Searle_Illocutionary-Acts

$&ODVVLILFDWLRQRI,OORFXWLRQDU\$FWV$XWKRUV-RKQ56HDUOH6RXUFH/DQJXDJHLQ6RFLHW\9RO1R$SUSS3XEOLVKHGE\&DPEULGJH8QLYHUVLW\3UHVV6WDEOH85/http://www.jstor.org/stable/4166848 .$FFHVVHGYour use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].. Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend access to Language in Society.http://www.jstor.org This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

Lang. Soc. 5, 1-23. Printed in Great Britain A classification of illocutionary acts' JOHN R. SEARLE Universityof CaliforniaABSTRA CT There areat least a dozen linguisticallysignificantdimensionsof differences between illocutionaryacts. Of these, the most importantare illocutionary point, directionof fit, and expressedpsychologicalstate. These three form the basis of a taxonomy of the fundamentalclasses of illocutionaryacts. The five basic kinds of illocutionaryacts are: representatives(or assertives), directives, commissives, expressives, and declarations. Each of these notionsis defined.An earlierattemptat constructinga taxonomyby Austin is defective for several reasons, especially in its lack of clear criteria for distinguishing one kind of illocutionary force from another. Paradigm performativeverbs in each of the five categoriesexhibit differentsyntactical properties.These areexplained.(Speech acts, Austin'staxonomy,functions of speech, implicationsfor ethnographyand ethnology; English.) 1. INTRODUCTIONOne of the crucialquestionsin studying languagein society is, 'How many waysof using language are there?' Most of the attempts to answer that questionsuffer from an unclarity about what constitutes a use of language in the firstplace. If you believe, as I do, that the basic unit of humanlinguisticcommunica-tion is the illocutionaryact, then the most importantformof the originalquestionwill be, 'How many categories of illocutionary acts are there?' This articleattemptsto answerthat question. The primarypurposeof this paper,then, is to developa reasonedclassificationof illocutionaryactsinto certainbasiccategoriesor types. Since anysuch attemptto developa taxonomymusttakeinto accountAustin'sclassificationof illocution-aryacts into his five basic categoriesof verdictive,expositive,exercitive,behabi-tive, and commissive,a second purposeof this paperis to assess Austin's classi-ficationto show in what respectsit is adequateand in what respectsinadequate.Furthermore, since basic semantic differences are likely to have syntactical consequences,a third purposeof this paperis to show how these differentbasic [i] This article was originally written for an audience of philosophers and linguists (it was first presented as a lecture at the Summer Linguistics Institute in Buffalo in I97I). It is published here in the belief that it may be of use to others interested in the special roles that language plays in human social behavior. I This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

LANGUAGE IN SOCIETYillocutionarytypes are realized in the syntax of a natural language such asEnglish. In what follows, I shall presupposea familiaritywith the generalpatternofanalysisof illocutionaryacts offeredin such worksas Austin, How to Do Thingswith Words,Searle, SpeechActs, and Searle, 'Austin on Locutionaryand Illo-cutionary Acts'. In particular, I shall presuppose a distinction between theillocutionaryforce of an utteranceand its propositionalcontent as symbolized F(p).The aim of this paperthen is to classify the differenttypes of F.II. DIFFERENT TYPES OF DIFFERENCES BETWEEN DIFFERENT TYPES OFILLOCUTIONARY ACTSAny taxonomicaleffort of this sort presupposescriteriafor distinguishingone(kind of) illocutionaryact from another.What are the criteriaby which we cantell that of three actual utterancesone is a report, one a predictionand one apromise?In order to develop higher ordergenera,we must first know how thespeciespromisep, redictionr, eport,etc., differonefromanother.Whenoneattemptsto answerthat question one discoversthat there are severalquite differentprin-ciples of distinction; that is, there are differentkinds of differencesthat enableus to say that the force of this utterance is different from the force of thatutterance.For this reasonthe metaphorof force in the expression'illocutionaryforce' is misleadingsince it suggests that differentillocutionaryforces occupydifferentpositions on a single continuum of force. What is actually the case isthat there are severaldistinct criss-crossingcontinua. A relatedsource of confusion is that we are inclined to confuse illocutionaryverbs with types of illocutionaryacts. We are inclined, for example, to thinkthatwherewe havetwo nonsynonymousillocutionaryverbsthey mustnecessarilymarktwo differentkindsof illocutionaryacts. In whatfollows, I shall try to keepa cleardistinctionbetweenillocutionaryverbs and illocutionaryacts. Illocutionsarea partof languageas opposedto particularlanguages.Illocutionaryverbs arealways part of a particularlanguage: French, German, English, or whatnot.Differencesin illocutionaryverbs are a good guide but by no meansa sure guideto differencesin illocutionaryacts. It seems to me there are(at least) twelve significantdimensionsof variationinwhich illocutionaryacts differone fromanotherand I shall- all too briskly- listthem:(i) Differencesin thepoint (orpurpose)of the (typeof) actThe point or purposeof an ordercan be specifiedby sayingthat it is an attemptto get the hearerto do something.The point or purposeof a descriptionis that 2This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

A CLASSIFICATION OF ILLOCUTIONARY ACTSit is a representation(true or false, accurateor inaccurate)of how somethingis.The point or purposeof a promiseis that it is an undertakingof an obligationbythe speaker to do something. These differences correspond to the essentialconditions in my analysis of illocutionary acts in Speech Acts (Searle 1969:Ch. 3). Ultimately, I believe, essential conditions form the best basis for ataxonomy, as I shall attempt to show. It is importantto notice that the termi-nology of 'point' or 'purpose'is not meantto imply, nor is it based on the view,that every illocutionaryact has a definitionallyassociatedperlocutionaryintent.For many, perhapsmost, of the most importantillocutionaryacts, there is noessential perlocutionaryintent associatedby definition with the correspondingverb, e.g. statements and promises are not by definition attempts to produceperlocutionaryeffects in hearers. The point or purposeof a type of illocution I shall call its illocutionarypoint.Illocutionarypoint is part of but not the same as illocutionaryforce. Thus, e.g.,the illocutionarypoint of request is the same as that of commands: both areattemptsto get hearersto do something. But the illocutionaryforces are clearlydifferent. In general, one can say that the notion of illocutionaryforce is theresultantof severalelements of which illocutionarypoint is only one, though, Ibelieve, the most importantone.(2) Differencesin the directionoffit betweenwordsand the worldSome illocutionshave as partof their illocutionarypoint to get the words (morestrictly- their propositionalcontent) to matchthe world, othersto get the worldto matchthe words.Assertionsarein the formercategory,promisesand requestsarein the latter.The best illustrationof this distinctionI knowof is providedbyMiss Anscombe(I957). Supposea man goes to the supermarketwith a shoppinglist given him by his wife on which are written the words 'beans, butter, bacon,and bread'. Suppose as he goes around with his shopping cart selecting theseitems, he is followed by a detective who writes down everything he takes. Asthey emerge from the store both shopper and detective will have identical lists.But the functionof the two lists will be quite different.In the case of the shop-per's list, the purpose of the list is, so to speak, to get the world to match thewords; the man is supposed to make his actions fit the list. In the case of thedetective,the purposeof the list is to makethe words matchthe world; the manis supposed to make the list fit the actions of the shopper. This can be furtherdemonstratedby observingthe role of 'mistake'in the two cases. If the detectivegets home and suddenly realizes that the man bought pork chops instead ofbacon, he can simply erase the word 'bacon' and write 'porkchops'. But if theshoppergets home and his wife points out he has bought pork chops when heshould have bought baconhe cannotcorrectthe mistakeby erasing'bacon'fromthe list and writing 'porkchops'. In these examplesthe list providesthe propositionalcontent of the illocution 3This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

