Must We Disimplicate It All? Stanley Louis Cavell on Herbert Paul Grice

Speranza

MUS T W E MEA N WHA T W E SAY? 1 by Stanley Cavell Universit y o f California , Berkeley Tha t wha t w e ordinaril y say an d mean may hav e a direc t an d dee p contro l ove r wha t w e can philosophicall y say an d mean i s a n ide a whic h many philosopher s fin d oppressive . I t migh t b e argue d tha t i n par t th e oppressio n result s fro m misunderstanding ; tha t th e ne w philosophy whic h proceed s fro m ordinary languag e i s no t that differen t fro m traditiona l method s o f philosophizing , an d tha t th e frequen t attack s upo n i t ar e misdirected . Bu t I shal l no t attemp t t o b e conciliatory , both becaus e I thin k th e ne w philosophy at Oxfor d i s criticall y differen t fro m tradi - tiona l philosophy , an d becaus e I thin k i t i s worth tryin g t o brin g out thei r difference s a s fully a s possible . Ther e is, afte r all , somethin g oppressiv e about a philosophy whic h seem s t o hav e uncanny informatio n about ou r mos t persona l philosophica l assumption s (those , fo r example , about whethe r w e can eve r kno w fo r certai n o f th e existenc e o f th e externa l world , o r o f othe r minds ; an d thos e w e mak e about favorit e distinction s betwee n "th e descriptiv e an d th e normative" , o r betwee n matter s o f fac t an d matter s o f language ) an d whic h inveteratel y nag s u s about them . Particularl y oppressiv e whe n that , philosoph y seem s s o ofte n merely t o na g an d t o try n o specia l answer s t o th e question s whic h posses s u s — unles s i t b e t o sugges t tha t w e si t quietl y i n a room . Eventually , I suppose , w e wil l hav e t o loo k at tha t sens e o f oppressio n itself: suc h feeling s can com e fro m a trut h about ourselve s whic h w e ar e holdin g off. M y hope s her e ar e modest . I shal l wan t t o say why , i n my opinion , som e o f th e •argument s Professo r Mate s bring s agains t th e Oxfor d philo - sopher s h e mention s ar e o n th e whol e irrelevan t t o thei r mai n concerns . An d thi s wil l requir e m e t o say somethin g about wha t I tak e t o b e t appearing alternately trivial and dogmatic. Perhaps that is only to be expected, given the depth and the intimacy of conflict between this way of proceeding in philosophy and the way I take Mates to be following. These ways of philosophy seem, like friends who have quarreled, to be able neither to tolerate nor to ignore one another. I shall frequently be saying something one could not fail to know; and that will appear trivial. I shall also be suggesting that something we know is being overemphasized and something else not taken seriously enough; and that will appear dogmatic. But since I am committed to this dialogue, the time is past for worrying about appearances. Professor Mates is less concerned to dispute specific results of the Oxford philosophers than he is to question the procedures which have led these philosophers to claim theln. In particular, he doubts that they have assembled the sort of evidence which their "statements about ordinary language" require. As a basis for his skepticism, Mates produces a disagreement between two major figures of the school over the interpretation of an expression of ordinary language — a disagreement which he regards as symptomatic of the shallowness of their methods.2 On Mates' account of it, the conflict is not likely to be settled successfully by further discussion. We are faced with two professors (of philosophy, it happens) each arguing (claiming, rather) that the way he talks is the right way and that what he intuits about language is the truth about it. But if this is what their claims amount to, it hardly seems worth a philosopher's time to try to collect evidence for them. To evaluate the disagreement between Austin and Ryle, we may distinguish among the statements they make about ordinary language, three types:3 (1) There are statements which produce instances of what is said in a language ("We do say .. . but we don't say "; "We ask whether... but we do not ask whether "); (2) Sometimes these instances are accompanied by explications—statements which make explicit what is implied when we say what statements of the first type instance us as saying ("When we say... we imply (suggest, say) '-'; "We don't say... unless we mean ."). Such statements are' checked by reference to statements of the first type. (3) Finally, there are generalizations, to be tested by reference to statements of the first two types. Since there is no special problem here about the testing of generalizations, we will be concerned primarily with the justification of statements of the first two types, and especially with the second. 173 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 Eve n withou t attemptin g t o b e mor e precis e about thes e differences , th e natur e o f th e das h betwee n Ryl e an d Austi n become s somewhat dearer . Notice , firs t o f all , tha t th e statemen t Mate s quote s fro m Austi n i s o f th e firs t type : "Tak e 'voluntarily*... : w e may.. . mak e a gift voluntarily... " — whic h I tak e t o b e materia l mod e for , "W e say , 'Th e gift wa s mad e voluntarily * ? . (Th e significanc e o f thi s shift o f "mode " wil l b e discussed. ) Onl y on e of th e many statement s Mate s quote s fro m Ryl e i s o f thi s type , viz., "I t make s sens e .. . t o as k whethe r a bo y wa s responsibl e fo r breakin g a window , bu t no t whethe r h e wa s responsibl e fo r finishin g hi s homework i n goo d time..." . Th e state - ment s o f Ryle' s whic h das h with Austin' s ar e different : "I n thei r mos t ordinary employment 'voluntary * an d 'involuntary * ar e used.. . a s adjec - tive s applyin g t o action s whic h ough t no t t o b e done . W e discus s whethe r someone' s actio n wa s voluntary o r no t onl y whe n th e actio n seem s t o hav e bee n hi s fault.. . etc. " These'd o no t produc e instances o f wha t w e say (th e way "W e say , 'Th e bo y wa s responsibl e fo r breakin g th e window' " does) ; the y ar e generalization s — a s th e phrase s "action s which " an d "onl y when " sho w — t o b e teste d by producin g suc h instances . I t i s tru e tha t th e instanc e quote d fro m Austi n doe s g o counte r t o Ryle' s generalization : makin g a gift i s no t alway s somethin g whic h ough t no t t o b e done , o r somethin g whic h i s alway s someone' s fault . Ther e i s dearl y a das h here . But i s our onl y intelligen t cours e at this point t o tak e a poll ? Woul d i t b e dogmati c o r unempirica l o f u s t o condud e simpl y tha t Ryl e i s wron g about this , tha t h e ha s settle d upo n a generalizatio n t o whic h an obviou s counte r instanc e ha s bee n produced ? I t is , moreover , an instanc e whic h Ryl e himsel f may wel l b e expecte d t o acknowledg e a s counter , t o hi s generalization ; indeed , on e whic h h e migh t hav e produce d fo r himself. Th e fac t tha t h e di d no t nee d indicat e onl y tha t h e wa s to o quic k t o accep t a generalization , no t tha t h e i s withou t (good ) evidenc e fo r it . On e o f Mates' objection s t o Ryl e can b e pu t thi s way : Ryl e is withou t evidenc e — anyway , withou t very goo d evidenc e — becaus e h e i s no t entitle d t o a statement o f th e firs t typ e (on e whic h present s an instance o f wha t w e say ) i n th e absenc e o f experimenta l studie s whic h demonstrat e it s occurrenc e i n th e language . T o se e tha t thi s objection , take n i n th e genera l sens e i n whic h Mate s urge s it , i s groundless , w e mus t bea r i n min d th e fac t tha t thes e state - ment s — statement s tha t somethin g i s sai d i n Englis h — ar e bein g mad e b y nativ e speaker s o f English . Such speaker s d o not , i n general, nee d evidenc e fo r wha t i s sai d i n th e language ; the y ar e th e sourc e o f suc h 17 4 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 evidence. It is from them that the descriptive linguist takes the corpus of utterances on the basis of which he will construct a grammar of that language. To answer some kinds of specific questions, we will have to engage in that "laborious questioning" Mates insists upon, and count noses; but in general, to tell what is and isn't English,, and to tell whether what is said is properly used, the native speaker can rely on his own nose; if not, there would be nothing to count. No one speaker will say everything, so it may be profitable to seek out others; and sometimes you (as a native speaker) may be unsure that a form of utterance is as you say it is, or is used as you say it is used, and in that case you will have to check with another native speaker. And because attending so hard to what you say may itself make you unsure more often than is normal, it is a good policy to check more often. A good policy, but not a methodological necessity. The philosopher who proceeds from ordinary language, in his use of himself as subject in his collection of data, may be more informal than the descriptive linguist (though not more than the linguistic theorist using examples from his native speech); but there is nothing in that to make the data, in some general way, suspect. Nor does this imply a reliance on that "intuition or memory" which Mates (p. 165) finds so objectionable. In claiming to know, in general, whether we do or do not use a given expression, I am not claiming to have an infallible memory for what we say, any more than I am claiming to remember the hour when I tell you what time we have dinner on Sundays. A normal person may forget and remember certain words, or what certain words mean, in his native language, but (assuming that he has used it continuously) he does not remember the language. There is a world of difference between a person who speaks a language natively and one who knows the language fairly well. If I lived in Munich and knew German fairly well, I might try to intuit or guess what the German expression for a particular phenomenon is. Or I might ask my landlady; and that would probably be the extent of the laborious questioning the problem demanded. Nor does the making of either of the sorts of statement about ordinary language I have distinguished rely on a claim that "[we have} already amassed... a tremendous amount of empirical information about the use of [our} native language" (Matesr-ibid.) That would be true if we were, say, making statements about the history of the language, or about its sound system, or about the housewife's understanding of political slogans, or about a special form in the morphology of some dialect. But for a native speaker to say what, in ordinary circumstances, is said when, no such special, information is needed or claimed. 175 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 Al l tha t i s neede d i s th e trut h o f th e propositio n tha t a natura l languag e i s wha t nativ e speaker s o f tha t languag e speak . Ryle' s generalization , however , require s mor e tha n simple , firs t leve l statement s o f instances ; i t als o require s statement s o f th e secon d type , thos e whic h contai n firs t leve l staterMnt s togethe r with a n "explication " o f them . Whe n Ryl e claim s tha t ".. . w e rais e question s o f responsibility onl y whe n someon e i s charged , justl y o r unjustly , with a n offence" , h e i s claimin g both , "W e say Th e bo y wa s responsibl e fo r breakin g a window 1 , bu t w e d o no t say Th e bo y wa s responsibl e fo r finishin g hi s homework i n goo d time' " , an d als o claiming , "Whe n w e say Th e bo y wa s responsibl e fo r (som e action)' • w e impl y tha t th e actio n wa s a n offence , on e tha t ough t no t t o hav e bee n done , on e tha t wa s hi s fault" . I wan t t o argu e tha t Ryl e is , i n general , a s entitle d t o statement s o f thi s secon d typ e a s he.i s t o statement s o f th e first' type ; althoug h i t i s jus t her e tha t th e particula r generalizatio n i n questio n misses . W e kno w Austin' s exampl e counter s Ryle' s clai m becaus e w e kno w tha t th e state - men t (o f th e secon d type) , "Whe n w e say , Th e gift wa s mad e volun - tarily * w e impl y tha t th e actio n o f makin g th e gift wa s on e whic h ough t no t t o b e done , o r wa s someone' s fault " i s false . Thi s i s clearl y knowledg e whic h Mate s wa s relyin g o n whe n h e produce d th e clas h betwee n them . I wil l tak e u p statement s o f th e second typ e i n a moment . Befor e proceedin g t o that , le t u s loo k at tha t das h a bi t longer : it s importanc e ha s altere d considerably . Wha t Austi n say s doe s no t g o full y counte r t o Ryle' s story . I t i s fundamenta l t o Austin' s account t o emphasiz e tha t w e canno t always say o f action s tha t the y wer e voluntary , eve n whe n the y obviousl y wer e no t involuntary either . Althoug h w e can (sometimes ) say , "Th e gift wa s mad e voluntarily" , i t i s specifically no t somethin g w e can say about ordinary , unremarkabl e case s o f makin g gifts . Onl y whe n th e actio n (o r circumstances ) o f makin g th e gift i s i n som e way unusua l (instea d o f hi s usua l Christma s bottle , yo u giv e th e neighborhoo d policeman a chec k fo r $1000) , o r extraordinary (yo u leav e you r heir s penniles s an d bequeath you r hous e t o your cat) , o r untowar d (yo u giv e your rockin g hors e t o your ne w friend , bu t th e nex t mornin g yo u cry t o hav e i t back) , can th e questio n whethe r i t wa s voluntary intelligibl y arise . Ryl e ha s no t completel y neglecte d this : hi s "action s whic h ough t no t b e done " an d hi s "actio n [which } seem s t o hav e been.. . [someone's ] fault " ar e dearl y example s o f action s whic h ar e abnormal , untoward , questionable ; s o h e i s righ t i n sayin g tha t abou t thes e w e (sometimes ) rais e th e questio n whethe r the y wer e voluntary . 