UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Wittgenstein on universals Fahrnkopf, Robert 1973

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-UBC_1973_A1 F33_7.pdf [ 15.88MB ]
Metadata
JSON: 831-1.0101096.json
JSON-LD: 831-1.0101096-ld.json
RDF/XML (Pretty): 831-1.0101096-rdf.xml
RDF/JSON: 831-1.0101096-rdf.json
Turtle: 831-1.0101096-turtle.txt
N-Triples: 831-1.0101096-rdf-ntriples.txt
Original Record: 831-1.0101096-source.json
Full Text
831-1.0101096-fulltext.txt
Citation
831-1.0101096.ris

Full Text

o-' WITTGENSTEIN ON UNIVERSAIS by ROBERT FAHRNKOPP B.A., Stanford Un i v e r s i t y , 1965 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in the Department of Philosophy We aocept t h i s thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA Ju l y , 1973 In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f the r e q u i r e m e n t s f o r an advanced degree a t the U n i v e r s i t y o f B r i t i s h Columbia, I agree t h a t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p urposes may be g r a n t e d by the Head o f my Department o r by h i s r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l n o t be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . Department o f Philosophy The U n i v e r s i t y o f B r i t i s h Columbia Vancouver 8, Canada Date ALMOST 7; /97S i i ABSTRACT In th i s thesis I dispute the commonly held opinion that, both i n his e a r l i e r and l a t e r philosophy, Wittgenstein opposes the view that there are such e n t i t i e s as universals. I contend, on the oontrary, that the early Wittgenstein inoludes properties and r e l a t i o n s , as well as p a r t i c u l a r s , among simple objects, and that the l a t e r Wittgenstein—though repudiating the Platonic ver-sion of realism found i n the T r a o t a t u s—imp l i c i t l y retains a moderate A r i s t o t e l i a n realism i n which at least s p e c i f i c shades of color are held to exist as objective e n t i t i e s i n the i r own r i g h t , though not apart from the-particular things they q u a l i f y . Chapters I through IV concern the early Wittgenstein; Chapters V through VII concern the l a t e r Wittgenstein. In Chapter I, I argue that Wittgenstein's equation of the •objects* of the Traotatus with Russell's, ' i n d i v i d u a l s ' . i n #46 of the Philosophical Investigations, i s compatible with a r e a l -i s t i c Interpretation of the Traotatus. sinoe Russell himself ( u n t i l 1913) included universals among i n d i v i d u a l s . In Chapter I I , I introduoe the d i s t i n c t i o n between inde-pendent and dependent symbols (where the symbolizing r o l e of the l a t t e r , but not the former, requires that they be i n r e l a t i o n to other symbols). With respect to thi s d i s t i n c t i o n , I trace Witt-genstein's views through a three-stage development. His f i r s t p o s i t i o n i s that both p a r t i c u l a r s and universals are represented by independent symbols; his second p o s i t i o n i s that only p a r t i c -i i i u l a r s are so represented, and that symbols which ostensibly stand for universals are, i n v i r t u e of their dependence on other symbols, not r e a l l y names of objects, and instead c o n t r i -bute only to a symbolization of the form of the proposition i n which they occur; his th i r d position—formulated f i r s t i n the Notebooks and retained i n the Traotatus—recognizes that every symbol i n a proposition has meaning only i n the context of the prop o s i t i o n , so that both universals and p a r t i c u l a r s are repre-sented by dependent symbols and are therefore e n t i t l e d to equal status as objects. In Chapter I I I , I advance additional textual arguments favoring a r e a l i s t i nterpretation of the Traotatus, and i n Chap-ter IV, I examine and rebut various counter-arguments. In Chapter V, I argue that neither Wittgenstein's family-resemblance doctrine nor his attack on over-simplified con-ceptions of the functions of words shows that he disallows the p o s s i b i l i t y that some words are names of universals; f u r t h e r -more, I argue that there i s evidence that colors are regarded as just such objeots. In Chapter VI, I trace the development of Wittgenstein's l a t e r views to show that his inte r e s t i n the problem of univer-sals oenters almost exclusively around the repudiation of his e a r l i e r Platonism, which helps, to explain his laok of an overt proclamation of h i s adoption of A r i s t o t e l i a n realism, occuring as i t does almost as a casual by-product of thi s repudiation of Platonism. In Chapter VII, I qual i f y my p o s i t i o n by arguing that, for i v the l a t e r Wittgenstein, language determines ontology, rather than the other way around, so that his r e a l i s t ontology amounts only to his playing or describing language-games appropriate to realism, where these language-games are not themselves j u s t i f i e d by appealing to any external standard such as 'the way r e a l i t y i s ' . I include two appendices. In Appendix A, I argue that Wittgenstein's contention that language-games have no ultimate j u s t i f i c a t i o n could have been u t i l i z e d to avoid a misplaced i attack on p r i v a t e objects. In Appendix B, I try to show that Wittgenstein had read Russell's The Problems of Philosophy and Our Knowledge of the External World. I also examine the i n -fluence of Ogden and Richards' The Meaning of Meaning on Witt-genstein's l a t e r work. V TABLE OF CONTENTS FOREWORD . v i i ABBREVIATIONS USED IN THE TEXT i x PART ONE: THE EARLY WITTGENSTEIN INTRODUCTION TO PART ONE 2 I RUSSELL'S 'INDIVIDUALS' . . . 3 II THE THREE STAGES OF WITTGENSTEIN'S DEVELOPMENT . . . 20 III ARGUMENTS FOR A REALISTIC INTERPRETATION 60 IV ARGUMENTS AGAINST A REALISTIC INTERPRETATION . . ... 96 PART TWO: THE LATER WITTGENSTEIN INTRODUCTION TO PART TWO 122 V WITTGENSTEIN'S MODERATE REALISM ;.. 125 VI FROM PLATO TO ARISTOTLE: THE TRANSITION TO THE LATER PHILOSOPHY 167 VII THE PRIMACY OF LANGUAGE .188 v i AFTERWORD . . 211 BIBLIOGRAPHY 212 APPENDIX A 218 APPENDIX B . . . 238 v i i FOREWORD It i s currently fashionable to regard Wittgenstein, both i n the e a r l i e r and l a t e r stages of his philosophical outlook, as being Ill-disposed i n p r i n c i p l e as well as i n pr a c t i c e to the r e a l i s t s o l u t i o n to the problem of univ e r s a l s . The contention of my thesis i s that this fashionable interpretation i s f a l s e . In both his e a r l i e r and l a t e r periods, I s h a l l argue, Wittgenstein includes non-particular items, i . e . , u n i v e r s a l s , In his onto-l o g i o a l inventory. In Part One, I try to make out thi s case with respect to the early Wittgenstein, and i n Part Two, with respect to the l a t e r Wittgenstein. In both p a r t s , I am almost exclusively concerned to give an exposition of his views rather than a c r i t i c a l evaluation of them, and my thesis i s therefore a work of scholarship, a study i n the history of philosophy, rather than a piece of c r e a t i v e , o r i g i n a l philosophy i n i t s own r i g h t . Nevertheless, Wittgenstein i s so often these days ap-pealed to i n support of thi s or that philosophical p o s i t i o n—hi s r o l e as an authority-figure among contemporary Anglo-American philosophers approaching that of A r i s t o t l e among the medievalists— that a thorough and accurate exposition of his views can be a p o s i t i v e and worthwhile contribution to the philosophical enter-p r i s e i t s e l f . Since my purpose i s mainly to make clear what Wittgenstein himself says, I have allowed myself to quote very l i b e r a l l y from his works, often reducing my own function to l i t t l e more than v i i l that of providing a running commentary. In many instances, paraphrases would probably have served just as well as direot quotes, but I oould see no compelling reason why my version of what Wittgenstein says should be preferred to his own version. Translations of passages from works not yet ava i l a b l e i n English ( s p e c i f i c a l l y , from Wittgenstein's Philosophisohe  Bemerkungen and Philosophisohe Grammatik. Waismann's Wittgen- s t e i n und der Wiener K r e l s . and Mueller's Ontologle i n Wittgen- steins 'Traotatus') are my own. In ad d i t i o n , while generally r e l y i n g on the standard translations of Wittgenstein's works where they do e x i s t , I have f e l t free to make occasional, minor modifications, especially as regards the Pears and McGuinness version of the Traotatus. I wish to thank Peter Remnant, Steven S a v i t t , Alan Jjoveland, and Andrew Lev i n e for t h e i r comments and c r i t i c i s m s . I am also g r a t e f u l to the philosophy department of the University of B r i t i s h Columbia, both for f i n a n c i a l support and for the unre-s t r l o t i v e atmosphere i n which I was allowed to carry out this p r o j e c t . I also wish to thank the Canada Council for two years of f i n a n c i a l support. ; F i n a l l y , I am indebted to my wife Pamela for her help and encouragement. ABBREVIATIONS USED IN THE TEXT Works by Russellt LK Logio and Knowledge OKEW Our Knowledge of the External World PMA P r l n o i p l a Mathematloa (with Whitehead) PM P r i n c i p l e s of Mathematics PP The Problems of Philosophy Works by Wittgenstein: BB The Blue Book BRB The Brown Book NB Notebooks. 1914-1916 NM "Notes Dictated to G. E. Moore, A p r i l 1914" NL "Notes on Logic, September 1914" PC On Certainty PI Philosophical Investigations PB Phllosophisohe Bemerkungen PG Phllosophisohe Grammatlk PT Proto-Traotatus T Traotatus Z Z e t t e l Miscellaneous: WWK Friedreich. Waismann, Ludwig Wittgenstein und d Wiener Krels "~ X ET Essays In Wittgenstein's Txactatus, ed. by I. M. Copl and R. W. Beard WPI Wittgensteins The Philosophical Investigations, ed• by George Pitcher I' PAET ONE THE EARLY WITTGENSTEIN 2 INTRODUCTION TO PART ONE Against the generally accepted view that the T does not i n -clude universals among simple objects—which I c a l l the nominal-i s t ^ i n t e r p r e t a t i o n—I prppose i n Part One to defend the r e a l i s t i n t e r p r e t a t i o n , i . e . , the view that the T does indeed include universals as well as p a r t i c u l a r s among objects. I mean to do t h i s , not merely by c i t i n g passages from the T i t s e l f which sup-port my view, but by trying to place the T i n i t s proper h i s t o r -i c a l context. This w i l l include a detailed study of the genesis of Wittgenstein's views, as evidenoed i n his pre-T writings as well as i n the works of Russell and Frege—especially the former— apart from whom the development of Wittgenstein's thought cannot be understood. In add i t i o n , I w i l l c i t e passages from Wittgen-stein's post-T writings which confirm my interpretation of the T. F i n a l l y , I w i l l conclude Part One by examining and rebutting various arguments which have been advanced i n favor of the nom-i n a l i s t Interpretation. -•-My use of the term 'nominalism' requires comment, since we cannot assume that proponents of this interpretation of the T want altogether to deny ontological status to properties and"" r e l a t i o n s , especially to the l a t t e r . To be sure, the sort of view I here c a l l 'nominalistio' denies that r e l a t i o n s exist as  objects t but at the same time i t seems to concede that r e l a t i o n s somehow e x i s t , either by having a kind of content d i f f e r e n t from that possessed by objects, or by way of contributing to the form (or s t r u c t u r e ) , though not the content, of f a c t s . Despite some misgivings, then, I s h a l l continue to use the convenient l a b e l 'nominalism' for thi s sort of p o s i t i o n , mainly because the onto-l o g i c a l implications of this p o s i t i o n are for me s u f f i c i e n t l y unclear that I do not know what else to c a l l i t . (Towards the end of Part One, I again b r i e f l y discuss this issue.) 3 CHAPTER I RUSSELL'S 1INDIVIDUALS1 I s h a l l begin by considering a point which seems to under-mine my in t e r p r e t a t i o n of the T from two d i f f e r e n t directions at once. In the introduction I stated both that an examination of Wittgenstein's r e l a t i o n to Russell and an examination of Wittgen-stein's own writings subsequent to the T w i l l oonfirm my view. Yet, the PI contains a passage which seems to go against my inte r p r e t a t i o n of the T, and indeed does so i n vi r t u e of Witt-genstein's comparison of his own e a r l i e r view with that of R u s s e l l . The passage i s as follows: What l i e s behind the idea that names r e a l l y s i g n i f y simples? —Socrates says i n the Theaetetus: "If I make no mistake, I have heard some people say t h i s : there i s no d e f i n i t i o n of the primary elements—so to speak—out of which we and everything else are composed; for everything that exists i n i t s own r i g h t can only "be named . . . " Both Russell's ' i n d i v i d u a l s ' and my 'objects' (Traotatus) were such primary elements (PI, #46). At least two commentators, Copi and Mueller, c i t e t h i s pas-sage to show that the early Wittgenstein considered only p a r t i c -ulars to be objects, since for Russell only p a r t i c u l a r s were i n d i v i d u a l s . I t seems, then, that we have i n effeot an admission from Wittgenstein himself that the nature of Russell's influence i n t h i s matter consisted i n Wittgenstein's adoption of just so much of Russell's ontology as to constitute a nominalistio p o s i t i o n . (I say 'just so much', beoause Russell was always w i l l i n g to include universals i n his own ontology.) This con-clusion i s soon rendered doubtful, however, when we consider the following: i n Moore's notes on Wittgenstein's lectures of 1930-4 1933 he t e l l s us that Wittgenstein spoke of colors as well as atoms, as i f they were * i n d i v i d u a l s ' i n Russell's sense.1 Since a color i s presumably not a p a r t i c u l a r , but a u n i v e r s a l , we can-not conclude from Wittgenstein's comparison of his 'objects' with Russell's ' i n d i v i d u a l s ' that Wittgenstein thought only of p a r t i c u l a r s as objects. This i n t e r p r e t a t i o n of what Russell means by an i n d i v i d u a l i s so unusual that we might be tempted to suppose that Wittgenstein i s here ta l k i n g about an instance of color rather than the color I t s e l f , i . e . , a p a r t i c u l a r red sense-datum rather than the universal redness.2 Such a strategy, though, can soaroely explain away Wittgenstein's remark i n the BB, where, i n the course of characterizing his e a r l i e r view of the T i n which he had held that i t was only the arrangement of elements, and not the elements themselves, which need not e x i s t , he says: . . . the elements, the individuals must e x i s t . I f redness, roundness, and sweetness did not e x i s t , we could not imagine them (BB, p. 31). Here i t i s c l e a r l y the universal redness—not a p a r t i c u l a r instance of a red sense-daturn—that i s called an i n d i v i d u a l , and thus Wittgenstein's comparison of his objects with Russell's i n d i v i d u a l s , far from supporting the contention that Wittgenstein conceived of his objects as p a r t i c u l a r s , seems rather to provide support, i n conjunction with Moore's report and the statement i n the BB, for the view that objects include u n i v e r s a l s . ^G. E. Moore, "Wittgenstein's Lectures i n 1930-33*" in. Philosophical Papers, p. 291. 2por t h i s d i s t i n c t i o n , see Russell's "On the Relation of Universals and P a r t i c u l a r s , " i n LK, p. 111. 5 But i f Wittgenstein does regard universals as i n d i v i d u a l s , what sense can we make of the puzzling s i t u a t i o n i n which a philosopher of his competence attr i b u t e s this view to Russell? After a l l , as Mueller says, Russell's individuals • • .are i d e n t i f i e d by their inventor with 'particulars* and circumspectly enough defended from every suspicion of u n i -v e r s a l i t y . . •1 whereupon he quotes Russell himself as sayings We s h a l l further define, ' i n d i v i d u a l s ' or 'particulars* as the objects that can be named by proper names.2 It would seem that I am driven to the unoomfortable a l t e r -natives of admitting that PI #46 shows either (a) that the T i s n o m i n a l i s t i o , or (b) that Wittgenstein i s unable to understand an u t t e r l y elementary view of one of the few philosophers whose works he had d i l i g e n t l y studied. Most philosophers, judging from the currently fashionable int e r p r e t a t i o n of the T, would have no qualms i n admitting (a), and some might take pleasure i n opting for (b), or even both together. Since I would l i k e to preserve Wittgenstein's reputation as well as a r e a l i s t i c i n t e r -pretation of the T, I would prefer a th i r d a l t e r n a t i v e ; f u r t h e r -more, I think one oan be found. The solution I propose w i l l not only guard Wittgenstein's reputation against the accusation of his having oommitted a gross blunder, but ought to enhance his reputation by showing that he surpasses his commentators i n his understanding of R u s s e l l . B r i e f l y , the solution i s that Wittgenstein correctly saw lAnselm Mueller, Ontologle i n Wittgensteins Traotatus, p. 81. 2Mueller, p. 81, quoting Bertrand R u s s e l l , Introduction to  Mathematical Philosophy, p. 142. 6 that Russell's p o s i t i o n as of the year 1912, when Wittgenstein commenced studying under him, was one which held both p a r t i c -u lars and universals to be Individuals. Russell's r e s t r i c t i o n of the expression 'Individual' to p a r t i c u l a r s was only a sub-sequent development, one Indeed for which Wittgenstein was responsible, though a r e s t r i c t i o n which, for other reasons, Wittgenstein himself ultimately saw no need to accept. Therefore, when Wittgenstein compares the objects of the T to Russell's i n d i v i d u a l s , he i s speaking of Russell's-position as i t stood from PM to PP—rthe Russell who had provided the starting-point for his own thought—and not Russell's p o s i t i o n as of the In t r o -duction to Mathematical Philosophy, written i n 1919, whose r e s t r i c t i o n of the expression ' i n d i v i d u a l ' to p a r t i c u l a r s , as quoted by Mueller, was a r e s u l t of Wittgenstein's work and not an i n s p i r a t i o n of i t . I proceed, therefore, to an examination of the considerations underlying Russell's use of the term ' i n d i v i d u a l ' , including the part played by Wittgenstein i n changing Russell's p o s i t i o n . This aooount, i n turn, w i l l set the stage for an account of the development of Wittgenstein's own views. Russell's pre-Wlttgensteinian account of what he means by an Individual i s summed up n i c e l y i n a paper of 1908, i n which he says: In an elementary proposition we can distinguish one or more terms from one or more oonoepts; the terms are whatever can be regarded as the subject of the proposition, while the oonoepts are the predicates or r e l a t i o n s asserted of these terms. The terms of elementary propositions we w i l l c a l l 7 i n d i v i d u a l s . ^ At f i r s t glance Russell indeed appears to be saying that only p a r t i c u l a r s are i n d i v i d u a l s , since i t would seem that the subjeot of a proposition must be a p a r t i c u l a r . But are we r e a l l y debarred from making a universal the subjeot of a proposition? Russell's own footnote following the f i r s t sentenoe of the above quote indicates that he feels we are not so debarred. This footnote, which i s presumably intended to throw further l i g h t on the d i s -t i n c t i o n between terms and concepts, r e f e r s us to #48 of PM, where Russell informs us that concepts have a curious two-fold use which allows them to function not only i n the i r predicative or r e l a t i o n a l oapacity, but also as terms, though not, of course, i n the same proposition. So, e.g., the two propositions 'Socrates i s human' and 'Humanity belongs to Socrates' are equivalent, but nevertheless d i s t i n c t , since the f i r s t proposition makes an assertion about Socrates while the second makes an assertion about humanity. In other words, the difference between the two propositions i s that the concept which functions i n the f i r s t proposition as a concept. functions i n the second proposition as a term.2 Things—i.e., p a r t i c u l a r s—ar e opposed to conoepts— i . e . , u n i v e r s a l s—in that the former can occur i n propositions only as subjects, whereas the l a t t e r can occur pre d i c a t i v e l y or ^-Bertrand R u s s e l l , "Mathematical l o g i c as based on the Theory of Types," i n LK, p. 76. 2I n #49 of PM Russell makes clear that the expressions 'human' and 'humanity', are meant to designate the same e n t i t y , and that t h e i r grammatical difference indicates only a difference i n external r e l a t i o n s between this entity and the other c o n s t i t -uents of the respective propositions. 8 r e l a t i o n a l l y i n addition to occurring as subjects, so that while a l l p a r t i c u l a r s are i n d i v i d u a l s , not a l l Individuals are p a r t i c u l a r s . Since Russell's paper of 1908 contains e s s e n t i a l l y the same account of the theory of types as presented i n volume I of PMA, which was .published just two years l a t e r , we have prima f a c i e grounds for believing that his view about individuals would not have changed In the meantime. The text of PMA i t s e l f i s unln-formative on t h i s Issue, where i t says merely that individuals are those ". • . objects which are neither propositions nor funotions . . . "^ and that an i n d i v i d u a l i s "• • • something which exists on i t s own account . . . "2 Nothing i n these s t a t e -ments should lead us to suppose that universals could not be i n d i v i d u a l s . S t i l l , i t i s so oommonly assumed that the i n d i v i d -uals spoken of i n PMA are p a r t i c u l a r s , that we would do well to consider the reasons for t h i s assumption.3 I think that there are c h i e f l y two: f i r s t , there are Russell's statements of 1PMA, p. 51. 2PMA, p. 162. ^Keyt, for example, takes t h i s for granted as t o t a l l y obvious In arguing that Wittgenstein's objects must be p a r t i c -u lars since his notation corresponds to Russell's. Thus: "• • . t h e r e i s Wittgenstein's notation i n which small Latin l e t t e r s from the beginning of the alphabet are used for names and from the end for name variables . • . which i s the Russellian way of denoting a p a r t i c u l a r . . • " (David Keyt, "Wittgenstein's Notion of an Object," i n ET, p. 291). Presumably Keyt i s thinking here of PMA, p. 51» which says: ". • . we w i l l use such l e t t e r s as a, b, c, x, y, z, w, to denote . . . i n d i v i d u a l s . " (Inciden-t a l l y , even i f Russell's use of these l e t t e r s i s not confined to representing p a r t i c u l a r s , Keyt's claim that t h i s i s Wittgen-stein's use of these l e t t e r s i s correct. Keyt's mistake, as we s h a l l see l a t e r , i s to suppose that only such l e t t e r s are used for names.) 9 l a t e r years, such as that previously quoted by Mueller, which may possibly influenoe one's interpretation of PMA; secondly, there i s an obscurity i n the exposition of the text i t s e l f , i nso-f a r as i t f a i l s to stress the difference between functions and oonoepts, where I mean by the l a t t e r , those e n t i t i e s having predioative and r e l a t i o n a l c a p a c i t i e s , i . e . , universals. I f we take an elementary proposition of the subject-predicate form, e.g., 'This i s red', and symbolize i t a f t e r the fashion of PMA as ' f ( a ) ' , where ' f ' and 'a' are both oonstants, we might be incli n e d to suppose that argument and function here correspond to term and concept. That i s , ' f ( a ) ' expresses i n l o g i c a l sym-bolism the same proposition which i s expressed i n words as 'This i s red', and sinoe 'a' stands for the same entity as i s desig-nated by the proper name ' t h i s ' , therefore 'f* must stand f o r the entity designated by the name 'red'.-1- Then, given that funotions are of a d i f f e r e n t type than Individuals, i t follows that oonoepts cannot be i n d i v i d u a l s , sinoe they correspond to the funotional part of propositions. Understandable as t h i s view i s , i t i s mistaken. We cannot equate concepts with funotions, because oonoepts are, but functions are not, constituents of elementary propositions, so there i s no suoh thing as the 'functional part* of a proposition. 3-In my discussion of Russell I w i l l assume that a propo-s i t i o n suoh as 'This i s red' i s t r u l y of the subject-predicate form, i . e . , that i t predicates a property of a thing, and i s not r e a l l y a r e l a t i o n a l proposition i n which ' i s ' serves to r e l a t e ' t h i s ' and 'red' as two terms. In other words, I w i l l assume i n t h i s discussion of PMA that a proposition such as 'This i s r e d ' has only two constituents. 10 That i s , the function i n ' f ( a ) ' i s not 'f* s I m p l i c i t e r . "but ' f ( x ) ' , where 'x' i s a variable standing for any argument Capable of s i g n i f i c a n t l y combining with the function to produce a true or f a l s e proposition, and so, as Russell and Whitehead note, I t should be remembered that a function i s not a c o n s t i t -uent i n one of i t s values; thus for example the function 'x i s human1 i s not a constituent of the proposition 'Socrates i s human'. 1 Likewise, i n the proposition 'This i s red', 'this* and 'red' — but not the function 'x i s red'—designate constituents. Regretably, Russell and Whitehead confuse the issue by speaking at one point as i f ' f ' and 'a' each stood independently for d i f f e r e n t constituents i n an elementary proposition. They say that • • whatever function si may be, there w i l l be argu-ments x with which j6x i s meaningless • • • , , t 2 a manner of speaking which misleadingly suggests that the function i s sym-bolized as V s i m p l l o i t e r . Shortly thereafter, they assert that . . . propositions which contain no apparent variables we c a l l elementary propositions, and the terms of such propositions, other than functions, we c a l l individuals.3 In the system of PMA i t i s of course possible for a function to be a term of a proposition by being an argument of that propo-s i t i o n , but the proposition w i l l then not be elementary, but of at least the seoond-order. The only function which can concern an elementary proposition i s a function of the proposition I t s e l f , but, as we have seen, a function cannot be a constituent i n one of i t s values, so no term of an elementary proposition IPMA, pp. 54-55. 2PMA. p. 161. 3PMA, p. 161. 