Category Differences Author(s): R. C. Cross Source: Proceedings of the Aristotelian Society, New Series, Vol. 59 (1958 - 1959), pp.
255-270 Published by: Blackwell Publishing on behalf of The Aristotelian Society Stable URL: http://www.jstor.org/stable/4544614 . Accessed: 30/09/2011 20:59
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
The Aristotelian Society and Blackwell Publishing are collaborating with JSTOR to digitize, preserve and extend access to Proceedings of the Aristotelian Society.
http://www.jstor.org
Meeting of the AristotelianSociety at 21, BedfordSquare, London,W.C.1, on 25th May, 1959, at 7.30 p.m.
XIlI-CATEGORY
DIFFERENCES
R. By PROFESSOR C. CROSS
The idiom of " categories ", " category differences ", "category mistakes" is familiar in recent philosophy. There have too been signs of dissatisfaction with it. Mr. Hampshire, in his review of The Concept of Mind, refers to " such notoriously obscure expressions as ' logical category ' ", and Mr. Warnock in his recent book English Philosophy since 1900 after a brief discussion concludes that "it is not ... unreasonably overscrupulous to be ill at ease with the use of an idiom which has none at all of the precise backing which it naturally implies ", and he suggests that it can be dispensed with. My present purpose is to consider again the use made of this idiom. Most of the ground is well-trodden, and I shall in the end come to a conclusion little different from that of Mr. Warnock; but the idiom itself is persistent, and what lies behind it important, and this is reason enough for further study. I begin with a brief historical excursus on the traditional Aristotelian use of the " category " terminology, not, to use a phrase of Ryle's, for purposes of philosophical paleontology, but to elicit some points in the Aristotelian tradition that may be useful in what follows. In the Categories a distinction is made between " things said without combination " (i.e., terms) and " things said in combination" (i.e., propositions), and in reference to the former a list of " categories " is then given-substance, quantity, quality, relation and so on, and these arethen illustrated"to sketch my meaning roughly, of quantity, such terms as 'two cubits long ' or ' three cubits long', of quality, such attributes as 'white ', ' grammatical '. ' Double ', ' half', ' greater ' fall under the category of relation ... .". The categories then, as the Greek word implies, seem to constitute a list of all the different sorts of predicates that can be predicated respectively of the various terms we use,1 and thus provide an
1
Cf. W. D. Ross, Aristotle, p. 23. 2 F
256
PROFESSORR. C. CROSS
inventoryof the differenttypes of terms. As Ryle has pointed all to at out,2Aristotleappears havearrived his list by considering the possible kinds of questionsthat can be asked in ordinary
speech-" what is it ?", " what like is it ? ", " when did it happen ? "
to according etc., and classifyingtermsas of differentcategories the interrogativeto which they are the appropriateresponse. " Thus " yesterday fits the " when" question, as "white " or " " grammatical could not; and equally " white or " gram" matical" fit the " whatlike " question,as " yesterday couldnot. " yesterday on the one hand, and " white" or " Thus " " grammatical on the other, are predicates of different categories. This is perhapsenough to recall the Aristoteliandiscussion; and now, ignoringmany other things that might be said, let us notice four points. (1) There is the point already indicated. Aristotle seems to have arrivedat his list by an analysisof the way we talk, by consideringthe possible sorts of questionthat the can be ordinarilyasked, and differentiating terms used in response to these questions into their appropriatecategories accordingto the questions to which they are sensible replies. So far then Aristotle seems to be concerned with the way languageworks, and so far in this sense we might say the categories have a purelylogical aspect. But (2) at the same time it seems quite clear that for Aristotle his categorieshad a metaand that he regardedthem as a list of the physicalsignificance different modesof existencethat the worlddisplayed. Theywere a list of the different modesof being,an indicationof the structure of the universe. Hence Ross's remark: " In its earliestform of the doctrine(of categories)was a classification the meanings of i.e., of the thingsmeant by, ' uncombinedwords', in other wordsan inventoryof the mainaspectsof reality,so far at leastas languagetakes accountof them."3 Hencetoo whatRobinsays:
" Elles (the categories) seraient donc dans l'Aristote'lismele lien qui unit le point de vue logique et le point-de vue ontologique, car elles sont des realite'spar soi. Ainsi 'I 'etre en tant qu'etre' se
2 "Categories ", P.A.S,
1938-39.