LANGUAGE IN SOCIETYand the illocutionaryforce determineshow that content is supposedto relatetothe world. I propose to call this differencea differencein directionof fit. Thedetective's list has the word-to-worlddirectionof fit (as do statements,descrip-tions, assertions, and explanations); the shopper's list has the world-to-worddirection of fit (as do requests, commands, vows, promises). I represent theword-to-worlddirectionof fit with a downwardarrowthus 4 and the world-to-word directionof fit with an upwardarrowthus t. Direction of fit is always aconsequenceof illocutionarypoint. It wouldbe veryelegantif we couldbuildourtaxonomy entirely aroundthis distinction in directionof fit, but though it willfigure largely in our taxonomy, I am unable to make it the entire basis of thedistinctions.(3) Differencesin expressedpsychologicasltatesA man who states, explains, asserts or claims that p expressesthebeliefthatp; aman who promises,vows, threatensor pledges to do a expressesan intentionto doa; a man who orders, commands, requests H to do A expressesa desire(want,wish)that H do A; a man who apologizesfor doing A expressesregretat havingdoneA; etc. In general, in the performanceof any illocutionaryact with a pro-positional content, the speaker expresses some attitude, state, etc., to thatpropositionalcontent. Notice that this holds even if he is insincere, even if hedoes not havethe belief, desire,intention,regretor pleasurewhich he expresses,he none the less expresses a belief, desire, intention, regret or pleasure in theperformanceof the speech act. This fact is markedlinguisticallyby the fact thatit is linguistically unacceptable(though not self-contradictory)to conjoin theexplicit performativeverb with the denial of the expressed psychologicalstate.Thus one cannot say 'I state that p but do not believe that p', 'I promise thatp but I do not intend that p', etc. Notice that this only holds in the first personperformativeuse. One can say, 'He statedthat p but didn't reallybelieve that p','I promisedthat p but did not reallyintend to do it', etc. The psychologicalstateexpressedin the performanceof the illocutionaryact is the sincerityconditionofthe act, as analyzedin SpeechActs, Ch. 3. If one tries to do a classificationof illocutionaryacts based entirelyon differ-ently expressed psychologicalstates (differencesin the sincerity condition) onecan get quite a long way. Thus, beliefcollects not only statements, assertions,remarksand explanations,but also postulations, declarations,deductions andarguments.Intentionwill collect promises,vows, threatsand pledges. Desireorwant will collect requests, orders, commands, askings,>prayers, pleadings,beggingsand entreaties.Pleasuredoesn'tcollect quiteso many- congratulations,felicitations,welcomes and a few others. In what follows, I shall symbolize the expressedpsychologicalstate with thecapitalizedinitiallettersof the correspondingverb,the B forbelieve,W for want,I for intend, etc. 4 This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

A CLASSIFICATION OF ILLOCUTIONARY ACTS These three dimensions- illocutionarypoint, direction of fit, and sinceritycondition- seem to me the most important,andI will build most of my taxonomyaroundthem, but there are severalothers that need remarking.(4) Differencesin theforce or strengthwithwhichtheillocutionarypointis presentedBoth, 'I suggestwe go to the movies' and 'I insist that we go to the movies' havethe same illocutionarypoint, but it is presentedwith differentstrengths.Analo-gously with 'I solemnly swearthat Bill stole the money' and 'I guess Bill stolethe money'. Along the same dimension of illocutionarypoint or purpose theremay be varyingdegrees of strength or commitment.(5)Differencesin thestatusorpositionof thespeakerandheareras thesebearon theillocutionaryforce of the utteranceIf the general asks the private to clean up the room, that is in all likelihood acommandor an order. If the privateasks the generalto clean up the room, thatis likely to be a suggestionor proposalor requestbut not an orderor command.This featurecorrespondsto one of the preparatoryconditions in my analysisinSpeechActs, Ch. 3.(6) Differencesin the way the utterancerelatesto the interestsof the speakerandthe hearerConsider, for example, the differences between boasts and laments, betweencongratulationsand condolences. In these two pairs, one hears the differenceas being between what is or is not in the interests of the speaker and hearerrespectively.This featureis anothertype of preparatorycondition accordingtothe analysisin SpeechActs.(7) Differencesin relationsto the restof the discourseSome performativeexpressions serve to relate the utteranceto the rest of thediscourse(and also to the surroundingcontext). Consider,for example 'I reply','I deduce','I conclude',and'I object'.These expressionsserveto relateutterancesto other utterances and to the surrounding context. The features they markseem mostly to involve utteranceswithin the class of statements. In additiontosimply statinga proposition,one may state it by way of objectingto what some-one else has said, by way of replyingto an earlierpoint, by way of deducing itfromcertainevidentiarypremises,etc. 'However','moreover'and'therefore'alsoperformthese discourse-relatingfunctions.(8) Differencesin propositionalcontentthat are determinedby illocutionaryforce-indicatingdevicesThe differences,for example, between a reportand a predictioninvolve the factthat a predictionmust be aboutthe futurewhereasa reportcan be aboutthe past SThis content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