11 can be a function. As i s elsewhere e x p l i c i t l y stated i n PMA, elementary propositions are those which not only contain no apparent v a r i a b l e s , but are those ". . . which contain no funotions • • • . T h e suggestion, then, that an elementary proposition can have two kinds of terms,2 v i z . , functions and i n d i v i d u a l s , can only be a careless s l i p , but one which leads the unwary reader closer to the assimilation of functions and concepts. The lack of ooncern i n PMA to d i s t i n g u i s h funotions from concepts probably stems from the fact that the theory of types can be explicated e n t i r e l y within the gramework of function and argument, so that the r e l a t i o n between funotions and those pecu-l i a r e n t i t i e s which are the predicative and r e l a t i o n a l c o n s t i t -uents of propositions can for p r a c t i c a l purposes be ignored. S t i l l , an account of the r e l a t i o n between functions and concepts from the standpoint of PMA would be desirable, though i t i s not my purpose to f i l l t h i s void here, other than to say that con-cepts are somehow included i n or implied by funotions, although not i d e n t i c a l with them. So, e.g., the function *x i s red' has something to do with the concept'redness, even though i t i s not i d e n t i c a l with t h i s concept. For our purposes, the important point i s that the difference between funotions and concepts 1PMA, p. 54. 2Here 'term' can only be taken i n the broad sense of 'constituent', since a funotion could obviously not be a term of an elementary proposition i n the narrower sense of ' l o g i c a l subject', though even i n the wider sense, the i n a b i l i t y of a funotion to be a term i n an elementary proposition i s obvious enough. 12 shows that we cannot conclude, from the difference i n type between functions and i n d i v i d u a l s , that there must be a d i f f e r -ence i n type between concepts and i n d i v i d u a l s . Therefore, the p o s s i b i l i t y that concepts can be Individuals remains open, and with no other evidence to suggest the contrary, we may j u s t i -f i a b l y conclude that at the time of the writing of PMA Russell s t i l l holds that concepts are included among i n d i v i d u a l s . Fortunately, we are not confined to mere speculation i n t h i s matter, since R u s s e l l , i n other writings contemporary with and subsequent to the publication of volume I of PMA i n 1910, makes i t clear that he regards concepts as capable of being l o g i c a l subjects of propositions. For example, i n Chapter VII of Philosophical Essays, written o r i g i n a l l y i n 1906 but revised fo r publication i n 1910, Russell propounds his theory of judge-ment, i n which one's judgement of the truth of a proposition i s not a r e l a t i o n of two terms—one's mind and the proposition (the l a t t e r considered as a complex object)—but rather a multiple r e l a t i o n between the mind and the constituents of the proposition, where a l l the constituents are regarded as terms of the judgement-r e l a t i o n . Thus, assuming that a l l propositions are r e l a t i o n a l (an assumption which Russell makes i n t h i s paper, though subse-quently r e t r a c t s ) , Every judgement i s a r e l a t i o n of a mind to several objects, one of which i s a r e l a t i o n ; the judgement i s true when the r e l a t i o n which i s one of the objects r e l a t e s the other objects, otherwise i t i s false.1 What thi s means i s that the r e l a t i o n l o v i n g , which funotions ^•Russell, Philosophioal Essays, p. 156. 13 r e l a t i o n a l l y i n the proposition 'John loves Mary'—as opposed to John and Mary, which are terms—ceases to function r e l a t i o n a l l y i n the proposition 'George judges that John loves Mary' and instead, beoomes a term, on a par with John and Mary, for the new r e l a t i o n of judging. As Russell expresses th i s point two years l a t e r , i n PP, i n connection with believing—and the same point oould be made for any other 'propositional a t t i t u d e ' such as hoping, f e a r i n g , wishing, etc.—the proposition 'Othello believes that Desdemona loves Cassio' contains the r e l a t i o n 'loving', not as the r e l a t i o n which creates the unity of the complex whole, but as ". . . one of the o b j e c t s—it Is a brick i n the structure, not the cement. The cement i s the r e l a t i o n 'be-l i e v i n g ' . "1 This shows that, as l a t e as 1912, Russell s t i l l regarded u n i v e r s a l s2 as having that curious two-fold use spoken of i n PM. So l i t t l e , i n f a c t , did he question the p o s s i b i l i t y of universals being i n d i v i d u a l s , that he seriously entertained the p o s s i b i l i t y , i n the paper "On the Relation of Universals and P a r t i c u l a r s " of 1911, that there were no such objects as p a r t i c u l a r s at a l l , so •"-PP, p. 128. 2I say 'universals' rather than 'concepts', sinoe by 1911 Russell recognized that the d i s t i n c t i o n he wanted to make between things, i . e . , p a r t i c u l a r s , and those e n t i t i e s capable of functioning predioatively and r e l a t i o n a l l y was more general than that between things and oonoepts. (See his paper "On the Relation of Universals and P a r t i c u l a r s , " i n IK, p. 106.) The expression 'concept' suggested to Russell that the entity must be the object of a oognitive a c t , and he wanted the d i s t i n c t i o n to extend to e n t i t i e s whioh might not happen to be related to cognitive a c t s , so the basic d i s t i n c t i o n became expressed as that between p a r t i c u l a r s and universals, a terminology which I w i l l likewise adopt i n the remainder of thi s discussion of Russell's p o s i t i o n . 14 that what we normally think of as 'things' are r e a l l y bundles of q u a l i t i e s and r e l a t i o n s . He went on to r e j e c t this view i n favor of the view that p a r t i c u l a r s and universals are two i r r e -duoibly d i f f e r e n t kinds of e n t i t i e s , but that he could even have considered the f i r s t view shows that he had no qualms about thinking of universals as i n d i v i d u a l s . I t i s i n s t r u c t i v e that his f i n a l characterization of the difference between p a r t i c u l a r s and u n i v e r s a l s , i n t h i s paper, i s that p a r t i c u l a r s only occur as subjects of predicates and as terms of r e l a t i o n s , while univer-sals can occur p r e d i c a t i v e l y and r e l a t i o n a l l y , whioh implies that universals oan also occur as subjects.1 E a r l i e r on, t h i s implication had been made e x p l i c i t , when Russell stated that the question—whioh he went on to answer affirmatively—was . . . whether non-relations are of two kinds, subjects and predicates, or rather terms which can only be subjects and terms which may be either subjects or predicates.2 So, at l e a s t through the publication of PP i n 1912, Russell included universals among i n d i v i d u a l s . Since i t i s equally clear that Russell subsequently affirmed that only p a r t i c u l a r s were i n d i v i d u a l s , we must inquire into the reason for t h i s change. Russell provides us with the answer i n his lectures on "The Philosophy of Logical Atomism" of 1918. Concerning the c r u c i a l passage, whioh i s over a page i n length, I here excerpt the g i s t : I pass on from p a r t i c u l a r s to predicates and r e l a t i o n s • • • A very great deal of what I am saying i n this course of lectures •'•Russell, "Universals and P a r t i c u l a r s , " p. 124. 2R u s s e l l , "Universals and P a r t i c u l a r s , " p, 109. 15 consists of ideas which I derived from my friend Wittgenstein, But I have had no opportunity of knowing how far his ideas have changed since August 1914. , . • Understanding a predicate i s quite a d i f f e r e n t thing from understanding a name . . . To understand a name you must be acquainted with the p a r t i c u l a r of which i t i s a name, and you must know that i t i s the name of that p a r t i c u l a r . You do not, that i s to say, have any suggestion of the form of a proposition, whereas i n understanding a predicate you do. . . . The importance of that i s i n connection with the theory of types . . . I t i s i n the fact that a predioate can never ooour except as a predioate. When i t seems to occur as a subject, the phrase wants amplifying and explaining, unless, of course, you are talking about the word i t s e l f . . • • Exactly the same thing applies to r e l a t i o n s . . . . A r e l a t i o n can never occur exoept as a r e l a t i o n , never as a subject. • . .The d i f -ferent sorts of words, i n f a c t , have d i f f e r e n t sorts of uses and must be kept always- to the r i g h t use and not to the wrong use, and i t i s f a l l a c i e s a r i s i n g from putting symbols to wrong uses that lead to the contradictions concerned with types.1 Prom t h i s passage we gather that i t was the influenoe of Wittgenstein, sometime before August 1914, that was responsible for bringing Russell to the view that only p a r t i c u l a r s could be the l o g i c a l subjeots of propositions. In f a c t , we can locate the date of Russell's conversion as being sometime before March 1914, since by t h i s time these new views had become subtly mani-f e s t i n the Lowell Lectures on OKEW. In these leotures, Russell does not proclaim outright'that his views have changed, but he meticulously avoids saying anything that would suggest that u n i -versals are i n d i v i d u a l s , almost as i f he feels that Wittgenstein must be r i g h t , but i s s u f f i c i e n t l y unsure about th i s new devel-opment—not yet being able f u l l y to assess the consequences for his own philosophy—so as to prefer to underplay the whole matter u n t i l he can get his own p o s i t i o n sorted out. So, e.g., Russell does not e x p l i c i t l y say that r e l a t i o n s cannot be l o g i c a l subjects, l R u s s e l l , "The Philosophy of Logical Atomism," i n LK, pp. 205-206. 16 but he does allow that When we say there are re l a t i o n s of more than two things, we mean there are single facts consisting i n a r e l a t i o n and more than two things • • . j1 and generally he i s careful to speak of properties and re l a t i o n s only i n connection with th e i r occurrence i n f a c t s , i . e . , where they are said to be predicated o f , or to r e l a t e , things. Also, Russell's views concerning our knowledge of general propositions have changed, from the po s i t i o n advocated i n PP to that whioh we find i n OKEW, which i s what we would expect, on the supposition that he oould no longer regard universals as l o g i c a l subjects. In PP, general propositions are thought to have only universals as constituents. In a proposition of l o g i c or arithmetic, the connection between universals i s a p r i o r i ; an empirical generalization, such as ' A l l men are mortal', i s also viewed as asserting a r e l a t i o n between the universals man and mort a l i t y , though i n this case the connection oan only be established i n d u c t i v e l y , by an examination of p a r t i c u l a r cases.2 In OKEW. Russell's account of general propositions scrupulously avoids any suggestion that universals could be l o g i c a l subjects. Regarding the a p r i o r i propositions of l o g i c , they are now seen as absolutely general, applying to a l l things and a l l properties, and thus true s o l e l y i n vi r t u e of thei r l o g i c a l form, sinoe they contain no constants.-^ Russell cannot analyze empirical general-1OKEW, p. 47. 2pP, pp. 105-106. ^They do not even include l o g i c a l constants, since Russell now follows Wittgenstein i n saying that a l l the apparent objects of l o g i c and mathematics, the so-called ' l o g i c a l constants', are not r e a l l y e n t i t i e s (OKEW. p. 162). "• 17 Izations i n the, same way, as their truth depends on more than just l o g i c a l form. He therefore considers the p o s s i b i l i t y of establishing the truth of ' A l l men are mortal' by examining atomic f a c t s , expressed by elementary propositions suoh as 'Jones i s mortal', 'Smith i s mortal', e t c . , which seems his only option i f he can no longer regard the examination of the connection between the universals man and mortality themselves as r e f l e c t i n g a proper analysis of the general proposition. He concludes, however, that a general proposition can never be equivalent to the l o g i c a l product of elementary propositions, since we need to know of these elementary propositions that they are a l l the true propositions (relevant to our inquiry) that there are, knowledge which i s i t s e l f expressible only by a general p r o p o s i t i o n .1 Knowledge of the truth of any general pr o p o s i t i o n , then, cannot be based solely on empirical evidence, since our knowledge of any p a r t i c u l a r atomic fact or facts can-not by i t s e l f show us that these are a l l the f a c t s , and Russell ends on the skeptioal note that he does not know whether, outside of l o g i c , there i s any general knowledge at a l l .2 Since this skeptioal conclusion does not r e s u l t when general propositions are viewed as involving no p a r t i c u l a r propositions i n t h e i r analysis and instead as concerning only r e l a t i o n s between u n i -v e r s a l s—t h i s being the way such propositions were viewed i n P P ~ Russell's acceptance of the skeptical r e s u l t shows his deter-mination to avoid at a l l costs the employment of universals as iQKEW. p. 50. 2OKEW. p. 51. 18 l o g i c a l subjects. We may safely conclude, then, that Russell's new view on uni v e r s a l s , which removes them from the realm of i n d i v i d u a l s , i s established by the time of the Lowell Lectures, though not openly advertised u n t i l the lectures on "The P h i l o -sophy of Logical Atomism" some four years l a t e r , where Russell for the f i r s t time e x p l i c i t l y equates individuals with p a r t i c -u l a r s . The part played by Wittgenstein i n Russell's conversion can be indicated i n terms of a rough chronological c l a s s i f i c a t i o n of the stages of Wittgenstein's development. Wittgenstein's views on the nature of objeots, I s h a l l argue, develop through three main stages, which we can c l a s s i f y aooording to the i r chrono-l o g i c a l order as: (1) from the summer of 1912 to January of 1913; (2) from January of 1913 to the l a t t e r part of 1914, from whence there i s a gradual t r a n s i t i o n to (3)t the mature view of the early Wittgenstein, whioh he continues to hold from the end of 1914 u n t i l sometime i n 1929 a f t e r the writing of "Some Remarks on Logical Form." A l l we know about (1) i s to be found, c r y p t i o a l l y expressed, i n a few l e t t e r s from Wittgenstein to R u s s e l l , and of these, only one of whioh i s r e a l l y h e l p f u l , v i z . , the l e t t e r of January, 1913* (2) f i r s t appears i n t h i s same l e t t e r and i s worked out i n d e t a i l i n the NL of September, 1913» and i n the NM of A p r i l , 1913. The sources for (3) are the NB of 1914-1916 and subsequent works. The change i n Russell's view of the status of unive r s a l s , i m p l i -c i t i n OKEW and e x p l i c i t i n the lectures of 1918, i s based on (2), or rather on Russell's faulty understanding of (2), since 19 he continued to suppose—what Wittgenstein had held i n (1) but repudiated i n ( 2 ) —that a theory of types was required to d i s -tinguish the syntaotio r o l e s of universals and p a r t i c u l a r s .1 I t i s unlikely that Russell was much influenced by (1), a p o s i t i o n which Wittgenstein does not s p e l l out i n the l e t t e r s at great length and which, i n f a o t , receives i t s only clear statement i n the l e t t e r i n which i t i s renounced i n favor of ( 2 ) . Concerning (3) , Russell could have known nothing u n t i l June of 1919» when he received a copy of the T, but i n any oase he never gives any in d i c a t i o n of being aware that there i s such a p o s i t i o n as (3) , though his ignorance i n thi s matter puts him i n the good company of almost everyone e l s e . Having said t h i s , I end the discussion of Wittgenstein's influence on Russell and turn to an examination of Wittgenstein's views themselves. iMore s p e c i f i c a l l y , OKEW was based on NL, while Russell presumably had the additional benefit of NM by the time of the 1918 leotures. 20 CHAPTER II THE THREE STAGES OF WITTGENSTEIN'S DEVELOPMENT . Wittgenstein's three positions can be distinguished by th e i r d i f f e r e n t attitudes towards the r e l a t i o n of p a r t i c u l a r s to u n i -versals with respect to a d i s t i n c t i o n which I s h a l l c a l l that between dependent and independent symbols. This d i s t i n c t i o n roughly corresponds to Russell's d i s t i n c t i o n between proper names, which have meaning i n i s o l a t i o n , and incomplete symbols, which have meaning only i n certai n contexts. The difference between Russell's d i s t i n c t i o n and mine i s that whereas incomplete symbols are mere technical devioes which are supposed to d i s -appear upon analysis and so do not r e a l l y symbolize anything at a l l , I do not wish to deny that dependent symbols have a genuine, indispensable, symbolizing r o l e . They d i f f e r from independent symbols i n that t h e i r symbolizing r o l e requires that they be i n r e l a t i o n to other symbols, a r e l a t i o n which for our purposes i n t h i s discussion w i l l always involve th e i r occurrence within the context of a proposition (or at least within the schema of a proposition, where variables or blank argument-places take the place of one or more oonstants). Given t h i s d i s t i n c t i o n between dependent and independent symbols, which w i l l hopefully become d e a r e r as we proceed, the development of Wittgenstein's thought oan be characterized as follows: (1) views both p a r t i c u l a r s and universals as represented by independent symbols, a l l such symbols therefore being names of 21 objects, (2) views only p a r t i c u l a r s as being represented by independent symbols and sees this as a reason for supposing that the symbols which ostensibly stand for un i v e r s a l s , since they are dependent symbols, are not r e a l l y names of objects, and that they instead contribute only to a symbolization of the form of the propositions i n which they occur,1 (3) recognizes "that every symbol i n a proposition has meaning only i n the con-text of the pr o p o s i t i o n , so that both universals and p a r t i c u l a r s are represented by dependent symbols, a s i t u a t i o n which removes any reason for supposing that only symbols for p a r t i c u l a r s are names of objects and therefore puts universals again on a par with p a r t i c u l a r s . . I s h a l l now support t h i s interpretation with a detailed examination of the relevant t e x t s . Unlike positions (2) and (3)» for which there are abundant t e x t u a l l sources, (1) must be gleaned from a few of Wittgenstein's l e t t e r s , of whioh, as I s a i d , only that of January, 1913, i s r e a l l y h e l p f u l . My presentation of (1) i s es s e n t i a l l y a recon-s t r u c t i o n , based on t h i s l e t t e r , but often going beyond anything e x p l i c i t l y stated there. Nevertheless, nothing i n my formulation of (1) i s the r e s u l t of mere guess-work, but consists rather i n plac i n g Wittgenstein's remarks i n that h i s t o r i c a l context which we can reasonably assume provides the background for the devel-opment of his own thought—the context i n t h i s case being pro-vided by the work of Wittgenstein's teacher, R u s s e l l . I begin l l s h a l l subsequently argue that Wittgenstein's character-i z a t i o n of (2) as the view whioh denies that any content i s symbolized by signs for properties and r e l a t i o n s i s belied by implications of po s i t i o n (2) i t s e l f . 22 by b r i e f l y r e c a p i t u l a t i n g the view of the r e l a t i o n between universals and p a r t i c u l a r s held by Russell through 1912. I f we suppose that 'Socrates i s mortal' i s an elementary proposition of the subject-predicate form, then, we w i l l r e c a l l , both 'Soorates' and 'mortality' are names of objects, a p a r t i c -u l a r and a u n i v e r s a l , r e s p e c t i v e l y . In this proposition, only 'Socrates* names an i n d i v i d u a l , but i n a d i f f e r e n t proposition 'mortality' could also name an in d i v i d u a l i f i t were to oocur i n t h i s proposition as l o g i c a l subjeot. Of the two names, only 'Socrates* i s a proper name, sinoe i t can only occur as naming an i n d i v i d u a l , whereas 'mortality* can oocur either as naming an in d i v i d u a l or as naming an object i n i t s predicative capacity. Against t h i s view, Wittgenstein f e e l s that 'Socrates* and 'mortality', i n th e i r occurrence i n the proposition 'Socrates i s mortal', both name individuals and are indeed both proper names. By this view, Wittgenstein i s not maintaining that 'mor-t a l i t y ' i s here functioning as a l o g i c a l subject, l e t alone that i t can only occur as the subject of a proposition. On the con-t r a r y , his view i s that 'mortality* can only ocour pred i c a t i v e l y and never as subject, the d i f f e r e n t syntactic functions of 'So-crates* and 'mortality* being insured by a difference i n l o g i c a l type of th e i r respective objects. He nevertheless prefers to regard both 'Socrates' and 'mortality' as proper names (and presumably, therefore, the named objects as individuals) because he does not see the difference i n syntactic function of the two names as r e f l e c t i n g a fundamental difference i n the kind of object named. 23 This i s i n contrast to R u s s e l l , who reserves the expression 'proper name' for names of objects which can only occur as sub-j e c t s , because he wants to d i s t i n g u i s h such objects, i . e . , par-t i c u l a r s , from universals i n v i r t u e of a peculiar c h a r a c t e r i s t i c which he supposes to belong solely to the l a t t e r kind of object. This i s the c h a r a c t e r i s t i c of being a constituent of a propo-s i t i o n i n such a way as to bring together the other constituents of the proposition so as to constitute a unity rather than a mere aggregate, an a b i l i t y which Russell attributes to the con-sti t u e n t i n i t s predicative or r e l a t i o n a l r o l e , a r o l e which he sees as the unique province of u n i v e r s a l s . The difference between a unity arid an aggregate i s most easily i l l u s t r a t e d by considering a r e l a t i o n a l p r o p o s i t i o n , e.g., 'John loves Mary'. I f we suppose th i s to be an elementary proposition of two terms and a r e l a t i o n , represented by the three names 'John', 'loving', and 'Mary', then the proposition i t s e l f i s not just a l i s t of these names, but a u n i f i c a t i o n of them into one proposition, where 'loving' actually functions r e l a t i o n a l l y to r e l a t e 'John* to 'Mary' and furthermore to r e l a t e them i n a s p e c i f i c d i r e c t i o n , i . e . , where the love i s said=to proceed from John to Mary and i s 1 not said to proceed from Mary to John. In PP, Russell speaks of t h i s r e l a t i o n a l function, of r e l a t i o n s as that of u n i t i n g , or 'knitting together', the terms between which the r e l a t i o n holds into one complex whole and giving this complex a s p e c i f i c sense, or d i r e c t i o n .2 l i k e w i s e , i n the proposition 'Socrates i s mortal', ISee PM, #54 and #135-136. 2PP, p. 127. 2k assuming i t to be of the subject-predicate form, the predicative function of 'mortality* performs t h i s same job of combining the two names into a prepositional u n i t y , which i s , however, not quite so obvious as i n a r e l a t i o n a l p r o p o s i t i o n , since i n a subject-predicate proposition, containing as i t does only one term, no problem about the ' d i r e c t i o n ' of suoh a proposition a r i s e s . Nevertheless, i t too i s not a mere l i s t of the names 'Socrates' and 'mortality', but i s a unity for which Russell feels the pred-i c a t i v e constituent i s responsible. A oonsequence of Russell's view i s that analysis i s always i n some measure a f a l s i f i c a t i o n , sinoe predicative and r e l a t i o n a l constituents r e t a i n t h e i r unifying function only while they are ac t u a l l y combined with the other constituents i n the proposition. As he says i n PM, " . . . analysis gives us the t r u t h , and nothing but the t r u t h , yet i t can never give us the whole t r u t h . "1 Wittgenstein avoids t h i s d i f f i c u l t y—a n d i n so doing mini-mizes the difference between universals and particulars—by a t t r i b u t i n g the unifying power, not to the predicative or r e l a -t i o n a l constituents i n the proposition, but rather to a speoial copula, which i s introduced only for this purpose and i s not an additional constituent. Thus, e.g., 'Socrates i s mortal' i s 1£ M , #138. This feature of Russell's conception of analysis has no doubt been obscured by the assumption that the analysis of propositions In,PMA into funotions and arguments preserves prop-o s i t i o n a l u n i t y . As we have seen, however, th i s i s not an a n a l -y s i s of propositions into t h e i r constituents, and so the account of PMA gives us no reason for supposing that Russell had by then renounced his e a r l i e r view of PM concerning the necessary short-comings of thi s l a t t e r kind of a n a l y s i s . 