17 6 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 His error lies in characterizing these actions incompletely, and in wrongly characterizing those about which the question cannot arise. Normally, it is true, the question whether satisfactory, correct or admirable performances are voluntary does not arise; but this is because there is usually nothing about such actions to question; nothing has gone wrong. Not seeing that the condition for applying the term "voluntary" holds quite generally — viz., the condition that there be something (real or imagined) fishy about any performance intelligibly so characterized — Ryle construes the condition too narrowly, supposes that there must be something morally fishy about the performance. He had indeed sensed trouble where trouble was: the philosophical use of "voluntary" stretches the idea of volition out of shape, beyond recognition. And his diagnosis of the trouble was sound: philosophers imagine, because of a distorted picture of the mind, that the term "voluntary" must apply to all actions which are not involuntary (or unintentional), whereas it is only applicable where there is some specific reason to raise the question. The fact that Ryle fails to specify its applicability precisely enough no more vitiates his entire enterprise than does the fact that he indulges a mild form of the same vice he describes: he frees himself of the philosophical tic of stretching what is true of definite segments of what we do to cover everything we do (as epistemologists stretch doubt to cover everything we say), but not from the habit of identifying linguistic antitheses with logical contradictories:4 in particular, he takes the question, "Voluntary or not?" to mean, "Voluntary or involuntary?", and seems to suppose that (responsible) actions which are not contemptible must be admirable, and that whatever I (responsibly) do either is my fault or else is to my credit. These antitheses miss exactly those actions about which the question "Voluntary or not?" really has no sense, viz., those ordinary, unremarkable, natural things we do which make up most of our conduct and which are neither admirable nor contemptible; which, indeed, could only erroneously be said to go on, in general, in any special way.5 Lacking firmness here, it is not surprising that Ryle's treatment leaves the subject a bit wobbly. Feeling how enormously wrong it is to remove "voluntary" from a specific function, he fails to sense the slighter error of his own specification.6 I have said that the ordinary language philosopher is also and equally entitled to statements of the second type I distinguished, which means that he is entitled not merely to say what (words) we say, but equally 177 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 t o say wha t w e shoul d mean i n (by ) sayin g them . Le t u s tur n t o state - ment s o f thi s typ e an d as k wha t th e relatio n i s betwee n wha t yo u explicitl y say an d wha t yo u imply ; or , t o avoi d beggin g th e question , as k ho w w e ar e t o account fo r th e fac t (supposin g i t t o b e a fact ) tha t w e onl y say o r as k A (" X i s voluntary" , o r "I s X voluntary?" ) wher e B i s th e cas e (somethin g is , o r seems , fishy about X). 7 Th e philosophica l proble m abou t thi s arise s i n th e followin g way : Philosopher s wh o pro - cee d fro m ordinary languag e ar e likel y t o insis t tha t i f yo u say A wher e B i s no t th e case , you wil l b e misusin g A , o r distortin g it s meaning . Bu t anothe r philosophe r wil l no t wan t t o allo w that , becaus e i t make s th e relatio n betwee n A an d B appear t o b e a logica l on e (I f A the n B ; an d i f notB the n not-A) ; wherea s logica l relation s hol d onl y betwee n state - ments , no t betwee n a statemen t an d th e world : that relatio n i s "merely " conventiona l (or , even , causal?) . S o th e occasio n o n whic h w e (happe n to? ) use a statemen t canno t b e considere d par t o f it s meanin g o r logic . Th e solutio n i s the n t o cal l th e latte r th e semantic s o f th e expressio n an d th e forme r it s pragmatics . Bu t i f w e can forge t fo r a moment tha t th e relatio n betwee n A an d B cannot b e a logica l one , w e may com e t o fee l ho w implausibl e i t i s t o say tha t i t i s no t logical ; o r rather , t o say tha t nothin g follows about B fro m th e utteranc e o f A . I t i s implausibl e becaus e w e d o no t accep t a questio n lik e "Di d yo u d o tha t voluntarily? " a s appropriat e about any an d every action . I f a perso n ask s you whethe r yo u dres s th e way yo u d o voluntarily , yo u wil l no t understan d hi m t o b e curiou s merely about you r psychologica l processe s (whethe r your wearin g the m "proceed s fro m fre e choice...") ; yo u wil l understan d hi m t o b e implyin g o r suggestin g tha t your manne r o f dres s i s i n som e way peculiar . I f i t b e replie d t o thi s tha t "voluntary " doe s no t mean "peculiar " (o r "special " o r "fishy" ) an d henc e tha t th e implicatio n o r suggestio n i s par t merel y o f th e pragmatic s o f th e expression , no t par t o f it s meaning (semantics) , my rejoinde r i s this : tha t repl y i s relevan t t o a differen t clai m fro m th e on e urge d here ; i t i s worth sayin g here onl y i f yo u ar e abl e t o account fo r th e relation betwee n th e pragmatic s an d th e semantic s o f th e ex - pression . I n th e absenc e o f such an account , th e repl y i s empty . Fo r con - sider : I f w e us e Mates' formul a fo r computin g th e pragmati c valu e o f a n expressio n — "H e wouldn' t say tha t unles s he... " — the n i n th e describe d situatio n w e wil l complet e i t wit h somethin g like , ".. . unles s h e though t tha t my way o f dressin g i s peculiar" . Gil l thi s implicatio n o f th e utteranc e "pragmatic" ; th e fac t remain s tha t h e wouldn' t (couldn't ) say wha t h e di d withou t implyin g wha t h e did : h e MUS T MEA N tha t 17 8 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 my clothes are peculiar. I am less interested now in the "mean" than I am in the "must". (After all, there is bound to be some reason why a number of philosophers are tempted to call a relation logical; "must" is logical.) But on this, the "pragmatic" formula throws no light whatever. What this shows is that the formula does not help us account for the element of necessity ("must") in statements whose implication we understand. But it is equally unhelpful in trying to explain the implication of a statement whose use we do not understand (the context in which the formula enters Mates' discussion). Imagine that I am sitting in my counting house counting up my money. Someone who knows that I do that at this hour every day passes by and says, "You ought to do that". What should we say about his statement? That he does not know what "ought" means (what the dictionary says) ? That he does not know how to use the word? That he does not know what obligation is? Applying the formula, we compute: "He wouldn't say that unless he asks himself whenever he sees anyone doing anything, 'Ought that person to be doing that or ought he not?' ". This may indeed account for his otherwise puzzling remark; but it does so by telling us something we did not know about him; it tells us nothing whatever we did not know about the words he used. Here it is because we know the meaning and use of "ought" that we are forced to account in the way Mates suggests for its extraordinary occurrence. I take Mates' formula, then, to be expandable into: "Since I understand the meaning and use of his expression, he wouldn't say that unless he . . .". Perhaps Mates would consider this a distortion and take a different expansion to be appropriate: "He wouldn't say that unless he was using his words in a special way". But now "say that" has a very different force. The expanded form now means, "I know what his expression would ordinarily be used to say, but he can't wish to say that: I don't understand what he is saying". In neither of its expansions, then, does the formula throw any light on the way an expression is being used: in the one case we already know, in the other we have yet to learn. (Another expansion may be: "He wouldn't say that unless he was using X to mean Y". But here again, it is the semantics and pragmatics of Y which are relevant to understanding what is said, and the formula presupposes that we already understand Y.) Our alternatives seem to be these: Either (1) we deny that there is any rational (logical, grammatical) constraint over the "pragmatic implications" of what we say — or perhaps deny that there are any impli179 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 cations, o n th e groun d tha t th e relatio n i n questio n i s no t deductiv e — s o tha t unles s what I say i s flatl y fals e o r unles s I explicitl y contradic t myself, i t i s pointles s t o sugges t tha t wha t I say i s wron g o r tha t I mus t mea n somethin g othe r tha n I say ; o r els e (2 ) w e admi t th e constrain t an d say eithe r (a ) sinc e al l necessity i s logical , th e "pragmati c impli - cations " o f ou r utteranc e ar e (quasi-)logica l implications ; wit h o r withou t addin g (b ) sinc e th e "pragmati c implications " canno t b e con - strue d i n term s o f deductiv e (o r inductive ) logic , ther e mus t b e som e "thir d sort " o f logic ; o r w e say (c ) som e necessity i t no t logical . Non e o f thes e alternative s i s withou t it s obscurities , bu t the y ar e clear enoug h fo r u s t o se e tha t Mate s i s takin g alternativ e (I), 8 wherea s th e philo - sophe r wh o proceed s fro m ordinary languag e i s likel y t o fee l th e nee d o f som e form o f (2) . Alternativ e (2a ) bring s out par t o f th e reaso n behin d th e Oxfor d philosopher' s insistenc e tha t h e i s talkin g logic , whil e (2b ) make s explici t th e reaso n othe r philosopher s ar e perplexe d a t tha t claim. 0 Th e differenc e betwee n alternative s (1 ) an d (2 ) i s fundamental ; s o fundamental , tha t i t i s very difficul t t o argue . Whe n Mate s says , "Per - hap s i t i s tru e tha t ordinaril y I wouldn' t say ' I kno w i f unles s I fel t grea t confidenc e i n wha t I wa s asserting..." , wha t h e say s i s not , i f yo u like , strictly wrong ; bu t i t i s wron g — or , wha t i t implie s i s wrong . I t implie s tha t whethe r I confin e th e formul a " I know... " t o state - ment s abou t whic h I fee l grea t confidenc e i s up to me (rightly u p t o me) ; s o tha t i f I say " I know... " i n th e absenc e o f confidence , I hav e no t misuse d language , an d i n particula r I hav e no t stretche d th e meaning o f th e wor d "know" . An d yet , i f a chil d wer e t o say " I kno w ... " whe n yo u kno w th e chil d doe s no t know (i s i n n o positio n t o say h e knows ) yo u may reply , "Yo u don' t reall y mean (nb ) you know, yo u only mean yo u believe" ; o r yo u may say , "Yo u oughtn' t t o say yo u know whe n yo u onl y think so" . Ther e ar e occasion s o n whic h i t woul d b e usefu l t o hav e th e "semantic - pragmatic " distinctio n a t hand . If, fo r example , a philosophe r tell s m e tha t th e statement , "Yo u ough t t o d o so-and-so " expresse s privat e emo - tio n an d i s hortatory an d henc e not , strictl y speaking , meaningful , the n i t may b e worth replyin g tha t nothin g follow s abou t th e meanin g (semantics ) o f a statemen t fro m th e way i t i s use d (pragmatics) ; an d thi s repl y may spar e ou r havin g t o mak e u p specia l brand s o f meaning . Bu t th e tim e fo r tha t argumen t is , presumably , past. 1 0 Wha t need s t o others) to make certain inferences, draw certain conclusions. (This is part of what'you say when you say that you are talking about the logic of ordinary language.) Learning what these implications are is part of learning the language; no less a part than learning its syntax, or learning what it is to which terms apply: they are an essential part of what we communicate when we talk. Intimate understanding is understanding which is implicit. Nor could everything we say (mean to communicate), in normal communication, be said explicitly11 — otherwise the only threat to communication would be acoustical. We are, therefore, exactly as responsible for the specific implications of our utterances as we are for their explicit factual claims. And there can no more be some general procedure for securing that what one implies is appropriate than there can be for determining that what one says is true. Misnaming and misdescribing are not the only mistakes we can make in talking. Nor is lying its only immorality. I am prepared to conclude that the philosopher who proceeds from ordinary language is entitled, without special empirical investigation, to assertions of the second sort we distinguished, viz., assertions like, "We do not say 'I know...' unless we mean that we have great confidence"...", and like "When we ask whether an action is voluntary we imply that the action is fishy" (call this S). But I do not think that I have shown that he is entitled to them, because I have not shown what kind of assertions they are; I have not shown when such assertions should be said, and by whom, and what should be meant in saying them. It is worth trying to indicate certain complexities of the assertions, because they are easy to overlook. Something important will be learned if we realize that we do not know what kind of assertion S is. When (if) you feel that S is necessarily true, that it is a priori, you will have to explain how a statement which is obviously not analytic can be true a priori. (That S is not analytic is what (is all) that is shown by Mates' arguments about the "semantic-pragmatic" confusion; it is perfectly true that "voluntary" does not mean (you will not -find set beside it in a dictionary) "fishy".) When I am impressed with the necessity of statements like S, I am tempted to say that they are categorial — about the concept of an action iiberbaupt. (A normal action is neither voluntary nor involuntary, neither careful nor careless, neither expected nor unexpected, neither right nor wrong...) This would account for our feeling of their necessity: they are instances (not of 181 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 Formal , but ) o f Transcendenta l Logic . Bu t thi s i s reall y n o explanatio n unti l w e mak e deare r th e nee d fo r th e concep t o f an actio n i n general . Howeve r difficul t i t i s t o mak e ou t a cas e fo r th e necessity o f S , i t i s importan t tha t th e temptatio n t o cal l i t a priori no t b e ignored ; other - wis e w e wil l acquiesc e i n callin g i t synthetic , whic h woul d b e badl y misleading . Misleadin g (wrong ) becaus e w e kno w wha t woul d count a s a disproo f o f statement s whic h ar e syntheti c (t o indicat e th e willing - nes s t o entertai n suc h disproo f i s th e poin t o f callin g a statemen t syn - thetic) , bu t i t i s no t dea r wha t woul d coun t a s a disproo f o f S . Th e feelin g tha t S mus t b e syntheti c comes , o f course , partl y fro m th e fac t tha t i t obviousl y i s no t (likel y t o b e take n as ) analytic . But i t come s als o fro m th e eas e with whic h S may b e mistake n fo r th e statement , "'I s X voluntary?' implie s tha t X i s fishy " (T) , whic h doe s seem obviously .synthetic . Bu t S an d T , thoug h the y ar e tru e togethe r an d fals e together , ar e no t everywher e interchangeable ; th e identica l stat e o f affair s i s de - scribe d b y both , but a perso n wh o may b e entitle d t o say T , may no t b e entitle d t o say S . Onl y a nativ e speake r o f Englis h i s entitle d t o th e state - men t S , wherea s a linguis t describin g Englis h may , thoug h h e i s no t a nativ e speake r o f English , b e entitle d t o T . Wha t entitle s hi m t o T i s hi s havin g gathere d a certai n amoun t an d kin d o f evidenc e i n it s favor . Bu t th e perso n entitle d t o S i s no t entitle d t o that statement for th e sam e reason . H e needs n o evidenc e fo r it . I t woul d b e misleadin g t o say tha t h e has evidenc e fo r S , fo r tha t woul d sugges t tha t h e ha s don e th e sort o f investigatio n th e linguis t ha s done , onl y les s systematically , an d thi s woul d make i t seem tha t hi s dai m t o kno w S i s very weakly based . An d i t woul d b e equall y misleadin g t o say tha t h e doe s not hav e evidenc e fo r S , becaus e tha t woul d mak e i t appea r tha t ther e i s some - thin g h e stil l needs , an d suggest s tha t h e i s no t ye t entitle d t o S . But ther e i s nothin g h e needs , an d ther e i s n o evidenc e (whic h i t make s sense , i n general, t o say ) h e has : th e questio n o f evidenc e i s irrelevant . A n examinatio n o f wha t doe s entitl e a perso n t o th e statemen t S woul d b e require d i n any ful l account o f suc h statements . Such an examinatio n i s ou t o f