND INTENTIONS . . INTRODUCTION . . . . . . . . . . . . HOW TO TALK ABOUT INTENTIONS AND ACTIONS O O O O O C O O O O O AUSTIN ON INTENTIONS AND SPEECH ACTS O O O O C O O O O O O O O O O SPEAKER MEANING AND ILLOCUTIONARY FORCE O O O O O O O O O O C O O O STRAWSON'S ACCOUNT OF SPEAKER MEANING AND THE ANALYSIS OF ILLOCUTIONARY ACTS . . . . . . . ESSENTIAL AVOWABILITY . . . . . . . TOWARD DEFENDING AN ADAPTATION OF GRICE'S ANALYSIS OF MEANING TO AN ANALYSIS OF ILLOCUTIONARY ACTS THOUGHTS OF SPEAKERS AND HEARERS . . SUMMARY 0 O O C O O C O O O O O O 0 THE RELATION BETWEEN CONVENTION AND INTENTION IN THE PERFORMANCE OF SPEECH ACTS O U C C I C O O C O 0 INTRODUCTION . . . . . . . . . . . . AUSTIN ON THE RELATION BETWEEN INTENTION AND CONVENTION IN THE ANALYSIS OF ILLOCUTIONARY ACTS . STRAWSON AND THE "CONVENTIONAL MEANS" FOR PERFORMING ILLOCUTIONARY ACTS Page 55 59 67 81 84 84 86 94 101 112 116 119 134 143 147 147 149 152 vi Chapter Page SEARLE ON THE CONNECTION BETWEEN INTENTION AND CONVENTION FOR THE PERFORMANCE OF ILLOCUTIONARY ACTS . . . . . 156 LEWIS ON CONVENTION . . . . . . . . . . . . . 159 OBJECTIONS TO LEWIS'S MUTUAL KNOWLEDGE CONDITION AND THE WEAKENING OF THE CONDITION . . . . . . . . . 166 COMBINING LEWIS AND GRICE . . . . . . . . . . 177 SUB-GRICEAN CONDITIONS, CONVENTIONS AND THE PROBLEM OF COMMON KNOWLEDGE . . . . . . . . . . . . . . . . . 181 COMBINING GRICE AND AUSTIN . . . . . . . . . . 185 CONCLUSIONS . . . . . . . . . . . . . . . . . 188 BIBLIOGRAPHY O O O O O O O O O O O O O C O O O O O O O O 191 INTRODUCTION The central purpose of this thesis is to show a way that we can combine Austin's view that illocutionary acts are conventional with QMXEEHLPQQEWthéfi thatmspeaker intentiiqnsmars 1961.51.79 .i.n.._. .4??? minimise A illocmionary acts wewperform. The purpose, put more succinctly, is to show how we can View illocutionary acts as both conventional and intentional. The principal thesis that will be examined in the concluding chapter is that if illocutionary acts are conventional then speaker intentions have an essential function in the performance of illocutionary acts. The problem of the relation of convention and intention in the performance of illocutionary acts has its roots in the work of J. L. Austin. Austin claims that illocutionary acts, acts like arguing, stating, and warning, are always governed by conventions. The kinds of conven- tions he has in mind are first discussed by Austin in his papers on performative utterances. The first chapter of this thesis traces the discussion of performative utter- ances to the point at which Austin argues that we should view all utterances as performative, rather than view some utterances as performative and others as constative. Austin's view is defended against some recent objections 2 raised by Urmson who attempts to show that Austin's early theory which urges that statements are not performatives is basically more valid. The second chapter develops the views of Austin, Strawson, and Searle on illocutionary acts and convention. Austin's views are considered first since the terms of the discussion originate with Austin. Illocutionary acts (the central category intended to cover the performative aspect of any utterance) are viewed by Austin to be conventional. I argue that this in part derives from the conventional status of obvious performatives like, "I do," said in a marriage ceremony. Two objections to Agétifijs view are considered which cast doubt on the thesis that illocution- ary acts must be conventional. Chapter 2 concludes with Strawson's and Searle's views on the ways in which illocu- tionary acts may be conventional. Searle's view is developed at some length. The third chapter takes up the possibility that we may be able to utilize Grice's analysis of Speaker meaning as an analysis of illocutionary acts. It is argued first that talking about intentions to communicate in a quite general way is consistent with Austin's discussions of intention and intentions necessary for the happy perform- ance of certain speech acts. Having made this preliminary discussion, Grice's first account of speaker meaning is reviewed and a modifi- cation of Grice's theory, suggested by Strawson, is 3 discussed. Strawson's modification is intended to show that a further condition, namely, avowability, is needed in Grice's analysis. Strawson utilizes his modification of Grice's theory to elaborate on Austin's analysis of illocu- tionary acts by trying to show why insinuations cannot be illocutionary acts. Strawson's success in making his case is disputed on the grounds that his avowability condition is superfluous in showing that insinuating is not an illo- cutionary act. Grice's analysis is defended as an analysis of illocutionary act performance by showing that certain pro- posed counter examples to his analysis have no force against the claim that Grice's conditions for speaker mean- ing are conditions for illocutionary act performance. Finally, a possible weakening of the Gricean con— ditions offered by Bennett is considered. The weakening is considered in order to show how we could maintain an intentional analysis of illocutionary acts without commit- ting us to a theory which requires conscious awareness of speaker intentions in communication. In the final chapter of the thesis an attempt is made to specify the relation of intention and convention in the performance of illocutionary acts. The first sec- tions review the views of Austin, Strawson, and Searle on this tOpic. Strawson and Searle both offer views which center on the thesis that we use conventional means to Inake our communicative intentions manifest. This is an 4 important thesis but can be supplemented in such a way as to show a more fundamental link between convention and intention in speech acts. The essential link depends upon having an analysis of convention. The analysis of convention that is devel- oped and, to some extent, defended is taken from Lewis. Lewis and Bennett both argue that the proper analy- sis of convention shows that any linguistic behavior which is sustained by convention is intentional, intentional in the way Grice has claimed cgmmgnication_ishintentional. They do not argue that all communication is conventional, but the argument offered does indicate a one-way connection between convention and intention in the performance of illocutionary acts. This important argument points to a way to partially unify the views of Austin and Grice. It is plausible to argue that Lewis and Austin both take the ultimate purposes of linguistic behavior to be the securing of perlocutionary effects. For example, we argue in order to convince people; convincing peOple is a perlo- cutionary effect. Of course, for the success of perlocu- tionary acts, much more than the performance of illocution- ary acts may be needed. For example, when you make an argu- ment, whether or not anyone is convinced depends upon a wide range of things, such as, attitudes and prior beliefs among others. This, however, does not alter the fact that we normally perform illocutionary acts in the hOpe of pro- ducing perlocutionary effects. 