25 analyzed into the two constituents Soorates and m o r t a l i t y , and the completely general proposition—though at this time Witt-genstein thought of i t as a name rather than as a proposition— 1 (3x, y ) « i ( x , y ) ' , where '€,' represents the unifying copula and where the proposition as a whole represents that aspect of the o r i g i n a l proposition which constitutes i t s unity, i n abstraction from the p a r t i c u l a r content which i s u n i f i e d . The subscript to the copula i s presumably meant to indicate that the copula i s p r e d i c a t i v e , i . e . , that i t combines two constituents, since Wittgenstein writes the r e l a t i o n a l proposition 'aRb' as 'Ca,(a» R, b ) ' , where the changed subscript presumably means that the copula i s here doing i t s oombining work i n two places, between a and R and between R and b. This a n a l y s i s , which allows us to i s o l a t e a l l the constituents of a proposition and treat them simply as things, none of which have a special unifying r o l e , removes the outstanding r a t i o n a l e for recognizing two kinds of names, one kind for p a r t i c u l a r s and the other kind for u n i v e r s a l s , and Wittgenstein therefore regards both 'Socrates* and 'mortality* as proper names. At the same time, he does not want a method of analysis which makes i t look as i f 'Mortality i s Socrates' makes sense, which he seems to be i n danger of per-mitting i f both 'Socrates' and 'mortality' are proper names of things. His strategy, accordingly, i s to affirm that 'Mortality i s Socrates* i s nonsense—though not because 'Socrates* i s given a syntactic r o l e which names of p a r t i c u l a r s are incapable of f u l f i l l i n g because of the absence of a certain oapacity i n such objects which i s possessed by other objects, v i z . , u n i v e r s a l s , 26 and which allows universals but not p a r t i c u l a r s to act predi-c a t i v e l y and relationally—-but rather because universals and p a r t i c u l a r s are simply objects of d i f f e r e n t l o g i c a l types. This strategy insures that 'Socrates' can never occur as a predicate or r e l a t i o n , but i t also means that, by the same token, 'mortal-i t y ' can never occur as subjeot, and that propositions i n whioh i t s u p e r f i c i a l l y occurs as the grammatical subjeot must be cap-able of further analysis which shows that i t i s not the true l o g i c a l subject, R u s s e l l , of course, had held that functions and individuals were of d i f f e r e n t types, but that universals and p a r t i c u l a r s—bot h being included among individuals—were of the same type, though at the cost of having to account for the two-f o l d r o l e of universals by a t t r i b u t i n g to them a peculiar capa-c i t y . Wittgenstein prefers to do away with the d i f f i c u l t i e s brought on by Russell's p o s i t i o n by admitting a type-difference between universals and p a r t i c u l a r s . This ooncludes my exposition of (1). By January of 1913* Wittgenstein has come to a conclusion which requires the abandonment of (1) and whioh sets i n motion a t r a i n of thought whose culmination i n the T has been widely misunderstood to be a kind of nominalism. Though, as I s h a l l subsequently argue, the mature p o s i t i o n of the T, v i z . , (3)» i s olearly r e a l i s t i c , including as i t does properties and r e l a t i o n s among objects, the intermediate p o s i t i o n we are now about to consider appears—on the basis of some of Wittgenstein's statements i n ( 2 ) —to embrace an extreme version of nominalism. Even here, however, appearances are deceiving, and we s h a l l see 27 that (2) amounts to no more than the claim that symbols for properties and r e l a t i o n s are r a d i c a l l y d i f f e r e n t than symbols for things ( p a r t i c u l a r s )— whioh I express by saying that the former are dependent, and the l a t t e r independent, symbols— a claim whioh f a l l s f ar short of impugning the ontologioal status of u n i v e r s a l s . I proceed now with an account of (2), beginning with Wittgenstein's reason for abandoning (1). As we have seen, (1) depends on the assumption that things are of d i f f e r e n t types to explain the r e s t r i c t i o n s which must be placed on the syntactic roles of the names of these things so as to prevent nonsensical formulations suoh as 'Mortality i s So-crate s ' . By January of 1913* Wittgenstein has become convinced that things, i . e . , the simple objects represented by proper names, cannot be of d i f f e r e n t types. He does not give reasons for his change of heart—perhaps the invocation of types to d i f f e r e n t i a t e simple objects just looks suspiciously ad hoc— but, i n any case, the consequence of his r e j e c t i o n of t h i s solution i s that he must find some other way to insure that nonsensical formulations remain i l l e g i t i m a t e . His new solution i s to a t t r i b u t e differences i n the syntactic roles of symbols for p a r t i c u l a r s and symbols for un i v e r s a l s , not to differences i n the type of object but to differences i n the kinds of symbols themselves. Only symbols for p a r t i c u l a r s are now considered proper names; symbols for un i v e r s a l s , on the other hand, are viewed as having a d i s t i n c t i v e form which renders impossible, by the nature of the symbols themselves, the substitution of one kind of symbol for another. For example, 'Socrates i s 28 mortal' i s analyzed into the proper name 'Socrates* and the form * ( 3 x ) x i s mortal*, or generally into *x* and * ( 3 x ) ^ x * . A r e l a -t i o n a l proposition such as *John loves Mary' i s analyzed into 'John', 'Mary', and ' ( 3 x , y)R(x, y ) ' . On this a n a l y s i s , prop-e r t i e s and r e l a t i o n s are no longer represented by symbols whioh have meaning i n i s o l a t i o n , as do the names of p a r t i c u l a r s ; r a t h e r , the symbols for universals manage to symbolize at a l l only by symbolizing the whole form of the proposition i n which suoh properties or r e l a t i o n s are represented. This way of characterizing the difference between kinds of symbols, as stated i n the l e t t e r of January, 1913* remains e s s e n t i a l l y the same i n NM and NL. The one change i s that the former regards ' (3x)jrfx' and ' ( 3 x , y)R(x, y)* as simple symbols, indeed, as names—only not names of things, but names of forms. By the time of the NL, however, Wittgenstein has emended his po s i t i o n to allow that symbols suoh as those above are not names, since i t makes sense to negate these symbols, while i t makes no sense to negate a name. These symbols are now conceived to perform t h e i r r o l e , not by being names of forms, but by being forms themselves.1 In other words, these symbols are fao t s , which symbolize by corresponding with f a c t s , or at le a s t possible ^Wittgenstein undoubtedly has his e a r l i e r view i n mind when he says, i n NL: "It i s easy to suppose only such symbols are. complex as contain names of objects, and that accordingly * (x, jzOjrfx' or * ( 3 x , y)xRy* must be simple. I t i s then natural to c a l l the f i r s t of these the name of a form, the second the name of a r e l a t i o n . But i n that case what i s the meaning, e.g., of '—(Sac, y)xRy'? Can we put 'not' before a name?" (NB, p. 104) And again: "There i s no thing which i s the form of a propo-s i t i o n , and no name whioh- i s the name of a form" (NB, p. 99). 29 f a c t s , i n r e a l i t y . Thus, what symbolizes i n V x ' or i n 'R(x, y)'-a l t e r n a t i v e l y written as 'xRy'—is not V or 'R', conceived as a name, but rather the fact that V i s to the l e f t of a name-form ( i . e . , a variable i n the place of a proper name) or the fact that 'R* i s between the name-forms 'x' and 'y'. Such, in essence, i s p o s i t i o n (2). What can we conclude from a l l t h i s about the ontologioal status of properties and r e l a t i o n s ? The answer i s : (a) Wittgenstein, who has no metaphysical axe to grind, i s quite w i l l i n g to adopt a riominalistic a t t i t u d e , and thus to view words l i k e 'mortality' as having to do only with the form of a proposition and not with i t s oontent, as long as t h i s strategy succeeds i n rendering impossible the substitution of one kind of symbol for another, or at least makes this impos-s i b l e i n a perspicuous l o g i c a l symbolism; (b) nevertheless, the view he actually advances cannot be regarded as n o m l n a l i s t i c , even though he has no qualms about so conceiving i t , since i t does not r e a l l y expunge the symbolic oontent from words l i k e 'loving' and 'mortality' ( o r—in the l o g i c a l symbolism—from l e t t e r s l i k e V and 'R'). Rather, the most we can say about such words (or l e t t e r s )1 i s that t h e i r way of having a content 0 •^This might be the appropriate place to note that, while i n the T Wittgenstein himself distinguishes symbols from signs, the l a t t e r being the physical medium of the former, I do not need t h i s d i s t i n c t i o n for my purposes and s h a l l consider sym-bols and signs, i n d i f f e r e n t l y , as characterizing that which represents i n contrast to that which i s represented. So, while Wittgenstein would c a l l the l e t t e r ' R ^ i t s e l f a sign rather than a symbol (the l a t t e r involving not only a physical sign but a mental correlation of sign and objeot), I w i l l speak i n d i f f e r e n t l y of 'R' as a sign or a symbol, as a matter of s t y l i s t i c convenience. 30 i s d i f f e r e n t than the way proper names have a content, since the former kind of symbol, unlike the l a t t e r , has a representative capacity only i n the context of a proposition. This i s a l l that Wittgenstein need claim to provide for the difference i n the two kinds of symbols, and, i n the l a s t a n a l y s i s , t h i s i s a l l that he does claim. I s h a l l now elaborate on both parts of this answer. (a) I t i s clear that the demands of l o g i c shape Wittgen-stein's metaphysics and not vice-versa.1 The most s t r i k i n g i l l u s t r a t i o n of t h i s i s i n the l e t t e r of 1913* where, to accom-modate his r e j e c t i o n of the notion of type-differences among things, he moves at one stroke from a fu l l - f l e d g e d realism to a view which regards properties and re l a t i o n s as mere copulae, i . e . , as taking over the unifying r o l e that i n (1) had been given i t s own sp e c i a l symbol '••€'• A c t u a l l y , Wittgenstein's w i l l i n g n e s s , for the sake of a solution to his l o g i c a l d i f f i -c u l t i e s , to give up any claim to an :ontological status for prop-e r t i e s and r e l a t i o n s , extends even further than a concession that words such as 'mortality' or 'loving' symbolize only the form of a proposition and not i t s content, since, by regarding '(3x)jrfx* and *(3x)xRy' as themselves names of forms, i . e . , as simple, unanalyzable symbols, Wittgenstein i m p l i c i t l y allows the symbolic r o l e of V and 'R' to be reduced to the l e v e l of, e.g., the l e t t e r 'c* i n the name 'Socrates'. That i s to say, V and 'R' are l e f t with no i n t e g r i t y as in d i v i d u a l symbols at a l l . Even when he l a t e r r e t r a c t s this extreme view and admits that, e.g., iThe remark i n NL i s Instructive here: "fphilosophyj consists of l o g i c and metaphysics, the former i t s basis" (NB, p. 93). 31 '(3x, y)xRy' retains the complexity of the proposition whose form i s here symbolized,'so that there i s once again an oppor-tunity to view *R' as i n some sense a separable symbolizing component, he f e e l s oontent to a t t r i b u t e the bare minimum of symbolic i n t e g r i t y to the l e t t e r *R' by speaking at times as i f i t s only r o l e i s to occur between fx' and 'y* i n order that the arrangement as a whole might constitute a symbolizing f a c t , without any suggestion that 'R' has any other symbolic r o l e at a l l . Thus, i n NM we find the blunt assertion: " . . . that fR' stands between 'a* and *b* expresses a r e l a t i o n " (NB, p. 109). Yet a l l of this shows, not an antipathy towards realism, but merely Wittgenstein's willingness to adopt whatever ontology his l o g i c a l arguments suggest to be appropriate, and i t i s this same dispassion towards ontologioal issues that f a c i l i t a t e s his re-affirmatlon of realism, once he becomes aware that the nomi-n a l i s t flavor which pervades his statement of (2) i s not warranted by his own arguments. (b) The basis for denying that (2) i s nominalistio i s quite simple.1 I f we consider, e.g., the form ' (3x, y)xRy', where *R* l i t i s possible that someone who holds what I have called a 'nominalist' interpretation of the T would admit that signs for properties and r e l a t i o n s symbolize some kind of content, though a kind d i f f e r e n t from that possessed by objects. I do not profess to be able to make any sense of such a view as an inte r p r e t a t i o n of the T, but, i n any case, the argument which I am about to give i s not directed against t h i s kind of nomi-n a l i s t , or quasi-nominalist, interpretation either of the T or of p o s i t i o n (2).. The argument whioh follows i s intended to show only that (2) i s not nominalistio i n the extreme sense i n which signs for properties and r e l a t i o n s are regarded as sym-b o l i z i n g no content whatever. (That the T cannot be nominal-i s t i o even i n any more moderate sense w i l l be shown l a t e r . ) S t i l l , although I do not claim that t h i s argument i s s u f f i c i e n t 32 i s constant, the reason Wittgenstein oannot be said to have successfully contended that 'R' symbolizes no content, and instead only helps to symbolize a form, i s that *R', i n addition to contributing to the symbolization of a form, by being between 'x' and 'y', t e l l s us something else about any p a r t i c u l a r s which would be represented by 'x' and *y'; v i z . , we are told that these p a r t i c u l a r s are related by the r e l a t i o n symbolized by 'R' rather than the r e l a t i o n symbolized, say, by 'S'. That i s , both 'xRy' and 'xSy' — or 'aRb' and 'aSb*, where 'a' and 'b' are constants--exhibit the same form; i f they nevertheless say something d i f -f e r e n t , as they surely do, this i s because 'R* and 'S' symbolize d i f f e r e n t contents, i . e . , stand for d i f f e r e n t r e l a t i o n s . This i s such an obvious point to be made against a nomi-n a l i s t interpretation of (2) that i t might be supposed to have missed the subtlety of Wittgenstein's p o s i t i o n . But l e t us con-sider the a l t e r n a t i v e s . Any account of Wittgenstein's p o s i t i o n that i s advanced w i l l have to allow for a difference between 'aRb' and 'aSb'. If my interpretation of (2) i s mistaken, and 'R' and *S' do not symbolize d i f f e r e n t contents, so that the difference between the two propositions i s not to be explained by saying that 'R' and 'S' stand for d i f f e r e n t u n i v e r s a l s , then Wittgenstein i s presumably l e f t only the option of saying that 'R' and 'S' are used only to d i f f e r e n t i a t e the forms of the two to compel the adoption of a f u l l - f l e d g e d r e a l i s t ontology, I w i l l nevertheless subsequently contend that Wittgenstein's own return to realism i n p o s i t i o n (3) was, i n f a c t , occasioned by his r e a l i z a t i o n of the force of t h i s argument and the closely r e l a t e d argument concerning the p o s s i b i l i t y of generalizing functions. 33 propositions, i . e . , the manner in which the names are arranged i n the propositions so as to r e f l e c t a corresponding (possible) configuration of p a r t i c u l a r s i n r e a l i t y . I can think of two ways i n which one might try to follow up t h i s l i n e of reasoning. The f i r s t way—which to my knowledge has not had any advocates but i s nevertheless at least a remotely possible way of construing some of Wittgenstein's remarks—is to say that 'aRb' and 'aSb* are d i f f e r e n t propositions, but not because 'R* and 'S' symbolize d i f f e r e n t contents; rather, since what symbolizes i n each propos i t i o n , other than proper names, i s the fact that the symbols are i n such-and-such an arrangement, 'aRb* and 'aSb' say something d i f f e r e n t because the propositional signs constitute d i f f e r e n t facts—the f i r s t f a ct being that 'a' and *b* are i n a c e r t a i n r e l a t i o n to the l e t t e r *R* and the second fact being that 'a* and 'b' are i n this same r e l a t i o n to the l e t t e r 'S*. Therefore, we need not suppose that 'R* and 'S* symbolize contents to aocount for 'aRb* and 'aSb' symbolizing d i f f e r e n t f a c t s ; the propositional signs themselves provide a l l the differences that we need. Whatever s l i g h t merit this i n t e r p r e t a t i o n may have i s f a r outweighed by i t s obvious d i f f i c u l t i e s , and we can a t t r i b u t e i t to Wittgenstein only at the cost of admitting the utter untena-b i l i t y of his p o s i t i o n . That i s , conceding that the difference between *R' and 'S' as l e t t e r s i s s u f f i c i e n t to d i s t i n g u i s h the propositional signs 'aRb* and 'aSb1 as f a c t s , i t i s hard to see how this difference alone i n the two facts i s supposed to sym-bol i z e a difference i n the configuration of the actual p a r t i c -34 u l a r s . In other words, i f 'aRb' and *aSb' are to symbolize d i f f e r e n t (possible) configurations of p a r t i c u l a r s , the symbolic r o l e of 'R' and 'S' must be greater than t h e i r mere occurrence as l e t t e r s between the symbols 'a' and 'b'; i f suoh a difference were enough to constitute a symbolization of d i f f e r e n t propo-s i t i o n a l forms, as would have to be the case i f 'aRb* and 'aSb' were of d i f f e r e n t forms merely i n v i r t u e of the difference between 'R' and 'S' as l e t t e r s , then 'aRb' and 'cRd' would also be of d i f f e r e n t forms, i n v i r t u e of the d i f f e r e n t l e t t e r s com-p r i s i n g these two propositional signs. The d i f f i c u l t i e s of t h i s i n t e r p r e t a t i o n lead us to consider,a second i n t e r p r e t a t i o n , which grants that *R' and 'S' have a symbolic r o l e over and above their occurrence as d i f f e r e n t l e t t e r s i n propositional signs, and i t i s to t h i s l a t t e r Interpretation that I now turn. This second i n t e r p r e t a t i o n , unlike the f i r s t , has a number of adherents,1 though as an interpretation of the T i t s e l f , and not just of (2), since (2) and (3) are not distinguished by other commentators. The passage of central importance for t h i s i n t e r -pretation i s 3.1432 of the T, which says: Not: "The complex sign 'aRb' says that a stands i n the r e l a t i o n R to b", butr That 'a' stands i n a certain r e l a t i o n to *b' says that aRb. The crux of this passage i s taken to be that Wittgenstein omits mention of the l e t t e r 'R' i n discussing the r e l a t i o n of 'a' to 'b*, which i n turn i s taken to mean that i n a l o g i c a l l y per-•^This i n t e r p r e t a t i o n seems to have originated with I. M. Copi, and has been endorsed i n i t s essentials by G. E. M. Anscombe, W. S e l l a r s , J . J . Thomson, G. P i t c h e r , and A. Mueller, among others. 35 spiouous symbolism 'R' would be superfluous, so that the con-f i g u r a t i o n of p a r t i c u l a r s could be f u l l y symbolized by a corre-sponding arrangement of names. For example, *aRb*, 'aSb', and 'aTb' might be equivalent to •a' 'ab*, 'ba', and b r e s p e c t i v e l y . On this view, unlike the f i r s t i n t e r p r e t a t i o n , 'R', 'S*, and 'T' have a symbolic r o l e i n addition to their occurrence as l e t t e r s , but t h i s i s not the r o l e of standing for universals; r a t h e r , i t i s the r o l e of being a shorthand i n d i c a -t i o n of the way 'a* and 'b' would be arranged i n the perspicuous notation. Thus, the 'R' i n 'aRb* t e l l s us that •a' i s to the l e f t and adjacent to 'b', and so on. Admittedly, Wittgenstein's silence about the place of 'R' i n the symbolizing fact might encourage the view that 'R' does not help constitute t h i s f a c t , but this i n i t s e l f i s hardly evidence for a nominalistio interpretation of the T. We can account for the nominalistio tone of 3.1432 by considering the source from which the passage i s l i f t e d nearly verbatim, v i z . « NI (NB, p. 105), whose nominalistio tone we can acknowledge without having to conclude that i t i s this same tone that Witt-genstein wants to preserve i n 3»1^32. Rather, the context of 3.1432 suggests that Wittgenstein i s u t i l i z i n g his e a r l i e r remark (as, indeed, the whole T substantially consists i n a re-arrangement and re-working of such e a r l i e r remarks) only to emphasize that a propositional sign i s a fact as opposed to a complex name. That i s , according to Wittgenstein's decimal 36 notation, 3*1432 i s a comment on 3»l43i and this l a t t e r passage i s concerned only to make the point that a propositional sign i s a f a c t , not a name; thi s i s also the context of the remark i n NL corresponding to 3»l432. On my i n t e r p r e t a t i o n , then, the purpose of 3.1432 i s only to contrast symbolizing facts with names, and the nominalist tone of this passage—which could have been avoided altogether had Wittgenstein specified that the r e l a t i o n i n which "a" stands to "'b' consists i n the i r respective r e l a t i o n s to 'R'—is i n any case minimized by the r e a l i z a t i o n that the status of 'R* as- a name i s implied i n many other contexts i n the T. My subsequent account of (3)» "by providing independent grounds for supposing that the T regards 'R' as the name of an object, w i l l , by im p l i c a t i o n , bear.out my interpretation of 3»1432. What I want to consider now, however, i s why the popular i n t e r -p retation of the T, based on 3.1432, i s not the r i g h t i n t e r p r e -t a t i o n of (2) e i t h e r , whose ostensible nominalism makes such an int e r p r e t a t i o n at le a s t prima f a c i e p l a u s i b l e for thi s e a r l i e r p o s i t i o n . The reason for r e j e c t i n g such an int e r p r e t a t i o n for (2) i s that there i s no suggestion i n any of the textual sources of (2) that a l e t t e r l i k e *R' would disappear i n a l o g i c a l l y perspicuous symbolism; i n f a c t , there are clear statements to.the contrary. Thus, i n NM we f i n d : When we say of a proposition of £the) form *aRb' that what symbolizes i s that 'R' i s between 'a' and 'b', i t must be remem-bered that i n fact the proposition i s capable of further analysis because a, R, and b are not simples. But what seems certain i s that when we have analysed i t we s h a l l i n the end come to prop-ositions of the same form i n respect of the fact that they do consist i n one thing being between two others. . . . though we 3? don't know any unanalyzable propositions of t h i s kind,.yet we can understand what i s meant by a proposition of the form Ox, y, R)xRy (whioh i s unanalyzable . . • and unanalyzable proposition = one i n whioh only fundamental symbols = ones not capable of d e f i n i t i o n , oc.our) . • . (NB, p. 110). From this passage we can conclude, f i r s t , that 'R* i s not t h e o r e t i c a l l y superfluous and, secondly, that the form of the proposition i s r e a l l y that i n whioh a l l the l e t t e r s , including 'R', are apparent v a r i a b l e s , so that 'aRb' and 'aSb', where the constants 'R' and 'S' take the place of the apparent v a r i a b l e , are propositions of the same form, which they could not be i f 'R' and *S' were shorthand indications of d i f f e r e n t arrangements of 'a' and 'b'. I f , as seems the undeniable conclusion, we must admit that 'R'--where It i s a constant i n a proposition such as 'aRb' — symbolizes a s p e c i f i c r e l a t i o n , and does so preci s e l y i n v i r t u e of being that l e t t e r and no other, how do we explain the passage from NI corresponding to 3«l432, i n which the omission of 'R' from the description of the symbolizing fact appears deliberate? The answer i s that 'R\ i s not mentioned because, unlike 'a' and 'b.', 'R* i s not considered to stand for anything i n i s o l a t i o n ; i f , instead of saying that 'a* stands i n a certain r e l a t i o n to 'b', Wittgenstein had said that 'a' and 'b* stand i n a cert a i n r e l a t i o n to each other by standing i n a cert a i n r e l a t i o n to 'R', he might thereby have fostered the misimpression, by treating 'R' i n i s o l a t i o n , that 'R* was a symbol just l i k e 'a' and *b'— and, as we have seen, Wittgenstein i s i n s i s t e n t that i t i s not 'R* that symbolizes a r e l a t i o n , but rather that 'R' i s between 'a' and 'b'. So, even though the symbolizing fact includes 'R' 38 just as much as i t does 'a' and 'b', Wittgenstein does not mention 'R' i n his description of the fact i n order to avoid any possible confusion. Unfortunately, as the history of Witt-gensteinian exegesis has borne out, his innocent remark has occasioned l i t t l e else but confusion, a confusion whioh I hope my own aooount has done something to d i s p e l . To sum up: despite appearances, (2) i s not n o m i n a l i s t i c . We are misled i n t h i s d i r e c t i o n by Wittgenstein's exaggeration of the difference i n the symbolic r o l e of proper names, on the one hand, and words for properties and r e l a t i o n s , on the other. When he says that . . . i n 'aRb*, 'R' i s not a symbol, but that *R* i s between one name and another symbolizes . • . TN17 p» 108) he i s j u s t i f i e d only i n saying that 'R' Is not an Independent symbol. Matters are only obscured by the suggestion that 'R' i s not a symbol at a l l ; 'R' and 'S' can very well serve by them-selves to t e l l us that a d i f f e r e n t r e l a t i o n i s being symbolized i n each case, even though neither 'R' nor 'S' f u l l y exercises i t s symbolic function except i n the context of a proposition or a propositional schema. It i s obvious, then, that Wittgenstein's intention to est a b l i s h a r a d i c a l dichotomy between proper names and words for properties and r e l a t i o n s—the former kind of symbol having to do only with the content, and the l a t t e r only with the form of a proposition—cannot be realized within (2) i t s e l f , which allows at most a d i s t i n c t i o n between symbols whioh concern only a con-tent and symbols whioh concern both a content and a form. Witt-genstein's eventual r e a l i z a t i o n of t h i s marks his t r a n s i t i o n 39 from (2) to (3)» where he once again acknowledges that both p a r t i c u l a r s and universals constitute the content of a prop-o s i t i o n , i . e . , both are included among objeots, and that the symbols both of universals and of p a r t i c u l a r s are names of these objects. We might suppose that, despite this r e - a s s i m i l a t i o n of universals and th e i r symbols with p a r t i c u l a r s and their symbols under the general heading 'object and name', there remains at le a s t the difference between those names involved only i n sym-b o l i z i n g a content and those which help symbolize a form as w e l l . But even t h i s one remaining r a t i o n a l e for distinguishing between kinds of symbols i s eliminated by Wittgenstein's admission that names of p a r t i c u l a r s , just as names of unive r s a l s , symbolize a form as well as a content—i.e., names of both kinds of objects are dependent symbols. With th i s admission, the f a l s e dichotomy between symbols which are names and those which are forms i s t o t a l l y repudiated, and Wittgenstein thereby arrives at his d e f i n i t i v e view of the nature of names and objects—or at lea s t the view which pre v a i l s throughout the T and i s not again to be revised u n t i l 1929« I t i s to an exposition of (3)* the mature p o s i t i o n of the early Wittgenstein, that I now turn. From the time of NM i n A p r i l , 1914, i n which Wittgenstein unreservedly embraces (2), u n t i l the f i r s t entry i n NB on August 22 of the same year, we have no record of his development. We can say only that his willingness to embrace nominalism probably continues throughout t h i s time, since the f i r s t pas-sages whioh indicate a d i s a f f e c t i o n with (2)—or at l e a s t with 40 his u n c r i t i c a l l y nominalistic conception of i t—do not appear u n t i l the end of September, 1914, when his views on the status of properties and r e l a t i o n s begin a metamorphosis whioh, some weeks l a t e r , issues i n the d e f i n i t i v e p o s i t i o n of (3)» v i z . , that universals are objects on a par with p a r t i c u l a r s . I s h a l l now examine th i s transformation, recorded primarily i n pp. 6-2? of NB, beginning with Wittgenstein's discussion of the prob-lem of true and f a l s e propositions, where his desire to retreat from nominalism f i r s t becomes manifest. The g i s t of t h i s discussion, paraphrased and quoted, i s as follows: The proposition, which i s one f a c t (Sachverhalt). i s a l o g i c a l representation of i t s meaning (Bedeutung), which i s another f a c t , only i f there i s a c o r r e l a t i o n between the com-ponents of both f a c t s . Besides the c o r r e l a t i o n of names and things named, • . i t i s clear that a c o r r e l a t i o n of r e l a t i o n s too takes place somehow" (25»9»14).1 In this respect, a propo-s i t i o n i s l i k e a p i c t u r e . Both the proposition 'A i s fencing with B' and the p i c t u r e of A fencing with B are able to express a sense independent of t h e i r truth or f a l s i t y , which means that they are both able to represent the r e l a t i o n of fencing (29»9»14). But i f A and B are not actually fenoing, and the pioture never-theless portrays them as fencing, the pioture represents a r e l a t i o n that does not exist (30.9«>14). How i s that possible? Now once more i t looks as i f a l l r e l a t i o n s must be l o g i c a l i n order for t h e i r existence to be •"•Paraphrases or quotes from NB w i l l be cited by dates i n place of page references; t h i s w i l l have the advantage of showing the chronological development of Wittgenstein's thought. 