th e questio n here . Bu t sinc e I wil l wan t t o clai m tha t Mates' "tw o methods " fo r gatherin g evidenc e i n support o f "statement s about ordinary language " lik e S ar e irrelevan t t o wha t entitle s a perso n t o S , an d sinc e thi s obviously rest s o n th e dai m tha t th e concep t o f evidenc e is , i n general , irrelevan t t o the m altogether , le t m e say jus t this : Th e clu e t o understandin g th e sort o f statement S i s lie s i n appreciatin g th e fac t tha t "we" , whil e plural , i s firs t person . Firs t perso n singular form s hav e recentl y com e i n fo r a grea t dea l o f attention , an d the y hav e bee n 18 2 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 shown to have very significant logical-epistemological properties. The plural form has similar, and equally significant, properties; but it has been, so far as I know, neglected.12 The claim that in general we do not require evidence for statements in the first person plural present indicative, does not rest upon a claim that we cannot be wrong about what we are doing or about what we say, but only that it would be extraordinary if we were (often). My point about such statements, then, is that they are sensibly questioned only where there is some special reason for supposing what I say about what I (we) say to be wrong; only here is the request for evidence competent. If I am wrong about what he does (they do), that may be no great surprise; but if I am wrong about what I (we) do, that is liable, where it is not comic, to be tragic. Statements like T have their own complexities, and it would be unwise even of them to say simply that they are synthetic. Let us take another of Mates' examples: "'I know if is not (ordinarily) said unless the speaker has great confidence in it" (T')< Mates takes this as patently synthetic, a statement about matters of fact (and there is no necessary connection among matters of fact). And so it might be, said by a Scandinavian linguist as part of his description of English. But if that linguist, or if a native speaker (i.e., a speaker entitled to say, "We do not say 'I know it' unless...") uses T' in teaching someone to speak English, or to remind a native speaker of something he knows but is not bearing in mind, T' sounds less like a descriptive statement than like a rule. Because of what seems to be the widespread idea that rules always sort with commands and must therefore be represented as imperatives, this complementarity of rule and statement may come as something of a shock. But that such complementarity exists can be seen in writings which set out the rules for games or ceremonies or languages. In Hoyle's Rules of Games we find statements like, "The opponent at declarer's left makes the opening lead .. . Declarer's partner then lays his whole hand face up on the table, with his trumps if any on the right. The hand so exposed is the dummy. .. . The' object of play is solely to win tricks, in order to fulfil or defeat the contract"; in Robert's Rules of Order, the rules take the form, "The privileged motion to adjourn takes precedence of all others, except the privileged motion 'to fix the time to which to adjourn', to which it yields" (in Section 17, headed "To Adjourn"); taking a grammar at random we find, "Mute stems form the nominative singular by the addition of -s in the case of masculines and feminines .. . Before -s of the nominative singular, a labial mute (p, b) remains unchanged." These are all statements in the indicative, not the imperative, 183 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 mood . (Som e expression s i n eac h o f thes e book s tel l u s wha t w e must do ; other s tha t w e may. I wil l sugges t late r a reaso n fo r thi s shift. ) I n on e light , the y appea r t o b e descriptions ; i n anothe r t o b e rules . Wh y shoul d thi s b e so ? Wha t i s it s significance ? Th e explanatio n o f th e complementarity ha s t o d o with th e fac t tha t it s topi c i s actions . Whe n w e say ho w a n actio n i s don e (ho w t o act ) wha t w e say may repor t o r describ e th e way w e in fact d o i t (i f w e ar e entitle d t o say ho w "we " d o it , i.e. , t o say wha t w e do , o r say wha t w e say ) bu t i t may als o lay out a way o f doin g o r sayin g somethin g whic h i s t o b e followed. Whethe r remark s lik e T' — remark s "about " ordinary language , an d equall y about ordinary action s — ar e statement s o r rule s depend s upo n ho w the y ar e taken : i f the y ar e take n t o stat e fact s an d ar e suppose d t o b e believed , the y ar e statements ; i f the y ar e take n a s guide s an d suppose d t o b e followed , the y ar e rules . Such expression s ar e n o mor e "i n themselves " rule s o r (synthetic ) statement s than othe r expression s are , i n themselves , postulate s o r conclusion s o r definition s o r replies . W e migh t pu t th e relatio n betwee n th e tw o context s o f T' thi s way : Statement s whic h describ e a languag e (o r a gam e o r an institu - tion ) ar e rule s (ar e binding ) i f yo u wan t t o speak tha t languag e (play tha t game , accep t tha t institution) ; or , rather , when yo u ar e speakin g tha t language , playin g tha t game , etc . / / / / is TRU E to say "I know it' is not used unless you have great confidence in it", then, when you are speaking English, it is WRON G (a misure) to say "I know it"unless you have great confidence in it. No w th e philosophe r wh o proceed s fro m ordinary languag e assume s tha t h e an d hi s interlocutor s ar e speakin g fro m withi n th e language , s o tha t th e questio n o f whethe r you wan t t o speak tha t languag e i s pointless . Wors e tha n pointless , becaus e strictl y th e ordinary languag e philosophe r doe s not , i n general , assume tha t h e an d hi s interlocutor s ar e speakin g fro m withi n a give n (thei r native ) languag e — any mor e tha n the y speak thei r nativ e language , i n general , intentionally. Th e onl y conditio n relevan t t o suc h philosophizin g i s tha t yo u speak (no t thi s o r tha t language , but ) period . A t thi s poin t th e argument ha s becom e aporetic . "Statement s about ordinary language " lik e S , T an d T' ar e no t analytic , and the y ar e no t (i t woul d b e misleadin g t o cal l them ) syntheti c (jus t lik e that). 1 3 No r d o w e kno w whethe r t o say the y ar e a priori, o r whethe r t o account fo r thei r ai r o f necessity a s a dialectica l illusion , du e mor e t o th e motio n o f ou r argumen t tha n t o thei r own nature . Give n our curren t alternatives , ther e i s n o way t o classify suc h statements ; w e d o no t ye t kno w wha t the y are . . 18 4 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 Before searching for new ways into these problems) I should perhaps justify my very heavy reliance on the idea of context, because on Mates' description of what a statement of context involves, it should be impossible ever to make one. Let me recall his remarks: "We have all heard the wearying platitude that "you can't separate" the meaning of a* word from the entire context in which it occurs, including not only the actual linguistic context, but also the aims, feelings, beliefs, and hopes of the speaker, the same for the listener and any bystanders, the social situation, the physical surroundings, the historical background, the rules of the game, and so on ad infinitum" (p. 168). Isn't this another of those apostrophes to the infinite which prevents philosophers from getting, down to cases?1 * Of course if I have to go on about the context of "voluntary" ad infinitum, I would not get very far with it. But I would claim to have characterized the context sufficiently (for the purpose at hand) by the statement that something is, or is supposed to be, fishy about the action. Giving directions for using a word is no more prodigious and unending a task than giving directions for anything else. The context in which I make a martini with vodka is no less complex than the context in which I make a statement with "voluntary". Say, if you like, that these actions take place in infinitely complex contexts; but then remember that you can be given directions for doing either. It may be wearying always to be asked for a. story within which a puzzling remark can seriously be imagined to function; but I know no better way of maintaining that relevance, or sense of reality, which each philosopher claims for himself and claims to find lacking in another philosophy. At least it would spare us the surrealism of worries like, ""What time is it?* asserts nothing, and hence is neither true nor false; yet we all know what it means well enough to answer it";13 or like, "If we told a person to close the door, and received the reply, "Prove it!" should we not, to speak mildly, grow somewhat impatient?"16 In recommending that we ignore context in order to make "provisional divisions" of a subject and get an investigation started, Mates is recommending the wrong thing for the right reason. It is true that we cannot say everything at once and that for some problems some distinction of the sort Mates has in mind may be of service. My discontent with it is that it has come to deflect investigation — I mean from questions on which Oxford philosophy trains itself. Where your concern is one of constructing artificial languages, you may explain that you mean to be considering only the syntax (and perhaps semantics) of a language, and not its pragmatics. Or where it becomes1 important to emphasize a distinc185 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 tio n betwee n (wher e ther e ha s com e t o b e a distinctio n between ) scientifi c an d metaphysica l assertion , o r betwee n factua l repor t an d mora l rule , yo u may se t ou t a "theory " o f scientifi c o r factua l utterance . I n thes e case s yo u wil l b e restrictin g concern i n orde r t o dea l wit h certai n propertie s o f forma l systems , certai n problem s o f meaning , an d t o defeat certai n form s o f nonsense . Fla t .contradiction , metaphysica l assertio n masqueradin g a s scientifi c hypothesis , mer e whi m unde r th e postur e o f a n ethica l o r estheti c (o r psychologica l o r legal ) judgmen t — thes e perhap s nee d houndin g out . Bu t th e philosophe r wh o proceed s fro m ordinary languag e i s concerne d les s t o aveng e sensationa l crime s agains t th e intellec t tha n t o redres s its civi l wrongs ; t o steady any imbalance , th e tinies t usurpation , i n th e mind . Thi s inevitabl y require s reintroducin g idea s whic h hav e becom e tyrannica l (e.g. , existence , obligation , certainty , identity , reality , trut h ... ) int o th e specifi c context s i n whic h the y func - tio n naturally . Thi s i s no t a questio n o f cuttin g bi g idea s dow n t o size , bu t o f givin g the m th e exac t spac e i n whic h the y can move , withou t corrupting . No r doe s our wis h t o rehabilitat e rathe r tha n t o deny o r expe l suc h idea s (b y suc h sentence s as , "W e can neve r kno w fo r cer - tai n ..." ; "Th e tabl e i s no t rea l (reall y solid)" ; "T o tel l m e wha t I ough t t o d o i s alway s t o tel l m e wha t yo u wan t m e t o do " ) com e fro m a sentimenta l altruism . I t i s a questio n o f selfpreservation : fo r wh o i s i t tha t th e philosophe r punishe s whe n i t i s th e min d itsel f whic h assault s th e mind ? I wan t no w t o tur n tw o other , related , question s o n whic h Mate s find s himsel f a t issu e with th e Oxfor d philosophers . Th e firs t concern s thei r tendenc y t o introduc e statement s o f th e firs t sor t I distinguishe d no t wit h "W e d o say... " but wit h "W e can say... " an d "W e can't say..." . Th e secon d questio n concerns , a t las t directly , reason s fo r sayin g tha t w e "must " mean b y our word s wha t thos e word s ordinarily mean . Le t m e begi n b y fulfillin g my promis e t o expan d upo n my remark tha t Austin' s saying , "W e may mak e a gift voluntarily " i s "materia l mode " fo r "W e can say , Th e gift wa s mad e voluntarily"' . Th e shift fro m talkin g about , languag e t o talkin g abou t th e worl d occur s almos t imperceptibl y i n th e statement o f Austin' s whic h Mate s quote s — almos t a s thoug h h e though t i t di d no t muc h matte r which h e talke d about . Le t m e recal l th e passag e fro m Austin : ".. . tak e 'voluntarily * an d 'involuntarily' : w e may joi n th e army o r mak e a gift voluntarily , w e 18 6 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 may hiccup or make a small gesture involuntarily." He begins here by mentioning a pair of words, and goes on to tell us what we may in fact do. With what right? Why is it assumed that we find out what voluntary and involuntary actions are (and equally, of course, what inadvertent and automatic and pious, etc., actions are) by asking when we should say of an action that it is voluntary or inadvertent or pious, etc.? But what is troubling about this? If you feel that finding out what something is must entail investigation of the world rather than of language, perhaps you are imagining a situation like finding out what somebody's name and address are, or what the contents of a will or a bottle are, or whether frogs eat butterflies. But now imagine that you are in your armchair reading a book of reminiscences and come across the word "umiak". You reach for your dictionary and look it up. Now what did you do? Find out what "umiak" means, or find out what an umiak is? But how could we have discovered something about the world by hunting in the dictionary? If this seems surprising, perhaps it is because we forget that we learn language and learn the world together, that they become elaborated and distorted together, and in the same places. We may also be forgetting how elaborate a process the learning is. We tend to take what a native. speaker does when he looks up a noun in a dictionary as the characteristic process of learning language. (As, in what has become a less forgivable tendency, we take naming as the fundamental source of meaning.) But it is merely the end point in the process of learning the word. When we turned to the dictionary for "umiak" we already knew everything about the word, as it were, but its combination: we knew what a noun is and how to name an object and how to look up a word and what boats are and what an Eskimo is. We were all prepared for that umiak. What seemed like finding the world in a dictionary was really a case of bringing the world to the dictionary. We had the world with us all the time, in that armchair; but we felt the weight of it only when we felt a lack in it. Sometimes we will need to bring the dictionary to the world. That will happen when (say) we run across a small boat in Alaska of a sort we have never seen and wonder — what? What it is, or what it is called? In either case, the learning is a question of aligning language and the world.17 What you need to learn will depend on what specifically it is you want to know; and how you can find out will depend specifically on what you already command. How we answer the question, "What is X?" will depend, therefore, on the specific case of ignorance and of knowledge. 