5 Lewis's analysis of convention takes the serious interests of members of a population who are parties to conventions as the basis of his analysis. This fundamental point of agreement lends some support to the View that we can take Lewis's account of convention as being at least consistent with Austin's general views. Having combined Lewis and Grice, a way is provided to combine Austin and Grice. Chapter 1 THE GENESIS OF AUSTIN'S THEORY OF SPEECH ACTS INTRODUCTION The purpose of this chapter is to lay the groundwork for succeeding chapters which will be primarily a discussion of the relationship between "intentionalist" and "convention- alist" views on speech acts. We shall review Austin's initial treatment of the performative-declarative (consta- tive) distinction and show how the performative-declarative dichotomy gives way to a more general doctrine of performa- tive forces.1 The genesis of the doctrine of illocutionary acts, it is argued in this chapter, is important for two reasons: Austin's initial isolation of the illocutionary act as part of the total speech act is most clearly seen in this genesis, and Austin's claims, considered in Chapter 2, that illocutionary acts are essentially conventional, can be illuminated by considering his views on and examples of what he initially called performatives. lSearle, among others, remarks that in How To Do Things With Words the original distinctions between perform- atives and constatives is replaced by a general theory of speech acts. J. R. Searle, "Austin on Locutionary and Illocutionary Acts," Essays on J. L. Austin (London: Oxford University Press, 1973), p. 141. 7 The first section of this chapter sets forth one of Austin's most basic insights, that language may be fruit- fully understood as consisting of actions, in contrast to understanding language as a system of statements or proposi- tions, each of which is either true or false. Austin does not argue that analysis of statements or propositions is unimportant, but rather that a complete theory of language must include theorizing which attempts to provide an under- standing of the nature of language as a form of action. The second section traces the main elements in Austin's discussion of performative utterances and grounds the third section which brings out Austin's key points on the similarities that obtain between performative and declarative utterances. The fourth section deals with the shift in Austin's view from an emphasis on the performative-declarative (performative-constative) distinction to his introduction of the notion of performative force. It will be shown that the notion of performative force is more basic than either performative or declarative categories and that it is more basic because this notion is intended to capture one sense in which every utterance is a kind of performance. The final section of this chapter is a consideration of Urmson's misgivings about the plausibility of Austin's transition from the distinction of performatives and declar- atives to a doctrine centering around performative forces. Austin's position is defended against Urmson's criticism. LANGUAGE AS ACTION In "Other Minds," Austin remarks that language in its origins was not primarily descriptive. To view lan- guage on a model which construes language as primarily descriptive, Austin would argue, is misleading from the outset. Austin's remark in "Other Minds" is a good place to begin a discussion of Austin's most general view of language because it is the first such general thesis in Austin's work and it is a clear statement of one of Austin's most influential claims. Certain utterances, Austin points out, are the performance of an action and not the description of an event or an action performed. It is the utterance which is a performance of an action that Austin considers basic to linguistic behavior, and understanding such an utterance is basic to understanding language. Austin isolates the sort of utterance he has in mind as active when he claims that "utterances of obvious ritual phrases, in appropriate circumstances, is not describing the action we are doing, but doing it . . . ."2 Two points are made in "Other Minds" which are relevant here: language is active, and this is most clearly seen in ritual utterances. Two reasons for making these 2J. L. Austin, "Other Minds," Philosophical Papers, eds. J. O. Urmson and G. J. Warnock (London: Oxford University Press, 1970), p. 103. 9 points are, first, philosophers have tended to ignore the "action utterance" in favor of a concentration on the descriptive utterances, and, second, the idea that linguis- tic behavior is fundamentally active is developed as these points are refined. The first of the purposes just mentioned is stated clearly in "Performative Utterances."3 Although Austin applauds the mission of sorting out sense from nonsense, the movement of verificationism in the philosophy of lan- guage had the unfortunate tendency to ignore a major portion of our linguistic behavior. All interesting (cognitively significant) utterances, it had been argued by many, were either true or false and, at least in principle, in some weak sense verifiable. Without going into the pros and cons of verificationism here, the points Austin argued were simply that not all we say is either true or false, and the position that only those utterances which are either true or false have meaning is too restrictive from the outset. A discussion of those utterances which are neither true nor false will follow in the next section. As for the claim that verificationism was too restrictive, too much can be made of this point. Austin's approach to philoso- phizing about language does not rule out a concern with 3J. L. Austin, "Performative Utterances," Philosophical Papers, eds. J. O. Urmson and G. J. Warnock (London: Oxford University Press, 1970), p. 235. 