Reprinted in Logic and Language
(Second Series),ed. A. G. N. Flew. 3 Ross, loc. cit. p. 23. For the following quotation, see L. Robin, Aristote, p. 101.
CATEGORY DIFFERENCES
257
connaltpar les ' categories' qui en sont les aspectsreels, comme, sont chez Spinoza,les Attributs ce quel'onconnaft la Substance." de Robingoes on to suggestthat the table of categories in the end is established somekindof intuition,however a by important prelimof inaryrolemaybe playedby the analysis language. Atanyrate, for our immediatepurposesit is clear that in this Aristotelian us tradition languagewouldbe onlya clue,assisting in the graspof is thesefundamental aspectsof reality. (3) Thelist of categories a limited list. In fact, in different places in the Aristotelian writingsthe listvariesin detail,but it is alwaysa shortlist, i.e., the variousmodesof beingcan be allocatedto a quitelimitednumber of pigeon-holes. (4) On categorymistakes I cannot find that Aristotlehas anythingto say. If one were asked " how tall is Socrates" and answered" in the Peirmus this would presum", ably be a category mistake. But then one would hardly be likely to answer in this way, and little that is philosophically excitingappearsto emerge. What Aristotledoes indicateis that are categorydifferences irreducible:" Thosethingsare said to be ' other in genus' whose ultimate substratumis different,and which are not analysedthe one into the other nor both into the same thing (e.g., form and matterare differentin genus); and things which belong to differentcategoriesof being; for some of the things that are said to 'be' signify essence, others a we quality,othersthe othercategories have beforedistinguished; these also are not analysedeitherinto one anotheror into some one thing."4 So much then for the Aristotelianbackground. When we turn now to considerthe use of the notions of categoriesand in categorydifferences recentphilosophy,it is of courseparticularly with Professor Ryle that these are associated. I shall thereforerefer largely to his use of these notions and what he has to say about them. My purposeis not howeverpolemical, and therefore I am not concerned to trace historically any theremayhavebeen in his views,nor to ensurethat developments a fair and complete picture of them is given. For example, Ryle might want to modify some of the things he says in his earlypaperon " Categories or againwhenhe saysin the much ";
4 Arist. Met.
1024b 10 foll. (Oxfordtrans.). 2 F2
258
PROFcSSOR R. C. CROSS
later Dilemmas in talking of categories " this idiom can be associations helpful as a familiarmnemonicwith some beneficial It can also be an impediment,if creditedwith the virtues of a skeleton-key this seemsto indicatethat the idiom as he usesit ", should not be pressedtoo hard. For presentpurposeshowever this is in a senseirrelevant;for we are not primarily in interested Ryle'sviewsas such,but only in themin so far as they are useful in askingmore generalquestionsabout categoriesand category differences. Havingthus,I hope,justifiedthe use of Ryle as a sourcebook, I wish first to recall the very familiarexampleshe gives at the beginningof TheConcept Mind,whenhe is explaining what he of means by the expression " category-mistake One of the ". examples is enough, for instance that of the foreigner who, watchingcricketfor the firsttime,learnsaboutthejobs of bowlers, batsmen,and so on, and then asks who is left to contributethe elementof team-spirit. The foreignerin this case is committing a category-mistake, failing to recognize that the concept of team-spiritbelongs to a differentcategoryfrom the conceptsof bowling and so on. Now about this and the other examples (the University and the division) which Ryle uses two things maybe noted. First,in the contextsin whichthey arepresented,5 it is clear in all of them that somethinghas gone wrong. It is certainly queer to suppose that e.g., team-spiritis an extra operation,additionalto bowling and fieldingand so on, and to speak of e.g., a fielderas eitherfieldingor displaying team-spirit Ryle's explanationof what is wrong is in terms of category mistakes. To say that a memberof the side was eitherfielding or exhibitingteam-spirit would be to show that one had failed to recognizea categorydifferencebetweenthe two concepts of fieldingand team-spirit. " The mistakes", he says, " weremade by peoplewho did not knowhow to wieldthe conceptsUniversity, divisionand team-spirit. Their puzzles arose from inability to use certain items in the English vocabulary."'6But secondly, whatexactlyis meantbysayingthattheconcepts theseexamples in belong to differentlogical categories, and that the queerness
Analysis, Vol. 15, No. 6, p. 135. 6 The Concept of Mind, p. 17.