41 guaranteed by that of the sign (30.9«14). Prom thi s l a s t remark we may conolude that Wittgenstein i s tending towards the view that r e l a t i o n s have some kind of onto-l o g i o a l content, and i s driven back to the view that symbols for r e l a t i o n s represent no content—i.e., that they are merely ' l o g i c a l1, merely copulae1—only by the consideration that i f the r e l a t i o n a l component i n a f a l s e proposition has no corre-l a t i v e component i n r e a l i t y , then neither can any such content be represented i n a true pr o p o s i t i o n , since the truth or f a l s i t y of the proposition i s ir r e l e v a n t to i t s sense. We may also gather that the ontologioal status Wittgenstein would be w i l l i n g to give to r e l a t i o n s , were i t not for the d i f f i c u l t y raised by f a l s e propositions, i s something less than the status of p a r t i c -u l a r s , i n the sense that p a r t i c u l a r s exist on the i r own, while Wittgenstein apparently f e e l s that r e l a t i o n s cannot exist apart from the existence of p a r t i c u l a r s i n these r e l a t i o n s . Thus, e.g i f A i s not fencing with B, then the r e l a t i o n of fencing, with respect to A and B, does not e x i s t . I say 'with respect to A and B', because C and D may very well be fencing, so that the r e l a t i o n exists between them, though not between A and B. On th i s view, the existence or non-existence of a r e l a t i o n i s alway with respect to s p e c i f i c p a r t i c u l a r s , even though the r e l a t i o n i t s e l f can be represented i n the abstract, v i z . , i n a propo-!That i s , i f 'Qx, y)xRy* has no content, then i t s ' e x i s t -ence' i s guaranteed by that of the sign i n the same sense that the 'existence' of l o g i c a l constants i s so guaranteed. In other words, there i s no r e a l question of there being a repre-sentation of existent contents here at a l l . s i t i o n a l schema, where the other symbols are a l l variables (4.10.14) or In a pioture whose figures standing i n t h i s r e l a t i o n themselves represent no s p e c i f i c individuals (30.9«14). In any case, the question as to how Wittgenstein would view the status of r e l a t i o n s appears academic, since the prob-lem of finding a c o r r e l a t i v e component i n r e a l i t y for the r e l a t i o n a l component of a proposition forces his concession that r e l a t i o n s might just be ' l o g i c a l * a f t e r a l l . This con-cession, however, i s only temporary, for we find him a few days l a t e r again mulling over the p o s s i b i l i t y that there i s a corre-l a t i o n of r e l a t i o n a l components, only to.be discouraged once more by even further d i f f i c u l t i e s . Thus: If there were such a thing as an immediate c o r r e l a t i o n of r e l a t i o n s , the question would be: How are the things that stand i n these r e l a t i o n s correlated with one another i n t h i s case? Is there suoh a thing as a d i r e c t c o r r e l a t i o n of r e l a t i o n s without consideration of th e i r direction? -Are we misled into assuming "relations between r e l a t i o n s " merely through the apparent analogy between the expressions: "relations between things" and "relations between re l a t i o n s " ? (9.10.14) This passage shows that he regards a c o r r e l a t i o n between r e l a t i o n a l components, assuming i t were p o s s i b l e , to be between t w 0 contents only. That i s , If 'aRb' i s the proposition, the r e l a t i o n a l component of the proposition to be correlated with i t s counterpart i n r e a l i t y would be 'R' s l m p l l c l t e r . and since th i s i s o l a t i o n of 'R' from the propositional sign as a whole would destroy the representation of the sense ( i . e . , d i r e c t i o n ) of the r e l a t i o n , Wittgenstein once again concludes that perhaps the whole attempt to correlate r e l a t i o n s i s misconceived, having received i t s impetus only by a misleading grammatical analogy. 43 Yet he feels that t h i s conclusion, too, i s unsatisfactory, as i s any conclusion whioh undertakes—after the f a c t , as i t were— to s e t t l e the question concerning the p o s s i b i l i t y of r e l a t i o n s v. e x i s t i n g , since l o g i c must decide this p o s s i b i l i t y for us from the s t a r t . As he says,;"The question 'about the p o s s i b i l i t y of existence propositions does not come i n the middle but at the very f i r s t beginning of l o g i c • . ." (9.10.14), and he sees his previous attempts to deal with the question about the status of r e l a t i o n s from the vantage-point of the 'middle' of the inquiry as a "fundamental mistake" (9.10.14). I t i s t h i s observation which spurs him on to the f i n a l s o l u t i o n of the problem, though by a somewhat circuitous route. Thus, as another i l l u s t r a t i o n of the theme that questions of l o g i c must be decided from the beginning, he considers the axiom of i n f i n i t y , and says that there can be no l o g i c a l prob-lem, i n connection with the proposition which states that an i n f i n i t e number of things e x i s t , which i s not already solved i n the proposition statin g that one thing e x i s t s—si n c e , presumably, no further l o g i c a l difference i s r e f l e c t e d i n the number of things we assert to e x i s t .1 Prom thi s he i s led to a disoussion of i n f i n i t e numbers (11.10.14), and then to a consideration of the d e f i n i t i o n s of cardinal numbers generally, ". . . i n order to understand the r e a l sense of propositions, l i k e the Axiom of I n f i n i t y " (12.10.14). He then considers the proposition: •^ -His actual statement i s : " A l l the problems that go with the Axiom of I n f i n i t y have already to be solved i n the prop-o s i t i o n »(3x)x=x' " (9.10.14). 44 'There i s a class with only one member', which he symbolizes as 1 (3frf).\(3x): tfxifiy, t$z?> _y=z' • That i s , he expresses what i s meant here by 'one', by saying that there i s a property whioh some thing uniquely possesses. I t i s clear from th i s that, i n addition to p a r t i c u l a r s , universals are subject to e x i s t e n t i a l q u a n t i f i c a t i o n .1 As Wittgenstein r h e t o r i c a l l y puts the matter, Can we speak of numbers i f there are only things? I.e. i f for example the world only consisted of one thing and of nothing e l s e , could we say that there was ONE thing? (13.10.14) His conclusion i s that i t i s l i k e l y that ". • .we can only t a l k of 1 i f there i s a material funotion which i s s a t i s f i e d only by one argument" (13.10.14). As we have seen, Wittgenstein i n NM had already admitted that r e l a t i o n s can be thus generalized, v i z . , i n the form '(3x, y, R)xRy', but the implication of thi s p o s s i b i l i t y of turning 'R' into an apparent variable had obviously not struck him at the time. Now, however, he sees that i f $f6* can be turned into an apparent v a r i a b l e , then, when V i s constant, i t must symbolize some s p e c i f i c content. Hence, by 15.10.14 we fi n d him maintaining that V and 'a' are both elements of the proposition Va* which r e l a t e to their respective meanings (Bedeutungen)- i n r e a l i t y . There i s no doubt that 'Bedeutung' as here employed carries the f u l l weight of 'ontologioal content', sinoe he also speaks on 15.10 of the elements of a proposition being correlated with objects.2 Of •'•Unlike R u s s e l l , Wittgenstein sees no problem i n supposing that a function-sign stands for a property. This matter i s discussed further on pp.. 87-91, 2Wittgenstein does not e x p l i c i t l y say that V i s an element, but he implies i t , when he says: "I have here regarded 45 perhaps even greater s i g n i f i c a n c e i n the entry of 15*10 i s Wittgenstein's recognition that 'C3x, jrf)' contains a p l u r a l i t y of forms—presumably, though he does not e x p l i c i t l y say, a function-form and an argument-form—an-admission no doubt brought on by the r e a l i z a t i o n that both components are general-i z a b l e independently of.each other and must therefore by regarded as separate units regarding both content and form. By 15»10, then, Wittgenstein's notes contain the essentials of p o s i t i o n (3)* both universals and p a r t i c u l a r s are included among objects, and the; symbols for both kinds of object repre-sent a form as well, as a oontent. If his remarks at this stage are l i t t l e more than hints and a l l u s i o n s , he soon becomes more s p e c i f i c . By 17.10.14 he says that i n a description of the world by means of completely general propositions, we would need the proposition ' (36) ("Y)Y=tf*— supposing there were only one property i n the world—to i d e n t i f y that property, and he e x p l i c i t l y o a l l s the property an object. Numerous subsequent remarks throughout NB show that he continues to use the term 'object' to include u n i v e r s a l s , the most patent example being his statement on 16.6.15 that r e l a t i o n s and properties are objects. the r e l a t i o n s of the elements of the proposition to t h e i r mean-ings as f e e l e r s , so to say, by means of whioh the proposition i s i n contact with the outer world; and the generalization of a proposition i s i n that case l i k e the drawing i n of f e e l e r s ; u n t i l f i n a l l y the completely general proposition i s quite i s o -l a t e d " (15-10.14). Though he then goes on to question t h i s account of g e n e r a l i z a t i o n , even for t h i s account to be enter-tained requires that V be regarded as an element towards whioh a f e e l e r i s supposedly withdrawn. 46 Regarding the question of form and content, he says on 1.11.14 that The l o g i c a l form of the proposition must already be given by the forms of i t s component p a r t s . . . . In the form of the subjeot and of the predicate there already l i e s the p o s s i b i l i t y of the subjeot-predioate proposition • • • • That he continues to regard a l l components of a proposition as having a form i s confirmed i n the remark of 20.6.15* where he says that we find i n the theorems of mathematical physios " . . . neither things nor funotions nor r e l a t i o n s nor any other l o g i c a l forms of o b j e o t l "1 Wittgenstein's a l l u s i o n here to various l o g i c a l forms serves to remind us of an important point: i n spit e of the main thrust of (3)» which i s to re-assimilate universals with p a r t i c u l a r s as being equally objects, he nevertheless wants to preserve some d i s t i n c t i o n between the two, i f only a difference i n the l o g i c a l form of these two kinds of object, as i s manifest i n the d i f f e r e n t syntactic r o l e s of th e i r respective symbols. Even with respect to differences i n l o g i c a l form, however, there i s no :emphasis on the d i s t i n c t i o n between universals and p a r t i c -u l a r s , since within these categories themselves there are objects of various forms. Among un i v e r s a l s , the most obvious example would be the difference i n form of any of those objects which funotion p r e d i c a t i v e l y from any of those whioh funotion r e l a t i o n -a l l y . But even p a r t i c u l a r s , a l l of whioh seem to have at le a s t that aspect of form i n oommon whioh permits them to oocur only as l o g i c a l subjects of a proposition ( i . e . , only as subjects i n ^ e e also the disoussion of 14.6.15. 47 a subject-predicate p r o p o s i t i o n , or as terms i n a r e l a t i o n a l p r o p o s i t i o n ) , can nevertheless d i f f e r i n o v e r - a l l l o g i c a l form. Thus; I f , e.g., I c a l l some rod "A", and a b a l l "BM, I can say that A i s leaning against the w a l l , but not B. Here the i n t e r n a l nature of A and B oomes into view. . . . "The watch i s s i t t i n g on the table" i s senseless' (22.6.15) That the s u b s t i t u t i o n of the name of•one p a r t i c u l a r for the name of another can turn a proposition into nonsense shows that the syntactic r o l e s of these names d i f f e r , i . e . , that the p a r t i c -ulars have d i f f e r e n t l o g i c a l forms. S t i l l , the occurrence of p a r t i c u l a r s only as subjects and the occurrence of universals only p r e d i c a t l v e l y and r e l a t i o n a l l y does mark a basic d i s t i n c t i o n , and Wittgenstein's desire to acknowledge th i s d i s t i n c t i o n between universals and p a r t i c u l a r s , coupled with his desire to stress t h e i r common nature as objects, creates a minor tension whose main r e s u l t i s a f l u c t u a t i o n i n the use of the expressions 'name' and 'thing*, depending on whether he i s thinking of the difference between universals and p a r t i c u l a r s or Is thinking of t h e i r s i m i -l a r i t i e s . That i s , he sometimes speaks as i f only p a r t i c u l a r s are things and have names, while at other times he uses both expressions i n connection with universals as w e l l . So, e.g., by 17•10•14, he has admitted that objects inolude u n i v e r s a l s , but he s t i l l f e e l s compelled at t h i s time to q u a l i f y h i s character-i z a t i o n of completely general propositions by saying that we describe such propositions " . . . without using any name or any other denoting sign . . . ," where the phrase 'any other denoting s i g n ' presumably alludes to symbols for u n i v e r s a l s .1 By 22.10.14 •^-Anscombe's t r a n s l a t i o n reads: 'any sort of names or other 48 t h i s d i s t i n c t i o n between names and other denoting signs seems to have been abandoned, sinoe Wittgenstein now contrasts names only with the l o g i c a l , non-arbitrary features of representation i n a p r o p o s i t i o n , whioh suggests that the symbols whioh are a r b i t r a r i l y stipulated to stand for universals are names as w e l l . On 28.10.14 he says that What the pseudo-proposition "There are n things" t r i e s to express shows i n language by the presence of n proper names with d i f f e r e n t meanings. This suggests, f i r s t , that the d i s t i n c t i o n between universals and p a r t i c u l a r s i s not r e f l e c t e d i n a d i s t i n c t i o n between names and other denoting signs, but between d i f f e r e n t kinds of names, where one kind has the q u a l i f i c a t i o n 'proper'; second, that 'thing' i s synonymous with ' p a r t i c u l a r ' , since presumably only p a r t i c u l a r s have 'proper* names. By 3»H»14 the treatment of both symbols for universals and symbols for p a r t i c u l a r s as names of objects i s re-inforoed, when he says that "The arbitrary cor-r e l a t i o n of sign and thing s i g n i f i e d . . . " occurs i n the elem-entary proposition ". . . by means of names • . • ," where the oontext makes clear that by 'things s i g n i f i e d * he i s including a l l the components of a f a c t , universals as well as p a r t i c u l a r s . This quote also shows his widening use of the expression 'thing* to cover both kinds of components rather than just p a r t i c u l a r s , a p r a c t i c e whioh i s continued i n his remark on 4.11.14 that denoting signs*, which obscures the point of the q u a l i f i c a t i o n by leaving open the p o s s i b i l i t y that symbols for universals might just be one sort of name. Wittgenstein's own phrase 'irgend einen Namen', however, means 'any name' or 'any name at a l l ' , where there i s no suggestion that names themselves might be of d i f f e r e n t s o r t s . 49 One name i s representative of one thing, another of another t h i n g , ana they themselves are connected? i n thi s way the whole images the s i t u a t i o n— l i k e a tableau vivant. On 21.1.15 we again fi n d him using the expression 'thing* i n the narrower sense, i . e . , i n contrast with properties and r e l a t i o n s as d i f f e r e n t species of which objects are the genus, a use of the expression which re-occurs on 31.5*15* when he says that Names are necessary for an assertion that t h i s thing pos-sesses that property and so on. They l i n k the propositional form with quite d e f i n i t e objects.1 On 16.6.15 he once again uses the expression 'name* i n the nar-row sense, i n speaking of a proposition as containing ". • • names, r e l a t i o n s , e t c . • • • , ' • though by 17.6.15 he again speaks of men-tioning a property by name (". . • muesste i c h s l e namentlioh anfuehren • • . " ) , a use of 'name* whioh i s reconfirmed on 21.6.15* To complicate matters f u r t h e r , on 22.6.15 he uses the expression 'object* i t s e l f i n a narrow sense, i n speaking of the fact ex-pressed by the proposition *The watch i s l y i n g on the table* as one i n whioh two objects are i n a r e l a t i o n . That universals are not being denied status as objects i n the wide sense i s obvious from the context, where there i s no in d i c a t i o n that any dr a s t i c reversal i n outlook has taken place from only the day before, when he had spoken of the property mortality as an object. A l s o , when NB resumes nearly a year l a t e r (the entry of 22.6.15 marks the end of the second notebook and the entries of the third notebook begin on 15.4.16), he again e x p l i c i t l y uses 1See also the discussion from 19*6.15 through 21.6.15. 50 'object1 i n the wide sense—e.g., on 9«7«l6 he says that "The proposition f a speaks of p a r t i c u l a r objects • • ."—so that the d i f f e r e n t i a t i o n of objects and r e l a t i o n s on 22.6.15 c l e a r l y does not portend a return to nominalism. I have i l l u s t r a t e d Wittgenstein's fluctuating terminology at length to make the point that even when he r e s t r i c t s the expressions 'thing', 'name*, and even 'object' to p a r t i c u l a r s , there i s no suggestion that r e l a t i o n s and properties are not themselves e n t i t i e s . P a r t i c u l a r s and universals continue to stand on equal|pfooting as objects, i n the wide sense, so that the use of 'thing', 'name', or 'objeot* i n the narrow sense can be intended to do no more than c a l l attention to the d i s t i n c t i v e status of p a r t i c u l a r s as l o g i c a l subjects. What makes our recog-n i t i o n of thi s f l u c t u a t i n g terminology i n Ng, and a proper assess-ment of i t s s i g n i f i c a n c e , especially important i s that we thereby establish a precedent f o r explaining comparable passages i n the T, whioh gives us a reason for denying that such passages support a nominalist i n t e r p r e t a t i o n of T, despite any prominence they may give to p a r t i c u l a r s by the narrow use of these expressions.1 Before moving on to an examination of the T i t s e l f , I want to consider two more points i n connection with the development of (3) i n NB. The f i r s t point concerns the status of r e l a t i o n a l components i n f a l s e propositions, a problem whioh i n i t i a l l y provides a stumbling-block for Wittgenstein's acceptance of these components as symbols of c o r r e l a t i v e components i n r e a l i t y , •'•This matter i s again taken up on p. 96. 51 but whioh i s simply shoved aside when he r e a l i z e s that symbols fo r universals must represent contents i n order for us to have something from whioh to abstract when such symbols are made into apparent v a r i a b l e s . He does not e n t i r e l y forget about th i s problem, however; within a few weeks of th i s decisive development, which we placed at 17,10,14, he again addresses himself to the problem of the f a l s e p r o p o s i t i o n , though now with one Important dif f e r e n c e : whereas he had previously characterized a f a l s e proposition i n terms of the non-existence of a c o r r e l a t i v e r e l a t i o n a l oomponent i n r e a l i t y , he now characterizes such a proposition i n terms of the non-existence of connections between objects i n r e a l i t y . Thus: But when I say: the connection of the prepositional com-ponents must be possible for the represented things—does th i s not contain the whole problem? How can a non-existent connection between objeots be possible? "The connection must be possible" means: The proposition and the components of the s i t u a t i o n must stand i n a p a r t i c u l a r r e l a t i o n . Then i n order for a proposition to present a s i t u a t i o n i t i s only necessary for i t s oomponent parts to represent those of the s i t u a t i o n and for the former to stand i n a connection which i s possible for the l a t t e r ( 5 » H » l 4 ) . On t h i s aooount, the f a l s i t y of the proposition 'A i s fencing with B' i s due, not to the non-existence of a oomponent, but only to the non-existence of a certa i n connection, or configuration, of existent components. Even i f A i s not configured with B i n the r e l a t i o n of fencing, the components themselves—including fenc-i n g—exi s t nonetheless, so that f a l s e propositions give r i s e to no problems concerning r e l a t i o n a l oontents. With t h i s aooount of f a l s e propositions, Wittgenstein f i n a l l y disposes of the one obstaole that haa prevented him from according ontologioal status to r e l a t i o n s . 