10 truth conditions of statements. What Austin does hold out for is another way of philosophizing about language which will be able, in some way or another, to account for all those utterances which are not, or at least are not simply, either true or false. If any important common features can be shown to exist among all the kinds of utterances we make, part of the project of showing those common features is accomplished when we do not at the outset limit our concern to only one common kind of utterance--the declarative. In summary, Austin eschews theorizing which con- cerns itself with only one type of linguistic behavior (statement making). He does this because not all utterances are statements; furthermore, if there is a hope of finding what all utterances have in common, then restricting our- selves to consideration of only one kind of utterance is a mistake, no matter how important such considerations are in many other ways. PERFORMATIVE UTTERANCES Performative utterances, as initially conceived by Austin, are those which resemble statements grammatically, and should be classed with statements in that they are not nonsense but at the same time are not true or false. Grammatically these are often utterances in ". . . first 4 person singular present indicative active." And we would 4Austin, "Performative Utterances," p. 235. 11 say that a person, when using a performative, is primarily going rather than saying something.5 Examples of the kinds of utterances Austin has in mind can be seen in the follow- ing passage: Suppose, for example, that in the course of a marriage ceremony I say, as people will, 'I do'-- (sc. take this woman to be my lawful wedded wife). Or again, suppose that I tread on your toe and say 'I apol- ogize.‘ Or again, suppose that I have a bottle of champagne in my hand and say 'I name this ship the Queen Elizabeth.‘ Or sgppose I say 'I bet you Sixpence it will rain tomorrow.‘ The kinds of utterances involved in ritual activi- ties that Austin mentioned in "Other Minds" are then those which are involved in marriages and ship christenings, but also, in Austin's view, more mundane activities like apol- ogizing. These kinds of utterances are not nonsense, clearly are not in any simple way true or false, and fit Austin's grammatical characterization, that is, are in the "first person singular present indicative active." This initial identification of performatives is the first step in the development of Austin's general view of language. The next steps in this development are made when Austin details the comparison of performative and declara- tive types of utterances. 5Austin, "Performative Utterances," p. 235. 6Austin, "Performative Utterances," p. 235. 12 PERFORMATIVE COMPARED WITH DECLARATIVE OR CONSTATIVE UTTERANCES The comparison of performative and declarative utterances results in Austin's first stage in the detailed study of language. The comparison results in a general doctrine which eventually focuses on the performance of illocutionary acts. This is the principal reason to review the comparison. After the initial identification of performatives Austin develops his thesis in roughly three stages. First, he identifies rules for the successful performative. Second, grammatical criteria for the identification of per- formatives are postulated and investigated. These criteria are not adequate to identify performatives, but rather sug- gest the similarity of performative and constative or declarative utterances. Third, a comparison of performatives and constatives is made on the basis of comparing the sorts of failures to which each is subject. Austin argues that the failures that both performative and constative are subject to are so similar that there is reason to pursue a theory which exploits this common ground. The overall effect of this three-stage program is to emphasize the basic common features of performatives and declaratives, establishing that performatives do not deserve treatment as a special class of utterances as much as that all utterances should be understood as performative. 13 The general requirements that must be met for a successful performative are: (l) A conventional procedure for the performative utter- ance in question must exist and be accepted. (2) Circumstances for the invocation of the procedure [mentioned in (1)] must be appropriate. (3) The conventional procedure must be carried through correctly. (4) Some procedures require that the person invoking the procedure in their performance have certain appropri- ate beliefs, feelings, or intentions. These rules for a successful performative utter- ance are not claimed by Austin to be exhaustive, but certainly central. Using Austin's example of the ship christening we can see that indeed conventional procedures exist for the use of the performative, "I name this ship the Queen Elizabeth." But, of course, for such a perform- ance to have force you need a ship, the proper time, and other circumstantial prerequisites. The procedures need to be ordered correctly, the words must be uttered correctly, and the prOper person is necessary to utter the words. Of course, the performance could be done with or without sincerity. 14 The grammatical criteria7 which seem to identify performatives serve a twofold purpose. First, grammatical criteria serve to indicate a distinction between performa- tives and declaratives initially. Second, examination of the grammatical criteria ground the important comparison of performatives and declaratives. The grammatical criteria Austin suggests for identi- fying performatives (criteria briefly mentioned earlier) are: (1) Performatives are often in the ". . . first person singular present indicative active."8 (2) There is an asymmetry between the first person present tense use of performative verbs and the use of the same verbs with other persons and in other tenses. This second criterion is nicely illustrated by the example: "I promise to be at the cabin," which certainly has a different force than, "I promised to be at the cabin," or "He promises to be at the cabin." The asymmetry is that in the first case we have a promise being made while in the other cases a statement is made about an action of promising. 7Austin, "Performative Utterances," p. 241. This way of speaking, of course, depends on what kinds of consid- erations are included under the heading, ”grammatical." The adequacy of traditional grammatical classifications are called into question by Austin in Lecture VII in How To Do Things With Words, but this does not affect the present points. 8 Austin, "Performative Utterances," p. 235. 