But f. Professor Popper, " A note on the Body-Mind problem",
CATEGORY DIFFERENCES
259
arises from failing to realizethis categorydifference? We feel whenwe do? And is the queerness, whatexactlyis happening but to it helpful,i.e., genuinely explanatory, accountfor the queerness in termsof differing categories? It shouldbe noted that Ryle is of in the main content,in The Concept Mind,to explainwhat he by like means by a category-mistake meansof illustrations those that above, and to let one feel the mistake,the queerness arises, accountof what lies behind withoutgivingany furthertheoretical
the queerness. His earlier paper, " Categories ", to which I
havealready seeksmoredirectly explainin a theoretical to referred, " way what he meansby " categorydifferences and how they are detected, and I propose thereforeto look for help from that quarter. In this earlierpaper Ryle proposesa test for discriminating conceptsinto differentcategories. What he says, at any rate to are begin with, is this: " Two proposition-factors of different such that when categoriesor types, if there are sentence-frames the expressionsfor those factors are imported as alternative complementsto the same gap-signs,the resultantsentencesare in in significant the one caseand absurd the other."7 An example " willmakewhatis meantclear. " - was all out for two hundred would be a sentence-frame, dash being the gap-sign. We the can put into the blank indicated by the dash an expression such as " England" or " The English (i.e., proposition-factor) side ", and the resultis a significant sentence; but if we attempt
to put in the blank an expression like e.g., " Cowdrey ", the
resultingsentence is absurd. In such a case we can say that an expressionlike " Cowdrey" is of a different logical category like from an expression " The Englishside ". It shouldbe noted will that Ryle is not claimingthat the fact that two expressions both go significantly into the same blank in the same sentenceframeshowsthat they are of the samecategory-e.g., both " The Englishside " and " Cowdrey" can go into the blank in " -is batting now "-compare his own example of " I " and " the can writerof this paper." The claim is that if a sentence-frame be found such that of two expressions can go into the blank one in it to produce a significantsentence,while the other cannot
I
ogic and Language (Second Series), pp. 77-78.
260
PROFESSORR. C. CROSS
go into the same blank without producingabsurdity,the two expressions belong to different categories. It is also worth notingthat, as Ryle himselfpointsout, thereis a certainsimilarity between his method of detecting category differencesby this device of a gap-signin a sentence-frame, that of Aristotle, and since interrogative sentences, if one ignores their practical purpose,are like sentence-frames, the interrogative and wordsin them like gap-signswhich, for sense to result, must be filledby the appropriate type of response. In both cases then, to revert to the firstpointI madein discussing Aristotelian the background, we are concernedwith the way languageworks, and so far, in bothcases we might say that the categorieshave a purelylogical aspect. Now therearedifficulties this suggested in methodof detecting categorydifferences. First, let us considera criticismProfessor Smart has made.8 He suggests that this test, if pushed far enough,mightshow everyexpression be of a different to category from every other, in which case, he adds, " we should be wise not to take it too seriously". Thus, to use one of his own
examples, given the sentence-frame " the seat of the is hard ",
we obtaina significant sentence " chair or " bench" is putinto if the blank, but an absurd sentence if "table " or " bed " is insertedinstead. This would seem to show that "chair and " bench" belong to a differentcategory from "table and " bed ". But, as Smartsays, " if furniture words do not form a category,we may well ask what do." It might be, however,that this criticismis not so formidable as it looks, for it leaves out of account some other important things that Ryle says. In particular,it seems clear from the latter part of Ryle's paper that he does not regardhis test for categorydifferences merelyas a simplematterof insertingwords in the gap-signin an isolated sentence-frame looking at the and result. He has his eye not simplyon the individualconstituents (proposition-factors) a proposition, simplyon the proposiin nor tion itself as a unit, but on the proposition its possiblerelations in to other propositions. He insists that the logical types of the variousconstituentsin a propositionare in the end determined