52 The second point I want to consider, though not d i r e c t l y concerned with the question of the status of u n i v e r s a l s , i s relevant to our appraisal of NB as a source for understanding the T. On 18.10.14 Wittgenstein remarks: "Roughly speaking: before any proposition can make sense at a l l the l o g i c a l con-stants must have meaning (Bedeutung)." Anscombe translates 'Bedeutung' here as 'reference* to emphasize that the word i s being used i n the technical Fregean sense of *what the word stands f o r * , from which she oonoludes that . . . there i s a great contrast between his ideas at t h i s stage of the Notebooks and those of the Traotatus, where he denies that l o g i c a l constants or sentences have 'Bedeutung' (NB, p. 15* translator's footnote). If Wittgenstein's views at 18.10.14 are at the r e l a t i v e l y crude stage of development that Ansoombe would have us b e l i e v e , t h i s casts suspicion on contemporaneous passages stating that universals are objects. Might not suoh passages r e f l e c t an equally p r i m i t i v e hypostatization, a l l suoh errors being corrected by the vastly more sophisticated T? The faot i s , however, that i n the passage i n question Wittgenstein i s obviously not main-taining the naive view, attributed to him by Ansoombe, that l o g i o a l constants have reference. He i s on record as objecting to this naive view as f a r back as 22.6.12, where, i n a l e t t e r to R u s s e l l , he says: • • . whatever may turn out to be the proper explanation of apparent v a r i a b l e s , i t s consequences must be that there are n o loffioal constants. Logic must turn out to be a t o t a l l y d i f f e r e n t kind than any other science (NB, p. 119). So, r i g h t from the beginning of Wittgenstein's philosophical 53 development, he sees a problem In the view that symbols l i k e 'v', *s', and **-* stand for objects, i n the way that names do, and one of his pre-oocupatlons from then on i s to give an explanation of the nature of l o g i o a l symbols whioh does not suppose that these symbols r e f e r to spe c i a l ' l o g i o a l1 objeots. The view of the T that the i n t e r - d e f i n a b i l i t y of *v*, *3', e t c . , shows that they are not symbols for objects (5.42) has already received expression i n NL (NB, p. 101), and his view that we could say that the one l o g i o a l constant would be what a l l prop-osit i o n s have i n common (5*47) has already been worked out i n NM, where he says that . . • i n order to introduce so-called " l o g i c a l constants" properly, you must introduce the general notion of a l l possible combinations of them = the general form of a proposition (NB, p. 116). Therefore, by the time Wittgenstein makes his remark i n NB about the 'Bedeutung* of l o g i c a l constants, i t i s clear that he i s using this term i n a wide, non-technical sense, an interpretation whioh i s supported by the q u a l i f y i n g preface, v i z . , 'Roughly speaking', as well as by everything else that he has ever said on the subject, including the remark made only a week l a t e r , on 25*10.14, when he says: "With the l o g i o a l constants one need never ask whether they e x i s t , for they can even vanish'" (where he i s presumably a l l u d i n g to t h e i r i n t e r - d e f i n a b i l i t y , i n whioh, e»g«» vanishes when 'poq' becomes *p v ~ q ' ) . As for Anscombe's other oharge—viz., that NB allows, but the T denies, a 'Bedeutung' to p r o p o s i t i o n s—it i s true that i n the T KBedeuting' i s used only i n connection with objeots (except 54 i n 5.31, where 'Bedeutung' i s usea i n a wiae, non-technical sense); t h i s terminological r e s t r i c t i o n , however, i s not intended to r e -pudiate the notion that a proposition can stand for something, but only to emphasize that Mie r e l a t i o n of a name to that for which i t stands—viz., the o b j e c t—is d i f f e r e n t than the r e l a t i o n of a proposition to that for which i t stands—viz., the f a c t— since the proposition sucoeeds i n having something fo r whioh i t stands only when i t i s true. Hence, Anscombe's remark about the 'great contrast' between NB and the T with regard to t h i s question of the 'Bedeutung' of l o g i c a l constants and propositions cannot be taken s e r i o u s l y . This example i l l u s t r a t e s the danger of treating Wittgenstein's remarks i n i s o l a t i o n and merely at face-value, instead of trying to f i t them into t h e i r h i s t o r i c a l oontext. My own exegetioal approach has been to provide just such a background for under-standing the ontology of the T. The r e s u l t s of this approach can be b r i e f l y r ecapitulated: we have seen that the i n i t i a l realism of (1) i s questioned i n (2) by Wittgenstein's r e a l i z a t i o n that symbols f o r universals cannot perform th e i r function apart from the oontext of a pro p o s i t i o n , which he takes as a reason for distinguishing symbols into those dealing exclusively with contents and those dealing exclusively with forms. (3) i s his r e a l i z a t i o n that t h i s dichotomy i s unwarranted, since symbols for universals concern a oontent, as well as a form, and symbols for p a r t i c u l a r s concern a form as well as.a content.1 •kjntil now, the d i s t i n c t i o n between Wittgenstein's three positions has been made predominantly at the l e v e l of language 55 Now, my contention i s that the ontology of the T neither reverts to the professed nominalism of (2), nor moves on to a new nominalistio p o s i t i o n beyond (3), but i s simply a statement of (3) i t s e l f . If I am r i g h t i n my assessment of the importance of the conclusions of (3) i n providing a d e f i n i t i v e solution to the problems which had plagued Wittgenstein nearly through the end of 1914, we should expect to find these conclusions promi-nently displayed i n the T. And this i s pr e c i s e l y what we f i n d . The middle of the very f i r s t page of the text through the be-ginning of the third page, from 2.01 to 2.02, i s devoted exclu-s i v e l y to setting out the r a d i c a l idea that objects are essen-t i a l l y dependent, i n that they are not merely a content but also a form.l Thus, a f t e r t e l l i n g us that a state of a f f a i r s i s a rather than the l e v e l of ontology, i . e . , as concerning symbols rather than the e n t i t i e s symbolized. This emphasis on language has been to f a o i l i t a t e the exposition and not neoessarily to r e f l e c t Wittgenstein's own p r i o r i t y of i n t e r e s t s . That i s , the l i n g u i s t i c d i s t i n c t i o n between dependent and Independent sym-bols i s the common denominator according to whioh a l l three positions oan be characterized, whereas an ontologioal d i s -t i n c t i o n between dependent and independent e n t i t i e s oan be straightforwardly invoked only to contrast (1) and (3) , since at l e a s t on Wittgenstein's own view of (2) nothing at a l l cor-responds to dependent symbols on the ontologioal l e v e l . Although we avoid complications by keeping the comparison of the three positions at the l e v e l of language, the formulation of (3) i n the T i s i n f a c t made primarily at the l e v e l of ontology. Uni-versals and p a r t i c u l a r s are both dependent objects, which are therefore both represented by dependent symbols. Accordingly, the following discussion of the T w i l l focus on the ontologioal l e v e l , even though (3) oontinues~"to admit of an adequate characterization s o l e l y at the l e v e l of language. •^Actually, the status of objects as contents i s not even mentioned u n t i l 2.025» so exclusively i s the i n i t i a l focus on t h i s idea of an object's form. 56 combination (Verbindung) of objects (2.01), the f i r s t thing Wittgenstein wants to t e l l us about objects i s that i t i s esse n t i a l to an object to be able to be a constituent of a state of a f f a i r s (2.011), whioh he immediately supplements by saying that In l o g i c nothing i s acc i d e n t a l : i f a thing can occur i n a state of a f f a i r s , the p o s s i b i l i t y of the state of a f f a i r s must be written into the thing i t s e l f (2.012). These passages make clear from the s t a r t that objeots are not the sort of self - s u b s l s t e n t e n t i t i e s that Russell—or Wittgen-s t e i n i n (1) and (2)—had assumed them to be. The nature of t h i s dependence of objects i s then elaborated i n the following passages, which are oomments on 2.012: I f I can think of objeots combined i n states of a f f a i r s , I cannot think of them excluded from the p o s s i b i l i t y of such combinations (2.0121). Things are independent i n so far as they can occur i n a l l possible s i t u a t i o n s , but t h i s form of independence i s a form of connection with states of a f f a i r s , a form of dependence. (It i s impossible for words to appear i n two d i f f e r e n t r o l e s : by themselves, and i n propositions.) (2.0122) I f I know an object I also know a l l i t s possible occurrences i n states of a f f a i r s . (Every one of these p o s s i b i l i t i e s must be part of the nature of the object.) (2.0123) The key passage here f o r our purposes i s 2.0122, whioh e x p l i c i t l y a r t i c u l a t e s t h i s notion of dependence, saying i n ef f e c t that the dependence of objects—which consists i n th e i r a b i l i t y to occur only i n the context of a state of a f f a i r s—ha s i t s p a r a l l e l i n the dependence of symbols, which have meaning only i n the con-text of a proposition (this point about symbols being made again at 3 . 3 ) . He concludes th i s c r u c i a l section by s p e c i f i c a l l y i d e n t i f y i n g that aspect of an object whioh makes for i t s depend-57 ence as the object's form ( 2 . o l 4 l ) . Now, although I t oould be argued that t h i s doctrine of the dependence of objects does not by i t s e l f show that properties and r e l a t i o n s are regarded as objects, I have shown that i n NB t h i s strategy of extending the notion of form to p a r t i c u l a r s coincides with the admission of oontent to u n i v e r s a l s , and f a c i l i t a t e s the r e - a s s i m l l a t i o n of universals with p a r t i c u l a r s . Under these circumstances, i t would be remarkable, to say the l e a s t , i f Wittgenstein were to r e t a i n t h i s sophisticated and ingenious outlook o r i g i n a l l y developed to bridge the gap between universals and p a r t i c u l a r s , only to throw away the f r u i t s of t h i s solution for a re-affirmation of nominalism. This i s strong circumstantial evidenoe i n favor of the view that the T includes universals among objects. Most com-mentators, nevertheless, have adopted the view that the T holds that properties and r e l a t i o n s are merely the way i n which p a r t i c -u lars are arranged—i.e., that p a r t i c u l a r s are the only objects and that properties and r e l a t i o n s are their configuration—and they generally defend t h i s view i n a manner which suggests that there are l i t t l e or no textual grounds i n the T i t s e l f for any other i n t e r p r e t a t i o n . Keyt, for example, admits that NB treats universals as objeots, but then says " . . . I can find no pas-sage i n the Traotatus that either says or implies t h i s . " l •^David Keyt, "Wittgenstein's Notion of an Object," i n ET, p. 290. Keyt has informed me that his t o t a l opposition to a r e a l i s t i c i n t e r p r e t a t i o n of the T, as found i n this a r t i c l e as well as i n "Wittgenstein's Picture Theory of Language," (also i n ET), mellowed somewhat i n his subsequently written review, \ 58 Pitoher Is only s l i g h t l y more generous i n conceding that one passage i n the T_, 4.123, has the appearance of advocating that universals are included among objects, though Pitoher argues that even th i s appearance i s misleading, and he concludes: Everything else [^Wittgenstein} says i n the Traotatus seems to me to require that objects be p a r t i c u l a r s , that they not include universals.1 These are bold claims, and although not a l l proponents of a nominalistic i n t e r p r e t a t i o n go to t h i s extreme—e.g., Mueller admits that quite a few passages at l e a s t s u p e r f i o i a l l y (though, as i t turns out, not ultimately) lend themselves to a r e a l i s t i n t e r p r e t a t i o n2—the prevalent opinion i s that the textual evidenoe i s , at any r a t e , preponderantly weighted i n favor of nominalism. If this prevalent opinion were anywhere near being oorreot, my own argument would have to r e s t largely on the circumstantial evidence*,already given. In f a c t , however, not only the circum-s t a n t i a l evidence, but also the textual evidenoe d e f i n i t e l y supports a r e a l i s t i c i nterpretation and poses insuperable d i f -f i c u l t i e s for the view that universals are merely configurations of objeots and not objects themselves. In the next chapter, I s h a l l oonsider textual evidenoe from the £ i t s e l f , as well as from the PT and post-2 w r i t i n g s , which supports realism. F i n a l l y , "A New Interpretation of the Traotatus Examined," Philosophioal  Review 74 (1965). My o r i t i o i s m of Keyt w i l l be confined to the views advanced i n his two e a r l i e r a r t i c l e s . •'-George P i t c h e r , The Philosophy of Wittgenstein, p. 114. 2Mueller, p. 82ff. 59 I s h a l l conclude Part One by examining, i n Chapter IV, passages whioh might be taken to support nominalism, but whioh—as I s h a l l show—do not r e a l l y provide any such support. CHAPTER III ARGUMENTS FOR A REALISTIC INTERPRETATION The f i r s t evidence whioh I offer i n favor of a r e a l i s t i c i n t e r p r e t a t i o n of the T concerns those passages i n which Witt-genstein says that a state of a f f a i r s i s composed of objects (2.01, 2.0272), or—to make the same point at the l e v e l of language—that an elementary proposition consists of names (4,22). The use of the p l u r a l implies that i n a proposition of the subject-predicate form, e.g., ' f ( a ) ' , both ' f ' and 'a' are names,1 each of which symbolizes an object—a property and a p a r t i c u l a r , r e s p e c t i v e l y . I f an opponent of the r e a l i s t i c i n t e r p r e t a t i o n concedes this i m p l i c a t i o n , he can evade i t s consequences only by maintaining that according to the T there are no propositions of the subject-predicate form, and that propositions whioh seem to concern only one p a r t i c u l a r are r e a l l y r e l a t i o n a l propositions oonoerning two or more p a r t i c -u l a r s . This view, advocated notably by Anscombe, i s given a concise statement by another advocate, Pitchers Wittgenstein could p e r f e c t l y well have held that an object's having a property i s not a matter of i t s being con-figured with a u n i v e r s a l , but rather of i t s being configured with other simple p a r t i c u l a r s . He could have maintained, for instance," that to say that a i s red i s not to say that a i s configured with the universal "redness", but rather to say that a i s configured with b, c, and d — a l l simple p a r t i c u l a r s -•LWhen I write a function-sign as a single l e t t e r , say ' f ' i t should be understood that a s t r i c t l y correot representation inoludes a set of empty parentheses and/or a variable to indicate the argument-place. 61 i n a oertain way.-1 On thiB view, the ' f i n ' f ( a ) ' has the dual-purpose of repre-senting the three p a r t i c u l a r s b, o, and d, as well as represent-ing the configuration of these p a r t i c u l a r s . This supposition, v i z . , that the symbol for a property i n an ostensibly subject-predicate proposition can represent p a r t i c u l a r s other than the one e x p l i c i t l y named i n the argument-place, i s probably one reason why Anscombe fe e l s no need to confine the r o l e of the symbol for a r e l a t i o n , i n a proposition suoh as 'aRb' which i s ostensibly r e l a t i o n a l , to that of symbolizing the configuration only of the e x p l i c i t l y named objects a and b. l i k e the 'f* i n ' f ( a ) ' , the 'R' i n 'aRb' might equally well represent other p a r t i c u l a r s i n addition to the configuration of a l l of them, so that while • • the objeots a and b oocur ' i n the sense' of the proposition 'aRb' . . . 50 or 1000 or an i n f i n i t y of other objects may oocur i n that sense as w e l l . "2 •••Pitoher, p. 117. 2G. E. M. Ansoombe, "Mr. Copi on Objects, Properties, and Relations i n the 'Traotatus'," i n £ £ , p. 187. In her book on the T, Ansoombe attempts to support the claim that 'aRb' might represent i n f i n i t e l y many objeots by arguing that, although Wittgenstein speaks of propositions of the form V ( x ) ' or *tf(x, y ) ' i n 4.24, ". . . i t must not be supposed from this that Wittgen-s t e i n intends V ( x , y)* to represent an atomic fact consisting of three objeots. He has only Just remarked (4.2211): 'Even i f the world i s i n f i n i t e l y oomplex, so that every faot consists of i n f i n i t e l y many atomic facts and every atomio fact i s composed of i n f i n i t e l y many objects, even so there must be objeots and atomic f a o t s . ' So when he writes V ( x , y ) * , nothing whatever i s indicated about how many names may be covered by the sign Of the funotion; there might, on the hypothesis that he has Just mentioned, be an i n f i n i t e number" (Ansoombe, A_n Introduction to  Wittgenstein's Traotatus, p. 99). I would argue, on the other hand, that the only conclusion we oan draw from 4.2211 i s that there may not be propositions of the form ' f ( x ) ' or V ( x , y ) ' . I t 62 This interpretation of the T., however, whioh allows that symbols for funotions also oovertly represent p a r t i c u l a r s , can-not p l a u s i b l y be maintained. F i r s t of a l l , the idea that a proposition such as ' f ( a ) ' might r e a l l y symbolize a configura-t i o n of the p a r t i c u l a r s a, b, o, and d, i s wholly a l i e n to Wittgenstein's desire that his symbolic language—his B e g r l f f -s o h r l f t—be perspiouous. I f we consider his meticulous e f f o r t to capture i n his symbolism, without recourse to the sign of i d e n t i t y , what we mean by saying, e.g., 'Only one x s a t i s f i e s f ( )'—his rendering being ' (3x)fx:~C3x, y)fx.fy'—the whole s p i r i t of t h i s pains-taking enterprise would seem to be mooked i f he r e a l l y means, but neglects to mention, that the one par-t i c u l a r s a t i s f y i n g this funotion i s of course joined by the i s merely taken for granted i n 4.24 (that there are subjeot-predioate propositions and r e l a t i o n a l propositions of two terms, but as Wittgenstein admits i n 5»55* "Since . . . we are unable to give the number of names with d i f f e r e n t meanings, we are also unable to give the composition of elementary propositions." A l s o , "It would be completely a r b i t r a r y to give any s p e c i f i c form" (5»554). That i s , he does not claim to know that there are propositions of the form V ( x , y ) ; for a l l he knows, every proposition might be a r e l a t i o n a l proposition with i n f i n i t e l y many terms, the point of 4.2211 being only that even i f thi s i s the case, the prooess of analysis must f i n a l l y lead us to objeots and states of a f f a i r s . S i g n i f i c a n t l y , the version of 4.2211 found i n the PT puts the matter hypothetioally, when i t says that even i f the world were i n f i n i t e l y complex, e t c , there would s t i l l have to be objeots and states of a f f a i r s . This shows even more d e a r l y than i n the case of the T that PT 5«4103 (which corresponds to 4.2211 of the T) is~not intended to imply that the forms mentioned i n PT 4.2212 (whioh corre-sponds to the relevant part of T. 4.24) might actually represent i n f i n i t e l y many objects. A l s o , we may observe that the close proximity i n the T of 4.2211 to 4.24, which Ansoombe uses to j u s t i f y her in t e r p r e t a t i o n of 4.24, does not oocur i n the PT, where the two remarks occur i n ent i r e l y d i f f e r e n t contexts. I conclude that Ansoombe has no basis for supposing that Wittgen-s t e i n regards a proposition of the form V ( x , y) as involving something more than two p a r t i c u l a r s and a r e l a t i o n . 63 three other p a r t i c u l a r s (or.whatever number) symbolized by ' f ' i t s e l f . That Wittgenstein could a t t r i b u t e any such symbolic r o l e to ' f ' i s extremely unlikely i n l i g h t of his claim that a proposition i s a p i c t u r e , with the same l o g i o a l m u l t i p l i c i t y as the state of a f f a i r s i t represents. Granted, we could learn to use t h i s one poor symbol ' f ' to express the fact that, be-sides the p a r t i c u l a r represented i n the argument-plaoe, there are three other p a r t i c u l a r s—b , o, and d—configured thus and so, but i f the pi c t u r e theory amounts to no more than t h i s , It seems a gratuitous gesture to bother dividing a proposition into function and argument at a l l . Why not symbolize the whole prop-o s i t i o n by a single l e t t e r ? Even supposing that we could reconcile such a symbolic r o l e for * f with Wittgenstein's desire for a perspicuous B e g r i f f - s o h r i f t , there i s the problem of explaining the unique status of the one p a r t i c u l a r which happens to be the argument to th i s f unction. If 'f(a ) * symbolizes a configuration of a, b, c, and d, what i s the symbolic import of regarding one of these p a r t i c -ulars as the argument rather than any of the others? We might suppose that there i s no i n t r i n s i c difference between a and the other p a r t i c u l a r s , so that *f(a)* could be equivalent to some other proposition 'g(b)', where *g' symbolizes a, c, and d, i n addition to symbolizing the same configuration of a l l four par-t i c u l a r s—t h e only difference between 'f(a)' and 'gCb)1 being the manner i n which the r e s p o n s i b i l i t y for symbolizing the four p a r t i c u l a r s i s divided up between the function-sign and the argument-sign. I f , however, we deny any special symbolic import 64 to the e x p l i c i t occurrence of a p a r t i c u l a r i n the argument-pl a c e , and say that the same p a r t i c u l a r could just as well have been symbolized by a function-sign, we remove any reason for the d i s t i n c t i o n of propositions of the form V ( x ) ' from those of the form V ( x , y ) ' ; i . e . , propositions of either form sup-posedly symbolize a configuration of more than one p a r t i c u l a r , so there i s no point at a l l i n d i f f e r e n t i a t i n g the two forms, as Wittgenstein does i n 4.24, unless the e x p l i c i t mention i n the argument-place of two p a r t i c u l a r s , i n the one case, and of only one p a r t i c u l a r , i n the other, r e f l e c t s some specia l feature of the p a r t i c u l a r s so honored. The T offers no clue as to what th i s special feature might be, which i s not r e a l l y s u r p r i s i n g , since neither i s there anywhere a suggestion that function-signs can represent p a r t i c u l a r s at a l l . The whole problem i s a bogus one, easily dissolved by the recognition that the function-sign and the argument-sign i n a subject-predicate proposition each represent a single object—a universal and a p a r t i c u l a r , respec-t i v e l y . D i f f i c u l t i e s In the nominallstic account of functions do not end here. The view whioh attributes to the function-sign the dual-purpose of representing p a r t i c u l a r s and their c o n f i g -uration has the additional curious consequence that, except for p a r t i c u l a r s e x p l i c i t l y represented i n the argument-place, we can-not generalize the p a r t i c u l a r s without generalizing the config-u r a t i o n . Thus, i f ' f ( a ) ' represents a configuration of a, b, c, and d, the symbolism does not permit a representation of t h i s s p e c i f i c configuration apart from a representation of these 65 s p e c i f i c p a r t i c u l a r s , except for the p a r t i c u l a r representee! by the l e t t e r ' a1, whioh can be replaced by an apparent variable while ' f ' remains constant. I f we try to generalize the remain-ing three p a r t i c u l a r s by turning ' f ' into an apparent v a r i a b l e , we thereby abstract from the s p e c i f i c configuration as w e l l . But surely the most elementary prerequisite of Wittgenstein's symbolism i s that its should be able to represent a constant configuration without representing any s p e c i f i c p a r t i c u l a r . For t h i s reason, he could not have tolerated a symbolism giving t h i s dual-role to funotion-signs, and that he does r e f e r , with-out qualms, to propositions of the form V ( x ) ' shows that V cannot be intended to designate both the configuration and par-t i c u l a r s i n the configuration. At best, V symbolizes a con-f i g u r a t i o n , but since only one p a r t i c u l a r can be symbolized i n V ( x ) ' , we are led to conclude that V designates neither any p a r t i c u l a r nor a configuration, but rather designates a prop-e r t y , i . e . , an object whioh i s a u n i v e r s a l . This problem, i t i s true, need a r i s e only for propositions of the form 'jrf(x)*, since i n propositions which are overtly r e l a t i o n a l , i . e . , which have two or more argument-places, we can suppose that the function-sign i s concerned solely with representing the configuration and that a l l the p a r t i c u l a r s occur e x p l i c i t l y as arguments. So, at least for propositions of the form V ( x , y,...n)', the p a r t i c u l a r s can be generalized while the configuration remains constant. But even i f we con-f i n e ourselves to r e l a t i o n a l propositions, there remains a f u r -ther problem for the view that funotion-signs represent oonfig-66 urations of p a r t i c u l a r s , v i z . , the problem of explaining the nature of the form-content d i s t i n c t i o n as suoh. That i s , i f , i n the proposition *R(a, b )1, 'a* and *b' represent the content and *R' represents the form,^" then the form, i n t h i s sense, would s t i l l seem to have some kind of content. Suppose, e.g., 'R(a, b ) ' expresses the fact that a i s to the r i g h t of b, and 'S(a, b )1 expresses the fact that b i s to the r i g h t of a. Must we not say that *R* and 'S1 thereby represent d i f f e r e n t contents, just as surely as these l e t t e r s would i f they were conceded to have the status of names? The form of a proposition i s what remains when we abstraot from the content by turning constants into v a r i a b l e s , and from this i t should be obvious that a function-sign cannot be regarded by Wittgenstein as represent-ing a form, since he allows that function-signs, too, oan be turned from oonstants into v a r i a b l e s , which implies that such signs represent a content no l e s s than do signs f o r p a r t i c u l a r s . Surely, the form of a state of a f f a i r s i s mirrored by the form of the proposition and not represented by i t s own i n d i v i d u a l symbol, so that the same form continues to be represented ISome commentators would say that *R* represents the s t r u c - ture, and not the form, but I would oontend that t h i s i s , i n any case, a mistake, sinoe the structure of a f a c t i s determined, not only by the way the objects are configured, but also by whioh objeots are i n configuration, so that *aRb* and 'cRd* would be propositions of the same form but of d i f f e r e n t s t r u c -tures. To t h i s extent I agree with MoGuinness (though he s t i l l goes on to contend that, i n 'aRb', 'R* represents the form) when he says: "A fact and i t s p i c t u r e may have the same form (must have, indeed) but cannot have the same struoture. Each fact that we are aware of has i t s own structure: that these objeots stand i n t h i s arrangement constitutes t h i s structure1' (B. F. MoGuinness, "Pictures and Forms i n Wittgenstein's 1 Traotatus1," i n ET, pp. l44-145). . ~ " 67 whether the funotion-sign remains constant or i s turned into an apparent v a r i a b l e . I t i s Wittgenstein's consideration of thi s t o p io, v i z . , the p o s s i b i l i t y of generalizing function-signs, which provides the oontext for a remark i n the T whioh cl e a r l y implies that function-signs are names of objects. In 5*526 he says that We can describe the world completely by means of f u l l y generalized propositions, i . e . without f i r s t c o r r e l a t i n g any name with a p a r t i c u l a r object. Then, i n order to a r r i v e at the customary mode of expres-s i o n , we simply need to add, a f t e r an expression l i k e , 'There i s one and only one x suoh that...', the words, 'and that x i s a'. The second paragraph, which shows the replacement of an apparent variable by a name only for the case of 'a', a l e t t e r custom-a r i l y used to represent an argument ( i . e . , a p a r t i c u l a r ) , i s s l i g h t l y misleading, because the point of talking about com-p l e t e l y generalized propositions—as i s made clear by 5.526I as well as by the comparable discussion of completely generalized propositions i n NB (p. l l f f . ) — i s that i n suoh propositions even the function i s generalized. Thus, we can desoribe the world not merely through propositions suoh as '(3x)fx* or '(x)fx' (these being the examples just previously given i n 5*521 of propositions embodying g e n e r a l i t y ) , where the function remains constant, but, as 5*5261 shows, we can describe the world even through propositions such as 'Ox, jrf)jrfx', i n which a l l oonstants are generalized. The po i n t , then, of the statement i n 5.526 that the description of the world can take place 'without f i r s t c o r r e l a t i n g any name with a p a r t i c u l a r object' i s surely to stress that even those objects which are univer s a l s , and thus 68 represented by function-signs, can be represented by apparent variables rather than by constants. I t i s i n s t r u c t i v e to com-pare 5*526 to the passage whioh i s i t s counterpart i n NB; Yes, the world could be completely described by completely general propositions, and hence without using any names or any other denoting signs. And i n order to a r r i v e at ordinary l a n -guage one would only need to - introduce names, etc. by saying, a f t e r an * (3x)*, "and thi s x i s A" and so on (17.10.14). In this passage, the expressions 'other denoting signs' arid ' e t c ' are meant to cover the case where functions are general-i z e d , since at t h i s stage of NB Wittgenstein i s not yet ready to accord function-signs the status of names (even though suoh signs are admitted to represent obj e c t s ) , a soruple which our examin-at i o n of NB has shown him to abandon soon afterwards. By the time the T i s wri t t e n , he f u l l y accepts function-signs as being names; hence, i n 5*526, the q u a l i f y i n g expressions found i n the comparable passage i n NB are no longer needed. We may conclude, then, that 5*526 constitutes incontrovertible evidence—about as straightforward as one could ever hope to glean from the cr y p t i c prose of the T—that objeots inolude un i v e r s a l s . I want now to examine some passages whioh show that Wittgen-s t e i n regards universals as objeots by showing how he proceeds to effeot t h e i r r e - a s s i m i l a t i o n with p a r t i c u l a r s . A common view i s that even i f we grant that universals and p a r t i c u l a r s are both speoies of objeots, they are at any rate quite d i f f e r e n t kinds of objects, and commentators have used t h i s view to sup-port a nominalistio i n t e r p r e t a t i o n of the T by arguing that, i f Wittgenstein had meant by 'objects* things as r a d i c a l l y d i s t i n c t 69 as universals and p a r t i c u l a r s , he would have said so.1 To be sure, I hope to have shown i n my previous arguments that he does say, or at l e a s t c l e a r l y imply, that universals are i n -cluded among objeots, but i n any case suoh an objection to the r e a l i s t interpretation misses the p o i n t , which i s that he wants to show that universals and p a r t i c u l a r s are not r a d i c a l l y d i s -t i n c t . I have suggested that t h i s re-assimilation i s aohieved by Wittgenstein's view that a l l objeots have both form and con-tent, so that the d i f f e r e n t syntactic roles enjoyed by the names 'Socrates' and 'mortality', supposing f o r purposes of i l l u s t r a t i o n that they are names of simple objects, can be explained at the most comprehensive l e v e l simply by saying that the objects repre-sented by these names have d i f f e r e n t forms, this being an explan-ation which covers not only the differenoe between universals and p a r t i c u l a r s but also differences between objeots within each cat-egory. On this account, a d i v i s i o n of objects into universals and. p a r t i c u l a r s ceases to be of much s i g n i f i c a n c e , except per-haps to d i s t i n g u i s h those objeots whose forms share the common ch a r a c t e r i s t i c of permitting suoh objects to oocur only as l o g i c a l subjects from those objeots whose forms determine that such objects occur predioatively or r e l a t i o n a l l y . I t might be s a i d , however, that even apart from their syntaotic roles there i s another r e -spect i n which universals and p a r t i c u l a r s d i f f e r r a d i c a l l y , which i s that while at l e a s t some p a r t i c u l a r s are looatable i n space •^ See Pitoher, pp. 116-117, and also Ansoombe, Introduction, p. 109, where she expresses the contrast as being between functions and i n d i v i d u a l s . 