15 One standard performative form then is the first person form where there exists an asymmetry between this first person use and other uses. In addition to the first person formula for perform- atives, Austin notes that many performatives are in the passive voice and the second or third person. Examples Austin mentions are: "Passengers are warned to cross the line by the bridge only," or the frequently used locution, "You are hereby authorized to . . . ." Now a suggestive problem, as Austin sees it, is that performatives do not necessarily always have either of these two forms. It raises the question as to whether or not we could always render performatives in "standard" forms. Performatives do not always fit the standard forms as they stand, but perhaps, as Austin observes, any perform- ative could be expanded or reformed in such a way as to fit a standard form. The idea that performatives could always be revised to a standard form is the first premise in the comparison of performatives with declaratives, a comparison which, as mentioned before, issues in a shift in emphasis from per- formatives as a special class of utterances to an emphasis on the performative nature of all utterances. Austin observes that within "primitive" or natural language, the kinds of actions we are performing in utter- ing many of the things we utter is in large part made evi- dent by devices like tone of voice, situation, and various 16 things which are not part of the corresponding written locution.9 In a more developed language the indications as to what kind of action is performed in making an utter- ance are an explicit part of the utterance itself. In an ideal language, one of Austin's devices for introducing his important general theory, we could expect to have a completely explicit language, a language where there were no ambiguities in speech and no reliance on things like tone of voice to make clear performative force. Further, we could expect, if the ideal were reasonable, that we could eventually classify the kinds of actions that can be per- formed by performative utterances by comparing and listing the various performative verbs. Austin, in fact, worked in this direction. The idea of making explicit the performative aspect of utterances has two sides. In the preceding, there is a tendency to talk about the performative aspect of a special class of utterances, "the performatives," as well as a hint that the performative aSpect of utterances is a quite general concern. These two sides to the issue of explicit performative form can be summarized in two important 91 qualify this by using "written locution" because tone of voice is clearly a part of the phonological aspect of locutions. It may be that tone of voice is nearly as unambiguous as is ever actually required in communication, but the concern with a reform which does not depend upon tone of voice and things of this sort for making clear the particular uses of utterances facilitates the development of Austin's position. 17 premises which support Austin's shift from concern with performatives as a special class to the view that all utterances have a performative aspect. The two premises are: (l) Performative utterances can be rendered in an explicit form, i.e., "Class dismissed," could be rendered, "I hereby authorize you to go." (2) Explicit performative formulae can be developed for utterances that are not performative in the original sense identified.10 If both (1) and (2) are feasible, then a distinction between performatives and declaratives (or constatives) is not supported on the grammatical basis that performatives have special forms. In carrying out the business of making explicit performatives, Austin finds support for point (2). Consta- tives, as Austin calls them, can be put into standard form using a number of different verbs: "I state that so and so," "I argue that so and so." Austin argues that: In fact in general we may remind ourselves that 'I state that . . .' does not look so very different from, 'I warn you that . . .' or 'I promise to . . .' It makes clear surely that the act that we are performing is an act of stating, and 5? functions just like 'I warn' or 'I order' . . . .1 10That is, explicit performative forms may be given to utterances not originally identified as doing an action, but which are rather identified as just "saying something." llAustin, "Performative Utterances," p. 247. 18 It does not appear then that there is a strong reason to maintain that performatives are special, at least not on the basis of the grammatical features initially identified. "I state that . . ." looks to be as performative as "I promise that . . . ." To summarize thus far: (1) Performatives often have forms in which the action being performed is manifest; these are identified by Austin as explicit performative forms. (2) Those performatives that are not in an explicitly performative form could be put into one. (3) Constatives rarely have a form in which the action aspect of the utterance is explicit. (4) Still, constatives too can be given a form in which the action being performed is explicit. But even though there does not seem to be any good reason to distinguish performatives and constatives on the basis that there are grammatical differences, we must see whether or not there are perhaps still important substantive differences between them. THE SUBSTANTIAL COMPARISON OF PERFORMATIVE AND CONSTATIVE UTTERANCES Statements are liable, Austin argues, to all the same sorts of failures as are performatives. His support for this is brought out clearly in his paper, "Performative— Constative" from which the following is in large part l9 extracted. Constatives may fail depending on what they presup- pose, imply, or entail, according to Austin.12 These failures are parallel to the main ways performatives can be unhappy. The parallels may be brought out by examining cases which Austin offers in order to demonstrate his point. I will pair the cases according to the similarity of their failure: Cl Someone says 'All John's children are bald, but [or 'and'] John has no children;' or perhaps he says 'All John's children are bald,‘ when, as a matter of fact, John has no children. P 'I bequeath my watch to you, but [or 'and'] I haven't got a watch;' or perhaps someone says 'I bequeaih my watch to you' when he hasn't got a watch. 3 In both of these cases Austin argues that there is a cer- tain presupposition invoked and that in each of the cases the presupposition doesn't hold. In the first, the failure is due to the failure of meeting the condition that John has a child. The second fails due to the lack of a watch; it is presupposed that we have what we give away. The next pair of examples fail due to a failure to v‘ *7 12For a discussion of the different criteria of entailment, implication,and presupposition, see J. L. Austin, How To Do Things With Words, eds. J. O. Urmson and M. Sbisa (Cambridge: Harvard University Press, 1962), pp. 48-50. 13J. L. Austin, "Performative-Constative," The Philosophy of Languagg, ed. J. R. Searle (London: Oxford University Press, 1971), pp. 17-18. 