LANGUAGE IN SOCIETYor present. These differencescorrespondto differencesin propositionalcontentconditions as explainedin SpeechActs.(g) Differencesbetweenthoseacts thatmustalwaysbespeechacts,andthosethat canbe, butneednot beperformedas speechactsFor example,one may classifythings by saying'I classifythis as an A and this asa B'. But one need not say anything at all in order to be classifying; one maysimply throwall the A's in the A box and all the B's in the B box. Similarlywithestimate,diagnoseand conclude. I may makeestimates,give diagnosesand drawconclusionsin saying 'I estimate', 'I diagnose',and 'I conclude' but in ordertoestimate, diagnose or conclude it is not necessaryto say anythingat all. I maysimply stand beforea buildingand estimateits height, silently diagnoseyou as amarginalschizophrenic, or conclude that the man sitting next to me is quitedrunk.In these cases,no speech actsnot even an internalspeechact, is necessary.(so) Differencesbetweenthoseacts that requireextra-linguisticinstitutionfsor theirperformanceand thosethat do notThere are a large number of illocutionaryacts that require an extra-linguisticinstitution,andgenerallya special positionby the speakerand the hearerwithinthat institution in order for the act to be performed.Thus, in order to bless,excommunicate,christen, pronounceguilty, call the base runnerout, bid threeno-trump,or declarewar,it is not sufficientfor any old speakerto say to any oldhearer 'I bless', 'I excommunicate',etc. One must have a position within anextra-linguisticinstitution.Austin sometimestalksas if he thoughtall illocution-aryactswerelikethis, but plainlythey are not. In orderto makea statementthatit is rainingor promise to come and see you, I need only obey the rules of lan-guage. No extra-linguistic institutions are required. This feature of certainspeech acts, that they require extra-linguisticinstitutions, needs to be distin-guished from feature (5), the requirementof certainillocutionaryacts that thespeaker and possibly the heareras well have a certain status. Extra-linguisticinstitutionsoften conferstatusin a way relevantto illocutionaryforce, but not alldifferencesof status derive from institutions. Thus, an armed robber in virtueof his possessionof a gun mayorderasopposedto, e.g., request,entreat,orimplorevictims to raisetheir hands. But his status here does not derive from a positionwithin an institution but from his possession of a weapon.(i i) Differencesbetweenthoseacts wherethe correspondinigllocutionaryverbhas aperformativeuseand thosewhereit doesnotMost illocutionaryverbs have performativeuses - e.g., 'state','promise','order','conclude'. But one cannot perform acts of, e.g., boasting or threatening, bysaying 'I hereby boast', or 'I hereby threaten'. Not all illocutionaryverbs areperformativeverbs. 6 This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

A CLASSIFICATION OF ILLOCUTIONARY ACTS(12) Differencesin the style of performanceof the illocutionaryactSome illocutionaryverbs serve to markwhat we might call the special style inwhich an illocutionaryact is performed.Thus, the differencebetween, for ex-ample,announcingand confidingneed not involve any differencein illocutionarypoint or propositionalcontent but only in the style of performanceof the illo-cutionaryact.III. WEAKNESSES IN AUSTIN S TAXONOMYAustinadvanceshis five categoriesverytentatively,moreas a basis for discussionthan as a set of establishedresults. 'I am not', he says (I962: I5I), 'putting anyof this forwardas in the very least definitive.' I think they form an excellentbasis for discussion but I also think that the taxonomy needs to be seriouslyrevisedbecauseit containsseveralweaknesses.Here are Austin'sfive categories: VerdictivesT. hese 'consist in the deliveringof a finding, officialor unofficial,upon evidence or reasonsas to value or fact so far as these are distinguishable'.Examples of verbs in this class are: acquit, hold, calculate, describe, analyze,estimate, date, rank,assess, and characterize. Exercitives.One of these 'is the giving of a decision in favor of or against acertaincourse of action or advocacyof it. . .', 'a decision that somethingis to beso, as distinctfromajudgmentthatit is so'. Someexamplesare:order,command,direct, plead, beg, recommend,entreat and advise. Request is also an obviousexample, but Austin does not list it. As well as the above, Austin also lists:appoint, dismiss, nominate, veto, declare closed, declare open, as well as an-nounce, warn, proclaim,and give. Commissives'.The whole point of a commissive',Austin tells us, 'is to committhe speakerto a certain course of action.' Some of the obvious examples are:promise, vow, pledge, covenant, contract, guarantee,embrace, and swear. Expositivesare used in acts of exposition involving the expoundingof views,the conductingof argumentsand the clarifyingof usages and reference'.Austingives many examplesof these, among them are: affirm,deny, emphasize, illus-trate,answer,report,accept,objectto, concede, describe,class, identifyand call. BehabitivesT. his class, with which Austin was very dissatisfied('a shocker',hecalledit) 'includesthe notion of reactionto other people's behaviorand fortunesand of attitudesand expressions of attitudes to someone else's past conduct orimminent conduct'. Amongthe examplesAustin lists are: apologize,thank, deplore,commiserate, 7This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

LANGUAGE IN SOCIETYcongratulate,felicitate,welcome, applaud,criticize,bless, curse,toast anddrink.But also, curiously: dare, defy, protest, and challenge. The firstthing to notice aboutthese lists is that they are not classificationsofillocutionaryacts but of Englishillocutionaryverbs.Austin seems to assumethata classificationof differentverbsis eoipsoa classificationof kindsof illocutionarvacts, that any two non-synonymous verbs must mark different illocutionaryacts. But thereis no reasonto supposethat this is the case. As we shall see, someverbs, for example,markthe mannerin which an illocutionaryact is performed,for example 'announce'.One may announceorders, promises and reports, butannouncingis not on all fourswith ordering,promisingandreporting.Announc-ing, to anticipatea bit, is not the name of a type of illocutionaryact, but of theway in which some illocutionaryact is performed.An announcement is neverjust an announcement.It must also be a statement,order, etc. Evengrantingthatthe lists areof illocutionaryverbsandnot necessarilyof dif-ferent illocutionaryacts, it seems to me, one can level the following criticismsagainst it. (a) First, a minor cavil, but one worth noting. Not all of the verbs listed areeven illocutionary verbs. For example, 'sympathize', 'regard as', 'mean to','intend', and 'shall'. Take 'intend': it is clearly not performative. Saying 'Iintend' is not intending; nor in the third person does it name an illocutionaryact: 'He intended. . .' does not report a speech act. Of course there is an illo-cutionary act of expressingan intention,but the illocutionaryverb phrase is:'express an intention', not 'intend'. Intending is never a speech act; expressingan intention usually, but not always, is. (b) The most importantweaknessof the taxonomyis simply this. There is noclearor consistentprincipleor set of principleson the basis of which the taxon-omy is constructed. Only in the case of Commissives has Austin clearly andunambiguouslyused illocutionarypoint as the basis of the definitionof a cate-gory. Expositives, in so far as the characterizationis clear,seem to be definedintermsof discourserelations(my feature(7)). Exercitivesseemto be at leastpartlydefined in terms of the exercise of authority.Both considerationsof status (myfeature (5) above) as well as institutional considerations(my feature (lo)) arelurkingin it. Behabitivesdo not seem to me at all well defined(as Austin, I amsure, would have agreed)but it seems to involve notions of what is good or badfor the speakerand hearer(my feature(6)) as well as expressionsof attitudes(myfeature(3)). (c) Becausethere is no clear principle of classificationand because there is apersistentconfusion between illocutionaryacts and illocutionaryverbs, there isa great deal of overlapfrom one categoryto anotherand a great deal of hetero-geneity within some of the categories.The problemis not that there areborder-line cases - any taxonomy that deals with the real world is likely to come up 8 This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