8 B.J. for Phil. of Sc., Vol. IV, pp. 227-28.
CATEGORY DIFFERENCES
261
by the logical relationsof the propositionitself to otherpropositions. It is these relationsthat revealthe formalproperties both of the propositionand of its constituents. Thus he says: " Let us give the label 'liaisons' to all the logical relations of a proposition,namelywhat it implies,what it is impliedby, what it is compatible with, and whatit is incompatible with. Now any respectin whichtwo propositionsdifferin form will be reflected in differences their liaisons. So two propositionswhich are in formally similar in all respects save that one factor in one is factorin the other, different typefroma partially in corresponding will have liaisons which are correspondingly dissimilar." And he goes on to add " The operationof extractingthe type of a factor cannot exclude the operation of revealingthe liaisons of propositionsembodyingit. In essence they are one operation."9 It might then be arguedagainstSmartthat this helps in the "chair" and " bed " example. For, it might be said, whatever may be thoughtaboutthe sentence"The seat of the bedis hard" " bed " propositions and " chair" propositions have similar logical liaisons, and thus from this we can recognizethat their factorsare not of different constituent logicaltypes. The trouble about this howeveris that rightfrom the startthe sentence" the seat of the bed is hard" seemsin some sense absurd. We tend
to be stopped immediately by a response like " that's nonsense ", or " I don't understand what you are talking about " or " what
do you mean? Beds haven'tgot seats." And given this initial check,it is not easy to go on to arguethat thoughthe sentenceis absurd, still, for purposesof placing the logical categoryof its constituents,we can see that its logical relationswould be like that of the respectable sentence" the seat of the chairis hard". From the beginningthe propositiondoes seem in some sense absurd, and thereforeit looks as though, since this is so, it satisfies Ryle's test, and " bed " and " chair" are words of differentcategories. The word " absurd" however is a difficult one. When Ryle is discussingwhat complementcan go into the gap in a he given sentence-frame, observes that there are two sorts of
'ogic and Language (Second Series), pp. 79-80.
262
PROFESSORR. C. CROSS
" " can " involved. Thereis a grammatical can ", where,to use requiresin the his own example,"- is in bed" grammatically blank a noun or noun-equivalent. Thus in this sense of " can " " Saturday can go into the gap, i.e., " Saturdayis in bed " is " correct. Yet, he says, this sentenceis absurdgrammatically " in the sense of " can" in which he is interested" Saturday cannot go into the gap. To put it in is to commit a logical absurdity-it logically cannot go in. The question arises howeverwhethertwo sorts of " can " are enough. If we look at Smart's example again, grammaticallythe sentence-frame " the seat of the - is hard" requiresin the blank a noun or cannotgo in. a noun-equivalent-grammaticallyverbor adjective But again " bed" cannot go in, because to put it in produces absurdity. Yet, as Smarthas pointed out, we would not want that to say that this is the sort of absurdity is a clue to a difference of categorybetween" chair", which will go into the blank,and " bed " which will not. One is tempted to say that " bed " cannot go into the blankbecausebeds are not in fact thingswith seats. If anyone tried to put " bed" in the blank, his mistake would arise because beds are not made that way; and one is thusfurther temptedto say thatthis kindof " can " or " cannot", basis. here,has a directlyempirical if thesewordsare appropriate we On the other hand, if we considerthe sentence-frame had earlier-" was all out for two hundred", the sense in which " England" can, and " Cowdrey" cannot go into the gap feels is different. If we put " Cowdrey" in the gap, the sentence again absurd,but this time, one wants to say, logicallyabsurd. It is of truethat we may say the mistakearisesfromignorance cricket, of casethe mistake arosefromignorance furniture, as in the former sort of but it seemsequallytrue that in the end this is a different " mistake, which reveals that " England" and " Cowdrey express proposition-factors of different logical types-one logically cannot use the expression" Cowdrey" in that sort of
way. It may help to illustrate the point if we consider a different set of examples, for instance the sentence-frame " He - the book in carefully ", and try inserting theblankthe three verbs" spliced",