20 meet an implication requirement. Austin argues that the performative and constative again look alike. C2 Someone says 'The cat is on the mat but [or 'and'] I don't believe it is;' or perhaps he says 'The cat is on the mat,‘ when, as a matter of fact, he does not believe it is. P2 'I promise to be there, but [or 'and'] I have no intention of being there;' or perhaps someone says 'I promise to be there' when he doesn't intend to be there. 4 We do generally think that people believe what they say. To assert some matter of fact does seem to involve committing oneself to a certain attitude toward the proposition. This, Austin argues, is parallel to a promise where we are expected to have a certain attitude toward our promises, the intention to do the thing promised. Both C2 and P2 involve sincerity. It would appear that performatives and constatives are exactly alike with regard to the failures to which they are subject; however, this is not quite the case. The grounds for criticizing performatives are claimed to be only similar but not identical on the entailment condition. The entailment condition is the requirement that one avoid contradicting oneself in discourse. Of course, this will happen sometimes almost unnoticeably; however, a blatant case is certainly ruled out of normal discourse. For example: Austin, "Performative-Constative," pp. 17—18. 21 C3 Someone says 'All the guests are French, and some of them aren't;' or perhaps he says 'All the guests are French,‘ and then afgerwards says 'Some of the guests are not French.‘l Austin claims that the sort of problem which makes the above sort of case an abuse of language is very much like cases of welcoming someone to your home and then abusing the person or, better still, failing to abide by your own stipulative definitions. Austin claims only that these last examples are analogous to C and are not identi- 3 cal, as the pairs C and P C and P2, were claimed to be. 1 1' 2 Further parallels between constatives and perform- atives can be made. For example, Austin argues that just as certain performatives require that the performer be in a special state or hold a special position to be successful, so do statements. This is surprising in that we usually, as Austin notes, think that no special warrants are needed for making statements. Austin suggests that where state- ments are concerned we can be challenged if we do not have either proper evidence for our statements or the appropriate position with regard to the purported information. To say there are fifty peOple in the next room is not to impart a bit of information unless the speaker has some evidence for this. What Austin is pointing out here is that in normal discourse we take it for granted that "factual" claims have some evidential basis and that this is more or less 15Austin, "Performative-Constative," pp. 17-18. 22 acknowledged by all. When factual claims are made with no warrant behind them there is a failure of authority somewhat like the failures of authority performatives are subject to. Ordering a peer to do something fails because we are not in a position to order peers. Stating that there are fifty people in the next room with no evidential basis is inappro- priate because we are simply not in the right sort of posi- tion to make that statement. For you to state that I am not bored when I have claimed I am is likewise inapprOpriate.16 Discussion of the parallels between performatives and constatives suggests further questions like, "Can performatives be true or false in some ways or in some cases?" How performatives may relate to matters of fact is not entirely clear, although Searle has quite bluntly claimed, ". . . Austin's careful researches show, certain performatives can be assessed as true or false (for example, warnings), . . . ."17 Austin does commit himself to this view, even though with a hedge, when he argues: . . . that there is no necessary conflict between (a) our issuing the utterance being the doing of some- thing; (b) our utterance being true or false. For that matter compare, for example, 'I warn you that it is going to charge; where likewise it is both a warning and true or false that it is going to charge; and that l6Austin does not use the word "evidence" in his discussion at this point and he does not venture into any detail on the privacy of feelings. Austin, "Performative- Constative," pp. 19-20. 17Searle, "Austin on Locutionary and Illocutionary Acts," p. 142; see also Austin, How To Do Things With Words, p. 136. 23 comes in in appraising the warning just as much as, though not guite in the same way as, in appraising the statement.1 This brief exposition has developed the positive points that performatives and constatives allow explicit formulation with regard to the type of action intended in issuing the utterance, and has presented the program of showing the similarity of the failures that performatives and constatives are subject to. Finally, we see that even holding out for a clear distinction between those utterances which are true or false and performatives may break down in some cases, like acts of warning. URMSON'S ATTEMPT TO RESIST THE LIBERALIZING OF THE CONCEPT OF A PERFORMATIVE What comes to be called illocutionary force in Austin's more developed View is mentioned first in "Performative Utterances." On the basis of the comparisons, both grammatical and otherwise already summarized, Austin concludes that: "What we need besides the old doctrine about meanings is a new doctrine about all possible forces of utterances, . . . ."19 The force of an utterance can be made explicit by rendering that utterance in an explicit performative form. If Austin is right and this can be done for any utterances, 18Austin, How To 29 Things With Words, pp. 135-136. 19Austin, "Performative Utterances," p. 251. 24 we can, in a sense, begin again with the study of language with this new emphasis on the performative nature of our utterances. However welcome this new emphasis is and however important as a basis for the views to be developed by Austin in How To Do Things With Words (and discussed in succeeding chapters here), this emphasis has not gone with- out criticism. Urmson has recently argued that the View of performatives held by Austin in How To Do Things With Words is less satisfactory than the doctrine of performatives that Austin earlier entertained20 and, further, that the earlier, more satisfactory, view is not subsumable under the theory of illocutionary forces.21 The first premise of Urmson's argument is that Austin, in using the performative-constative dichotomy as a steppingstone to his later views, ignores the actual basis of the distinction between performatives and consta- tives. Urmson charges that Austin forgets his original characterization of performatives in his more developed 20A doctrine, according to Urmson, of which there is scant written evidence, the only written indications being in "Other Minds." J. O. Urmson, "Performative Utterances," Midwest Studies in Philosophy, Volume II: Studies in the Philosophy of Language, eds. P. A. French, T. E. Uehling, Jr., and H. K. Wettstein (Morris: The University of Minnesota, 1977), p. 121. 