A CLASSIFICATION OF ILLOCUTIONARY ACTSwith borderlinecases - nor is it merely that a few unusual cases will have thedefiningcharacteristicsof morethan one category;rather,a verylargenumberofverbs find themselvessmackin the middle of two competing categoriesbecausethe principlesof classificationare unsystematic.Consider,for example,the verb'describe',a very importantverb in anybody'stheory of speech acts. Austin listsit as both a verdictiveand an expositive. Given his definitions, it is easy to seewhy: describingcan be both the deliveringof a findingand an act of exposition.But then any 'act of exposition involvingthe expoundingof views' could also inhis ratherspecial sense be 'the deliveringof a finding,officialor unofficial,uponevidence or reasons'.And indeed, a look at his list of expositives (pp. I6I-2) issufficientto show that most of his verbs fit his definitionof verdictivesas well asdoes describe.Consider'affirm','deny', 'state', 'class', 'identify', 'conclude',and'deduce'. All of these are listed as expositives, but they could just as easily havebeen listed as verdictives. The few cases which are clearly not verdictives arecases where the meaning of the verb has purely to do with discourse relations,e.g. 'begin by', 'turn to', or where there is no question of evidence or reasons,e.g. 'postulate','neglect', 'call', and 'define'.But then that is really not sufficientto warrant a separate category, especially since many of these -'begin by','turnto', 'neglect'- are not names of illocutionaryacts at all. (d) Not only is there too much overlap from one category to the next, butwithin some of the categoriesthere arequite distinctkindsof verbs.Thus Austinlists 'dare', 'defy' and 'challenge', alongside 'thank', 'apologize', 'deplore' and'welcome' as behabitives.But 'dare', 'defy' and 'challenge'have to do with thehearer'ssubsequent actions, they belong with 'order', 'command'and 'forbid'both on syntacticaland semantic grounds, as I shall argue later. But when welook for the family that includes 'order', 'command'and 'urge', we find theseare listed as exercitives alongside 'veto', 'hire' and 'demote'. But these, againas I shall arguelater, arein two quite distinct categories. (e) Relatedto these objectionsis the furtherdifficultythat not all of the verbslisted within the classes really satisfy the definitionsgiven, even if we take the definitionsin the ratherloose and suggestivemannerthat Austin clearlyintends. Thus 'nominate','appoint'and 'excommunicate'are not 'giving of a decision in favorof or againsta certaincourse of action', much less arethey 'advocating'it. Ratherthey are,as Austin himselfmight havesaid,performanceosf these actions, not advocaciesof anything. That is, in the sense in which we might agree that ordering, commanding and urging someone to do something are all cases of advocatingthat he do it, we can not also agreethat nominatingor appointingis also advocating.When I appointyou chairman,I don't advocatethat you be or become chairman;I makeyou chairman. In sum, there are (at least) six relateddifficultieswith Austin's taxonomy; in ascendingorderof importance:there is a persistentconfusionbetweenverbsand acts, not all the verbs are illocutionaryverbs, there is too much overlap of the 9 This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

LANGUAGE IN SOCIETYcategories, too much heterogeneity within the categories, many of the verbslisted in the categories don't satisfy the definition given for the category and,most important, there is no consistent principle of classification. I don't believe I have fully substantiated all six of these charges and I will notattempt to do so within the confines of this paper, which has other aims. Ibelieve, however, that my doubts about Austin's taxonomy will have greaterclarity and force after I have presented an alternative. What I propose to do istake illocutionary point, and its corollaries, direction of fit and expressed sincerityconditions, as the basis for constructing a classification. In such a classification,other features - the role of authority, discourse relations, etc. - will fall into theirappropriate places.IV. ALTERNATIVE TAXONOMYIn this section, I shall present a list of what I regard as the basic categories ofillocutionary acts. In so doing, I shall discuss briefly how my classificationrelates to Austin's. Representatives.The point or purpose of the members of the representative classis to commit the speaker (in varying degrees) to something's being the case, tothe truth of the expressed proposition. All of the members of the representativeclass are assessable on the dimension of assessment which includes true andfalse. Using Frege's assertion sign to mark the illocutionary point common to alland the symbols introduced above, we may symbolize this class as follows: F4B(p).The direction of fit is words to the world; the psychological state expressed isBelief (that p). It is important to emphasize that words such as 'belief' and'commitment' are here intended to mark dimensions; they are so to speak deter-minable rather than determinates. Thus, there is a difference between suggestingthat p or putting it forward as a hypothesis that p on the one hand and insistingthat p or solemnly swearing that p on the other. The degree of belief and commit-ment may approach or even reach zero, but it is clear or will become clear, thathypothesizing that p and flatly stating that p are in the same line of business in away that neither is like requesting. Once we recognize the existence of representatives as a quite separate class,based on the notion of illocutionary point, than the existence of a large number ofperformative verbs that denote illocutions that seem to be assessable iantheTrue-False dimension and yet are not just 'statements' will be easily explicablein terms of the fact that they mark features of illocutionary force which are inaddition to illocutionary point. Thus, for example, consider: 'boast' and 'com-plain'. They both denote representatives with the added feature that they havesomething to do with the interest of the speaker (condition (6) above). 'Conclude' I0 This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