187 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 I t sometime s happen s tha t w e kno w everythin g ther e i s t o kno w abou t a situatio n — wha t al l o f th e word s i n questio n mean , wha t al l o f th e relevan t fact s are ; an d everythin g i s i n fron t o f our eyes . An d ye t w e fee l w e don' t kno w something , don' t understan d something . I n thi s situation , th e questio n "Wha t i s X? " i s very puzzling , i n exactly th e way philosoph y i s very puzzling . W e fee l w e wan t t o as k th e question , an d ye t w e fee l w e already hav e th e answer . (On e migh t say w e hav e al l th e elements o f an answer. ) Socrate s say s tha t i n such a situatio n w e hee d t o remin d ourselve s o f something . S o doe s th e philosophe r wh o proceed s fro m ordinary language : w e nee d t o remin d ourselve s o f what we should say when} 6 But wha t i s th e poin t o f remindin g ourselve s o f that ? Whe n th e philosophe r asks , "Wha t shoul d w e say here?" , wha t i s meant is , "Wha t woul d b e th e norma l thin g t o say here?" , o r perhaps , "Wha t i s th e mos t natura l thin g w e coul d say here? " An d th e poin t o f th e questio n i s this : answerin g i t i s sometime s th e onl y way t o tel l — tel l other s an d tel l fo r ourselve s — wha t th e situatio n is. Sometime s th e onl y way t o tell . But when ? Th e natur e o f th e Oxfor d philosopher' s question , an d th e natur e o f hi s conceptio n o f philosophy , can b e brough t out i f w e tur n th e questio n upo n itself, an d thu s remin d ourselve s o f whe n i t i s w e nee d t o remin d ourselve s o f what w e shoul d say when . Ou r questio n the n becomes : Whe n shoul d w e as k ourselve s whe n w e shoul d (an d shoul d not ) say "Th e x i s F " i n orde r t o fin d ou t wha t an F(x ) is ? (Fo r "Th e x i s F " rea d "Th e actio n i s voluntary (o r pious)" , o r "Th e Statement i s vagu e (o r false)" , o r "Th e questio n i s misleading". ) Th e answe r suggeste d is , whe n yo u hav e to . Whe n yo u hav e mor e fact s tha n you kno w ho w t o mak e of, o r whe n you d o no t kno w wha t ne w fact s woul d show . When , tha t is , yo u nee d a clear vie w o f wha t you already know . Whe n yo u nee d t o d o philosophy. 1 9 Euthy - phr o doe s no t nee d t o learn any ne w facts , ye t h e need s t o learn some - thing : yo u can say eithe r tha t i n th e Euthypro Socrate s wa s findin g ou t wha t "piety " mean s o r findin g ou t wha t piety is . Whe n th e philosophe r wh o proceed s fro m ordinary languag e tell s us , "Yo u can' t say such-and-such" , wha t h e mean s i s tha t yo u canno t say tha t here an d communicat e this situatio n t o others , o r understan d i t fo r yourself. 2 0 Thi s i s sometime s wha t h e mean s b y callin g certai n ex - pression s "misuses " o f language , an d als o make s dea r th e consequence s o f suc h expressions : the y break ou r understanding . Th e normativenes s 18 8 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 which Mates felt, and which is certainly present, does not lie in the ordinary language philosopher's assertions about ordinary use; what is normative is exactly ordinary use itself. The way philosophers have practiced with the word "normative" in recent years seems to me lamentable. But it is too late to avoid the word, so even though we cannot now embark on a diagnosis of the ills which caused its current use, or those which it has produced, it may be worth forewarning ourselves against the coufusions most likely to distract us. .The main confusions about the problem of "normativeness" I want to mention here are these: (l) the idea that descriptive utterances are opposed to normative utterances; and (2) that prescriptive utterances are (typical) instances of normative utterances. We have touched upon these ideas in talking about rule-statement complementarity; here we touch them at a different point. In saying here that it is a confusion to speak of some general opposition between descriptive and normative utterances, I am not thinking primarily of the plain fact that rules have counterpart (descriptive) statements, but rather of the significance of that fact, viz., that what such statements describe are actions (and not, e.g., the movements of bodies, animate or inanimate). The most characteristic fact about actions is that they can — in various specific ways — go wrong, that • they can be performed incorrectly. This is not, in any restricted sense, a moral assertion, though it points the moral of intelligent activity. And it is as true of describing as it is of calculating or of promising or plotting or warning or asserting or defining... These are actions which we perform, and our successful performance of them depends upon our adopting and following the ways in which the action in question is done, upon what is normative for it. Descriptive statements, then, are not opposed to ones which are normative, but in fact presuppose them: we could not do the thing we call describing if language did not provide (we had not been taught) ways normative for describing. The other point I wish to emphasize is this: if a normative utterance is one used to create or institute rules or standards, then prescriptive utterances are not examples of normative utterances. Establishing a norm is not telling us how we ought to perform an action, but telling us how the action is done, or how it is to be done.21 Contrariwise, telling us what we ought to do is not instituting a norm to cover the case, but rather presupposes the existence of such a norm, i.e., presupposes that there is something to do which it would be correct to do here. Telling us what we ought to do may involve appeal to a pre-existent rule or stan189 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 dard , bu t i t canno t constitut e th e establishmen t o f tha t rul e o r standard . W e may expec t th e retor t her e tha t i t i s jus t th e appeal whic h i s th e sensitiv e normativ e spot , fo r wha t w e ar e reall y doin g whe n w e appea l t o a rul e o r standar d i s tellin g somebody tha t the y ough t t o adher e t o it . Perhap s thi s wil l b e followe d b y th e query , An d suppos e the y don' t accep t th e rul e o r standar d t o whic h yo u appeal , wha t then ? Th e retor t i s simpl y false . An d t o th e query on e may repl y tha t thi s wil l no t b e th e firs t tim e w e hav e bee n tactless ; no r can we , t o avoi d oversteppin g th e bound s o f relationship , follo w every statemen t by , ".. . i f yo u accep t th e fact s an d th e logi c I do" , no r every evaluatio n b y ".. . i f yo u accep t th e standard s I do" . Such caution s wil l finall y sugges t appendin g t o everythin g w e say ".. . i f yo u mean b y your word s wha t I mean b y mine" . Her e th e pantomim e o f cautio n concludes . I t i s tru e tha t w e sometime s appea l t o standard s whic h our interlocuto r doe s no t accept ; bu t thi s doe s no t i n th e leas t sho w tha t wha t w e ar e ther e reall y doin g i s attemptin g t o institut e a standar d (o f ou r own) . No r doe s i t i n th e leas t sho w tha t w e ar e (merely ) expressin g our own opinio n o r feelin g o n th e matter . W e o f cours e may expres s ou r privat e opinio n o r feelin g — w e normall y d o s o wher e i t i s no t dea r wha t (o r tha t any ) rul e o r standar d fit s th e cas e at han d an d wher e w e ar e therefor e no t willin g o r abl e t o appea l t o any . Th e practic e o f appealin g t o a nor m can b e abused , a s can any othe r o f our practices . Sometime s peopl e appea l t o a rul e whe n w e deserve d mor e intimat e attentio n fro m them . Jus t a s sometime s peopl e tel l u s wha t w e ough t t o d o whe n al l the y mean i s tha t the y wan t u s to . Bu t thi s i s a s muc h an abus e wher e the contex t i s mora l a s i t i s wher e th e contex t i s musica l ("Yo u ough t t o accent th e appogiatura") , o r scientifi c ("Yo u ough t t o us e a contro l grou p here") , o r athleti c ("Yo u ough t t o sav e your win d o n th e firs t tw o laps") . Privat e persuasio n (o r persona l appeal ) i s no t th e paradig m o f ethica l utterance , bu t represent s th e breakdow n (o r th e transcending ) o f mora l interaction . W e can , to o obviously , becom e morall y inaccessibl e t o on e another ; bu t t o tel l u s tha t thes e ar e th e moment s whic h reall y constitut e th e mora l life wil l onl y ad d confusio n t o pain . I f not , then , by sayin g wha t action s ought t o b e performed , ho w do w e establis h (o r justify o r modify o r drop ) rule s o r standards ? Wha t genera l answe r can ther e b e t o thi s genera l questio n othe r than , I n variou s ways , dependin g o n th e context ? Philosopher s wh o hav e imagine d tha t th e questio n ha s on e answe r fo r al l case s mus t b e tryin g t o assimilat e th e member s o f Footbal l Commissions , o f Chil d Developmen t Researc h 19 0 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 Teams, of University Committees on Entrance Requirements, of Bar Association Committees to Alter Legal Procedures, of Departments of Agriculture, of Bureaus of Standards, and of Essene Sects, all to one "sort" of person, doing one "sort" of thing, viz., establishing (or changing) rules and standards. Whereas the fact is that there are, in each case, different ways normative for accomplishing the particular normative tasks in question. It has in recent years been emphasized past acknowledgment that even justifications require justification. What now needs emphasizing is that (successfully) justifying a statement or an action is not (cannot be) justifying its justification.22 The assumption that the appeal to a rule or standard is only justified where that rule or standard is simultaneously established or justified can only serve to make such appeal seem hypocritical (or anyway shaky) and the attempts at such establishment or justification seem tyrannical (or anyway arbitrary). It would be important to understand why we have been able to overlook the complementarity of rule and statement and to be content always to sort rules with imperatives. Part of the reason for this comes from a philosophically inadequate (not to say disastrous) conception of action; but this inadequacy itself will demand an elaborate accounting. There is another sort of reason for our assumption that what is binding upon us must be an imperative; one which has to do with our familiar sense of alienation from established systems of morality, perhaps accompanied by a sense of distance from God. Kant tells us that a perfectly rational being does in fact (necessarily) conform to "the surpreme principle of morality", but that we imperfectly rational creatures are necessitated by it, so that for us it is (always appears as) an imperative. But if I understand the difference Kant sees here, it is one within the conduct of rational animals. So far as Kant is talking about (the logic of) action, his Categorical Imperative can be put as a Categorial Declarative (description-rule), i.e., a description of what it is to, act morally: When we (you) act morally, we act in a way we would regard as justified universally, justified no matter who had done it. (This categorial formulation does not tell us how to determine what was done; neither does Kant's categorical formulation, although, by speaking of "the" maxim of an action, it pretends to, or anyway makes it seem less problematical than it is.) Perhaps it is by now a little dearer why we are tempted to retort, "But suppose I don't want to be moral?"; and also why it would be irrelevant here. The Categorial Declarative does not tell you what you ought to do // you want to be moral (and hence is untouched by the feeling that no imperative can really be categorical, can bind us no 191 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 matte r what) ; i t tell s you (par t of) wha t yo u i n fac t d o whe n yo u are moral . I t canno t — nothin g a philosophe r say s can — insur e tha t you wil l no t ac t immorally ; bu t i t i s entirel y unaffecte d b y what yo u d o o r d o no t want . I am no t sayin g tha t rule s d o no t sometime s sort with imperatives , bu t onl y denyin g tha t the y alway s do . I n th e Britannic a articl e (11t h edition ) o n chess , onl y on e paragrap h o f th e twenty o r s o whic h describ e th e gam e is " heade d "Rules" , an d onl y her e ar e w e tol d wha t w e must do . Thi s paragrap h deal s with suc h matter s a s th e conventio n o f sayin g "j'adoube " whe n yo u touc h a piec e t o straighte n it . I s th e differenc e betwee n matter s o f thi s kin d an d th e matte r o f ho w piece s move , a differenc e betwee n penaltie s (whic h ar e impose d fo r misplay ) an d move s (whic h ar e accepte d i n orde r t o play a t all ) — s o tha t w e woul d cheer - full y say tha t w e can play (ar e playing ) ches s withou t th e "j'adoube " convention , bu t les s cheerfully tha t w e can play withou t followin g th e rul e tha t "th e Quee n move s i n any direction , squar e o r diagonal , whethe r forwar d o r backward" ? Thi s woul d sugges t tha t w e may thin k o f th e differenc e betwee n rul e an d imperativ e a s on e betwee n thos e action s (o r "parts " o f actions ) whic h ar e eas y (natural , normal ) fo r us , an d thos e w e hav e t o b e encourage d t o do . (Wha t I d o a s a rul e yo u may hav e t o b e mad e o r directe d t o do. ) We " ar e likel y t o forge t t o say "j'adoube" , s o w e hav e t o b e made (t o remember ) t o d o it ; bu t w e d o no t hav e t o b e made t o mov e th e Quee n i n straight , unobstructe d paths. 2 3 Thi s furthe r suggest s tha t wha t i s though t o f a s "alienation " i s some - thin g whic h occur s within mora l systems ; sinc e thes e ar e profoundl y haphazar d accumulations , i t i s n o surpris e tha t w e fee l part o f som e region s o f th e system an d fee l apart fro m othe r regions. 