21The doctrine of illocutionary forces will be developed in detail in succeeding chapters. At this point suffice it to say that illocutionary forces are those called performative in the way just introduced. 25 presentations of the performative-constative distinction. In early statements of the doctrine Austin claimed that performatives were the "utterance of obvious ritual phrases . . . ." It is this that Urmson thinks Austin forgot and that suggests to Urmson a way of critically evaluating the prOgress of Austin's discussion. Urmson develops his argument on the basis of a consideration of conventions which govern speech acts in contrast to conventions which govern performatives. Some wholly conventional acts, Urmson argues, are such that the governing conventions are not linguistic even though the governing conventions may require that certain linguistic conventions be utilized in the performance of the action. A marriage ceremony would be such an activity according to Urmson. The fact that the uttering of certain words helps to correctly carry out a marriage is not a matter of linguistic convention, according to Urmson, but rather non- linguistic convention. According to Urmson, an act (linguistic or other- wise) is conventional if it is non-natural, that is, it is not an act which could be the kind that it is in the absence of rules, principles, 1aws,or conventions. Kicking a goal would be an example of a non-natural activity; eating would be an example of a natural activity. An act is more specifically conventional rather than rule-governed in that you can have a breach of a rule but not a convention. This second way of characterizing a conventional act is simply to 26 say you cannot break a convention. You can fail to perform a convention-governed action, but you cannot violate the convention in doing this; you are, properly speaking, failing to invoke the convention. With this idea of conven- tion, Urmson argues that even though speech acts are perhaps wholly conventional in his sense and although performatives are wholly conventional, the primary or relevant conven- tions in the two cases differ. The conventions governing speech acts are linguistic; the conventions governing performatives are non-linguistic. One example which Urmson counts toward proving his point is this: If I were in Turkey and wished to state that the cat was on the mat I might reasonably ask how one says that in Turkish. I wish to perform a speech-act and so inquire about the linguistic conventions of Turkish. But if, being in Turkey, I wish to get married my initial question must be, not "How do I get married in Turkish?", but "How do I get married in Turkey?" Primarily I need to know the legal, social and, perhaps, religious conventions of Turkey, a civil community, not She linguistic conventions of a linguis- tic community.2 Austin's theory goes astray, according to Urmson, by Austin's failure to keep mindful of this sort of point. Perhaps remembering the "ritual" characterization of per- formatives would have helped Austin avoid mistaking some utterances (which are not genuine performatives) for performatives. Further, this would have perhaps led Austin to think that performatives are not speech acts at all, 22Urmson, "Performative Utterances," p. 124. 27 which Urmson, in fact, thinks is correct. The reason Austin fails to maintain a strict dis- tinction between performative and other kinds of utterances, Urmson believes, is importantly linked with Austin's "mis- take" of calling acts of warning performative.23 Warnings are not rituals and warnings are not governed by non- linguistic conventions; for these reasons Urmson considers Austin's assimilation of warnings to the class of performa- 24 . . Further, once warnings are con51dered tives a mistake. to be performative it begins to appear as though there are only two classes of utterances, performative and non- performative (or constative), Urmson argues, and this he believes is misleading. Urmson argues that Austin's early intuitions led him to believe that there were many important types of utterances, among them performatives. The view discussed in the preceding sections depends largely upon distinguish- ing and comparing two types of utterances, the performative and the declarative. In Urmson's opinion, the dichotomy between performatives and constatives, treated this way, is misleading and would never have been utilized as it was by Austin had he refrained from thinking that warnings and other non-constative utterances were performative. This is 23Austin, How To Do Things With Words, p. 55. 24It should be noted that warnings are not in any simple way true or false, and so one characterization of performatives may fit warnings. 28 an important-sounding criticism, but it is actually unimpor- tant. It seems important because Austin does in fact introduce his central thesis, that all utterances have a performative aspect, by considering the comparisons of performatives and constatives. If these categories are not correctly drawn to begin with, the whole project would seem threatened. Urmson's point is unimportant, however, because the dichotomy between performative and constative was never claimed to be perfectly clear nor exhaustive. Another attack Urmson brings against Austin's treatment of the performative-constative distinction con- cerns Austin's belief that we could make explicit performa- tive formulae for any of our utterances. Urmson believes that Austin was mistaken in thinking that we could attach a phrase, "I hereby . . ." onto any utterance, thereby making clear that all utterances are performative.25 Urmson argues that using "hereby" phrases to signal the performative aspect of any utterance fails because (1) some genuine performatives do not permit reformulations using a "hereby" phrase, and (2) some genuine speech acts are rendered rather absurd by the insertion of a "hereby" phrase. An example of the first sort of problem would be, "I, John, hereby take you, Mary, . . . ." This is not the correct formula for getting married. An example of the 25Urmson, ”Performative Utterances," p. 125. 