A CLASSIFICATION OF ILLOCUTIONARY ACTSand 'deduce' are also representativeswith the added feature that they markcertain relations between the representativeillocutionary act and the rest ofthe discourse or the context of utterance(condition (7) above). This class willcontainmost of Austin's expositivesand many of his verdictivesas well for the,by now I hope obvious, reason that they all have the same illocutionarypointand differonly in other featuresof illocutionaryforce. The simplest test of a representativeis this: can you literally characterizeit(interalia) as true or false. I hasten to add that this will give neither necessarynor sufficientconditions, as we shall see when we get to my fifth class. These points aboutrepresentativeswill, I hope, be clearerwhen I discuss mysecond class which, with some reluctance, I will call Directives.The illocutionarypoint of these consists in the fact that they areattempts (of varyingdegrees, and hence, more precisely, they are determinatesof the determinablewhich includes attempting)by the speakerto get the hearerto do something.They may be very modest 'attempts'as when I invite you to doit or suggest that you do it, or they may be very fierceattemptsas when I insistthat you do it. Using the shriekmarkfor the illocutionarypoint indicatingdevicefor the membersof this class generally,we have the following symbolism: ! t W (H does A)The direction of fit is world-to-words and the sincerity condition is want (orwish or desire).The propositionalcontent is alwaysthat the hearerH does somefutureactionA. Verbsdenotingmembersof this class areask,2order,command,request, beg, plead, pray, entreat, and also invite, permit, and advise. I thinkalsothatit is clearthat dare,defy andchallenge,which Austinlists as behabitives,arein this class. Many of Austin's exercitivesare also in this class. CommissivesA. ustin's definitionof commissivesseems to me unexceptionable,and I will simply appropriateit as it standswith the cavilthat severalof the verbshe lists as commissive verbs do not belong in this class at all, such as 'shall','intend', 'favor', and others. Commissives then are those illocutionary actswhose point is to commit the speaker(again in varyingdegrees)to some future course of action. Using C for the members of this class, generallywe have thefollowing symbolism: C T I (S does A) The directionof fit is world-to-wordsand the sincerityconditionis Intention. The propositionalcontent is alwaysthat the speakerS does some future action A. Since the directionof fit is the same for commissivesand directives, it would [2] Questions are a species of directives since they are attempts by S to get H to answer - i.e. to perform a speech act. II This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

LANGUAGE IN SOCIETYgive us a simplertaxonomyif we could show that they arereallymembersof thesame category. I am unable to do this because, whereasthe point of a promiseis to commit the speakerto doing something (and not necessarilyto try to gethimselfto do it), the point of a requestis to try to get the hearerto do something(and not necessarilyto commit or obligate him to do it). In orderto assimilatethe two categories,one would have to show that promisesare reallya species ofrequeststo oneself(this hasbeen suggestedto me by JulianBoyd),oralternativelyone would haveto show that requestsplacedthe hearerunderan obligation(thishas been suggested to me by William Alston and John Kearns). I have beenunable to make either of these analyses work and am left with the inelegantsolution of two separatecategorieswith the same direction of fit. A fourth categoryI shall call ExpressivesT. he illocutionarypoint of this class is to expressthe psychologicalstate specified in the sincerity condition about a state of affairsspecified in thepropositionalcontent. The paradigmsof Expressiveverbs are'thank','congratu-late', 'apologize','condole', 'deplore',and 'welcome'.Notice that in expressivesthere is no direction of fit. In performingan expressive,the speakeris neithertrying to get the world to match the words nor the words to match the world,ratherthe truth of the expressedpropositionis presupposed.Thus, for example,when I apologizefor havingstepped on your toe, it is not my purposeeither toclaim that your toe was stepped on or to get it steppedon. This fact is neatlyreflected in the syntax (of English) by the fact that the paradigm-expressiveverbs in their performativeoccurrencewill not take that clauses but requireagerundivenominalizationtransformation(or some other nominal). One cannotsay: *I apologizethat I stepped on your toe;ratherthe correct English is, I apologizefor stepping on your toe.Similarly,one cannot have: *1 congratulateyou that you won the racenor *I thank you that you paid me the money.One must have: I congratulateyou on winningthe race(congratulationson winning the race). I thankyou for payingme the money (thanksfor payingme the money). These syntacticalfacts, I suggest, areconsequencesof the fact that there is no I2 This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

A CLASSIFICATION OF ILLOCUTIONARY ACTSdirection of fit in expressives. The truth of the proposition expressed in anexpressive is presupposed. The symbolization therefore of this class mustproceedas follows: E 0 (P) (SIH+ property)Where E indicates the illocutionarypoint common to all expressives, o is thenull symbolindicatingno directionof fit, P is avariablerangingoverthe differentpossible psychologicalstates expressed in the performanceof the illocutionaryacts in this class, and the propositional content ascribes some property (notnecessarilyan action) to either S or H. I can congratulateyou not only on yourwinning the race, but also on your good looks. The property specified in thepropositionalcontent of an expressive must, however, be related to S or H. Icannot without some very special assumptions congratulateyou on Newton'sfirst law of motion. It would be economicalif we could include all illocutionaryacts in these fourclasses, and to do so would lend some furthersupport to the generalpatternofanalysisadoptedin SpeechActs, but it seems to me the classificationis still notcomplete. There is still left an importantclass of cases, where the state of affairsrepresentedin the propositionexpressedis realizedor broughtinto existence bythe illocutionaryforce-indicatingdevice, cases where one bringsa state of affairsinto existence by declaringit to exist, cases where, so to speak, 'sayingmakesitso'. Examplesof these cases are'I resign', 'You'refired','I excommunicateyou','I christen this ship, the battleship Missouri', 'I appoint you chairman', and'War is hereby declared'.These cases were presented as paradigmsin the veryearliest discussions of performatives,but it seems to me they are still not ade-quatelydescribedin the literatureandtheir relationto otherkindsof illocutionaryacts is usuallymisunderstood.Let us call this class Declarations.It is the defining characteristicof this class that the successfulperformanceof one of its members brings about the correspondencebetweenthe propositionalcontent and reality,successful performanceguaranteesthatthepropositionalcontent correspondsto the world: if I successfully perform theact of appointingyou chairman,then you arechairman;if I successfullyperformthe act of nominatingyou as candidate,then you area candidate;if I successfullyperform the act of declaring a state of war, then war is on; if I successfullyperformthe act of marryingyou, then you are married. The surfacesyntacticalstructureof many sentences used to performdeclara-tions concealsthis point from us because in them there is no surfacesyntacticaldistinctionbetween propositionalcontent and illocutionaryforce. Thus, 'You'refired' and 'I resign' do not seem to permit a distinction between illocutionaryforce and propositionalcontent, but I think in fact that in their use to performdeclarationstheir semantic structureis: 13 This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