70 and a l l p a r t i c u l a r s are locatable i n time—unless we want to make an exception for instants themselves—universals would seem to exist i n a world apart, outside of space and time. R u s s e l l , i n PP, expresses the difference as follows: The 'idea' .justice i s not i d e n t i c a l with anything that i s just: i t i s something other than p a r t i c u l a r things, whioh p a r t i c u l a r things partake o f . Not being p a r t i c u l a r , i t cannot i t s e l f exist i n the world of sense. Moreover i t i s not f l e e t i n g or changeable l i k e the things of sense: i t i s eternally i t s e l f , immutable and i n d e s t r u o t i b l e . l Russell feels the difference between objects which exist i n time and those whioh exist outside of time to be Elufficiently great that he prefers to say that objeots of the l a t t e r kind, v i z . , u n i v e r s a l s , do not exist at a l l ; rather, they subsi s t , or have being. Thus, universals and p a r t i c u l a r s comprise, as i t were, two d i s t i n c t worlds: The world of being i s unchangeable, r i g i d , exact . . . The world of existence i s f l e e t i n g , vague, without sharp boundaries, without any clear plan or arrangement . . • .2 Suoh a b i f u r c a t i o n of universals and p a r t i c u l a r s , as pro-posed by R u s s e l l , i s indeed an obstacle to the programme for thei r mutual r e - a s s i m i l a t i o n . Therefore, given my contention that this i s Wittgenstein's programme, i t i s not surprising to find that the T addresses i t s e l f to t h i s problem, and hardly more surprising that the language i n which Wittgenstein's own solution i s framed shows that he probably has Russell's views s p e c i f i c a l l y i n mind.^ Thus, against the view that universals are immutable and p a r t i c u l a r s are f l e e t i n g and changeable, •"•PP, P. 92. 2PP, p. 100. ^For additional arguments that Wittgenstein actually read PP, see Appendix B. 71 Wittgenstein says that The permanent (Feste), the existent (Bestehende), and the object are one (2,0277^ The object i s what i s permanent, existent; the c o n f i g -uration i s what i s changing, unstable (2.0271). In other words, Russell's characterization of universals i s extended by Wittgenstein to universals and p a r t i c u l a r s a l i k e . That i s , not only u n i v e r s a l s , but p a r t i c u l a r s as w e l l , are out-side of space and time. To see how thi s i s p o s s i b l e , we need only consider Wittgen-stein's remark that objeots are colorless (2.0232), for i n this respect a p a r t i c u l a r ' s status of having a s p a t i a l and temporal loc a t i o n i s on a par with i t s status of being oolored. Suppose, then, that there i s a state of a f f a i r s i n whioh a p a r t i o u l a r , which we w i l l c a l l 'John', i s white. Now, according to 2.0231 there i s a sense i n which we can say that John i s c o l o r l e s s . This seemingly enigmatic remark becomes re a d i l y understandable i f we look at 2.0231, whioh says: The substance of the world can only determine a form, and not any material properties. For these material properties are represented only by propositions, formed only by the configuration of objects. In other words, our acquaintance with John does not include know-ledge of John's whiteness, whioh i s , a f t e r a l l , only an external property whioh John could cease to have without ceasing to be John. What we do know, as Included i n John's form, i s that John i s the sort of object that must have a c o l o r , though which one i s something that can only be determined by extending our know-ledge beyond the single object John, with i t s p o s s i b i l i t i e s of configuration with other objects, to the state of a f f a i r s i n 72 whioh John i s actually configured with whiteness. Sinoe John and whiteness are two separate objects, each having i t s own form and content, obviously neither content i s part of the other, and so John i s i n a sense c o l o r l e s s ; at most, i t i s part of John's nature to be required to be configured with some color or other and to be able to be configured s p e c i f i c a l l y with whiteness as one of the options. Applying what has been said about color to space and time, i f John i s a s p a t i a l and temporal object (and i f i t i s the former, i t must be the l a t t e r as w e l l ) , t h i s means that John's form requires that i t be con-figured with some s p a t i a l point and some temporal instant, though whioh s p e c i f i c point and instant this should be i s exter-nal to John's nature and i s determined only by the actual con-f i g u r a t i o n of objeots. In thi s sense, just as John i s c o l o r l e s s , so i s John outside of space and time. Therefore, when Wittgenstein characterizes objeots as being permanent or unalterable, he does not mean that they continue to exist unchanged throughout every moment of time, but rather that even i f an object should exist only for the b r i e f e s t temporal span, i t i s eternal i n the sense that whatever r e l a t i o n i t has to any p a r t i c u l a r time i s external to i t s own nature. This i s not to deny that there may also be objeots which are eternal i n the sense of p e r s i s t i n g throughout every moment of time—and Wittgenstein may well have thought that the material points of physios were just suoh objects1—but the permanence which he i c f . PM, #440. 73 attributes to objects generally, i n 2.027, i s of a d i f f e r e n t , more fundamental kind. By this strategy, then, the basis for distinguishing between universals and p a r t i c u l a r s as r a d i c a l l y d i f f e r e n t kinds of objects i s removed. Although I think that my account of Wittgenstein's strategy for dealing with space and time does provide evidence that he regards universals as objeots, by showing that his programme i s indeed to assimilate universals with p a r t i c u l a r s and, i f any-thing, to model the l a t t e r on the former, I admit that the ao-oount, as i t stands, begs the question insofar as i t r e l i e s on the example of color to i l l u s t r a t e the strategy of separating a s p a t i a l and temporal object from the points and instants with whioh i t can be configured. It would be desirable i f independent evidence could be found for this treatment of space and time; suoh evidenoe would not only make possible an explanation of Wittgen-stein's assimilation of p a r t i c u l a r s with universals without pre-supposing that colors are objeots, but would i n addition oonfirm that this interpretation of his remark about colorless objects i s correct, since we have reason to suppose that his treatment of color p a r a l l e l s his treatment of space and time, given that space, time, and color (being colored) are grouped together i n 2.0251 as being forms of objects. The one difference w i l l be that while those objeots whioh are capable of being configured with a p a r t i c u l a r i n v i r t u e of i t s s p a t i a l and temporal form w i l l themselves be (or at least function as) p a r t i c u l a r s , v i z . , points and in s t a n t s , those objeots capable of suoh a conf i g -uration i n v i r t u e of the p a r t i c u l a r ' s form of coloredness w i l l 74 be universals, v i z . , c o l o r s . Therefore, the task at hand i s to produce evidenoe for supposing that an object's s p a t i a l and temporal form amounts to i t s c o n f i g u r a b i l i t y with points and i n s t a n t s . This evidenoe, I think, oan be found i n 2.0121: Just as we are unable to think of s p a t i a l objects outside of spaoe, or temporal objeots outside of time, so i t i s that we oan think of no object apart from the p o s s i b i l i t y of i t s com-bination with others. C l e a r l y , the reference to space and time i s here intended as a s p e c i f i c i l l u s t r a t i o n of the general notion of an object's form, i . e . , i t s p o s s i b i l i t y of combining with other objeots, from which we can only conolude that just suoh a p o s s i b i l i t y of com-bination with other objeots i s indicated by saying that the object i s s p a t i a l or temporal. What could these other objeots be, i n the case of s p a t i a l and temporal objects, i f not points and i n s t a n t s ?1 It i s i n s t r u c t i v e to oompare my interpretation of Wittgen-stein's views on spaoe and time with the following passages from PM: Among terms which appear to e x i s t , there are . . . four great classes: (1) i n s t a n t s , (2) p o i n t s , (3) terms whioh occupy instants but not p o i n t s , (4) terms whioh occupy both points and i n s t a n t s .2 Change i s due, u l t i m a t e l y , to the fact that many terms have re l a t i o n s to some parts of time whioh they do not have to others. But every term i s e t e r n a l , timeless, and immutable; the r e l a t i o n s i t may have to parts of time are equally immutable. I t i s merely xWe may note, i n passing, that no matter what Wittgenstein means by an object's being temporal, the notion of a temporal object i s pleonastic, unless there are objects which are not tem-p o r a l . The nominalist interpretation allows us to suppose only that there are two kinds of particulars—temporal and non-temporal— while the r e a l i s t i nterpretation provides us with a more pl a u s i b l e candidate f o r a non-temporal object, v i z . , a u n i v e r s a l . 2PM, #438. 75 the fact that d i f f e r e n t terms are related to d i f f e r e n t times that makes the difference between what exists at one time and what exists at another. And though a term may cease to e x i s t , i t cannot cease to be • • • .1 Prom these passages we may reasonably conclude that Wittgenstein's treatment of space and time i n r e l a t i o n to p a r t i c u l a r s which are i n space and time d i r e c t s i t s e l f against Russell's p o s i t i o n i n 2 PP by adopting the viewpoint of Russell's e a r l i e r work. The influence of PM on the T i s also apparent i n the related problem of the individuation of objects. According to PM, an object's properties or r e l a t i o n s to other objeots, including i t s l o c a t i o n i n space and time, cannot ultimately serve to d i f f e r -entiate one object from another object, because the properties and r e l a t i o n s of thi s object are external to i t s essential nature and are, indeed, objeots i n their own r i g h t . Russell i s thus forced to say that ". . . any two simple terms simply d i f -f e r immediately—they are two, and this i s the sum t o t a l of thei r differences."3 This notion that there i s no basis for distinguishing objects apart from their being d i f f e r e n t i s taken over by Wittgenstein, with one modifications even when we con-XPM, #443. 20ne could argue that PP, though changing the emphasis of th i s e a r l i e r p o s i t i o n , i s not r e a l l y a change of view, sinoe, i n PP, Russell also characterizes space and time as a c o l l e c t i o n of p a r t i c u l a r points and instants;(PP, p. 93 and 146) and char-acterizes the existenoe of an object as i t s r e l a t i o n to one or more instants (PP, p. 100). S t i l l , i n an a r t i c l e roughly con-temporaneous with PP, Russell speaks as i f p a r t i c u l a r s them-selves are mutable and i s w i l l i n g to count p a r t i c u l a r s as indes-t r u c t i b l e substances only i f they continue to p e r s i s t throughout time. (See "On the Relation of Universals and P a r t i c u l a r s , " p. 122.) Prom PM to PP, then, Russell's views on thi s matter have undergone considerable modification. 3PM. #428. 76 sider an object apart from i t s external properties, Wittgenstein recognizes that there are s t i l l i n t e r n a l properties which can dist i n g u i s h this object from others, so that the necessity of appealing solely to the 'immediate d i f f e r e n c e ' of two objects as two arises only when these objeots have the same int e r n a l prop-e r t i e s , i . e . , are of the same l o g i c a l form. Thus: Two objects of the same l o g i o a l form are—apart from t h e i r external properties—distinguished from one another only i n that they are d i f f e r e n t (2.0233). The PM, therefore, has evidently l e f t i t s mark on Wittgenstein's discussion of objeots i n the T, presumably extending to a com-parable treatment of points and ins t a n t s . A simple i d e n t i f i -cation of the two views, however, would be a mistake, because Wittgenstein also has had the benefit of reading OKEW,1 i n whioh Russell follows Whitehead i n regarding points and instants as l o g i o a l constructions rather than as simple objeots i n th e i r own r i g h t . So Wittgenstein at lea s t has been exposed to thi s way of thinking about spaoe and time, and he probably has thi s develop-ment i n mind when, i n NJB, he says: MA singular l o g i o a l mani-p u l a t i o n , the p e r s o n i f i c a t i o n of time!" followed almost imme-diately by the remark: MCan we regard a part of space as a thing? In a certain sense we obviously always do thi s when we talk of s p a t i a l t h ings."2 The f i r s t remark i s admittedly c r y p t i c , iFor a defense of this olaim, see Appendix B. 2These remarks are made on 13»5.15» and slnoe i t i s around t h i s time that a pre-occupation with topics discussed i n OKEW i s evidenced by various remarks i n NB (see Appendix B), we can plausibly suppose that OKEW inspires these remarks on space and time as w e l l . 77 but, taken In conjunction with the second remark, the general theme l i k e l y concerns the status of points and instants as e n t i t i e s . I take the •pe r s o n i f i c a t i o n1 of time to be that pro-cess whereby instants are given the appearance of being genuine e n t i t i e s , though they are r e a l l y only l o g i c a l constructions defined i n terms of events,1 whioh i s why Wittgenstein c a l l s this p e r s o n i f i c a t i o n a ' l o g i c a l manipulation*. What i s not clear i s whether, i n c a l l i n g this a singular (elgentuemlloh) l o g i o a l manipulation, he means to suggest, with a touch of irony, that instants cannot r e a l l y be defined i n thi s way at a l l , and that they just are simple objeots, or whether he admits that instants are l o g i c a l constructions and i s simply struck by the p e c u l i a r i t y of this process. The intent of the second remark i s also unclear, i n that we cannot be sure whether the phrase * i n a certain sense* i s meant to suggest that parts of spaoe--presumably, s p a t i a l p o ints— though treated as things, are r e a l l y l o g i o a l constructions, or whether he regards points as genuine objeots and inserts the qua l i f y i n g phrase only to allow that there are other contexts i n whioh i t would be odd to regard parts of space themselves as things, but that t h i s i s nevertheless the proper way to regard them when considering their r e l a t i o n to s p a t i a l objects. So, although these remarks about spaoe and time are quite l i k e l y prompted by his reading of OKEW. i n whioh the status of points and instants as e n t i t i e s i s questioned, we cannot t e l l from 1See NB, p. 93ff. 78 these remarks whether or not Wittgenstein himself i s w i l l i n g to admit that points and instants are possibly just constructions. A olue towards determining Wittgenstein's own outlook i s perhaps provided by a passage i n NB, following these f i r s t r e -marks by about a month, i n whioh he says: I f a point i n space does not e x i s t , then i t s co-ordinates do not exist e i t h e r , and i f the co-ordinates exist then the point exists too. That i s how i t i s i n l o g i c (21.6.15). This passage suggests that he may have regarded a point as a complex object constructed out of i t s co-ordinates, an i n t e r -pretation supported by the oontext, which i s a discussion of the functioning of complex objects as simples. He might be saying, then, that points ' e x i s t ' i n the sense that l o g i c oom-pels us to say, of any complex, that i t exists when i t s parts are appropriately assembled, even though this existenoe of the complex i s r e a l l y nothing over and above the existenoe of i t s parts (in th e i r appropriate c o n f i g u r a t i o n )—in this oase the existence of a s p a t i a l point being nothing more than the e x i s t -ence of i t s co-ordinates. Such a view of the nature of the com-p l e x i t y of p o i n t s— i f this i s indeed Wittgenstein's view (and we should not be too hasty to suppose that he seriously holds the admittedly peculiar view that co-ordinates exist as constituent o b j e c t s )—is at any rate far d i f f e r e n t from Russell's view i n OKEW of the nature of thi s complexity. Russell defines a s p a t i a l point ". . . as a cer t a i n class of s p a t i a l objeots, namely a l l those . . . which would naturally be said to contain the point • • • ,1 , 1 i . e . , • • a l l the objects whioh enclose members of 1OKEW, p. 92. 79 a given punctual enclosure-series,"1 This r e q u i r e s , ", . . to prevent t r i v i a l exceptions . , , , "2 that ", • • there are to be instances of enclosure, i . e . there r e a l l y are to be objects of which one encloses the other."3 A l l of this implies (and Indeed Russell e x p l i c i t l y adopts the view**) that space i s r e l a -t i v e to s p a t i a l objects; without suoh objeots there would be no poi n t s , and therefore no space. Wittgenstein, on the other hand, seems to treat space as being absolute, depending on no s p a t i a l object for i t s own existence, but rather, l i k e a graph, requiring only i t s oo-ordinates to define any s p a t i a l p o i n t . That he r e -gards space as absolute i s supported i n the T by 2.013: Each thing i s , as i t were, i n a space of possible states of a f f a i r s . I can think of this space as being empty, but I cannot think of the thing without the space. This assertion that spaced could be thought of as empty—whioh implies that i t i s l o g i c a l l y possible that space could be empty-i s incompatible with Russell's d e f i n i t i o n of s p a t i a l points i n OKEW and shows that i f Wittgenstein does go beyond Russell's e a r l i e r view of PM i n whioh points are straightforwardly regarded as simple objeots, the kind of complexity Wittgenstein allows points to have leaves his oonoeption of space much closer to this e a r l i e r , a b s o l u t i s t i c view than i s Russell's own subsequent, r e l a t i v i s t i o view i n OKEW. We may conclude, at any r a t e , that IQKEW. p. 92. 2pKEW, p. 93. 30KEW, p. 92. ^OKEW. p. 116. ^That his remark i s meant to apply to space i n the usual sense, and not just to a metaphorical ' l o g i c a l ' space, i s shown by the s p e c i f i c i l l u s t r a t i o n of 2.0131. 80 Wittgenstein does not hesitate to treat s p a t i a l p o i n t s , for a l l p r a o t i o a l purposes, as though they were objeots i n t h e i r own r i g h t . Even i f he might concede that, i n the l a s t a n a l y s i s , suoh an attitude towards s p a t i a l points r e f l e c t s a hypostati-z a t i o n , this i s a hypostatization i n whioh he nevertheless seems to be w i l l i n g to indulge.1 There i s one passage i n the T which could be taken to be evidence against my claim that s p a t i a l points are treated as objects. This i s the remark that "A s p a t i a l point i s an argument-place" (2.0131). If this passage i s construed as a remark about 1The only commentator who discusses the T's conception of space and time i n any d e t a i l i s Mueller, who contends that the T holds a r e l a t i v e , rather than an absolute, view of space and time. (See Mueller, Ontologie, p. 107.) The only evidence he offers from the T i t s e l f , however, i s 6.36II, whioh says that ". . . the description of a temporal course (Verlauf) i s pos-s i b l e only i f i t i s based on another process. Something quite analogous holds for spaoe." Whatever Wittgenstein i s trying to say here, I see nothing that implies that spaoe and time are r e l a t i v e . That "We cannot compare a process with 'the passage of time'—there i s no suoh thing—but only with another process • . ." (6.36II), i s an observation quite compatible with a view of time as absolute, for we oan easily suppose that Wittgenstein admits the existence of instants without also wanting to admit the existence of the 'passage' of these i n s t a n t s . Mueller also c i t e s the l e t t e r to Russell of January, 1914, i n whioh Wittgen-s t e i n puts forward the idea that the law of causality implies the r e l a t i v i t y of spaoe and time, because i f there were one s o l i t a r y p a r t i c l e i n the world which suddenly began to move, we see that ". . • no a p r i o r i insight makes suoh events seem impossible to us, except i n the case of spaoe and time's being r e l a t i v e " (NB, p. 129). But t h i s l e t t e r cannot be taken s e r -iously as evidenoe for the views of the T. The tentativeness of i t s conclusion i s made clear by the fact"~that i n the T the law of causality i s no longer regarded as capable of excluding any-thing a p r i o r i . "What can be described can also happen . . . " (6.362T, and since the T would therefore regard as possible that which the l e t t e r to" Russell describes, but regards as im-p o s s i b l e , Wittgenstein's basis i n the l e t t e r for advocating the r e l a t i v i t y of space and time has no r a t i o n a l e i n the oontext of the T. 81 the manner i n whioh s p a t i a l points are represented i n the sym-bolism, then we would have to say that points are not being treated here as objeots, beoause an object ( i . e . , an object whioh i s a p a r t i c u l a r ) i s represented as an argument which f i l l s an argument-place, and not as thi s argument-place i t s e l f , which i s only a set of parentheses and/or a variable indicating where the argument i s to appear. Black does construe 2.0131 i n thi s way, and he concludes: Since space i s a form of objects . . . we may i n f e r that •spaoe' alludes to a •formal concept*. • • Thus 2.0131a ex-presses an int e r n a l *feature*. . . of s p a t i a l f a c t s . In a sati s f a c t o r y notation, t h i s i n t e r n a l feature would be shown by an i n t e r n a l property of the sign used—perhaps by the occurrence i n i t of a (three-dimensional) v a r i a b l e . When we tal k about a 'sp a t i a l p o i n t ' (or a 'position') we are not r e f e r r i n g to a thing, but rather to a certain common in t e r n a l feature of prop-ositions 'about* that position.1 Blaok*s i n t e r p r e t a t i o n , however, i s untenable. He f a i l s to appreciate that the form of an object, though i t s e l f i n t e r n a l to the object, e s s e n t i a l l y involves the configuration of this object with other objects, whioh cannot themselves be represented merely by an in t e r n a l feature of the symbolism; henoe, i n attempting to characterize s p a t i a l points as just suoh in t e r n a l features—the s p a t i a l object being regarded as the argument and the s p a t i a l point as the argument-plaoe—he negleots to t e l l us whioh objeots are those able to be configured with a s p a t i a l object i n vi r t u e of i t s form, i f these objeots are other than s p a t i a l points themselves. Furthermore, the view that a s p a t i a l point i s an argument-place means that there i s no difference i n s p a t i a l •l-Max Black, A Companion to Wittgensteln*s ' Traotatus *. p. 50. 