29 second would be "I hereby warn you that . . . ." In sum, Urmson claims: It is not true that all performative utterances may properly contain the word 'hereby' or that, whenever 'hereb ' is inserted, an utterance becomes performa- tive.2% Urmson's points here are misleading because it is not clear that Austin's doctrine of illocutionary forces and types depends very much on his use of "hereby" phrases in making illocutionary forces explicit. In addition, Austin is careful to point out that the use of the "hereby" formula is ". . . rather too formal for ordinary pur- poses."27 This formality would only be appropriate in an ideal reformed language and in such a language there is really no problem conceiving a marriage ceremony as being an even "more explicit" performance, nor is there anything particularly odd in a formal language about making clear that some act like a warning is taking place. What my argument hinges on is the contention that Austin is using "hereby" phrases to make a point, but the point is not that colloquial English presently uses "hereby" phrases in just this or that way. If I am right, Urmson is mistaking a heuristic doctrine of Austin's for a report on possible correct English usages involving "hereby." Austin is care- ful in his remarks on "hereby" as the following indicates: 26 . Urmson, "Performative Utterances," p. 126. 27 . . . Austin, How To Do Things With Words, p. 61. 3O . . . in writing at least the word 'hereby' is often and perhaps can always be inserted; this serves to indicate that the utterance (in writing) of the sentence is, as it is said, the instrggent effecting the act of warning, authoriZing, etc. Two points can be made based on these claims by Austin. The strangeness of a "hereby" in an act of warning that Urmson has pointed to is almost entirely abated when the written context is brought up. The second point brings up a point made earlier in this paper. The written model seems to dominate much of Austin's thinking on the subject of explicit performatives. Most of the ambiguities which can be eliminated by explicit performative formulae are appro- priate for written statements. Whether or not "hereby" can always or only some- times be inserted in an utterance is not as important as Urmson indicates. In reviewing the breakdown of the performative-constative dichotomy, as Austin carried it out, we were able to outline the doctrine that perhaps any utterance, statement,or otherwise, has a performative aspect. This doctrine does not rely as much on Austin's discussion of "hereby" as it does on his opinions on per- formative verbs. This later part of Austin's view is certainly dominant in his presentation just as its initial phase was dominant in the exposition here. If anything could be argued as most important for Austin's change to a doctrine of illocutionary forces, it was not his 28Austin, How To Do Things With Words, p. 57. 31 understanding of "hereby" but rather his observation that "I state that . . ." does not look so very different from, "I warn you that . . . ."29 The most striking points Urmson has attempted to develop concern his preferred criteria for identifying performatives, that they are governed by non-linguistic conventions. Analysis of this portion of Urmson's argument will depend upon a careful assessment of the necessity of linguistic conventions for any illocutionary acts (kinds of linguistic performances). This tepic, as it has emerged in the recent literature, involves an analysis of the relation between intention and convention in speech acts and will be initiated in succeeding chapters with a discussion of Strawson and others on this tOpic. As mentioned before, in Urmson's view, performatives are not speech acts because they are not governed by linguistic conventions. In order to judge this issue correctly, certain possibilities must be kept in mind. Urmson notwithstanding, the primary con- ventions for many speech acts may well be non-linguistic. And, further, many speech acts may be the type they are with no conventions specifically fixing them as objections, entreaties, etc. Strawson argues for this last point.30 29Austin, "Performative Utterances," p. 247. 30P. P. Strawson, "Intention and Convention in Speech Acts," The Philosophy of Language, ed. J. R. Searle (London: Oxford University Press, 1971), pp. 26-27. 32 In short, part of the rescue of the doctrinecaf performative forces from Urmson's attack will depend on issues which themselves need further elaboration. The rescue of the doctrine will involve some disagreement with Austin, as well as Urmson, on the issue of conventional aspects of speech acts. Urmson's criteria for pure performatives suggests that we can call an utterance a performative if non- linguistic conventions govern the utterance. Urmson is, however, prepared to allow that promises are performative.31 I would not disagree with the claim that promises are performative, but wonder at the appropriate non-linguistic governing conventions that are operative with promising. If we could specify these conventions here I am sure we could find ways to do the same for many utterances to which Urmson does not wish to grant performative status. Going back to an example mentioned earlier, we can ask how one makes a promise in Turkey as well as how to promise in Turkish; either question could be appropriate. Urmson's criteria for classifying performatives probably will not work as these examples help to indicate. If this is so, his attack on Austin is not effective. Urmson's belief that only performatives are governed by non-linguistic conventions should have led him to a much more liberal assignment of performative status to utterances 31 Urmson, "Performative Utterances, p. 120. 33 than it has. Clearly, "Pass the salt" (perhaps a request or an order) will be governed by certain conventions which are not primarily linguistic, such as rules of etiquette. This sort of example will not settle the issue but suggests again that the issue will need basis in a more careful analysis of convention, and the kinds of conventions which play a part in determining our linguistic performances. In sum, Urmson attacks the generality of the doctrine of illocutionary forces by attacking Austin's program for breaking down the distinction between performa- tives and constatives. Urmson's attack fails for the reasons discussed, but brings up the need for further analy- sis of certain key issues mentioned here. In the following chapter some of the ways our linguistic performances may be conventional will be explored. S UMMA RY Austin's View that we can fruitfully examine language in terms of the kinds of actions we perform linguistically is grounded in his doctrine of performative utterances. The later view is grounded in the earlier in at least two ways. First, the comparison of performatives to constatives indicates more similarities than differences and, therefore, points the way to a general doctrine. Second, specific views on performatives seem to influence Austin's views on illocutionary acts. This second point is 34 discussed in Chapter 2. The