LANGUAGE IN SOCIETY I declare:your employmentis (hereby)terminated. I declare: my position is (hereby)terminated. Declarationsbring about some alternationin the status or condition of thereferred-toobject or objects solely in virtue of the fact that the declarationhasbeen successfully performed. This feature of declarationsdistinguishes themfrom the other categories. In the history of the discussion of these topics sinceAustin's first introduction of his distinction between performativesand con-statives, this feature of declarations has not been properly understood. Theoriginaldistinctionbetween constativesand performativeswas supposed to be adistinction between utterances which are sayings (constatives, statements,assertions,etc.) and utteranceswhich are doings (promises,bets, warnings,etc.).What I am calling declarationswere includedin the class of performatives.Themain theme of Austin's maturework, How to Do Thingswith Words,is that thisdistinction collapses. Just as saying certain things constitutes getting married(a 'performative')and saying certain things constitutes making a promise(another'performative'),so saying certainthings constitutesmakinga statement(supposedlya 'constative').As Austin saw but as many philosophersstill fail tosee, the parallel is exact. Making a statement is as much performingan illo-cutionary act as making a promise, a bet, a warning or what have you. Anyutterancewill consist in performingone or more illocutionaryacts. The illocutionary force-indicating device in the sentence operates on thepropositionalcontent to indicateamongotherthings the directionof fit betweenthe propositionalcontentand reality.In the case of representatives,the directionof fit is words-to-world,in the case of directivesand commissives,it is world-to-words; in the case of expressivesthere is no directionof fit carriedby the illo-cutionaryforce because the existence of fit is presupposed.The utterancecan'tget off the groundunless there alreadyis a fit. But now with the declarationswediscovera very peculiarrelation.The performanceof a declarationbrings abouta fit by the very fact of its successfulperformance.How is such a thing possible? Notice that all of the examples we have considered so far involve an extra-linguisticinstitution,a system of constitutiverulesin additionto the constitutiverules of language,in order that the declarationmay be successfully performed.The mastery of those rules which constitutes linguistic competence by thespeakerandheareris not ingeneralsufficientforthe performanceof a declaration.In addition,there must exist an extra-linguisticinstitution and the speakerandhearermust occupy special places within this institution. It is only given suchinstitutions as the Church, the law, private property, the state and a specialposition of the speaker and hearer within these institutions that one can ex-communicate, appoint, give and bequeath one's possessions or declare war.The only exceptions to the principle that every declarationrequiresan extra-linguistic institution are those declarationsthat concern languageitself, as for I4 This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

A CLASSIFICATION OF ILLOCUTIONARY ACTSexample,when one says, 'I define, abbreviate,name, call or dub'.3Austin some-times talks as if all performatives(and in the general theory, all illocutionaryacts) required an extra-linguistic institution, but this is plainly not the case.Declarationsare a very special categoryof speech acts. We shall symbolizetheirstructureas follows: D : 0(p) Where D indicates the declarationalillocutionarypoint; the direction of fitis both words-to-worldand world-to-wordsbecause of the peculiarcharacterofdeclarations;there is no sincerity condition, hence we have the null symbol inthe sincerity condition slot; and we use the usual propositionalvariablep. The reasonthere has to be a relationof fit arrowhere at all is that declarationsdo attemptto get languageto match the world. But they do not attempt to do iteither by describing an existing state of affairs(as do representatives)nor bytryingto get someone to bringabouta futurestate of affairs(as do directivesandcommissives). Some membersof the class of declarationsoverlapwith membersof theclassofrepresentatives.This is because in certain institutional situations we not onlyascertainthe facts but we need an authorityto lay down a decision as to whatthefacts are afterthe fact-findingprocedurehas been gone through.The argumentmust eventuallycome to an end and issue in a decision, and it is for this reasonthat we havejudges and umpires. Both, the judge and the umpire, makefactualclaims; 'you are out', 'you are guilty'. Such claims are clearly assessablein thedimensionof word-worldfit. Washe reallytaggedoff base?Did he reallycommitthe crime?They areassessablein the word-to-worlddimension.But, at the sametime, both have the force of declarations.If the umpire calls you out (and isupheld on appeal),then for baseballpurposesyou are out regardlessof the factsin the case, and if the judge declaresyou guilty (on appeal),then for legal pur-poses you areguilty. There is nothing mysteriousaboutthese cases. Institutionscharacteristicallyrequireillocutionaryacts to be issued by authoritiesof variouskinds which have the force of declarations.Some institutions requirerepresen-tativeclaimsto be issuedwith the forceof declarationsin orderthat the argumentover the truth of the claim can come to an end somewhereand the next institu-tional steps which wait on the settling of the factualissue can proceed: the pris-oner is released or sent to jail, the side is retired, a touchdown is scored. Theexistenceof this classwe may dub 'Representativedeclarations'.Unlike the otherdeclarations,they share with representativesa sincerity condition. The judge,jury and umpire can, logically speaking, lie, but the man who declareswar ornominates you cannot lie in the performance of his illocutionary act. Thesymbolismfor the class of representativedeclarations,then, is this:(3] Another class of exceptions are supernatural. When God says 'Let there be light' that is a declaration. I5This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

LANGUAGE IN SOCIETY D, I t B(p)Where D. indicates the illocutionarypoint of issuing a representativewith theforce of a declaration,the first arrowindicatesthe representativedirectionof fit,the second indicates the declarationaldirection of fit, the sincerity condition isbelief and the p representsthe propositionalcontent.V. SOME SYNTACTICAL ASPECTS OF THE CLASSIFICATIONSo far I havebeen classifyingillocutionaryacts and have used facts about verbsfor evidence and illustration. In this section I want to discuss explicitly somepoints about English syntax. If the distinctionsmarkedin section IV are of anyreal significancethey are likely to have various syntactical consequences and Inow proposeto examinethe deep structureof explicit performativesentences ineach of the five categories;that is I want to examinethe syntacticalstructureofsentences containing the performativeoccurrenceof appropriateillocutionaryverbsappropriateto each of the five categories.Since all of the sentenceswe willbe consideringwill contain a performativeverb in the main clause, and a sub-ordinate clause, I will abbreviate the usual tree structures in the followingfashion: The sentence, e.g., 'I predictJohn will hit Bill', has the deep structureshown in Figure x. I will simply abbreviatethis as: I predict+ John will hit Bill.Parentheseswill be used to markoptionalelementsor elementsthatareobligatoryonly for restrictedclass of the verbs in question. Wherethere is a choice of oneof two elements, I will put a stroke between the elements, e.g. I/you. RepresentativesT. he deep structureof such paradigmrepresentativesentences NP V NP / /N VI predict S NP VP N Aux V NP ]l I ~~N Jo n will hit Bill FIGURE 1. I6This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