2 * S o th e subjec t o f responsibility , o f obligatio n an d commitment , open s int o th e se t o f question s havin g t o d o with difference s betwee n doin g a thin g wrongl y o r badl y (strangely , ineptly , inexactly , partially... ) an d no t doin g th e thin g a t all . Thes e difference s tak e u s int o a furthe r regio n o f th e concep t o f a n action : w e hav e note d tha t ther e ar e many (specific ) way s i n whic h an actio n can g o wron g (a t leas t a s many a s th e myria d excuse s w e ar e entitle d t o proffe r whe n wha t w e hav e don e ha s resulte d i n som e unhappiness) ; bu t i t woul d b e incorrec t t o suppos e tha t w e ar e obligated t o se e t o i t (t o tak e precaution s t o insure) , whenever w e undertak e t o d o anything , tha t non e o f thes e way s wil l com e t o pass . Ou r obligatio n i s t o avoi d doin g somethin g at a tim e an d plac e o r i n a way whic h i s likely t o resul t i n som e misfortune , o r t o avoi d bein g careles s wher e i t i s easy t o be , o r t o b e especially carefu l wher e th e 19 2 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 action is dangerous or delicate, or avoid the temptation to skip a necessary step when it seems in the moment to make little difference. If for all excuses there were relevant. obligations, then there would be no excuses and action would become intolerable. Any particular excuse may be countered with a specific obligation; not even the best excuse will always get you off the hook (That is no excuse; you should have known that was likely to result in an accident, you ought to have paid particular heed here, etc.). Without pretending to give an account of (this part of) obligation, what I think the foregoing considerations indicate is this: a statement of what we must do (or say) has point only in the context (against the background) of knowledge that we are in fact doing (or saying) a thing, but doing (saying) it — or running a definite risk of doing or saying it — badly, inappropriately, thoughtlessly, tactlessly, self-defeatingly, etc.; or against the background of knowledge that we are in a certain position or occupy a certain office or station, and are behaving or conducting ourselves inappropriately, thoughtlessly, self-defeatingly .. . The. same is true of statements about what we may do, as well as those containing other "modal auxiliaries" — e.g., about what we should do. or what we are or have to do, or are supposed to do, and about one sense of what we' can do; these are all intelligible only against the background of what we are doing or are in a position (one sense of "able") to do. These "link verbs" share the linguistic peculiarity that while they are verb-like forms they cannot stand as the main verb of a sentence. This itself would suggest that their use is not one of prescribing some new action to us, but of setting an action which is antecedently relevant to what we are doing or to what we are — setting it relevantly into the larger context of what we are doing or of what we are.23 "You must (are supposed, obliged, required to) move the Queen in straight paths..." or, "You may (can, are allowed or permitted to) move the Queen in straight paths ... " say (assert) no more than "You (do, in fact, always) move the Queen in straight paths..."; which of them you say on a given occasion depends not on any special motive or design of yours, nor upon any special mode of argument. There is no question of going from "is" to "must", but only of appreciating which of them should be said when, i.e., of appreciating the position or circumstances of the person to whom you are speaking. Whatever makes one of the statements true makes them all true, though not all appropriate. To tell me what I must do is not the same as to tell me what I ought to do. I must move the Queen in straight paths (in case I am absent193 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 minde d an d continu e movin g i t lik e th e Damsel , cf. n . 23) . Wha t woul d i t mean t o tel l m e tha t I ought t o mov e th e Quee n i n straigh t paths ? "Ought" , unlik e "must" , implie s tha t ther e i s an alternative ; "ought " implie s tha t yo u can , i f yo u choose , d o otherwise . Thi s doe s not mean merel y tha t ther e i s somethin g els e t o d o whic h i s i n your power (" I can mov e th e Quee n lik e th e Knight ; jus t watch!" ) bu t tha t ther e i s one 1 withi n you r rights. Bu t i f I say trul y an d appropriately , "Yo u must... " the n i n a perfectl y goo d sens e nothin g yo u the n d o can prov e m e wrong . Yo u CA N push the little object called the Queen i n many ways , a s yo u can /// / i t o r throw i t acros s th e room ; no t al l o f thes e wil l b e moving the Queen. Yo u CA N ask , "Wa s your actio n voluntary " an d say t o yourself, "Al l I mean t o as k i s whethe r h e ha d a sensatio n o f effort jus t befor e h e moved" , bu t tha t wil l no t b e findin g out whethe r th e actio n wa s voluntary . Again , i f I hav e borrowe d mone y the n I must (unde r norma l circumstances ) pay i t bac k (eve n thoug h i t i s rathe r painful). 2 8 I t make s sens e t o tel l m e I ought t o pay i t bac k only i f ther e i s a specifi c reaso n t o suppose , say , tha t th e perso n fro m who m I go t th e mone y meant to give i t t o m e rathe r tha n merel y lend i t (neverthe - les s h e need s i t badly , wors e tha n I know) , o r i f ther e i s a reaso n t o pay i t bac k tomorro w instea d o f nex t week , whe n th e deb t fall s du e (I'l l sav e interest ; I'l l onl y spen d i t an d hav e t o mak e anothe r loan) . Th e differenc e her e resemble s tha t betwee n doin g a thin g an d doin g th e thin g wel l (thoughtfully , tactfully , sensibly , graciously...) . Thi s differenc e may b e mad e deare r b y considerin g on e way principle s diffe r fro m rules . Rule s tel l yo u wha t t o d o whe n yo u d o th e thin g a t all ; principle s tel l yo u ho w t o d o th e thin g well , wit h skil l o r under - standing . I n competitiv e games , actin g wel l amount s t o doin g th e sort o f thin g tha t wil l win , s o th e principle s o f game s recommen d strategy . "N o rais e shoul d [nb ] b e give n t o partner' s sui t withou t at leas t Q-x-x , J-10-x , K-x-x , A-x-x , o r any four trumps..." . Bu t yo u coul d fai l t o adop t thi s an d stil l play bridge , eve n play well . I t i s a principl e strateg y i n Culbertson' s system; 2 7 bu t anothe r expert may hav e a different understandin g o f th e gam e an d develo p principle s o f strategy whic h ar e equall y successful . Principle s g o wit h understanding . (Havin g an under - standin g o f a gam e i s no t knowin g th e rules ; yo u migh t fin d a boo k calle d Principle s o f Economic s o r Psychology , bu t non e calle d Rule s o f Economics , etc. ) Understanding a principl e involve s knowin g ho w an d wher e t o appl y it . But som e move s see m s o immediatel y t o b e calle d fo r b y th e principle s o f strategy , tha t thei r formulation s com e t o b e though t o f a s rules : shoul d w e say , "Th e thir d han d play s high... " 19 4 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 or "The third hand should play high..."? You may, strictly speaking, be playing bridge if you flout this, but you won't be doing the sort of thing which will win (and therefore not really playing? When is not doing a thing well not really doing the thing?). All players employ maxims (which may be thought of as formulating strategies as though they were moves) in order to facilitate their play; like everything habitual or summary, maxims have their advantages and their dangers. Both the rules which constitute playing the game, and the "rules" or maxims which contribute to playing the game well have their analogues in ordinary moral conduct. I think it is sometimes felt that drawing an analogy between moral conduct and games makes moral conduct seem misleadingly simple (or trivial?), because there are no rules in moral conduct corresponding to the rules about how the Queen moves in chess.28 But this misses the point of the analogy, which is that moves and actions have to be done correctly; not just any movement you make will be a move, or a promise, a payment, a request. This does not mean that promising is (just) following rules. Yet if someone is tempted not to fulfill a promise, you may say, "Promises are kept", or "We keep our promises (that is the sort of thing a promise is)", thus employing a rule-description — what I have called a categorial declarative. You may say, "You must keep this promise" (you are underestimating its importance; last time you forgot). This is not the same as, "You ought to keep this promise", which is only sensible where you have a reason for breaking it strong enough to allow you to do so without blame (there is a real alternative), but where you are being enjoined to make a special effort or sacrifice. (This is partly why "You ought to keep promises" is so queer. It suggests that we not only always want badly to get out of fulfilling promises, but that we always have some good (anyway, prima. facie) reason for not keeping them (perhaps our own severe discomfort) and that therefore we are acting well when we do fulfill. But we aren't, normally; neither well nor ill.) "Ought" is like "must" in requiring a background of action or position into which the action in question is set; and, like "must", it does not form a command, a pure imperative. All of which shows the hopelessness of speaking, in a general way, about the "normativeness" of expressions. The Britannica "rules" tell us what we must do in playing chess, not what we ought to do // we want to play. You (must) mean (imply), in speaking English, that something about an action is fishy when you say, "The action is voluntary"; you (must) mean, when you ask a person, "Ought you to do that?" that there is some 195 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 specific way i n whic h wha t h e is doin g migh t b e don e mor e tactfully , carefully , etc . .. . Ar e thes e imperatives ? Ar e the y categorica l o r hypo - thetical ? Hav e yo u i n n o way contradicte d yoursel f i f yo u flout them ? (cf. n . 26. ) Tha t "moda l imperatives " ("must" , "suppose d to" , "ar e to" , "hav e to " ... ) requir e th e recognitio n o f a backgroun d actio n o r positio n int o whic h th e relevant actio n i s placed , indicate s a portentou s differenc e betwee n thes e form s o f expressio n an d pur e imperatives , commands . Whethe r I can comman d depend s onl y upo n whethe r I hav e powe r o r authority , an d th e onl y characteristic s I mus t recogniz e i n th e objec t o f th e comman d ar e thos e whic h tel l m e tha t th e objec t i s subjec t t o my powe r o r authority . Employin g a moda l "imperative" , however , require s tha t I recogniz e th e objec t a s a person (someon e doin g somethin g o r i n a certai n position ) t o whos e reasonablenes s (reason ) I appea l i n usin g th e secon d person . (Compare , "Open , Sesame! " wit h "Yo u mus t open , Sesame." ) Thi s i s on e reaso n tha t commands , pur e imperatives , ar e no t paradigm s o f mora l utterance , bu t represen t a n alternativ e t o suc h utterance . Withou t pretendin g tha t my argumen t fo r i t ha s bee n nearl y ful l o r clear enough , le t me , by way o f summary, - flatl y stat e wha t i t i s I hav e trie d t o argu e about th e relatio n betwee n wha t yo u say an d wha t yo u (must ) mean , i.e. , betwee n wha t yo u (explicitly ) say an d wha t sayin g i t implie s o r suggests : I f "wha t A (a n utterance ) means " i s t o b e under - stoo d i n term s o f (o r even a s directl y relate d to ) "wha t i s (mus t be ) mean t i n (by ) sayin g A", 2 0 the n th e meanin g o f A wil l no t b e give n b y it s analyti c o r definitiona l equivalents , o r b y it s deductiv e implications . Intensio n i s no t a substitut e fo r intention . Althoug h w e woul d no t cal l th e statemen t "Whe n w e say w e kno w somethin g w e impl y (mean ) tha t w e hav e confidence , tha t w e ar e i n a positio n t o say w e know... " analytic , ye t i f th e statement i s tru e i t i s necessarily tru e i n jus t thi s sense : i f i t i s true , the n whe n yo u as k wha t th e statement suppose s yo u t o as k yo u (must ) mean wha t th e statemen t say s yo u (must ) mean . Necessary an d no t analytic : i t wa s — apar t fro m th e parody o f Kan t — t o summarize , an d partl y explain , thi s peculiarity tha t I calle d suc h statement s categoria l declaratives : declarative , becaus e somethin g i s (authoritatively ) mad e known ; categorial , becaus e i n tellin g u s wha t w e (must ) mean b y assertin g tha t (o r questionin g whether) x i s F , the y tel l u s wha t i t i s fo r an x t o b e F (a n actio n t o b e moral , a statemen t claimin g knowledg e t o b e a statemen t expressin g knowledge , a move - 19 6 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 ment to be a move).30 Shall we say that such statements formulate the rules or the principles of grammar — the moves or the strategies of talking? (And is this, perhaps, to be thought of as a difference between grammar and rhetoric? But becoming clearer about this will require us to see more clearly the difference between not doing a thing well (here, saying something) and not doing the thing; and between doing a thing badly and not doing the thing.) The significance of categorial declaratives lies in their teaching or reminding us that the "pragmatic implications" of our utterances are (or, if we are feeling perverse, or tempted to speak carelessly, or chafing under an effort of honesty, let us say must be) meant; that they are an essential part of what we mean when we say something, of what it is to mean something. And what we mean (intend) to say, like what we mean (intend) to do, is something we are responsible for. Even with this slight rehabilitation of the notion of normativeness, ' we can begin to see the special sense in which the philosopher who proceeds from ordinary language is "establishing a norm" in employing his second type of statement. He is certainly not instituting norms, nor is he ascertaining norms (see note 21); but he may be thought of as confirming or proving the existence of norms when he reports or describes how we (how to) talk, i.e., when he says (in statements of the second type) what is normative for utterances instanced by statements of the first type. Confirming and proving are other regions of establishing. I have suggested that there are ways normative for instituting and for ascertaining norms; and so are there for confirming or proving or reporting them, i.e., for employing locutions like, "We can say...", or "When we say... we imply ". The swift use made of them by the philosopher serves to remind mature speakers of a language of something they know; but they would erroneously be employed in trying to report a special usage of one's own, and (not unrelated to this) could not be used to change the meaning of an expression. Since saying something is never merely saying something, but is saying something with a certain tune and at a proper cue and while executing the appropriate business, the sounded utterance is only a salience of what is going on when we talk (or the unsounded when we think); so a statement of "what we say" will give us only a feature of what we need to remember. But a native speaker will normally know the rest: learning it was part of learning the language. Let me warn against two tempting ways to avoid the significance of this: (1) It is perfectly true that English might have developed differ197 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 entl y tha n i t ha s an d therefor e hav e impose d differen t categorie s o n th e worl d tha n i t does ; an d i f so , i t woul d hav e enable d u s t o assert , describe , question , define , promise , appeal , etc. , i n way s othe r tha n w e do . Bu t usin g Englis h no w — t o convers e wit h other s i n th e language , o r t o understan d th e world , o r t o thin k b y ourselve s — mean s knowin g whic h form s i n wha t context s ar e normativ e fo r performin g th e activi - tie s w e perform b y usin g th e language . (2 ) I t i s n o escap e t o say : "Stil l I can say wha t I like ; I needn' t alway s us e norma l form s i n sayin g wha t I say ; I can speak i n extraordinary ways , an d you wil l perfectly wel l understan d me. " Wha t thi s call s attentio n t o i s th e fac t tha t languag e provide s u s with way s fo r (contain s form s whic h ar e normativ e for ) speakin g i n specia l ways , e.g. , fo r changin g th e meanin g o f a word , o r fo r speaking , on particular occasions, loosel y o r personally , o r paradoxi - cally , cryptically , metaphorically.. . D o yo u wis h t o clai m tha t yo u can speak strangel y ye t intelligibl y — an d thi s o f cours e mean s intelligibl y t o yoursel f a s wel l — i n way s no t provide d i n th e languag e fo r speakin g strangely ? I t may b e fel t tha t I hav e no t ye t touche d on e o f Mates' fundamenta l criticisms . Suppos e yo u gran t al l tha t has-bee n sai d about