" read" and " saw ". " He spliced the book carefully" is absurd in the sense that books are not like that-one splices
CATEGORY DIFFERENCES
263
ropes or fishing rods, but books are constructeddifferently. " He readthe book carefully" is impeccable. On the otherhand " He sawthe book carefully is againabsurd,but again,it seems, " in a differentway from the first sentence. Here we want to say one logicallycannotuse " see " in this sort of sentence-frame ", whichcontainsan adverblike " carefully whereasone can use both " splice" and "read" in such a frame. It is true that there is an absurdityin using " spliced" here, but a different in. The and sort of absurdity, not the sort that Ryleis interested difficultywith the adverb " carefully" in the " see " sentence between" see" and either" splice" points to a logical difference or " read "; for one can both splicecarefullyand readcarefully but one can never see carefully. Grantedhowever that " He " saw the book carefully is absurdin the sense in which we are interested,it is still not easy to be clearwhat the sense is. Ryle
says he adopts the word ' absurd ' in preference to " nonsensical " or " meaningless " partly because it " has helpful associations
ad withthe reductio absurdum, evenits nuanceof ridiculousness and is useful ratherthan the reverse,for so many jokes are in fact This is not over-helpful, nor does recourseto type-pranks."10 the OxfordEnglishDictionary help mucheither-" incongruous, unreasonable, illogical. In modernuse, esp. plainlyopposedto
reason, and hence, ridiculous, silly ", leaves us wide scope. In
that this difficultyabout " absurd" we must howeverremember though we seem on occasion to recognize the logical sort of we absurdity arelookingfor at oncewithinan isolatedproposition, that we and though indeed earlierRyle seemedto be suggesting in detect categorydifferences just this sort of way, presumably neverthelessthe absurdityis in the end connected with the normalbehaviourof the conceptin questionas exhibitedby the in liaisonsof propositions whichit normallyoccurs,as contrasted with its presence in an alien proposition. But here again it would seemthat thereis no logicallyexact standard availablefor the determining possibleliaisons of a propositionitself, such as would reveal in a once-for-all way the type-features of its constituents.
In The Concept of Mind Ryle says, " I try to use reductio
10 Loc.
cit. p. 76.
264
PROFESSORR. C. CROSS
ad absurdumargumentsboth to disallowoperationsimplicitly recommendedby the Cartesianmyth and to indicate to what logical types the concepts under investigation ought to be allocated."11But it is clearthat many of the arguments used in the book are less rigorousthan this, as Ryle himselfrecognizes when he goes on to say " I do not howeverthinkit improperto use from time to time arguments a less rigoroussort ". Thus of it does not seemthat thereis any simple,logicallyexactstandard, such as for examplecontradiction, providethe crucialtest for to revealingcategorymistakeswhen one is examiningthe liaisons between propositions. Thus whether within a proposition or between propositionsthere does not appear to be any precise andsimpletest of absurdity;andRylehimselfof courserecognizes this both when he stressesthe importanceof " ratiocinationor argumentation"in formulatingthe liaisons of a proposition,12 andby the questionwithwhichheendsthepaperon " Categories': "But what are the tests of absurdity?" At any rate, whateverthe difficulties may be, it will be noted " thatthroughout earlypaperon " Categories Ryle'sapproach the is a logical approach. To revert for a moment to my second it point aboveconcerning Aristotelian the background, was noted there that for Aristotle the categorieswere featuresof the real world. Languagewas a useful clue to their discovery,in that it reflectedthese differences the real; and behind the logical in differences revealedin languagewere the modes of being which made language what it was. Ryle's approach on the other hand begins and ends in logic-category differences deterare mined by the behaviourof concepts within propositions and by the relationsbetweenpropositions. I wish now to develop this point, but beforeI do so two notes shouldbe added. First, it has been pointed out13 that there are placesin The Concept of MindwhereRyleseemsto ask himselfthe questionwhethersome does or doesnot " standfor " or " designate" puzzlingexpression something. Thisway of talkingwouldsuggestthatitwaspossible to have some sort of priorknowledgeof the natureof thingsand to assessand discriminate in language the light of that knowledge.