purpose of this chapter has been to show how the general view of Austin's is in fact elicited from his consideration of the special class of utterances, the performative, which intrigued Austin in his early work on language. Chapter 2 ILLOCUTIONARY ACTS AND CONVENTION: AUSTIN, STRAWSON AND SEARLE INTRODUCTION The hypothesis that language is conventional is taken by some to be so obvious that little argument seems required for its support. However obvious it may appear to be, a closer look reveals that it is not a clear hypothesis nor is it above controversy in any of its forms. Two of the forms the claim may take are these: (1) There is something arbitrary about "il pleut," "It is raining," etc. The meaning may not be arbitrary, but the particular words which express the meaning are arbitrary. Besides this arbitrary aSpect, languages are regular and systematic. These two factors can only be accounted for, it is argued, if languages are con- ventional.1 (2) There are conventions governing the sorts of things one does with the sentences one utters. The type of argument in (1) will be ignored here. The discussion of (2) will be the focus of this chapter. 1F. P. Dinneen, An Introduction to General Linguistics (New York: Holt, Rinehart and Winston, 1967), p. 9. 35 36 The claim of (2) is not precise but will be taken as the claim that there are special conventions governing what you do with sentences which are different from con- ventions of meaning. Strawson makes the distinction which is important here in pointing out that: We must refer, Austin would say, to linguistic con- ventions to determine what locutionary act has been performed in the making of an utterance, to determine what the meanin of the utterance is. The doctrine now before us is the further doctrine that where force is not exhausted by meaning, the fact that an utterance has the further unexhausted force it has is a matter of convention; or, where it is exhausted by meaning, the fact that it is, is a matter of convention. This chapter will be developed in four stages. First a preliminary identification of illocutionary acts will be derived from Austin. Second, Austin's views on convention and illocutionary acts will be developed in terms of both his discussion of performatives and in terms of passages where he claims that illocutionary acts are conventional. Third, Strawson's objections to Austin's views on illocutionary acts will be discussed primarily in terms of Austin's views on the kinds of conventions impor- tant in understanding performatives. Finally, Searle's approach to the topics of illocutionary acts and convention will be discussed. Searle attempts to reduce the study of illocutionary acts to a study of those conventions which determine the meanings of our sentences. In the case of 2P. F. Strawson, "Intention and Convention in Speech Acts," in The Philosophy of Language, ed. J. R. Searle, (London: Oxford University Press, 1971), p. 25. 37 illocutionary acts this depends upon being able to render any illocutionary act in an explicit performative form and then studying the meaning of that performative. Much of interest in Searle's theory is beyond the scope of this chapter and so, for the most part, other than pointing out certain similarities and differences between his and Austin's general approach, criticisms are left until later. Two further preliminaries should help forestall potential confusions that may arise in this chapter. In the last chapter it was argued that a fruitful approach to the study of language could be based upon Austin's insight that the performative aspect of language is basic: basic in the sense that any utterance can be considered some type of doing. This basic insight has generated the discussion of speech acts. The notion of a speech act is not entirely clear because it has been used in a variety of ways. For the purposes of this discussion it will be used as an over- arching term designating a complex of actions termed by Austin, locutionary, illocutionary and perlocutionary acts. The notion of speech acts is not always used in this way; for example, Searle claims that "speaking a language is performing speech acts, acts such as making statements, giving commands, asking questions, making promises and 38 so on."3 Searle's usage would seem to equate speech acts and illocutionary acts because his list is a list of what Austin termed illocutionary acts. Another potential source of confusion in the dis- cussion of speech acts is rooted in the way the concern originates in Austin's work. As detailed in the preceding chapter, Austin's discussion of speech acts is grounded in his discussion of performative utterances; because of this it is not clear just how far to assimilate the discussion of illocutionary acts (his principal later concern) to a discussion of performatives. One example of this possible source of confusion will illustrate this point. In identi- fying performatives Austin "prefers" the criterion that there is an asymmetry between first person and other uses of performative verbs;4 however, in distinguishing illocu- tionary acts from other aspects of speech acts, Austin puts emphasis on the criterion that illocutionary acts are conventional whereas perlocutionary acts at least are not. The examples of performatives Austin gives are conventional and at least many types of performatives are considered types of illocutionary acts, and so the doctrine that illocutionary acts are conventional seems in large part to 3J. R. Searle, Speech Acts: An Essay in the Philosophy of Language (London: Cambridge University Press, 1969), P- 16. 4J. L. Austin, How To Do Things With Words, eds. J. O. Urmson and M. Sbisa (Cambridge: Harvard University Press, 1962), pp. 63-67. 39 be a development from the discussion of performatives. Whether or not these later facts indicate that we should think all the criteria used to identify performatives are relevant to the discussion of illocutionary acts is not clear. Specifically it is not clear that the doctrine con- cerning asymmetry is crucial for the discussion of illocu- tionary acts even though Austin does use this criterion when suggesting how we might develop a list of illocutionary act verbs. I LLOC UTIONARY ACTS : PRELIMINARY IDENTIFICATION According to Searle the illocutionary act is the basic unit of linguistic communication. Acts such as stating, asserting, describing, warning, remarking, com- menting, promising, are illocutionary acts.5 Further, the study of language which takes illocutionary acts as the basic unit of concern attempts to study characteristics of actual linguistic behaviors. To quote Searle: It is not, as has generally been supposed, the symbol or word or sentence, or even the token