an ordinary us e bein g normativ e fo r wha t anyon e says . Wil l yo u stil l wis h t o ask : "Doe s i t follo w tha t th e ordinary use s whic h ar e normativ e fo r wha t professor s say ar e th e sam e a s th e ordinary use s whic h ar e normativ e fo r wha t butcher s an d baker s say? " O r perhaps : "I s an ordinary us e fo r a professo r an ordinary ordinary use? " I s tha t a sensibl e question ? T o determin e whethe r i t is , w e mus t appreciat e wha t i t i s t o tal k together . Th e philosopher , understandably , ofte n take s th e isolate d man ben t silentl y ove r a boo k a s hi s mode l fo r wha t usin g languag e is . Bu t th e primary fac t o f natura l languag e i s tha t i t i s somethin g spoken , spoke n together . Talkin g togethe r i s actin g together , no t makin g motion s an d noise s a t on e another , no r transferrin g unspeakabl e message s o r essence s fro m th e insid e o f on e dose d chambe r t o th e insid e o f another . Th e difficultie s o f talkin g togethe r are , rather , real ones : th e activitie s w e engag e i n b y talkin g ar e intricat e an d intricatel y relate d t o on e an - other . I suppos e i t wil l b e grante d tha t th e professo r an d th e bake r can tal k together . Conside r th e mos t obviou s complexitie s o f cooperativ e activity i n whic h the y engage : ther e i s commentin g ("Nic e day") ; commending , persuading , recommending , enumerating , comparin g ("Th e pumpernickl e i s good , bu t th e whol e whea t an d th e ry e eve n better") ; 19 8 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 grading, choosing, pointing ("I'll have the darker loaf there"); counting, making change, thanking; warning ("Careful of the step"); promising ("Be back next week") ...; all this in addition to the whole nest or combination of actions which comprise the machinery of talking: asserting, referring, conjoining, denying, .. . Now it may be dearer why I wish to say: if the professors and the baker did not understand each other, the professors would not understand one another either. You may still want to ask: "Does this mean that the professor and baker use particular words like 'voluntary1 and 'involuntary', or 'inadvertently* and 'automatically1 the same way? The baker may never have used these words at all." But the question has now become, since it is about specific expressions, straightforwardly empirical. Here Mates' "two methods" (pp 165 ff.) at last become relevant. But I am at the moment less interested in determining what empirical methods would be appropriate to investigate the matter than I am in posing the following questions: What should we say if it turned out, as it certainly might, that they in fact do use the words differently? Should we, for example, say that therefore we never have a right to say that people use words in the same way without undertaking an empirical investigation; or perhaps say that therefore they speak different languages? What should make us say that they do not speak the same language? Do we really know what it would be like to embark upon an empirical investigation of the general question whether we (ordinarily, ever) use language the way other people do? . There is too much here to try to unravel. But here are some of the threads: The words "inadvertently" and "automatically", however recondite, are ordinary; there are ordinary contexts (nontechnical, nonpolitical, nonphilosophical contexts) which are normative for their use. It may be that half the speakers of English do not know (or cannot say, which is not the same) what these contexts are. Some native speakers may even use them interchangeably. Suppose the baker is able to convince us that he does. Should we then say: "So the professor has no right to say how uwe use" "inadvertently", or to say that "when we use the one word we say something different from what we say when we use the other"?" Before accepting that conclusion, I should hope that the following consideration would be taken seriously: When "inadvertently" and "automatically" seem to be used indifferently in recounting what someone did, this may not at all show that they are being used synonymously, but only that what each of them says is seperatery true of the person's action. The decanter is broken and you did it. You may 199 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 say (an d i t may b e importan t t o conside r tha t yo u ar e already embarrasse d an d flustered ) either : " I di d i t inadvertently " o r " I di d i t automatically" . Ar e yo u sayin g th e sam e thing ? Well , yo u automaticall y grabbed the cigarette whic h ha d falle n o n th e table , an d inadvertentl y knocked over the decanter. Namin g action s i s a sensitiv e occupation. 3 1 I t i s easy t o overloo k th e distinctio n becaus e th e tw o adverb s ofte n g o togethe r i n describin g action s i n whic h a sudde n movemen t result s i n som e mishap . Suppos e th e bake r doe s no t accep t thi s explanation , bu t replies : " I us e •automatically' an d 'inadvertently' i n exactly th e sam e way . I coul d jus t a s wel l hav e said : ' I grabbe d th e cigarett e inadvertentl y an d knocke d ove r th e decante r automatically"' . Don' t w e fee l th e temptatio n t o reply : "Yo u may say this , bu t yo u can' t say i t an d describ e th e sam e situation ; yo u can' t mean wha t yo u woul d mean i f yo u sai d th e other" . Bu t suppos e th e bake r insist s h e can ? Wil l w e the n b e prepare d t o say : "Wel l yo u can' t say th e on e an d mea n wha t I mean b y th e other" ? Grea t car e woul d b e neede d i n claimin g this , fo r i t may loo k lik e I am saying , " I kno w wha t I mean an d I say the y ar e different" . Bu t why i s th e bake r no t entitle d t o thi s argument ? Wha t I mus t no t say is : " I kno w what word s mean i n my language" . Her e th e argumen t woul d hav e pushe d m e t o madnness . I t may tur n ou t (dependin g upo n jus t wha t th e dialogu e ha s bee n an d wher e i t wa s stopped ) tha t w e shoul d say t o th e baker : "I f yo u cooke d th e way yo u talk , yo u woul d forg o specia l implement s fo r differen t jobs , an d peel , core , scrape , slice , carve , cho p an d saw , al l wit h on e knife . Th e distinctio n i s there , i n th e languag e (a s implement s ar e ther e t o b e had) , an d yo u jus t impoveris h wha t yo u can say b y neglectin g it . An d ther e i s somethin g yo u aren' t noticin g about th e world". 3 2 Bu t t o a philosophe r wh o refuse s t o acknowledg e th e distinctio n w e shoul d say somethin g more : no t merel y tha t h e impoverishe s wha t h e can say abou t actions , bu t tha t h e i s a poo r theoris t o f wha t i t i s t o d o something . Th e philosophe r wh o ask s abou t everythin g w e do , "Volun - tary o r not? " ha s a poo r vie w o f actio n (a s th e philosophe r wh o ask s o f everythin g w e say , "Tru e o r false? " o r "Analyti c o r synthetic? " ha s a poo r vie w o f communication) , i n somethin g lik e th e way a ma n wh o ask s th e coo k about every piec e o f food , "Wa s i t cut o r not? " ha s a poo r vie w o f preparin g food . Th e cook wit h onl y on e knife i s i n muc h bette r conditio n tha n th e philosophe r wit h onl y "Voluntar y o r involun - tary? " t o us e i n dividin g actions , o r "Tru e o r false? " t o us e iri hackin g ou t meaningfu l statements . Th e coo k can ge t o n wit h th e preparatio n o f th e mea l eve n i f h e mus t improvis e a metho d her e an d there , an d 20 0 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 makes more of a mess than he would with more appropriate implements. But the philosopher can scarcely begin to do his work; there is no job the philosopher has to get on with; nothing ulterior he must do with actions (e.g., explain or predict them), or with statements (e.g., verify them). What he wants to know is what they are, what it is to do something and to say something. To the extent that he improvises a way of getting past the description and division of an action or a statement, or leaves a mess in his account — to that extent he laves his own job undone. If the philosopher is trying to get dear about what preparing a meal is and asks the cook, "Do you cut the apple or not?", the cook may say, "Watch me!", and then core and peel it. "Watch me!" is what we should reply to the philosopher who asks of our normal, ordinary actions, "Voluntary or not?" and who asks of our ethical and esthetic judgments, "True or false?" Few speakers of a language utilize the full range of perception which the language provides, just as they do without so much of the rest of their cultural heritage. Not even the philosopher will come to possess all of his past, but to neglect it deliberately is foolhardy. The consequence of such neglect is that our philosophical memory and perception become fixated upon a few accidents of intellectual history. I have suggested that the question of "[verifying} an assertion that a given person uses a word in a given way or with a given sense" (Mates, ibid, my emphasis) is not the same as verifying assertions that "We say... " or that "When we say... we imply ". This means that I do not take the "two basic approaches" which Mates offers in the latter part of his paper to -be directed to the same question as the one represented in the title he gives to his paper (at least on my. interpretation of that question). The questions are designed to elicit different types of information; they are relevant (have point) at different junctures of investigation. Sometimes a question is settled by asking others (or ourselves) what we say here, or whether we ever say such-and-such; on the basis of these data we can make statements like, "'Voluntary1 is used of an action only where there is something (real or imagined) fishy about it". I take this to be a "statement about ordinary language" (and equally, about voluntary action). But surely it is not, under ordinary circumstances, an assertion about how a word is used by me (or "some given person"); it is a statement about how the word is used in English. Questions about how a given person is using some word can sensibly 201 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 aris e onl y wher e ther e i s som e specifi c reaso n t o suppos e tha t h e i s usin g th e wor d i n an unusua l way . Thi s poin t can b e pu t th e othe r way around : th e statemen t " I (o r som e give n person ) us e (used ) th e wor d X i n such-and-suc h a way " implie s (dependin g o n th e situation ) tha t yo u inten d (intended ) t o b e usin g i t i n a specia l way , o r tha t someon e els e i s unthinkingl y misusin g it , o r usin g i t misleadingly , an d s o on . Thi s i s anothe r instanc e o f th e principl e tha t action s whic h ar e norma l wil l no t tolerat e any special description . I n a particular cas e you may realiz e tha t word s ar e no t t o b e take n normally , tha t som e wan t o r fear o r specia l intentio n o f th e speake r i s causin g a n abberratio n i n th e drift o f hi s words . A littl e gir l wh o says t o he r brother , "Yo u can hav e hal f my candy " may mean , "Don' t tak e any!" ; th e husban d wh o scream s i n fury , "Stil l n o buttons! " may reall y b e saying , "I f I wer e honest , I' d d o wha t Gaugi n did" . A knav e o r a criti c o r a n heires s may say , " X i s good " an d mean " I wan t o r expec t o r comman d yo u t o lik e (o r approv e of) X" ; an d we , eve n withou t a specia l burde n o f malice , o r o f taste , o r o f money , may sometime s fin d ourselve s imitatin g them . Mate s interpret s Ryle' s assertio n tha t th e ordinary us e o f "voluntary " applie s t o action s whic h ar e disapprove d t o mean tha t "th e ordinary ma n applie s th e wor d onl y t o action s o f whic h h e disapproves " (p . 169) ; thi s apparentl y involve s a referenc e t o tha t man' s persona l "aims , feelings , beliefs , an d hopes" ; an d these , i n turn , ar e supposedl y par t merel y o f th e pragmatic s (no t th e semantics ) o f a word . I t i s therefor e a mistake , Mate s concludes , t o clai m tha t th e philosophe r i s usin g th e wor d i n a "stretched , extraordinary sense" (ibid) , my emphasis ) merely o n th e groun d tha t h e may no t happe n t o fee l disapprovin g abou t an actio n h e call s voluntary . Th e mistake , however , i s t o suppos e tha t th e ordinary us e o f a wor d i s a functio n o f th e interna l stat e o f th e speaker . (I t i s sometime s t o emphasiz e tha t your remark s about "use " ar e no t remark s abou t suc h state s tha t yo u wan t t o say yo u ar e talkin g about th e logic o f ordinary language. ) Anothe r reaso n fo r th e tenacity o f th e ide a tha t a statemen t o f wha t w e mean whe n w e say so-and-s o ( a statement o f th e secon d type ) mus t b e syntheti c i s tha t w e suppos e i t t o b e describing th e menta l processe s o f th e perso n talking . T o gai n perspectiv e o n tha t idea , i t may b e o f hel p t o conside r tha t instea d o f sayin g t o th e chil d wh o sai d h e knew (whe n w e kne w h e ha d n o righ t t o say so) , "Yo u mea n yo u think so" , w e migh t hav e said , "Yo u don't kno w (or , Tha t i s no t wha t i t i s t o kno w something) ; you jus t thin k so" . Thi s say s neithe r mor e no r les s tha n th e formulatio n about wha t h e means, an d neithe r o f the m i s a descriptio n o f wha t i s goin g o n insid e th e child . 20 2 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 They are both statements which teach him what he has a right to say, what knowledge is. Mates tells us (ibid) that his "intensional approach" is meant, in part, "to do justice to the notions (1) that what an individual means by a word depends at least in part upon what he. wants to mean by that word, and (2) that he may have to think awhile before he discovers what he 'really* means by a given word". With respect to the first notion, I should urge that we do justice to the fact that an individual's intentions or wishes can no more produce the general meaning for a word than they can produce horses for beggars, or home runs from pop flies, or successful poems out of unsuccessful poems.33 This may be made clearer by noticing, with respect to the second notion, that often when an individual is thinking "what he 'really* means" (in the sense of having second thoughts about something), he is not thinking what he really means by a given word. You have second thoughts in such cases just because you cannot make words mean what you wish {by wishing); it is for that reason that what you say on a given' occasion may not be what you really mean. To say what you really mean you will have to say something different, change the words; or, as a special case of this, change the meaning of a word. Changing the meaning is not wishing it were different. This is further confirmed by comparing the locutions, "X means YZ" and "I mean by X, YZ". The former holds or fails to hold, whatever I wish to mean. And the latter, where meaning does depend on me, is performative;34 something I am doing to the word X, not something I am wishing about it. What these remarks come to is this: there is no such activity as myfinding-out-what-I-mean-by-a-word. But there obviously is finding-outwhat-a-word-means. You do this by consulting a dictionary or a native speaker who happens to know. There is also something we may call finding-out-what-a-word-really--means. This is done when you already know what the dictionary can teach you,' when, for some reason or other, you are forced into philosophizing. Then you begin by recollecting the various things we should say were such-and-such the case. Socrates gets his antagonists to withdraw their definitions not because they do not know what their words mean, but because they do know what they (their words) mean, and therefore know that Socrates has led them into paradox. (How could I be led into a paradox if I could mean what I wished by my words? Because I must be consistent? But how could I be inconsistent if words would mean what I wanted them to mean?) What they had not realized was what they were saying, or, what they 203 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 wer e really saying , an d s o no t know n what they meant. T o thi s extent , the y ha d no t know n themselves , an d no t know n th e world . I mean , o f course , th e ordinary world . Tha t may no t b e al l ther e is , bu t i t i s im - portan t enough : morality i s i n tha t world , an d s o ar e forc e an d love ; s o i s ar t an d a par t o f knowledg e (th e par t whic h i s about tha t world) ; an d s o i s religio n (whereve r Go d is) . Som e mathematic s an d science , n o doubt , ar e no t Thi s i s wh y yo u wil l no t fin d out wha t "number " o r "neurosis " o r "mass " o r "mas s society " mean i f yo u onl y liste n fo r our ordinary use s o f thes e terms. 