"The Conceptof Mind,p. 8. 12 " Categories", LanguageandLogic, p. 80. 13 E.g., by Hampshire,Mind, 1950 p. 241 foll.
CATEGORY DIFFERENCES
265
Categorydifferences would be differences the real, whichwere in subsequently reflected language; and this wouldbe reminiscent in of the Aristoteliantraditionwe have discussed. But in fact it is clearthat Ryle does not wishto suggestthis, and thereis certainly no such suggestion in the earlier paper on " Categories ". the Secondly,it mayperhaps addedherethat granted investigabe tion into categoriesand categorydifferences withinlanguage, lies it would not at all follow that no light was thrown therebyon the natureof things,in some sense or other of the latterphrase. I should entirely agree with Ryle's remarkthat " If a child's perplexity why the Equatorcan be crossedbut not seen, or why the CheshireCat could not leave its grin behindit, is perplexity about the ' natureof things', then certaincategory-propositions will give the requiredinformationabout the nature of things. ".14 And the samewill hold good of less frivoloustype-perplexities To returnnow to the point that, in contrastwith Aristotle, " Ryle's " categories lie wholly within logic, I wish to refer briefly to a recent paper by Professor Manley Thompson.15 In this paper Thompson defends the traditional Aristotelian positionconcerning categorydifferences the problemsarising and therefromagainst various modern views. He thinks that in like understanding use of conceptssomething the Aristotelian the notion of categorydifferences indispensable. Thus we have is to usean ontological wellas a linguistic idiom,andconsequently as we cannot evade the traditionalproblemsconcerningthe senses of " existence" or the making of some form of ontological statements. In his developmentof this generalthesis there are two points that are perhapsparticularly relevantto the present
discussion.
First, there is the way in which Thompson detects a category-difference.Take, for example,a fox and a crow. We can use a commonpredicateof both of these-a fox and a crow are both animals. The predicate" animal" separatesboth off from a wide range of other things e.g., from a cruiserand a destroyer,which are both ships, and opens up a whole groupof questionscommon to both, just as in the second case thereis a groupof questionscommonto ships, and so on. Suchgroupsof
14
6
15 "
Categories", Logic andLanguage(Second Series)p. 81. On CategoryDifferences". Phil. Review, 1957, pp. 486 foll.
266
PROFESSORR. C. CROSS
are questions eachmatterfor the special sciences, thequestions e.g., aboutanimalsfor zoology,those aboutshipsfor marineengineercan ing. Wherea commonpredicate be found whichenablesus to ask meaningful questionsabout the subjectsof the predicate, there is no differenceof category between the subjects-for example between foxes and crows. But now consider a fox and a number. In this case we can find no common predicate analogousto " animals" in the formercase. The best we can is do by way of commonpredicate to say that they are " objects"
or " things " or " entities " or " existents ". But if this is all
they have in common,then they reallyhave nothingin common which would mark them off from any other pair, for we can use " thing", " entity" etc. as a predicate any objectwhatever. of Wherethis happens,i.e., wheretheobjects havenothing concerned in commonin the senseexplained, have a categorydifference. we In such cases,if we attemptto use a commonpredicate, have we to use a word like " entity" etc., i.e., in Thompson'sphrase " entity" etc. occur essentially. In such a case too no area has been delimitedfor any particularscience or branchof studycontrastthe fox-crow-animal case. If any light is to be thrown on a statementsuch as " a fox and a numberare both entitiesor existents" it has to be done by explainingthat " entity" or " existent" is being used in a differentsense in each of the two cases, that foxes and numbersare differentkinds of things or existents. This leads us directlyto the traditionalAristotelian problems, which on the traditional view are problems for philosophyas distinctfromthe problems the specialsciencesor of disciplines. It may be noted here in passingthat Thompson'sexamples are confinedto nouns of the speciesand genus type and that he does not discuss other sorts of expressions-for exampleverbs. It may be that in his context he would wish to ignore these; but others might want to say that there is a differenceof " category", in some sense of that word, betweenfor example " reading" and " seeing"; and it is not altogetherclear how these could be fitted into his scheme. This raises a further when one point, that he givesno detailedcriteriafor determining has found a straightforward genuine common predicate, as opposed to the suspect type (" entity" etc.) which indicatesa