A CLASSIFICATION OF ILLOCUTIONARY ACTSas 'I statethatit is raining'and'I predicthe will come'is simply, I verb(that)+ S.This class,as a class,providesno furtherconstraints;thoughparticularverbsmayprovidefurtherconstraintson the lower node S. For example, 'predict'requiresthat an Aux in the lower S must be future or, at any rate, cannot be past. Suchrepresentativeverbs as 'describe','call', 'classify', and 'identify'take a differentsyntacticalstructure,similar to many verbs of declaration,and I shall discussthem later. Directives.Such sentences as 'I order you to leave' and 'I command you tostand at attention'have the following deep structure: I verb you+ you Fut Vol Verb (NP) (Adv)'I orderyou to leave'is thus the surfacestructurerealizationof 'I orderyou+ youwill leave'with equi NP deletion of the repeated'you'. Notice that an additionalsyntacticalargumentfor my including 'dare', 'defy', and 'challenge',in my listof directive verbs and objecting to Austin's including them with 'apologize','thank','congratulate',etc., is that they havethe same syntacticalform as do theparadigmdirective verbs 'order', 'command',and 'request'. Similarly, 'invite',and 'advise'(in one of its senses) have the directivesyntax. 'Permit'also has thesyntax of directives,though giving permissionis not strictly speakingtrying toget someoneto do something,ratherit consists in removingantecedentlyexistingrestrictionson his doing it. CommissivesS. uch sentences as 'I promise to pay you the money', and 'Ipledge allegianceto the flag',and 'I vow to get revenge',have the deep structure I verb (you)+ I Fut Vol Verb (NP) (Adv).Thus, 'I promise to pay you the money' is the surface structure realizationofI promiseyou+ I will payyou the money, with equi NP deletion of the repeatedI. We hearthe differencein syntaxbetween 'I promiseyou to come on Wednes-day' and 'I orderyou to come on Wednesday'as being that 'I' is the deep struc-ture subject of 'come' in the first and 'you' is the deep structure subject of'come' in the second, as requiredby the verbs 'promise'and 'order'respectively.Notice that not all of the paradigm commissives have 'you' as an indirectobject of the performativeverb. In the sentence 'I pledge allegianceto the flag'the deep structureis not 'I pledge to you flag+ I will be allegiant'.It is I pledge+ I will be allegiantto the flag. Whereasthere are purely syntacticalargumentsthat such paradigmdirectiveverbs as 'order',and 'command',as well as the imperativemood require 'you'as the deep structuresubject of the lower node S, I do not know of any purelysyntacticalargumentto show that commissivesrequire'I' as the deep structure '7This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

LANGUAGE IN SOCIETYsubject on their lower node S. Semantically,indeed, we must interpret suchsentences as 'I promisethat Henry will be here on Wednesday'as meaning I promisethat I will see to it that Henry will be here next Wednesday,in so far as we interpretthe utteranceas a genuine promise,but I know of nopurely syntactical arguments to show that the deep structure of the formersentence containsthe italicizedelements in the latter. Expressives.As I mentioned earlier, expressives characteristicallyrequire agerundivetransformationof the verb in the lower node S. We say: I apologizefor stepping on your toe, I congratulateyou on winning the race, I thankyou for giving me the money,The deep structureof such sentences is: I verb you+ I/you VP =>.GerundiveNom.And, to repeat, the explanationof the obligatorygerundiveis that there is nodirection of fit. The forms that standardly admit of questions concerningdirection of fit, that clauses and infinitives, are impermissible. Hence, theimpossibilityof *I congratulateyou that you won the race, *1 apologizeto step on your toe.However, not all of the permissiblenominalizationtransformationsare gerun-dive; the point is only that they must not produce that clauses or infinitivephrases,thus, we can have either I apologizefor behavingbadly,or I apologizefor my bad behavior,but not, *I apologizethat I behavedbadly, *I apologizeto behave badly. Before considering Declarations, I want now to resume discussion of thoserepresentativeverbs which have a differentsyntax from the paradigmsabove. Ihave said that the paradigmrepresentativeshave the syntacticalform I verb (that)+ S.But, if we considersuch representativeverbs as 'diagnose','call' and 'describe', I8 This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

A CLASSIFICATION OF ILLOCUTIONARY ACTSas well as 'class', 'classify'and 'identify',we find that they do not fit this patternat all. Consider'call', 'describe'and 'diagnose',in such sentences asI call him a liar,I diagnosehis case as appendicitisandI describeJohn as a Fascist.and in generalthe form of this isI verb NP1+ NP1 be pred.One cannot say*I call that he is a liar,*I diagnosethat his case is appendicitis(perversely,some of my students find this form acceptable.),*I describethat John is a Fascist. There, therefore,seems to be a very severe set of restrictionson an importantclass of representativeverbs which is not sharedby the other paradigms.Wouldthis justify us in concluding that these verbs werewronglyclassed as represen-tativesalongwith 'state','assert','claim'and'predict'andthatwe need a separateclass forthem?It might be arguedthat the existence of these verbs substantiatesAustin'sclaimthatwe requirea separateclass of verdictivesdistinctfromexposi-tives, but that would surely be a very curious conclusion to draw since Austinlists most of the verbswe mentionedaboveas expositives.He includes'describe','class', 'identify'and 'call' as expositivesand 'diagnose'and 'describe'as verdic-tives. A common syntax of many verdictives and expositives would hardlywarrantthe need for verdictives as a separateclass. But leaving aside Austin'staxonomy,the question still arises, do we requirea separatesemantic categoryto accountfor these syntacticalfacts?I think not. I thinkthere is a much simplerexplanationof the distributionof these verbs. Often, in representativediscourse,we focus our attentionon some topic of discussion.The questionis notjust whatis the propositionalcontent we are asserting, but what do we say about theobject(s) referredto in the propositional content: not just what do we state,claim, characterize,or assert, but how do we describe, call, diagnoseor identifyit, some previouslyreferredto topic of discussion.When, for example,thereis aquestion of diagnosing or describing it is always a question of diagnosing aperson or his case, of describinga landscapeor a party or a person, etc. TheseRepresentativeillocutionaryverbsgive us a device for isolatingtopics from whatis said about topics. But this very genuine syntacticaldifferencedoes not marka semanticdifferencebig enough to justify the formationof a separatecategory.Notice in support of my argumenthere that the actual sentences in which the '9 This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