82 p o i n t s , since one blank argument-plaoe i s the same as any other, unless we want to adopt Black's ad hoc solution that the s p a t i a l object i t s e l f carries the burden of in d i c a t i n g i t s own s p a t i a l p o s i t i o n , i n whioh case s p a t i a l points are rendered completely superfluous. A l s o , supposing that the s p a t i a l form of an objeot were symbolized merely by the argument-plaoe, how would, say, the object's temporal form be symbolized—by the same blank argument-plaoe or, i f by something e l s e , what? The obscurities into which we are led by Black's i n t e r -pretation constitute a good case for supposing that Wittgenstein i s not, i n 2.0131, using the expression 'argument-place' i n the usual sense of designating an int e r n a l feature of the symbolism. A pl a u s i b l e a l t e r n a t i v e Interpretation i s that he i s using the expression 'place* i n the sense of 's p a t i a l p o s i t i o n ' , and 'argument* i n the sense of ' p a r t i c u l a r ' , so that to say that a s p a t i a l point i s an argument-place i s simply to say that i t i s a p o s i t i o n i n space occupiable (but not necessarily occupied) by a s p a t i a l p a r t i c u l a r . This interpretation i s borne out by the context. The passage i n question actually occurs i n 2.0131 as a parenthetical remark, so that the entire paragraph reads: A s p a t i a l object must be situated i n i n f i n i t e space. (A spatial-point i s an argument-place.) That i s , a s p e c i f i c point i s one of an i n f i n i t e number of plaoes that a s p a t i a l object may occupy.1 2.0131, In turn, i s a comment •^That Wittgenstein intends to make the second, parenthetical sentenoe of 2.0131 a comment on the f i r s t sentence i s confirmed by the corresponding version of the PT, which reads: "Suppose that things are material points with i n f i n i t e space around them. It i s clear that a material point i s unthinkable without i n f i n i t e 83 on 2.013, which a r t i c u l a t e s the notion of an object being i n a 'space* of possible states of a f f a i r s— t h i s being another ex-pression for the object*s p o s s i b i l i t i e s of configuration with other objects—so that the o v e r a l l oontext favors my claim that s p a t i a l p o i n t s , or p l a c e s ,1 are not i n t e r n a l features of the symbolism, but are themselves objeots (at l e a s t aocording to the treatment they receive i n 2.0131) whioh are configurable with s p a t i a l objeots so as to comprise states of a f f a i r s . And, of course, i f the r e l a t i o n of a s p a t i a l object to s p a t i a l points i s an i l l u s t r a t i o n of the theme of the c o n f i g u r a b i l i t y of one object with others, then the other examples discussed i n 2.0131— the r e l a t i o n of a patoh i n the v i s u a l f i e l d to oolor-spaoe, a sound to pitch-space, or an object of the sense of touch to space (PT 2.0141). On this i n t e r p r e t a t i o n , a spatial-point i s an argument-plaoe"0 (PT 2.01411). Here, the status of the remark about argument-plaoes as an elucidation of the notion of an objeot's being situated i n i n f i n i t e space i s obvious, f i r s t , from the f a c t that the second remark i s subsumed under the f i r s t i n v i r t u e of their respective decimal numbers, and secondly, by the occurrence of the expression *on t h i s i n t e r p r e t a t i o n ' , whioh must re f e r PT 2.01411 back to the interpretation advanced i n PT 2.0141. On Black's view of what Wittgenstein means by 'argument-place', PT 2.01411 would be t o t a l l y i r r e l e v a n t as a comment on PT 2.014T. •^In the essay "Totali t y and System," included i n Appendix A of WWK, we again find Wittgenstein describing s p a t i a l points as 'places', where i t i s even clearer than i n the T that he i s ta l k i n g about s p a t i a l positions rather than features of the symbolism, sinoe he says 'spatial-place' rather than 'argument-place ' . Thus: "Can we describe a s p a t i a l point by specifying which objeots are found at t h i s spatial-place? No, for we do not know how we should a r r i v e at t h i s s p a t i a l p o i n t . I t i s essential to a s p a t i a l description that i t gives us the way that we are to a r r i v e at a s p a t i a l - p l a c e . Specifying a s p a t i a l point means specifying a method for a r r i v i n g at the s p a t i a l point" (WWK, p. 215). 84 hardness-space—imply that c o l o r s , p i t c h e s , and hardnesses are objects as w e l l . So, to sum up the resu l t s of our examination of the status of s p a t i a l p o i n t s , we oan say that the form of a s p a t i a l object i s i t s i n t e r n a l property of being able to combine with s p a t i a l p o i n t s , which are themselves objects. By par i t y of treatment, we can assume that temporal objeots have the form of being con-figurable with those objeots whioh are instants; this renders temporal objects themselves 'timeless1 and thus provides for the assimilation of p a r t i c u l a r s to un i v e r s a l s . F i n a l l y , given that similar treatment should be accorded to co l o r , since 2.0251 groups together spaoe, time, and color as forms of objects (and from 2.0131 we see that p i t c h and hardness could also be i n -cluded as such forms), we may conclude that objects of this l a t t e r form are configured with c o l o r s , which, l i k e points and ins t a n t s , are themselves objects, though universals rather than p a r t i c u l a r s . Therefore, a proper understanding of Wittgenstein's views on space and time oonfirms, from two directions at once, a r e a l i s t i c i n t e r p r e t a t i o n of the T. I turn next to a group of passages which provides evidence that universals are included among objects by the contrast made between them and so-called 'l o g i o a l constants*. F i r s t , there i s 4.0312: The p o s s i b i l i t y of propositions i s based on the p r i n c i p l e of the representation of objeots by signs. My fundamental idea i s that the ' l o g i c a l constants' do not represent anything; that the l o g i c of facts permits no suoh representation. 85 Now, the objection iaade here to the r e i f i o a t i o n of l o g i c a l con-stants i s r e l a t i v e l y mild oompared to the view that no s i g n , other than the sign for a p a r t i c u l a r , represents an object. Therefore, i f Wittgenstein actually held this more extreme view, we would expect him to. say so here, rather than to place exclu-sive emphasis on the milder view; that he does not do this i s an in d i c a t i o n that the intended contrast i s between l o g i c a l con-stants and a l l other signs—including function-signs as well as argument-signs—which, as long as they are pr i m i t i v e signs (Urzelohen), are a l l names of objects. That th i s i s Wittgenstein's intention i s subsequently con-firmed i n his extended discussion of l o g i c a l constants s t a r t i n g at 5*4. In 5»42 and 5»46l, taken together, we see that l o g i o a l constants suoh as 'v' and 'o' are only pseudo-relations, as op-posed to r e a l r e l a t i o n s suoh as r i g h t and l e f t , and that, since l o g i o a l constants are in t e r - d e f i n a b l e , only the signs for r e a l r e l a t i o n s are p r i m i t i v e (Urzeiohen). Now, a case could be made for saying that Wittgenstein uses the expression 'primitive s i g n ' synonymously with 'name',1 which would show d i r e c t l y that the intended contrast i s one between l o g i c a l constants and other signs, a l l of whioh are names. Even apart from this consideration, his intent i s made clear i n 5«44: "'"See 3»26, The one exception would be Wittgenstein's reference to the general form of the proposition—the one genuine l o g i o a l constant—as a pr i m i t i v e sign (5»46), which could, however, be argued to be an uncharacteristic and even f i g u r a t i v e use of the expression, given the unique status of the one l o g i o a l constant as representing only a form and no oontent. 86 Truth-functions are not material funotions. . . . The proposition *~~p' i s not about negation, as i f negation were an object; to be sure, however, the p o s s i b i l i t y of negation i s already contained i n a f f i r m a t i o n . And i f there were an object called '-»', '--p' would have to say something other than what 'p' says. For the one prop-o s i t i o n would then be about — , and the other would not. Here the assertion that truth-functions are not material functions i s explained by a statement to the effect that '«-*' i s not the name of an object, whioh would be a pointless i l l u s -t r a t i o n unless the sign for a material function were such a name. Wittgenstein's discussion of l o g i o a l constants gives ad d i t i o n a l evidence that n o t s a l l objects represented by p r i m i -t i v e signs are p a r t i c u l a r s by implying that some pri m i t i v e signs stand for oonoepts. which, i f regarded as objeots at a l l — and the context shows that he does so regard them—must be con-sidered to be universals rather than p a r t i c u l a r s . The relevant passages here are 5*45 and 5»451« If there are p r i m i t i v e l o g i o a l signs, a correct l o g i c must make d e a r th e i r r e l a t i o n (Stellung) to one another and j u s t i f y t h e i r existence (5*45)• If l o g i o has p r i m i t i v e concepts (Grundbegrlffe). they must be independent of one another. If a p r i m i t i v e concept i s i n t r o -duced, i t must be introduced i n a l l the combinations i n whioh i t ever occurs. . . . (In short, Frege's remarks about introducing signs by means of d e f i n i t i o n s . . . also apply, mutatis mutandis, to the i n t r o -duction of p r i m i t i v e signs.) (5.45ll If there were l o g i o a l Urzeiohen. they would stand for l o g i o a l objeots—as i s shown by 5*4, 5»42, and 5«44—so the fa c t that Wittgenstein elaborates upon the supposition that there are suoh p r i m i t i v e signs by supposing that logic^has p r i m i t i v e oonoepts shows that he i s w i l l i n g to include oonoepts among objeots. A l s o , the remark that concepts must be introduced i n 87 a l l combinations i n whioh they can occur echoes the discussion i n the 2.01'a concerning the l o g i o a l form of objeots, whioh gives us a further reason to suppose that he regards conoepts as objects. As i t turns out, of course, l o g i o a l constants are not p r i m i t i v e signs and so are not signs for concepts, but we are l e f t with the impression that he does inolude concepts among the Bedeutungen of pr i m i t i v e signs. The r e a l i s t i c i n t e r p r e t a t i o n of concepts i s oonfirmed by 4.126, i n whioh Wittgenstein c r i t i c i z e s the 'old1 l o g i c— undoubtedly that of Prege and Russell—only for regarding formal oonoepts as conoepts proper (elgentliohe B e g r l f f e ) . which implies that he i s i n general accord with Frege's and Russell's treatment of conoepts proper. Sinoe t h e i r treatment i s r e a l i s t i c , we may conolude that Wittgenstein's i s r e a l i s t i c as w e l l ; i f his t r e a t -ment were nominalistio, and thus a r a d i c a l departure from Frege c and R u s s e l l , the r e l a t i v e l y mild o r i t i o i s m advanced i n 4.126 of the old l o g i c would be a puzzling understatement. This i s not to say that, apart from Wittgenstein's d i s -t i n c t i o n between formal ooncepts and oonoepts proper, his own view of the nature of conoepts proper i s i n complete accord with the views of Frege and R u s s e l l , nor even that Frege and Russell are themselves i n oomplete accord, though the T admittedly does underplay those differences whioh exist between the three, as when, i n 4.126 and 4.1272, Wittgenstein implies that i n repre-senting oonoepts proper by means of functions he i s merely following Frege and R u s s e l l . What Wittgenstein means by saying that he represents conoepts proper by functions i s innocent 88 enough; supposing that 'John i s white' i s a proposition of the subject-predicate form, then he i s saying only that t h i s prop-o s i t i o n oan be symbolized as ' f ( a ) * , where the argument-sign represents John and the function-sign represents the concept, or property, of whiteness.1 This manner of symbolizing the proposition follows the praotioe of Prege and R u s s e l l , and to th i s extent a l l three views are the same. Wittgenstein, however, obsoures an important difference between Frege and Russell con-cerning the r e l a t i o n of concepts to funotions. As f a r as Frege i s concerned, i t i s s t r i c t l y inaccurate to say that oonoepts are represented by functions, because oonoepts are a species of function, v i z . , those funotions of one argument-place whose value i s always a truth-value. Concepts, then, are represented expressions for just those kinds of funotion, i . e . , by funotion-signs and not by functions themselves. R u s s e l l , unlike Frege, regards concepts as capable of assuming the r o l e of logioal.subject i n a proposition (or at l e a s t he assumes th i s i n PMA), which they could not do i f they were themselves 'unsaturated' or incomplete, as functions would have to be i f they were e n t i t i e s . Since oonoepts oannot be lj.n the 'old' l o g i c , concept and property amount to the same thing. I oan say either that the object John f a l l s under the concept whiteness or that the object John has the property whiteness. (See Frege, "On Concept and Object," i n P h i l o - sophical Writings of Gottlob Frege, ed. by Peter Geaoh and Max Black, p. 51.) That Wittgenstein follows Frege here oan be seen by 4.126, where he implies that concepts are represented by means of functions and that properties are expressed by means of functions, both i n contrast to formal oonoepts and properties. Unless we want to read some subtle d i s t i n c t i o n into the expressions 'represents' and 'expresses', for whioh there seems to be no b a s i s , the evidenoe of 4.126 i s that Wittgenstein equates oonoepts with pr o p e r t i e s . 89 functions and there i s no reason to assume that propositions somehow involve e n t i t i e s suoh as functions i n addition to con-oepts, functions are relegated to the status of merely a sym-b o l i c device—provided that we discount those few care l e s s l y -worded passages i n PMA whioh seem to impute to functions an ontologioal status—enabling us to represent a proposition i n abstraction from some speoifio part of i t s content. Unlike the sign for a oomplete proposition, whioh, i f the proposition i s true, has as i t s ontologioal counterpart the appropriate com-plex of objects and concepts i n r e a l i t y , a propositional function, whioh i s merely the schema which remains when part of the content of a proposition i s abstracted, has no 'schematic' ontologioal counterpart. Thus, the d i s t i n c t i o n between sign and thing s i g n i f i e d , with respect to functions, i s i n a p p l i c a b l e , so that insofar as the notation of PMA can be said to provide for the representation of oonoepts at a l l ,1 there can be no scruple against assigning the symbolic r o l e to functions themselves rather than to function-signs, since both amount to the same thi n g . Wittgenstein agrees with Frege against Russell i n allowing that,, not just symbols, but also the e n t i t i e s they represent are, as i t were, 'unsaturated', so that we might also expect Wittgen-s t e i n to follow Frege i n distinguishing between functions and •'•This problem was previously discussed on p. 9 f f » Wittgen-s t e i n seems not to have noticed this d i f f i c u l t y for the Russellian view, probably because no s i m i l a r d i f f i c u l t y exists for his own view, i n whioh individ u a l s are quite unproblematically repre-sented by unsaturated signs. 90 funotion-signs. I r o n i c a l l y , however, Wittgenstein's extension of the feature of unsaturatedness to a l l objects, and not just to concepts (and r e l a t i o n s , i f we want to distinguish between functions with one argument-place and those with more than one), drives him instead towards Russell's view that functions are a symbolic device rather than a kind of object. That i s , i f a l l e n t i t i e s are unsaturated, the main point i n speaking about funotions w i l l be to d i s t i n g u i s h those signs whioh overtly sym-bo l i z e the unsaturatedness of the object they represent—by containing within them an empty set of parentheses, or one f i l l e d only with a v a r i a b l e , whioh shows where another sign may be i n -serted—from those signs whioh make no suoh overt display but whioh are nevertheless equally unsaturated, i . e . , dependent symbols i n the sense of 2.0211. The former we c a l l funotion-signs and the l a t t e r argument-signs. We may, of course, con-tinue to observe the function/argument d i s t i n c t i o n at the l e v e l of ontology i n addition to the corresponding d i s t i n c t i o n between funotion-signs and argument-signs at the l e v e l of symbolism (and I have indeed done so i n my exposition of Wittgenstein's views), as long as we recognize that, by the o r i t e r i o n of unsaturated-ness, e n t i t i e s whioh we c a l l arguments could be called functions as w e l l , so that i f we continue to di s t i n g u i s h between them, this i s only to express their d i f f e r e n t syntactic r o l e s , a difference I l l u s t r a t e d at the l e v e l of symbolism by confining overt unsat-uratedness to those signs whioh behave predioatively or r e l a t i o n -a l l y , i . e . , function-signs. Of course, sinoe we also have other d i s t i n c t i o n s to mark the d i f f e r e n t syntaotic roles of e n t i t i e s 91 (or, r a t h e r , of the names whioh represent them)—e.g., the d i s -t i n c t i o n between thing and property, or between object and con-cept (where the terms 'thing' and 'object' are used here i n the narrow sense of ' l o g i c a l subject*, as opposed to the wide sense i n whioh oonoepts and r e l a t i o n s are things, or objeots, as well) — the function/argument d i s t i n c t i o n has i t s p a r t i c u l a r usefulness pr i m a r i l y at the l e v e l of symbolism rather than at the l e v e l of ontology. I t i s therefore understandable that Wittgenstein should s l u r over Frege*s d i s t i n c t i o n between functions and function-signs i n expressing his own p o s i t i o n as being that oon-oepts are represented by functions, though i t i s s t i l l a piece of carelessness to a t t r i b u t e t h i s view to Frege himself. In any case, whatever differences may exist between the views of Frege, R u s s e l l , and Wittgenstein regarding the nature of oonoepts and th e i r r e l a t i o n to functions, the important point for our purposes i s that Wittgenstein does follow Frege and Russell i n regarding oonoepts as objects.1 A s t r i k i n g confirmation of Wittgenstein's r e a l i s t i c i n t e r -p r e t a t i o n of concepts i s found i n Appendix A of WWK. In this appendix we find Wittgenstein taking pains to prevent the sort of misunderstanding of his p o s i t i o n that has nevertheless be-come prevalent i n the nominalistio int e r p r e t a t i o n of the T, a •*-It should be noted that, although i n the sense i n whioh Frege uses the term 'object', conoepts are never objeots because (according to Frege) conoepts are, and objeots are not, unsat-urated, nevertheless Frege would grant that oonoepts are objects i n the sense that oonoerns us here, v i z . , concepts are included among those e n t i t i e s whioh constitute the meaning (Bedeut ng) of p r i m i t i v e signs. 92 misunderstanding epitomized i n the following passage of Anscombe's: What • • • has become of Frege's 'concepts' i n Wittgen-stein's theory? They seem to have disappeared e n t i r e l y ; a c t u a l l y , however, instead of making concepts or universals into a kind of objeots . . . Wittgenstein made the gulf between oonoepts and objeots much greater than Frege ever made i t . So f a r as concerns the oontent of a functional expression, that w i l l consist i n the objeots covered by i t . But i n respect of having argument-places, conoepts go over e n t i r e l y into l o g i o a l forms. In the 'completely analysed proposition' . . . the Fregean 'concept', the thing with holes i n i t , has become simply the l o g i o a l form.l In the section of Appendix A of WWK,? appropriately t i t l e d "Con-cept and Form," Wittgenstein expressly sets out to combat thi s c o n f lation of oonoepts with forms. Thus: The form of a proposition i s produced by abstracting from the meaning of the words, by turning them into v a r i a b l e s . A subject-predicate proposition has a d i f f e r e n t form than a r e l a -t i o n a l p r o p o s i t i o n , a symmetrical r e l a t i o n a l proposition a d i f -ferent form than one whioh i s assymetrical. The state of a f f a i r s i s a combination of things. The things are represented In the proposition by signs, but the form of the state of a f f a i r s i s not; i t i s shown by the form of the proposition. A concept must be explained, the form of a proposition shows i t s e l f . The form i s not describable, for the description portrays the form. To have a form means to be a p i c t u r e , to think or to speak means to p i c t u r e . Conoepts are expressed through signs. The form, the p r o p o s i t i o n a l - p i c t u r e , shows i t s e l f . The form i s not a generalization and not a oommon property of a olass of prop-o s i t i o n s . Symmetry and assymetry show themselves i n the prop-o s i t i o n s , are contained i n the description—they are not prop-e r t i e s l i k e yellow and hard, whioh are expressed through a propositional-function as names.2 1Ansoombe, Introduction, pp. 108-109. 2WWK. p. 220. Concerning the dating of this m a terial, McGuinness remarks: "Although i t i s probable that the material found i n Appendix A was typed and duplicated i n l a t e 1930 . . . i t Is nevertheless possible that the conversations from whioh th i s material was derived took place before December, 1929" (B. F. McGuinness, In editor's foreword to WWK. pp. 20-21). A perusal of Appendix A shows thi s p o s s i b i l i t y to be a high proba-b i l i t y , since the outlook, and even the wording, of these notes i s extremely close to the T, whereas by 1930 Wittgenstein's 93 Unless we quite a r b i t r a r i l y assume that the r e a l i s t i c p o s i t i o n expressed i n t h i s passage r e f l e c t s a r a d i c a l departure from the view of the T—an assumption which i s not merely implausible but i s u t t e r l y gratuitous when we consider that t h i s passage i n other respects, e.g., concerning the state of a f f a i r s as a combination of things, or the proposition as a pioture of t h i s state of a f f a i r s , f a i t h f u l l y embodies the outlook of the T—we. oan only conclude that this passage provides decisive evidenoe that the T includes concepts, or p r o p e r t i e s , among objeots. In addition to t h i s passage from Appendix A of WWK, there are a number of other passages i n Wittgenstein's post-T writings that at l e a s t i n d i r e c t l y oonfirm the r e a l i s t i c i n t e r p r e t a t i o n of the T, i n that t h e i r intense concern to combat thi s sort of realism suggests that they are intended as a repudiation of a p o s i t i o n held e a r l i e r by Wittgenstein himself. Many such pas-sages w i l l be discussed i n Part Two, especially i n Chapter VI. For now I w i l l confine myself to examining only those few pas-sages which are especially noteworthy. In PG, Wittgenstein observes that One oan, quite understandably, speak of combinations of colors  with shapes (for instance, the colors red and blue with the shapes square and c i r c l e ) i n the same way as one speaks of combinations of d i f f e r e n t figures or bodies. And here i s the root of the i l l -chosen expressions the fact i s a complex of objeots. That a man i s s i c k here becomes compared with the putting together of two things, of which one would be the man and the other would be siokness (PG_, p. 58)• views had begun to change r a d i c a l l y . A l s o , MoGuinness himself t e l l s us, on p. 20 of WWK. that EngHcmann's reoently discovered copy of these notes bears the t i t l e "Dictated o r a l l y by L. W. (recorded before 1930)." 94 Although the T nowhere contains prec i s e l y the phrase 'the fact i s a complex of objects', other remarks, e.g., 2.01, are suf-f i c i e n t l y close that we can reasonably assume that the T Is the subject of c r i t i c i s m . I f thi s assumption i s correct, then the implication i s that the a t t r i b u t i o n of a property to a thing i s indeed treated i n the T as the configuration of two e n t i t i e s , one of them being of a predicative nature. In an a r t i c l e e n t i t l e d "Complex and Fact," written i n 1931 and published as an appendix both to PB and PG., t h i s sameecon-cern to di s t i n g u i s h between facts and complexes issues i n the following series of remarks: To say that a red c i r c l e consists of redness and c i r c u l a r i t y , or i s a complex of these constituents, i s a misuse of these words and i s misleading. . . . I t i s likewise misleading to say that the fact that this c i r c l e i s red (that I am weary) i s a complex of the constituents red and c i r c l e (of me and weariness). The house i s also not a complex out of brioks and the i r s p a t i a l r e l a t i o n s . That i s , t h i s also goes against the correot use of words (PG, p. 200). Als o , the chain consists of i t s members, not of them and thei r s p a t i a l r e l a t i o n s . The fact that these members are so com-bined 'consists' of nothing at a l l (PG, p. 201). Since the denial that a faot 'consists' i n anything i s c l e a r l y aimed at the T (e.g., 2.05, 2.14, 2.141), which confirms that the T i s the subject of c r i t i c i s m here, the obvious implication once again i s that properties and r e l a t i o n s are regarded i n the T as configurable objects.1 Ll should mention, i n a n t i c i p a t i o n of my argument i n Part Two, that nothing said by Wittgenstein i n either of these two passages implies that at the time of the i r writing he altogether denies object-hood t o , say, c o l o r s . From his statement that "A complex consists of i t s p a r t s , which are things of the same kind (gleichartigen Dingen) . . ." (PG, p. 200), I take his point to be only that colors and shapes, or bricks and s p a t i a l r e l a t i o n s , 95 It would be possible to c i t e further passages, both from the T and from the l a t e r w r i t i n g s , that lend support to my i n t e r p r e t a t i o n , but none would be as decisive as those already c i t e d , and, anyway, there i s no need to belabor the i s s u e . More than enough favorable evidenoe has already been accumulated. What now remains to be done i s to examine the counter-arguments, sinoe the popularity of the nominalistic interpretation shows that there must, a f t e r a l l , be some passages i n the T which are at l e a s t ostensibly favorable to such an i n t e r p r e t a t i o n . Before my i n t e r p r e t a t i o n oan be regarded as established, these counter-arguments w i l l have to be met. are not objeots of a s u f f i c i e n t l y similar kind to be properly spoken of as constituting oomplexes. Possibly there i s the further implication that not only are colors and shapes suf-f i c i e n t l y unlike one another to d i s q u a l i f y them from being fellow-constituents i n the same complex, but that no properties or r e l a t i o n s , just because they are not p a r t i c u l a r things, should be allowed the status of constituents of complexes. In either case, the denial that, e.g., the color red i s a con-st i t u e n t i n a complex i s not the denial that this property i s an object of some kind. I t i s true that Wittgenstein does go on to say that "The root of this confusion i s the muddled use of the term 'object' . . ." (PG, p. 201), but the context i n -dicates that the muddle does not oonsist i n applying the term 'object' to properties and r e l a t i o n s , but rather i n thinking of the fact as 'consisting' of i t s members, i . e . , i n treating the f a c t i t s e l f as a kind of (complex) object. 