CATEGORY DIFFERENCES
267
category difference. To use our example again, some might supply for "reading" and " seeing" the common predicate
" processes or " activities ", taking this as a genuine common
a predicatedemarcating common field, while others would say that to do this was to obliterate a category difference and generate spurious problems. At any rate it looks as though here too, as in the case of Ryle's sentence-frames, there is a about this apparentlysimple, immediatetest. certain difficulty I do not howeverwish to discussthis furtherhere, but to come to my secondmain point from Thompson's paper. The way of talkingthat we havejust been discussing leads to the traditional problemsabout " existence", and these problems have provedintractable. The questionthus ariseswhetherthere is something and wrongwiththe problems themselves, Thompson discusses several modern attempts to avoid them. I confine myselfto one of these attempts,whichseemsdirectlyrelevantto our earlierdiscussion,and to threepoints in it. (a) First, there is the attemptitself. It beginsfrom the view that the errorin the view traditional positionarisesfromanover-simplified of language, namelythe view that all commonnouns (" entity" etc. of course included)are names of kinds of things, and differences the in use of words must be explainedby a difference the things to in which the words apply. Positively,it arguesinstead,-and here, I since Thompsonputs the position admirably, use his words" As with most adjectives,verbs, adverbs,and prepositions,we explain the meaningof common nouns by exhibitingdiscourse in which the words are used far more than we do by exhibiting things to which the words apply.... One thus learns to ask sensiblequestionsaboutanimalsandinstitutions to recognize and the absurdityof seekingtheir common genus even though one has neverthought of tryingto expressthe latterpoint by saying that ' animal' and 'institution' apply to different kinds of entities. Such a statement would add nothing to what one already knows about the use of ' animal' and 'institution'." And he later adds, still of coursesimplyputtingthe case for this view, " once we have got rid of the principlethat a difference in the use of commonnouns is alwaysdue to a difference the in entities signified, traditional problems about the senses of ' existence' are of course no longer to be taken as theoretical.
268
PROFESSORR. C. CROSS
While it is true that when we say ' Animals and institutions exist' we are to some extent using ' existence ' in different senses, this fact is not subject matter for a theory distinct from theories about animals and institutions."'16 (b) Thompson himself is unable to accept this as all that need be said. He thinks there also comes a point at which we have to say in addition, if the talk is for instance of institutions, that these are abstract entities. We must know this as part of the minimal knowledge requiredfor the correct use of " institutions ", i.e., we must have this minimal theoretical information about the objects of discourse. In this way we are back to the entity language, which we have seen is part of the traditional view of categories, modes of existence, and, though Thompson himself displays much wariness about this, of ontology. Here again there are many interestingpoints in what Thompson says on which one would like to comment-for example his handling of the question, " Why is it legitimate to operate only thus and so with ' number ' and ' institution '?"17 And again, against Thomspon's point that we require some minimal information about the subject matter in addition to what we collect from examining the logic of our discourse, one might follow the line suggested by Waismann in " Language Strata "18 when he says " It was hitherto the custom to refer to what I have called ' strata ' by indicating their subject matter ... What I now suggest we do ... is to reverse the whole situation.... If we carefully study the texture of the concepts which occur in a given stratum, the logic of its propositions, the meaning of truth, the web of verification, the senses in which a description may be complete or incomplete-if we consider all that, we may thereby characterize the subject matter." But detailed discussion would not be appropriate here, and for present purposes I proceed to my third point, in which I agree with Thompson. (c) He observes that by the attempt to dispose of the traditional problems that we have just been discussing " we not only dispose of the traditional problems as theoretical problems about kinds of entities, but we also get rid of the idea of category
Loc. cit. pp. 498 and 499. cit. p. 502. 18 Logic and Language (Second Series),p. 30.