3 5 Bu t yo u wil l neve r fin d ou t wha t voluntary actio n i s i f yo u fai l t o se e whe n w e shoul d say o f a n actio n tha t i t i s voluntary . On e may stil l fee l th e nee d t o say "Som e action s are voluntary an d som e ar e involuntary . I t woul d b e convenien t {[for what? } t o cal l al l action s voluntary whic h ar e no t involuntary . Surely I can cal l the m anythin g I like ? Surely wha t I call the m doesn't affec t wha t the y are?" Now : ho w wil l yo u tel l m e wha t "they " are? 3 6 Wha t w e nee d t o as k ourselve s her e is : I n wha t sort o f situation s doe s i t mak e n o differenc e wha t I cal l a thing ? or : A t wha t poin t i n a dialogu e doe s i t becom e natura l o r prope r fo r m e t o say , " I (you ) can cal l i t wha t I (you ) like" ? A t thi s poin t i t may b e safe t o say tha t th e questio n i s (ha s become ) verbal. 3 7 I f yo u reall y hav e a way o f tellin g jus t wha t i s denote d b y "al l action s whic h ar e no t involuntary" , the n yo u can cal l the m anythin g yo u like . I jus t trie d t o characteriz e th e situatio n i n whic h w e ordinaril y ask , "Wha t doe s X mean?" , an d t o characteriz e th e different situatio n i n whic h w e ask , "Wha t doe s X reall y mean? " Thes e question s neithe r conflic t no r substitut e fo r on e another , thoug h philosopher s ofte n tak e th e secon d a s a profoun d versio n o f th e firs t — perhap s t o consol e them - selve s fo r thei r lac k o f progress . Isn' t thi s par t o f th e troubl e about Synonymy ? "Doe s X really mean th e sam e a s Y? " i s no t a profoun d versio n o f "Doe s X mean th e sam e a s Y? " I t (it s occasion) is , thoug h relate d t o th e firs t i n obviou s an d deviou s ways , differen t Th e sam e goe s fo r th e pair : "Wha t di d h e do? " an d "Wha t di d h e reall y (literally ) doln ; an d fo r th e pair : "Wha t d o yo u see? " an d "Wha t d o yo u really (immediately ) see?"; an d fo r th e pair : "I s th e tabl e solid? " an d "I s th e tabl e really (absolutely ) solid? " Sinc e th e member s o f th e pair s ar e obviously different , philosopher s wh o d o no t se e tha t th e differenc e i n th e secon d member s lie s i n thei r occasions , i n wher e an d whe n the y ar e 20 4 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 posed, handsomely provide special entities, new worlds, for them to be about. But this can only perpetrate — it will not penetrate — a new reality. The profoundest as well as the most superficial questions can be understood only when they have been placed in their natural environments. (What makes a statement or a question profound is not its placing but its timing.) The philosopher is no more magically equipped to remove a question from its natural environment than he is to remove himself from any of the conditions of intelligible discourse. Or rather, he may remove himself, but his mind will not follow. This, I hope it is clear, does not mean that the philosopher will not eventually come to distinctions, and use words to mark them, at places and in ways which depart from the currently ordinary lines of thought.38 But it does suggest that (and why) when his recommendations come too fast, with too little attention to the particular problem for which we have gone to him, we feel that instead of thoughtful advice we have been handed a form letter. Attention to the details of cases as they arise may not provide a quick path to an all-embracing system; but at least it promises genuine instead of spurious clarity. Some philosophers will find this program too confining. Philosophy, they will feel, was not always in such straits; and it will be difficult for them to believe that the world and the mind have so terribly altered that philosophy must relinquish old exitements to science and to poetry. There, it may be claimed, new uses are still invented by profession, and while this makes the scientist and the poet harder to understand initially, it enables them eventually to renew and to deepen and to articulate our understanding. No wonder the philosopher will gape at such band ^wagons. But he must sit still. Both because, where he does not wish to invent (hopes not to invent), he is not entitled to the rewards and licenses of those who do; and because he would otherwise be running from his peculiar task — one which has become homelier perhaps, but still quite indispensable to the mind. The "unwelcome consequences" (Mates, p. 164) which may attend using words in ways which are (have become) privately extraordinary is just that our understanding should lose its grasp. Not only is it true that this can happen without our being aware of it, it is often very difficult to become aware of it — like becoming aware that we have grown pedantic or childish or slow. The meaning of words will, of course, stretch and shrink, and they will be stretched and be shrunk.-One of the great responsibilities of the philosopher lies in appreciating the natural and the normative ways in which 205 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 suc h thing s happen , s o tha t h e may mak e u s awar e o f th e on e an d capabl e o f evaluatin g th e other . I t i s a wonderfu l ste p towar d under - standin g th e abutmen t o f languag e an d th e worl d whe n w e se e i t t o b e a matte r o f convention . But thi s idea , lik e every other , endanger s a s i t release s th e imagination . Fo r som e wil l the n suppos e tha t a privat e meanin g i s no t mor e arbitrary tha n on e arrive d at publicly , an d tha t sinc e languag e inevitabl y changes , ther e i s n o reaso n no t t o chang e i t arbitrarily . Her e w e nee d t o remin d ourselve s tha t ordinary languag e i s natura l language , an d tha t it s changin g i s natural . (I t i s unfortunat e tha t artificia l languag e ha s com e t o see m a genera l alternative t o natura l language; 3 9 i t would , I suggest , b e bette r though t o f a s on e o f it s capacities. ) Som e philosophers , apparently , suppos e tha t becaus e natura l languag e i s "constantly " changin g i t i s to o unstabl e t o support on e exac t thought , le t alon e a dea r philosophy . But thi s Heraditea n anxiety i s unnecessary : linguisti c chang e i s itsel f a n objec t o f respectabl e study . An d i t misse s th e significanc e o f tha t change . I t i s exactl y becaus e th e languag e whic h contain s a cultur e change s wit h th e change s o f tha t cultur e tha t philosophica l awarenes s o f ordinary languag e i s illuminating ; i t i s tha t whic h explain s ho w th e languag e w e travers e every day can contai n undiscovere d treasure . T o se e tha t ordinary languag e i s natura l i s t o se e tha t (perhap s eve n se e why ) i t is' normativ e fo r wha t can b e said . An d als o t o se e ho w i t i s by searchin g definition s tha t Socrate s can coax th e min d dow n fro m self-assertio n — subjectiv e assertio n an d privat e definitio n — an d lea d i t back , throug h th e community , home . Tha t thi s als o renew s an d deepen s an d articulate s ou r understandin g tell s u s somethin g about th e mind , an d provide s th e consolatio n o f philosophers . Professo r Mates , a t on e poin t i n hi s paper , put s hi s doubt s about th e significanc e o f th e daim s o f ordinary languag e thi s way : "Surel y th e poin t i s no t merel y tha t i f you us e th e wor d 'voluntary' jus t a s th e philosophe r does , yo u may fin d yoursel f entangle d i n th e philosophi c proble m o f th e Freedo m o f th e Will " (p . 164) . Perhap s th e reaso n h e think s thi s i s a negligibl e consequenc e i s tha t h e hear s i t o n analog y wit h th e assertion , "I f yo u us e th e ter m 'space-time " jus t a s th e physicis t does , yo u may fin d yoursel f entangle d i n th e philosophi c prob - le m o f simultaneity" . Th e implicatio n i s tha t th e proble m mus t simply b e faced , no t avoided . I , however , hea r th e remark differently : I f yo u us e alcoho l jus t a s th e alcoholi c does , o r pleasur e a s th e neuroti c dots , yo u may fin d yoursel f entangle d i n th e practica l proble m o f th e freedo m o f th e will . Downloaded By: [University of Alberta] At: 21:30 25 February 2011 NOTES 1 This is a later, greatly expanded, version of the paper read as part of the symposium mentioned in Mates' first note. Since writing the relevant portions of this paper, I have seen three articles which make points or employ arguments similar to those I am concerned with: R. M. Hare, "Are Discoveries About the Uses of Words Empirical?", Journal of Philosophy, November 1957; G. E. M. Anscombe, "On Brute Facts", Analysis, January, 1958; S. Hampshire and H. L. A. Hart, "Decision, Intention and Certainty", Mind, January, 1958. But it would have lengthened an already lengthy paper to have tried to bring out more specifically than will be obvious to anyone' reading them their relevance to what I have said. 2 I am too conscious of differences in the practices of Oxford philosophers to be happy about referring, in this general way, to a school. But nothing in my remarks depends on the existence of such a school - beyond the fact that certain problems are common to the philosophers mentioned, and that similar questions enter into their attempts to deal with them. It is with these questions (I mean, of course, with what I understand them to be) that I am concerned. 3 Perhaps I should say "ideal" types. The statements do not come labeled in the discourse of such philosophers, but I am going to have to trust that my placing of statements into these types will not seem to distort them. 4 The harmfulness of this habit is brought out in Austin's "A Plea for Excuses" (Proceedings of the Aristotelian Society, 1956-7, pp. 16-19). Pp. 7-12 of this paper contains an elaborate defense of (anyway Austin's version of) "ordinary language philosophy". No one concerned with the general subject of the present symposium (or, in particular, with the possibility of budging the subject of moral philosophy) should ( = will) neglect its study. 5 Austin's discovery (for our time and place, anyway) of normal action is, I think, important enough to bear the philosophical weight he puts upon it — holding the clue to the riddle of Freedom. (Cf. "Excuses", p. 6.) A case can also be made out — as I try to do in a paper I hope soon to publish - that it was failure to recognize such action which produced some of the notorious paradoxes of classical Utilitarianism: what neither the Utilitarians nor their critics seem to have seen dearly and constantly is that about unquestionable (normal, natural) action no question is (can be) raised; in particular not the question whether the action ought or ought not to have been done. The point is a logical one: to raise a question about an action is to put the action in question. It is partly the failure to appreciate this which makes the classical moralists (appear?) so moralistic, allows them to suppose that the moral question is always appropriate (except, of course, where the action is unfree (caused?)). But this is no better than the assumption that the moral question is never appropriate (because we are never really free). Such mechanical moralism has got all the punishment it deserves in the recent mechanical antimoralism, which it must have helped inspire. 5 At the same time, Ryle leaves "involuntary" as stretched as ever when he allows himself to speak of "the involuntariness of [someone's] late arrival" (Concept of Mind, p. 72). 207 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 7 I realiz e tha t th e poin t i s controversia l an d tha t i n puttin g s o muc h emphasi s o n i t I ma y b e doin g som e injustic e t o di e poin t o f vie w I a m tryin g t o defend . Ther e may b e consideration s whic h woul d lea d on e t o b e mor e temperat e i n makin g th e point ; bu t agains t the poin t o f vie w Mate s i s adopting , i t seem s t o m e t o deman d al l th e attentio n i t ca n get . 8 A s is mos t clearl y show n wher e h e say s (p . 169) , ".. . 'Whe n I sa y 1 ma y b e wrong ' I d o no t imply ma t I hav e n o confidenc e i n wha t I hav e previousl y asserted ; I onl y indicat e it" . Wh y "only" ? Wer e h e willin g t o sa y "...bu t I d o (inevitably ) indicat e it" , ther e ma y b e no argumen t 9 Alternativ e (2b ) ha s bee n take n - fo r different , bu t no t unrelated , reason s - i n th e writing s o f Joh n Wisdo m (e.g. , "Gods" , i n Logic and Language, edite d b y A . G . N . Fle w (Oxford , 1951) , p . 196) , an d i n S . Toulmin' s The Place of Reason in Ethics (Cambridge , 1950) , p . 83 , an d i n S . Hampshire's , "Fallacie s i n Mora l Philosophy " (Mind, October , 1949) , pp . 470-471 . 1 0 I t wa s essentiall y th e argumen t with whic h th e pragmatist s attempte d t o subdu e emotiv e "meaning" . Se e Joh n Dewey' s "Ethica l Subject-Matte r an d Language " (Journal of Philosophy, 1945 , pp . 701-12) . 1 1 I thin k o f thi s a s a la w o f communication ; bu t i t woul d b e importan t an d instructiv e t o loo k fo r apparen t counte r instances . Whe n couldn't wha t i s sai d b e misunderstood ? M y suggestio n is , onl y whe n nothing , i s implied , i.e. , whe n everythin g yo u sa y i s sai d explicitly . (Shoul d w e add , o r whe n al l o f th e implication s o f wha t i s asserte d ca n b e mad e explici t in a certain way, e.g. , b y th e method s o f forma l logic ? I t may b e alon g suc h line s tha t utterance s i n logica l for m com e t o see m th e idea l o f under - standabl e utterances , tha t her e yo u ca n communicat e only wha t yo u say , o r els e more tha n yo u sa y withou t endangerin g understanding . Bu t w e migh t thin k o f forma l logi c no t a s th e guaranto r o f understandin g bu t a s a substitut e fo r it . (Cp . W . V . O . Quine , "Mr . Strawso n o n Logica l Theory" , Mind, October , 1953 , pp . 