LANGUAGE IN SOCIETYdescribing,diagnosing,etc., is done areseldom of the explicitperformativetype,but ratherareusuallyin the standardindicativeformswhich areso characteristicof the representativeclass. Utterancesof He is a liar, He has appendicitis, He is a Fascist,are all characteristicallystatementsi,n the makingof which we call, diagnoseanddescribe,as well as accuse, identify and characterize.I concludethen that thereare typically two syntacticalforms for representativeillocutionaryverbs; one ofwhich focusses on propositionalcontent, the other on the object(s) referredtoin the propositionalcontent, but both of which aresemanticallyrepresentatives. Declarations.I mention the syntacticalform I verb NP + NP1 be predboth to forestallan argumentfor erectinga separatesemanticcategoryfor themandbecausemanyverbsof declarationhavethis form. Indeed,thereappearto beseveral different syntactical forms for explicit performativesof declaration.Ibelieve the following three classes are the most important. (i) I find you guilty as charged. I now pronounceyou man and wife. I appoint you chairman. (2) War is hereby declared. I declarethe meeting adjourned. (3) You'refired. I resign. I excommunicateyou.The deep syntacticalstructureof these three, respectively,is as follows: (i) I verb NP1+ NP1 be pred.Thus, in our examples, we have I find you+ you be guilty as charged. I pronounceyou+ you be man and wife. I appoint you+ you be chairman. (2) I declare+ S.Thus, in our exampleswe have I/we (hereby)declare+ a state of war exists. I declare+ the meeting be adjourned. 20 This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

A CLASSIFICATION OF ILLOCUTIONARY ACTS This form is the purest form of the declaration: the speaker in authoritybrings about a state of affairsspecified in the propositionalcontent by saying ineffect, I declarethe state of affairsto exist. Semantically,all declarationsare ofthis character,though in class (i) the focussing on the topic producesan altera-tion in the syntaxwhich is exactlythe same syntaxas we saw in such representa-tive verbs as 'describe', 'characterize','call' and 'diagnose', and in class (3) thesyntax concealsthe semanticstructureeven more.(3) The syntax of these is the most misleading. It is simply I verb (NP)as in our examples,I fire you.I resign.I excommunicateyou.The semantic structure of these, however, seems to me the same as class (2).'You'refired',if utteredas performanceof the act of firingsomeone and not as areport meansI declare+ Yourjob is terminated.Similarly,'I hereby resign' meansI hereby declare+ My job is terminated.'I excommunicateyou' meansI declare+ Your membershipin the church is terminated.The explanationfor the bemusingly simple syntactical structure of the threesentencesseems to me to be thatwe havesome verbswhich in their performativeoccurrenceencapsulateboth the declarativeforce and the propositionalcontent.VI. CONCLUSIONWe are now in a position to drawcertaingeneralconclusions. (i) Many of the verbswe call illocutionaryverbs arenot markersof illocution-ary point but of some other featureof the illocutionaryact. Consider'insist' and'suggest'. I can insist that we go to the movies or I can suggest that we go to themovies; but I can also insist that the answeris found on p. i6 or I can suggestthat it is found on p. i6. The firstpairaredirectives,the second, representatives.Does this show that insisting and suggesting are different illocutionary actsaltogether from representativesand directives, or perhaps that they are bothrepresentativesand directives?I think the answerto both questions is no. Both'insist' and 'suggest' are used to mark the degree of intensity with which the 21This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

LANGUAGE IN SOCIETYillocutionarypoint is presented.They do not marka separateillocutionarypointat all. Similarly, 'announce', 'present' and 'confide' do not markseparateillo-cutionarypointsbut ratherthe style ormannerof performanceof an illocutionaryact. Paradoxicallyas it may sound, such verbs are illocutionaryverbs, but notnames of kinds of illocutionaryacts. It is for this reason,amongothers, that wemust carefully distinguish a taxonomy of illocutionaryacts from one of illo-cutionaryverbs. (2) In section IV I tried to classify illocutionaryacts, and in section V I triedto explore some of the syntactical features of the verbs denoting members ofeach of the categories.But I have not attemptedto classify illocutionaryverbs.If one did so, I believe the following would emerge. (a) First, as just noted some verbs do not markillocutionarypoint at all, butsome otherfeature,e.g. insist, suggest, announce,confide, reply, answer,inter-ject, remark,ejaculateand interpose. (b) Many verbs markillocutionarypoint plus some other feature,e.g. 'boast','lament','threaten','criticize','accuse'and 'warn'all addthe featureof goodnessor badnessto their primaryillocutionarypoint. (c) Some few verbs mark more than one illocutionarypoint, e.g. a protestinvolves both an expressionof disapprovaland a petition for change. Promulgating a law has both a declarationstatus (the propositionalcontentbecomes law) and a directivestatus (the law is directivein intent). The verbs ofrepresentativedeclarationfall into this class. (d) Some few verbs can take more than one illocutionarypoint. Consider'warn'and 'advise'. Notice that both of these take either the directivesyntax orthe representativesyntax. Thus,I warnyou to stay away from my wife! (directive)I warn you that the bull is about to charge. (representative)I advise you to leave. (directive)Passengersare hereby advised that the trainwill be late. (representative)Correspondingly,it seems to me, that warning and advising may be eithertelling you that somethingis the case (with relevanceto what is or is not in yourinterest) or telling you to do something about it (because it is or is not in yourinterest). They can be, but need not be, both at once. (3) The most importantconclusion to be drawnfrom this discussion is this.There are not, as Wittgenstein(on one possible interpretation)and many othershave claimed, an infinite or indefinite number of language games or uses oflanguage.Rather,the illusion of limitless uses of languageis engenderedby anenormous unclarityabout what constitutes the criteriafor delimiting one lan-guage game or use of languagefrom another.If we adopt illocutionarypoint asthe basic notion on which to classify uses of language,then there are a rather 22 This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions

A CLASSIFICATION OF ILLOCUTIONARY ACTSlimited numberof basic things we do with language:we tell people how thingsare, we try to get them to do things, we commit ourselves to doing things, weexpress our feelings and attitudes and we bring about changes through ourutterances.Often, we do morethan one of these at once in the same utterance. REFERENCESAnscombe, G. E. M. (957). Intention. Oxford: Basil Blackwell.Austin, J. L. (I962). How to do things with words. Harvard and Oxford.Searle, J. R. (I968). Austin on locutionary and illocutionary acts. Philosophical Review LXXVII. 405-424.- (I969). Speech acts: an essay in the philosophy of language. Cambridge.EDITOR S NOTEThe editor believes, as stated in his correspondencewith the author, that thisarticlewill be of use to linguisticethnographers.The workof identifyinglocallyvalid systems of illocutionaryacts will be stimulatedand aided by the clarity offocus attained here by Searle. In turn, ethnographicdiscoveries will test theuniversalityof the criteriaand kinds of illocutionaryacts, and enableus to beginto understandtypologicallydifferencesin hierarchyand markednessamonglocalsystems. (The supernaturalexceptions,noted in n. 3, arelikelyto proveparticu-larlyimportant.) 23This content downloaded from 132.235.75.160 on Thu, 4 Dec 2014 11:20:35 AM All use subject to JSTOR Terms and Conditions