16
17 Loc.
CATEGORY DIFFERENCES
269
differences ".19 This arises in the following way. The view recommendsthat we keep clear of ontological investigations, and that we explain the meaning of words by exhibitingtheir use in discourse,i.e., it arguesthat we should confine ourselves to logical investigations,in a broad sense of logical. Now in exhibitingand examining uses of wordsin this way, we shall the find some wordswhichin certainaspectsof theiruse differfrom other words in a way which the traditionalview puts by saying that the one group refers to abstract, the other to concrete entities. If we like, we can still use this idiom, providedwe are clear what we mean, and in particularprovidedwe recognize that in so doing we are indicatingonly one amongmanypossibly importantdifferences use. On the traditionalview the uses in reflected different categories of being, and the categories of being themselvesprovideda limited list. They also could, as studiedin ontology, providea theory of why the uses of words were limited as they were. On the new view howeverthere is no separateontologicalinvestigationof subjectmatter,and the fact that words like, for example, " fox " and " number are ' never used to distinguishspecies of a common genus, might be one of many possibly important differencesbetween them. Thompsonsums up the position neatly when he says " In sum, ' a phraselike 'category difference is neededonly in the context of a theory about differences the use of words, and yet the in impossibilityof any such theory is precisely the point being urged against the traditionalpositionr."20 Now the view,directedagainstthe traditional Aristotelianism, which we have just been discussing bears a considerable resemblance the actualpracticeof certainmodernphilosophers to like Ryle, who nevertheless describeshis practicein the idiom of " categories "-detecting " categorymistakes", showingthat " conceptsare of " different categories and so on. But if this is correct, the corollarywould seem to be that the " category" idiom is misleading. Of course other philosophers besides Aristotlehave used the idiom. Thereis, if one likes, a Kantian traditionas well as an Aristoteliantradition,and there are no trade rightsabout its use. But it does have strongAristotelian
19Loc. cit. p. 500. 20Loc. cit. p. 501.
270
PROFESSORR. C. CROSS
associations. For that reason it is misleading,in so far as it suggestsextra-logicalassociationswith the world as it really is as revealedto ontologicalinvestigationor insight (the second point in our earlierhistoricalexcursuson Aristotelianism); a or limited list of nameablecategories(the third point earlier);or again that as the categoriesare irreducible one another(the to fourthpoint earlier),so the words we use ought alwaysto keep to clearlydefinedpaths. If we can forgetthesesuggestions, is a it mere matterof terminologywhetherwe keep the idiom or not. is Whatis important patience tracingout the actualsubtleties in of language. One shouldnot expect that in their behaviourwords fall and defined exclusive willalways intosharp clearly and patterns, theirpatterns. test orthatthereis someeasyand simple to determine A crossingof patternsmay not alwaysproducenonsense-as a trivial example,while we mightwant for most purposesto disof tinguishthe logical patterns colour words and sound words,a like " loud colour" makes good sense; and some collophrase cationsof this sort can be highly strikingand illuminating. And mistakes" maygenerate againwhile" category myths,it mightbe a couldbeilluminating its own thatsometimes mythso generated in right; that is, it might be that not all " categorymistakes" were mistakes,howevermuch it may be true that most of them are. Ryle himself has noted the connexion between "category mistakes" andjokes (" so manyjokes are in fact type-pranks "); and it is not uninteresting, though no doubt the point cannot be pressed,that Koestlerin his book Insight Outlook and attempts to account not only for the comic, but also for discoveryand invention, by what he calls the " bisociation" of two fields of experiencewhich had not been connected togetherbefore.