444—45) . The n w e ca n expres s thi s "la w o f communication " thi s way : Wha t need s understandin g ca n b e misunderstood. ) Bu t whe n is everythin g sai d explicidy ? Whe n th e statemen t i s abou t sense-dat a rathe r tha n "physical " objects ? Whe n i t i s abou t th e (physical ) movement s I mak e rathe r tha n th e (nonphysical? ) action s I perform ? Perhap s th e opponent s o f the Quest fo r Certainty (whos e passio n seem s t o hav e atrophie d int o a fea r o f th e wor d "certain" ) hav e embarke d upo n a Ques t fo r Explicitness . Straw-son' s notio n o f presupposing is relevant here, since explicitness and presupposition vary inversely. See hi s "O n Referring" , Mind, 1950 ; reprinte d i n Essays in Conceptual Analysis, edite d b y A . Fle w (London , 1956) . 1 2 A n examinatio n o f suc h form s i s part o f a mor e extende d stud y o n whic h I a m no w working . 1 3 I f i t stil l seem s tha t statement s lik e S an d T must b e synthetic , perhap s i t wil l hel p t o realiz e tha t anywa y the y ar e no t just some more syntheti c statement s abou t voluntar y action , o n a pa r wit h a statemen t t o th e effec t tha t somebod y doe s (indeed ) dres s th e wa y h e doe s voluntarily . I t may b e tru e tha t i f th e worl d wer e differen t enough, th e statement s woul d b e false ; bu t tha t amount s t o sayin g tha t i f "voluntary " mean t somethin g othe r tha n i t does , th e statement s woul d no t mea n wha t the y d a - whic h i s no t surprising . Th e statement s i n questio n ar e mor e closel y relate d t o suc h a statemen t as , "Th e futur e wil l resembl e th e past" : thi s i s no t a (no t jus t another ) prediction , o n a pa r wit h statement s abou t whethe r i t wil l rain . Russell' s chicke n 20 8 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 (who was fed every day throughout its life but ultimately had its neck wrung) was so well fed that he neglected to consider what was happening to other chickens. Even if he had considered this, he would doubtless still have had his neck wrung; but at least he wouldn't have been outsmarted. He could have avoided that indignity because he was wrong only about one thing; as Russell very properly says, ".. . in spite of frequent repetitions there sometimes is a failure at the last" (Problems of Philosophy, p. 102, original edition). But if the future were not (in the general sense needed) "like" the past, this would not be a failure. The future may wring our minds, but by that very act it would have given up trying to outsmart us. 14 A complaint Austin voiced in the course of his William James Lectures, on Performatives, at Harvard in the Spring term of 1955. 1 5 John Hospers, Introduction to Philosophical Analysis, p. 69. My emphasis. 16 Charles Stevenson, Ethics and Language, p. 26. 17 For modern instruction in the complexities of this question, see Austin's and P. F. Strawson's contributions to the symposium, "Truth" (Proceedings of the Aristotelian Society, Supplementary Vol. XXIV (1950)); D . Pears, "Universal", and "Incompatibilities of Colors" (both in Logic and Language, second series, edited by A. G. N . Flew (Oxford, 1953); W. V. O. Quine, "Two Dogmas of Empiricism" (Philosophical Review, 1951 ; reprinted in From a Logical Point of View, Harvard University Press, 1953); John Wisdom, papers collected in Philosophy and Psycho-Analysis (Oxford, 1953), especially "Philosophical Perplexity", "Metaphysics and Verification", and "Philosophy, Metaphysics and Psycho-Analysis". 18 The emphasized formula is Austin's ("Excuses", p. 7). Notice that the "should" cannot simply be replaced by "ought to", nor yet, I believe, simply replaced by "would". It will not, that is, yield its secrets" to the question, "Descriptive or normative?" 19 This is part of the view of philosophy most consistently represented in and by the writings of John Wisdom. It derives from Wittgenstein. 2 0 Of course you can say (the words), "When I ask whether an action is voluntary I do not imply that I think something is special about the action". You can say this, but then you may have difficulty showing the relevance of this "voluntary" to what people are worrying about when they ask whether a person's action was voluntary or whether our actions are ever voluntary. We might regard the Oxford philosopher's insistence upon ordinary language as an attempt to overcome (what has become) the self-imposed irrelevance of so much philosophy. In this they are continuing — while at the same time their results are undermining — the tradition of British Empiricism: being gifted pupils, they seem to accept and to assassinate with the same gesture. 2 1 This latter distinction appears in two senses of the expression "establishing a rule or standard". In one it means finding what is in fact standard in certain instances. In the other it means founding what is to be standard for certain instances.' "Settle" and "determine" have senses comparable to those of "establish". 2 2 It is perfectly possible to maintain that any "justifications" w e offer for our conduct are now so obviously empty and grotesquely inappropriate that nothing w e used to call a justification is any longer acceptable, and that the immediate questions which face us concern the ultimate ground of justification itself. We have beard about, if we have not seen, the breaking down of convention, the fission of traditional values. But it is not a Continental dread at the realization that our standards have no ultimate 209 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 justificatio n whic h lend s t o s o muc h Britis h an d America n mora l philosophizin g it s hysterica l quality . (Suc h philosoph y ha s bee n abl e t o tak e th e deat h o f Go d i n it s stride. ) Tha t qualit y comes , rather , fro m th e assumptio n tha t th e questio n o f justifyin g case s i s o n a pa r wit h (appropriat e i n th e sam e contex t as ) th e questio n o f justifyin g norms . 2 3 Thoug h i n anothe r contex t w e migh t have . Imagin e tha t befor e ches s wa s introduce d int o ou r culture , anothe r gam e — cal l i t Ques t - ha d bee n popula r wit h us . I n tha t game , playe d o n a boar d wit h 6 4 squares , an d lik e ches s i n othe r respects , th e piec e calle d th e Damse l ha d a fickl e wa y o f moving : it s firs t move , an d ever y od d mov e afterwards , followe d th e rul e fo r th e Quee n i n chess ; it s eve n move s followe d th e rul e fo r th e Knight . I t ma y b e suppose d tha t whe n peopl e bega n t o pla y Quest , i t ofte n happene d tha t a gam e ha d t o b e stoppe d upo n rememberin g tha t severa l move s earlie r a Quee n wa s permitte d a Knight' s move . Th e rul e fo r th e Queen' s mov e migh t the n hav e bee n formulate d i n som e suc h wa y as : Yo u mus t mov e the Quee n i n straight , unobstructe d paths.. . 2 4 Perhap s thi s differenc e provide s a wa y o f accountin g fo r ou r tendenc y sometime s t o min k o f law s a s rule s an d a t othe r time s t o thin k o f the m a s commands . Thi s ma y (i n part ) depen d upo n wher e w e - i.e. , wher e ou r norma l action s - stan d (o r wher e w e imagin e the m t o stand ) wit h respec t t o th e la w o r syste m o f law s i n question . I t ma y als o b e significan t tha t whe n yo u ar e describin g a syste m o f laws , yo u ar c likel y t o thin k o f yoursel f a s externa l t o th e system . 2 3 Bu t thi s require s a grea t dea l o f work . W e mus t hav e a bette r descriptio n o f th e "class " an d th e functio n oE "moda l auxiliaries" , an d w e nee d a n understandin g o f wha t make s somethin g w e d o "another " actio n an d wha t make s i t "part " o f a n actio n i n progress . 2 6 "Must " retain s it s logica l forc e here . Kan t ma y no t hav e provide d a n analysi s sufficien t t o sustai n hi s sayin g tha t " a deposi t o f mone y mus t b e hande d bac k becaus e i f th e recipien t appropriate d it , i t woul d n o longe r b e a deposit" ; bu t Bergso n to o hastil y conclude s tha t Kant' s explanatio n o f thi s i n term s o f "logica l contradiction " i s "obviousl y jugglin g wit h words" . (Se e Bergson' s Tbe Two Sources of Morality and Religion, Ne w York , 1935 , p . 77. ) Th e differenc e betwee n you r depositing an d simpl y banding ove r som e mone y ha s i n par t t o d o wit h wha t yo u mea n o r inten d t o b e doin g - an d wit h wha t yo u can mea n o r inten d b y doin g wha t yo u d o i n th e wa y yo u d o i t i n tha t particula r historica l context . W e may , followin g a suggestio n o f H . P . Grice' s (i n "Meaning" , Philosophical Review, July , 1957) , thin k o f th e action s o f depositin g an d o f acceptin g a deposi t a s complicate d "utterances" : yo u inten d tha t wha t yo u d o shal l b e understood . The n i t wil l no t see m s o extraordinar y t o sa y tha t a late r "utterance " (viz. , appropriatin g th e entruste d money ) contradict s a forme r on e (viz., acceptin g a deposit) . 2 7 Cite d i n Hoyle' s Rules of Games, ed . Morehea d an d Mott-Smit h (Signe t Book , Ne w York , 1949) . 2 3 Som e philosopher s wh o emplo y th e notio n o f a rul e hav e give n the impressio n tha t ther e are . Wha t I a m tryin g t o argu e i s tha t ther e aren't , bu t tha t the analog y i s stil l a goo d one . On e o f th e claim s mad e fo r th e concep t o f a rul e i s tha t i t illuminate s th e notio n o f justification ; an d critic s o f the concep t argu e ma t i t fail s i n thi s an d tha t therefor e th e concep t i s unilluminatin g i n the attemp t t o understan d mora l conduct . I chin k bot h o f thes e claim s ar e improper , resultin g i n par t fro m th e failur e t o appre - " 21 0 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 dat e differences (1) between rules and principles, and (2) between performing an action and making some movements. The concept of rule does illuminate the concept of action, but not that of justified action. Where there is a question about what I do and I cite a rule in my favor, what I do is to explain my action, make clear what I was doing, not to justify it, say that what I did was well or rightly done. Where my action is in accord with the relevant rules, it needs no justification. Nor can it receive any: I cannot justify moving the Queen in straight, unobstructed paths. See John Rawls' study of this subject, "Two Concepts of Rules" (Philosophical Review, January, 1955). My unhappiness with the way in which the analogy is drawn does not diminish my respect for this paper. For a criticism (based, I think, on a misunderstanding) of the view, see H. J. McCloskey, "An Examination of Restricted Utilitarianism", The Philosophical Review, October, 1957. 2 9 Such an understanding of meaning is provided in Grice (op. cit), but I do not think he would be happy about the use I wish to put it to. A conversation we had was too brief for me to be sure about this, but not too brief for me to have added, as a result of it, one or two qualifications or clarifications of what I had said, e.g., the third point of note 32, note 33, and the independent clause to which the present note is attached. 3 0 If truth consists in saying of what is that it is, then (this sense or source of) necessary truth consists in saying of what is what it is. The question, "Are these matters of language or matters of fact?" would betray the obsession I have tried to calm. I do not claim that this explanation of necessity holds for all statements which seem to us necessary and not analytic, but at best for those whose topic is actions and which therefore display a rule-description complementarity. 3 1 Austin's work on Excuses provides a way of coming to master this immensely important idea. The way I have put the point here is due directly to it. 3 2 Three points about this conclusion need emphasizing: (1) It was reached where the difference concerned isolated words; where, that is, the shared language was left intact. (2) The tasks to be performed (scraping; chopping; excusing a familiar and not very serious mishap) were such as to allow execution, if more or less crude, with a general or common implement. (3) The question was over the meaning of a word in general, not over its meaning (what it was used to mean) on a particular occasion; there was, I am assuming, no reason to treat the word's use on this occasion as a special one. Wittgenstein's role in combatting the idea of privacy (whether of the meaning of what is said or of what is done), and in emphasizing the functions of language, scarcely needs to be mentioned. It might be worth pointing out that both of these teachings are fundamental to American pragmatism; but then we must keep in mind how different their arguments sound, and admit that in philosophy it is the sound which makes all the difference. 33 I am not, of course, denying that what you say depends upon what you intend to be saying. I am, rather, denying that intending is to be understood as a wanting or wishing. And I am suggesting that you could not mean one thing rather than another ( = you could not mean anything) by a given word on a given occasion without relying on a (general) meaning of that word which is independent of your intention on that occasion (unless what you are doing is giving the word a special meaning). For an analysis of meaning in terms of intention, see Grice, op. cit. 211 Downloaded By: [University of Alberta] At: 21:30 25 February 2011 3 4 O r els e i t i s a special report , lik e th e on e o n p . 4 3 (las t sentence) ; bu t i t i s stil l no t a descriptio n o f m y wishe s o r intentions . Unti l Austin' s Willia m Jame s Lecture s ar e published , th e bes t plac e t o fin d ou t wha t a "performative " i s remain s pp . 142-4 7 o f hi s "Othe r Minds" , i n Logic and Language, secon d series . 3 5 Thi s ma y b e summarize d b y sayin g tha t ther e i s n o suc h thin g a s finding out wha t a number , et c is . Thi s woul d the n provid e th e occasio n an d th e justificatio n fo r logica l construction . 3 6 Cf. D . F . Pears , "Incompatibilitie s of Colours " (Logic and Language, secon d series , p . 119 , n . 2) . 37 On e o f th e bes t way s t o ge t pas t th e ide a tha t philosophy' s concer n wit h languag e i s a concer n wit h word s (wit h "verbal " matters ) b t o rea d Wisdom . Fortunatel y i t i s a pleasan t way ; becaus e sinc e the ide a i s on e tha t yo u hav e t o ge t pas t agai n an d again , th e wa y pas t i t wil l hav e t o b e take n agai n an d again . 38 A s Austi n explicitl y says . Se e "Excuses" , p . 11 . . 3 9 Thi s sometime s appear s t o b e th e onl y substantiv e agreemen t betwee n th e philosopher s wh o procee d fro m ordinar y languag e an d thos e wh o procee d b y constructin g artificia l languages . Bu t thi s ma y wel l b e obscurin g thei r deepe r disagreements , whic h are , I believe , les s abou t languag e tha n abou t whethe r th e tim e ha s com e t o dra g fre e o f th e philosophica l traditio n establishe d i n respons e to , an d a s par t of, th e "scientifi c revolution " o f th e 16t h an d 17t h centuries . I hav e foun d instructio n abou t thi s i n conversation s wit h m y frien d an d colleagu e Thoma s S . Kuhn , t o who m I a m als o Indebte d fo r havin g rea d (an d force d th e rewritin g of) tw o shorte r version s o f thi s paper . Erratum: Vo L 1 , 1958 , No . 3 , p . 16 3 line s 9—1 1 read: "Th e differenc e betwee n th e ordinary us e an d th e philosophi c us e hinge s rathe r upo n th e allege d fac t tha t th e ordinary man applie s th e wor d "voluntary " almos t exclusively t o action s whic h ... "