96 CHAPTER IV ARGUMENTS AGAINST A REALISTIC INTERPRETATION Great weight i s given to a number of passages i n the T whioh use the expressions 'thing', 'object', or 'name' inwa manner p l a i n l y suggesting that only p a r t i c u l a r s are being con-sidered as things, or symbols for p a r t i c u l a r s as names, and therefore implying that objeots do not include u n i v e r s a l s . The most frequently cited passage of this sort i s 4.24: Names are the simple symbols: I indicate them by single l e t t e r s ('x', 'y', ' z ' ) . I write an elementary proposition as a function of names, i n the form: 'fx*, 'jrf(x, y ) * , e t c . Here names seem to be r e s t r i c t e d to argument-signs and to be opposed to functions (or, rather, function-signs, i f we want to maintain a d i s t i n c t i o n here between the l e v e l of ontology and of symbolism), leading many commentators to suppose that funotion-signs are not names and therefore do not represent objects.1 Just as 4.24 r e s t r i c t s names to argument-signs, 4.1211 r e s t r i o t s objects to what are represented by argument-signs: Thus one proposition *fa* shows that the object a occurs i n i t s sense, two propositions 'fa* and 'ga' show that the same object i s mentioned In both of them. A comparable use of the expressions 'object' and 'thing' i s found i n 5*5301, 5*5302, 5*5352, and 5*553* These passages have been generally regarded as compelling evidence of Wittgen-^See I . M. Copi, "Objects,. Properties, and Relations i n the 'Tractatus'," i n ET, p. 182, W. S e l l a r s , "Naming and Saying," i n ET. p. 261, Anscombe, Introduction, p. 99» MoGuinness, p. 149, Pi t c h e r , p. 114, and Mueller, p. 89. 97 stein's nominalism, and a r e a l i s t i c i nterpretation of the T must take them into account. In f a c t , the compatibility of these passages with realism i s ea s i l y shown. We need only assume that Wittgenstein here means to contrast e n t i t i e s , or their symbols, whioh funotion as l o g i c a l subjects, with those e n t i t i e s , or t h e i r symbols, which funotion p r e d i c a t i v e l y or r e l a t i o n a l l y ; i n order to emphasize th i s contrast, he r e s t r i c t s the expressions 'object', 'thing*, and 'name* to th i s f i r s t kind of entity or Its symbol, and c a l l s the other kinds of e n t i t i e s *properties*, 'concepts*, or ' r e l a t i o n s * , and c a l l s t h e i r symbols 'functions* or 'funotion-signs'. That properties or r e l a t i o n s are not objects i n th i s narrow sense would then be p e r f e c t l y compatible with th e i r being objects i n the wide sense of constituting the meanings of p r i m i t i v e signs; In t h i s same wide sense, the signs themselves, whether function-signs or argument-signs, would a l l be names. Granted that the assumption that Wittgenstein uses 'object', e t c . , i n both a narrow and a wide sense enables us to reconcile passages such as 4.24 with a r e a l i s t i c interpretation of the T, are there any grounds for making t h i s assumption? C l e a r l y , there are such grounds; the precedent for the use of these terms i n both a narrow and a wide sense has been established i n NB, as our examination has shown.1 Further confirmation i s found i n the PT, whioh i s more e x p l i c i t than the T i n this matter. Thus, corre-sponding to 4.24 of the T are the following passages of the PT: 1See p. 47ff. 98 In what follows I indicate names of objeots (Gegenstands-namen) by the l e t t e r s x, y, z, u, v, w (4,2211). Generally i n what follows I indicate elementary propositions by the l e t t e r s p, q, r , s, t , or else ( l i k e Frege) I write them as funotions of th e i r objects i n the form *tf(x)', lt ( X | y ) ' , etc . (4.2212) The expression 'Gegenstandsnamen' i s pleonastic, unless Wittgen-s t e i n i s using 'Gegenstand* i n the narrow sense, and, on one occasion (PT 4.102273) he uses the expression *names of concepts' (Begriffsnamen) as w e l l , which reinforces the supposition that the expression 'Gegenstandsnamen1 i s intended to single out names of one kind of entity among others, v i z . , names of objeots as opposed to names of oonoepts. Furthermore, that the r e s t r i c -t i o n of the term 'object' i n t h i s context to those e n t i t i e s represented by argument-signs i s not to be taken as an endorsement of nominalism i s shown by Wittgenstein's e x p l i c i t comparison of his view of elementary propositions with that of Frege, who also contrasts functions and objeots, while maintaining that funotions are e n t i t i e s i n th e i r own r i g h t . Unfortunately, the version of these passages which ultimately appears as 4,24 of the T has been revised i n a way whioh obscures i t s o r i g i n a l p r o - r e a l i s t i n t e n t , 'Gegenstandsnamen' has been changed to 'Namen', reference to Frege has been omitted, and the Fregean d i s t i n c t i o n between functions and objects—which i s a d i s t i n c t i o n between kinds of e n t i t i e s ana thus whioh implies that functions do have ontologioal status—has been replaced by the d i s t i n c t i o n between funotions and names, a d i s t i n c t i o n whioh i s no better than neutral as far as providing evidence i n favor of realism. Whatever prompted these r e v i s i o n s , we can see from the PT that i t was not Wittgenstein's o r i g i n a l intent to deny 99 that funotion-signs symbolize objeots, i n the wide sense, and unless one i s prepared to contend that the change from the text of the PT to that of the T marks the change from realism to nominalism—a gratuitous suggestion, given that the PT and the T are otherwise much too similar to allow the supposition that they maintain r a d i c a l l y d i f f e r e n t ontologies—the only possible conclusion i s that, l i k e PT 4.2211 and 4.2212, 4.24 of the T does not advocate nominalism, though i t preserves i t s compati-b i l i t y with realism by using 'name', rather than 'object*, i n the narrow sense. Admittedly, Wittgenstein's casual employment of these key terms opens the door to a great deal of confusion—as i s amply demonstrated by the many commentators who are unaware of the two senses of 'object', etc., and who therefore take passages l i k e 4.24 to be compelling evidenoe against realism—but, given the assiduously cultivated obsourity of the T, Wittgenstein's nonohalance i n a matter of this importance i s not r e a l l y too s u r p r i s i n g . We are, a f t e r a l l , warned i n the preface to the T that he has not written a text-book and that the book might be understood only by someone who has had thoughts along the same or si m i l a r l i n e s ; suoh a reader would not be misled by a few Isolated?passages, or rather by what might s u p e r f i c i a l l y be taken as th e i r i n t e n t . .Of oourse, with the benefit of exposure for over half a oentury, the T has become accessible to the gen-e r a l philosophical p u b l i c , and for the sake of a l l oonoerned, we can only regret that Wittgenstein was not more careful i n his employment of key terms. S t i l l , I hope to have shown that his 100 terminological carelessness does not prevent his r e a l i s t i c ontology from c l e a r l y emerging, as long as we have the time and the patienoe to f i t together the pieces of the puzzle. Of those passages other than 4.24 whose nominalistio appearance i s accounted for by our recognition of Wittgenstein's f l u c t u a t i n g terminology, 5*553 deserves mention as a spec i a l case because questions are also raised here about the use of the ex-pression ' i n d i v i d u a l ' . I t i s the f i r s t sentence of 5*553* i n which Wittgenstein i s characterizing a p o s i t i o n of Russell's, that concerns us. He observes that "Russell said that there were simple r e l a t i o n s between d i f f e r e n t numbers of things ( i n -d i v i d u a l s ) . " Mueller takes t h i s passage as implying that r e l a t i o n s are not things, an int e r p r e t a t i o n confirmed, he thinks, by Wittgenstein's i d e n t i f i c a t i o n of things with l n d i v i d u a l s .1 Now, even i f we contend, i n opposition to Mueller, that the contrast i s between r e l a t i o n s and things only i n the narrow sense of 'thing', so that i n the wide sense r e l a t i o n s are things as w e l l , there remains the'problem of explaining the parentheti-c a l reference to i n d i v i d u a l s , s i n ce, on my interpretation of what Wittgenstein means by ' i n d i v i d u a l ' , r e l a t i o n s are included among individuals rather than; contrasted with them. To solve t h i s problem we must look to the context i n whioh 5*553 appears. From 5*55 to 5*6 the discussion concerns the p o s s i b i l i t y of investigating and c l a s s i f y i n g l o g i o a l forms. In OKEW. which i s c l e a r l y the Russellian text to whioh 5.553 a l l u d e s , Russell M u e l l e r , pp. 90-91. 101 holds that • . the f i r s t business of l o g i c • . .-" i s . . a c l a s s i f i c a t i o n of the l o g i c a l forms of facts . . • According to R u s s e l l , a prime example of suoh c l a s s i f i c a t i o n i s the deter-mination of whether there are r e l a t i o n s of more than two terms, which he cautions us against confusing with the issue of whether a two-termed r e l a t i o n oan r e l a t e more than two d i f f e r e n t terms i n the sense i n which t h i s r e l a t i o n i s a constituent of two d i f f e r e n t f a c t s , as when a man i s the son of his father as well as a son of his mother. Thus: When we say that there are r e l a t i o n s of more than two terms, we mean that there are single facts consisting of a single r e -l a t i o n and more than two things. . . . the facts I am speaking of have no facts among thei r constituents, but only things and r e -l a t i o n s . For example, when A i s jealous of B on account of C, there i s only one f a c t , involving.three people; there are not two instances of jealousy, but only one. I t i s i n suoh oases that I speak of a r e l a t i o n of three terms, where the simplest possible fact i n which the r e l a t i o n oocurs i s one involving three things i n addition to the r e l a t i o n . And the same applies to r e l a t i o n s of four terms or f i v e or any other number. A l l such r e l a t i o n s must be admitted i n our inventory of the l o g i c a l forms of f a c t s : two facts involving the same number of things have the same form, and two which involve d i f f e r e n t numbers of things have d i f f e r e n t forms.2 From this passage we can see that Wittgenstein's parenthetical reference to individuals i n 5*553 i s simply the expression of Russell's requirement that the c l a s s i f i c a t i o n be of r e l a t i o n s as constituents of single f a c t s , i . e . , r e l a t i o n s which have only simple things (individuals) and not complexes as th e i r terms, sinoe complexes would themselves be further facts and so would v i o l a t e the requirement that the r e l a t i o n a l facts to be consid-ered are single facts ( i . e . , atomic f a o t s , states of a f f a i r s ) . OKEW. p. 47. 2QKEW, pp. 47-48. 102 Hence, the introduction of the expression ' i n d i v i d u a l s ' i n 5*553 bears only on the d i s t i n c t i o n between a thing's being simple or complex, and not on i t s being a p a r t i c u l a r or a u n i v e r s a l , so that 5*553 does not constitute counter-evidenoe to my claim that Wittgenstein includes universals among i n d i v i d u a l s .1 Thus far I have been arguing that the d i s t i n c t i o n between a wide and narrow sense of 'object' enables us to re c o n c i l e some seemingly nominalistio passages with a r e a l i s t i n t e r p r e t a t i o n . I n t e r e s t i n g l y , proponents of a, nominalist interpretation have also supposed that Wittgenstein uses 'object' In both a narrow and wide sense, and they have appealed to this d i s t i n c t i o n , i n the way that they draw i t , to support t h e i r own case. The pas-sage they c i t e i s 4.123: A property i s i n t e r n a l i f i t i s unthinkable that i t s object does not possess i t . (This blue color and that one stand i n the i n t e r n a l r e l a t i o n of l i g h t e r and darker eo i p s o . I t i s unthinkable that these two objeots should not stand i n this r e l a t i o n . ) (Here the s h i f t i n g use of the word 'object* corresponds to the s h i f t i n g use of the words 'property* and 'relation*.) Just as *darker* i s an i n t e r n a l r e l a t i o n between two c o l o r s , and not a r e l a t i o n proper, likewise—so runs the argument—the colors themselves are not objeots proper; talk of the s h i f t i n g use of 'object* i s meant to warn us that properties are objects only i n a wide sense, i . e . , only i n a loose way of speaking.2 ••-I have argued that i n fact Russell did quite probably equate individu a l s with p a r t i c u l a r s by the time of OKEW. but i t i s ques-tionable whether Wittgenstein was even aware that Russell had changed his p o s i t i o n ; i n any case, i t i s not t h i s aspect of Rus-sel*s p o s i t i o n that Wittgenstein i s trying to express by the parenthetical reference to individuals i n 5.553. 2This int e r p r e t a t i o n i s advanced by Ansoombe, Introduction, p. 122, P i t c h e r , p. 113, Mueller, P» 93» and Copi, p. l b l . 103 I t seems to me, on the other hand, that we oan make more sense out of this passage by supposing that the s h i f t here i s not between a wide and narrow use of 'object', but rather a s h i f t between talk of objects proper (in the narrow sense)— which are the bearers of genuine properties, i . e . , whioh f a l l under r e a l (elgentlioh) oonoepts, or whioh are related to one another by genuine, external r e l a t i o n s (these properties and re l a t i o n s themselves being objects proper, i n the wide sense)— and those so-oalled 'objects' whioh f a l l under formal concepts, or are related by in t e r n a l r e l a t i o n s . Another example of this same s h i f t i n g use of 'object* occurs shortly afterward, i n 4.126, where Wittgenstein says: That something f a l l s under a formal concept as one of i t s objeots cannot be expressed by means of a proposition. Instead i t i s shown by the sign of this object i t s e l f . (A name shows that i t s i g n i f i e s an object, a sign for a number that i t s i g -n i f i e s a number, etc.) In other words, both names and number-signs are signs of objects i n the sense that what each of them represents f a l l s under a formal concept, but only names represent objeots proper, i . e . , e n t i t i e s capable of f a l l i n g under r e a l concepts.1 .The s h i f t i n g use of 'object', therefore, i s only a s h i f t within what I have called the narrow sense of that expression, the sense i n which objects are contrasted with predicative e n t i t i e s ( i . e . , concepts, properties) as t h e i r subjects and with r e l a t i o n a l e n t i t l e s as th e i r terms. The s h i f t within t h i s narrow sense i s between •^ I am, of course, assuming that the d i s t i n c t i o n between 'con-cept* and "object* which runs throughout 4.126 means that 'object* i s here being used i n the narrow sense, i n oontrast to 'concept'. 104 objeots proper, whioh are q u a l i f i e d by external properties and related by external r e l a t i o n s , and what we may c a l l 'formal* ob-je o t s , which are q u a l i f i e d by i n t e r n a l properties and related by i n t e r n a l r e l a t i o n s . To say, then, that A i s darker than B (where A and B are two shades of blue) i s to treat A and B as two terms of a r e l a t i o n , but just as the r e l a t i o n i s i n t h i s case Internal, so, l i k e w i s e , are A and B objeots only i n the formal sense of being terms of an i n t e r n a l r e l a t i o n . So the s h i f t i n 4.123 i s only between proper and formal objects, which corresponds to a s h i f t between external and internal r e l a t i o n s . No doubt Wittgen-s t e i n could have expressed himself here more clearly—and 4.123 involves the additional complication that there i s not even the p o s s i b i l i t y that c o l o r s , whioh are p r e d i c a t i v e , could be objeots proper, whioh makes the warning about the s h i f t i n g use of 'object' almost superfluous, as i t would not have been had the example con-cerned p a r t i c u l a r s rather than colors—but when a l l i s sorted out 4.123 says nothing whioh impugns the ontologioal status of prop-e r t i e s as objects (in the wide sense of being predicative e n t i t i e s ) . With t h i s discussion of 4.123 I conclude the examination of those passages whose nominalistio appearance depends either upon the f a i l u r e to distinguish between a wide and narrow sense of the expressions 'object', 'thing', and 'name', or upon the f a i l u r e to make a judicious use of such a d i s t i n c t i o n . Of the remaining arguments against r e a l i s m , most have already been either e x p l i -c i t l y or i m p l i c i t l y answered i n the course of the exposition of my own i n t e r p r e t a t i o n , so that the majority of suoh arguments need by dealt with here only b r i e f l y . 105 F i r s t , there i s the argument that external properties can-not be objects because they are configurations. That i s , we know that formal, or i n t e r n a l , properties are obviously not objects, because we cannot represent them as we do objeots; n e i t h e r , however, oan material, or external, properties be ob-j e c t s , because, according to 2.0231, material properties are f i r s t formed by the oonflguratlon of objeots. Therefore, no properties are objects.1 The answer to th i s argument i s that, while i t i s true that formal (or internal) properties are not objeots, the notion that material (or external) properties are only configurations of objeots, rather than themselves objeots, i s based on a misunder-standing of 2.0231* As I have argued, what requires the c o n f i g -uration of objects i s the a t t r i b u t i o n of the property to the p a r t i c u l a r , since, apart from th i s configuration, the p a r t i c u l a r cannot be said to have the property, which i s , a f t e r a l l , ex-ternal to the p a r t i c u l a r ' s nature, unlike the form of the p a r t i c -u l a r , which i s i n t e r n a l . This account s a t i s f a c t o r i l y explains 2.0231, while allowing that both p a r t i c u l a r s and material prop-e r t i e s are objeots, and, as we have seen, suoh an account i s overwhelmingly confirmed by independent evidence. Next, there are those arguments which simply assume that properties and r e l a t i o n s are configurations of objeots, and whioh then argue from this assumption that properties and r e l a t i o n s cannot be objects as w e l l . The main argument along i T h i s argument i s advanced by Copi, p. I 8 3 , and. by Pitoher, P. 115. 106 these l i n e s i s the i n f i n i t e regress argument. According to Copi, i f properties were objeots, then 2.0231 . . . would assert material properties to be f i r s t formed by the configuration of material properties, themselves f i r s t formed by the configuration of material properties, and so on.1 Pitcher puts the same argument for r e l a t i o n s ; If r e l a t i o n s were objects, then when two or more p a r t i c -ulars were configured so as to form a state of a f f a i r s , the con-f i g u r a t i o n - - ! , e. , the r e l a t i o n amongst the particulars—would be another object, over and above the p a r t i c u l a r s . But then we would be l o s t i n an i n f i n i t e regress, for i f the configuration of two or more p a r t i c u l a r s were an object, then presumably the configuration of the p a r t i c u l a r s and the f i r s t configuration would also be an object, and the configuration of the p a r t i c -ulars and the f i r s t and second configurations would also be an object, and so on ad i n f i n i t u m .2 Needless to say, these i n f i n i t e regress arguments are generated i n the f i r s t place only i f we assume that properties and r e l a t i o n s are configurations of objeots. Onoe we r e j e c t this assumption— as being based on a misunderstanding of 2.0231 as well as being a f a i l u r e to understand that the way objeots are configured con-s t i t u t e s the structure (or form, depending on whether or not we specify which objeots are i n configuration), not the content, of a state of a f f a i r s , and therefore oonoerns only i n t e r n a l , and not external, properties—the p o s s i b i l i t y of an I n f i n i t e regress vanishes. The assumption that r e l a t i o n s are configurations of objects i s also behind Mueller's argument that i f no objects stand In a certain r e l a t i o n , then that r e l a t i o n does not exist and Is there-fore not part of the substance of the world, i » e . , i t i s not an Copi, p. I83. 2P i t o h e r , p. 116. 107 o b j e c t .1 We have already seen that Wittgenstein, i n NB, has raised the problem of the status of r e l a t i o n s i n f a l s e prop-os i t i o n s and eventually dealt with i t by holding that a f a l s e ( r e l a t i o n a l ) proposition happens to represent a non-existent configuration of an existent r e l a t i o n with existent p a r t i c u l a r s . That i s , the r e l a t i o n i s riot the configuration of p a r t i c u l a r s , but i s i t s e l f an objeot i n thi s configuration. So Mueller's objection f a i l s . Other miscellaneous arguments advanced, as far as I know, only by Mueller include the following: When, aooordlng to 3*314, "the variable name too" can be regarded as a propositional v a r i a b l e , what other expressions could be "transformed" into variables (3*315)? It can be a question only of functions, whioh therefore are not names and yet indicate "a form and a content" (3.31).2 Leaving aside the problem of what Mueller thinks 'content* means here, given his view that a funotion-sign represents the s t r u c - ture of a state of a f f a i r s , 3 he i s wrong i n supposing that the only possible contrast i s between names and functions (or function-signs)• The contrast might also be intended to be between names, which represent simples, and expressions, which may be treated as representing simples by being able to be turned into v a r i a b l e s , e t c . , but whioh may nevertheless represent com-plexes.^ Even i f Wittgenstein does Intend the contrast to be between names and function-signs, this need be taken as implying t e l l e r , pp. 86-87. 2Mueller, p. 84. 3Mueller,'p. 71ff* ^1 assume that by the Introduction of 'expression* Wittgen-s t e i n means to provide a term which, l i k e *name*, designates a unit of meaning, but whioh, unlike *name*, i s neutral to the question of whether that which i s represented i s simple or complex. 108 no more than that funotion-signs are not names i n the narrow sense. That funotion-signs are names i n the wide sense has been more than amply demonstrated. Another of Mueller's objections i s that the objeots referred to i n 2.025l7-where space, time, and color are given as forms of objeotsr?-can only be p a r t i c u l a r s , sinoe universals are not at a s p e c i f i c p l a c e , e t c .1 What Mueller says here i s true, but i r -relevant. Granted that the objects referred to i n 2.0251 are p a r t i c u l a r s , this does not exclude the p o s s i b i l i t y of there being other objects whioh are unive r s a l s , and, as I have argued, to say that spaoe, time, and color are forms of objects i s to say that suoh objeots oan be configured with other objeots—which, i n the case of objeots having the form of being colored, w i l l be u n i v e r s a l s , v i z . , c o l o r s . Mueller also argues that 4.013* whioh says that the p i c t o r i a l character of a proposition i s not destroyed by apparent i r r e g u -l a r i t i e s i n the notation, implies that funotion-signs are not r e a l l y names because the p a r a l l e l between the example given by 4.013 of such an apparent i r r e g u l a r i t y , v i z . , the use of a c c i -dentals (sharps and f l a t s ) i n a musical score, and propositions of the form ' f ( x ) ' i s . • only too obvious."2 That i s , sharps and f l a t s do not stand for no.tes^but have a representational r o l e only by being to the l e f t of a note, and the same thing oan be said for function-signs. The answer to thi s argument i s that, while i t i s true that ^•Mueller, p. 85. 2Mueller, pp. 85-86. 109 funotion-signs symbolize only by being to the l e f t of argument-signs, i t i s equally true that argument-signs symbolize only by being to the r i g h t of funotion-signs, which i s a consequence of Wittgenstein's view that a l l names are dependent symbols. This s i m i l a r i t y between function-signs and accidentals, then, cannot count against function-signs having the status of names, and i n other respects the p a r a l l e l between funotion-signs and acciden-t a l s i s l a c k i n g , since only the l a t t e r , and not the former, can claim to be notational i r r e g u l a r i t i e s . In sum, there i s no reason to suppose that 4.013, with i t s example of sharps and f l a t s , intends any p a r a l l e l to be drawn to function-signs at a l l , l e t alone that i t wants to show that function-signs are not names, I consider next two more arguments of Copi's, In the f i r s t argument, Copi r i g h t l y points out that 2,022 and 2.023 together imply that the objects which constitute the substanoe of the world must be the same i n a l l possible worlds and so cannot be contingent. He then oontends that Wittgenstein's claim i n 6.1233 that we could think of a world i n which the axiom of r e d u o i b i l i t y does not hold—which i s equivalent to the claim that we could think of a world i n whioh certain properties whioh are assumed to exist i n v i r t u e of the axiom of r e d u o i b i l i t y do not exist—shows that Wittgenstein regards the existenoe of properties as contin-gent. Therefore, since objeots are not contingent, properties cannot be objeots.1 In response to this argument, I should f i r s t point out that l-Copi, p. 182. 110 Copi i s on precarious ground In trying to build any kind of argument around Wittgenstein's attitude towards the axiom of r e d u o i b i l i t y , because Wittgenstein's attitude i s not a l l that c l e a r . Copi's assumption—which admittedly seems to be borne out by 6.1232 and. 6.1233—is that Wittgenstein regards the axiom of r e d u o i b i l i t y as a contingent proposition, so that, were this proposition f a l s e , there would be at least one higher-order function for which there was no formally equivalent predioatlve function. But surely the whole thrust of Wittgenstein's d i s -t i n c t i o n between functions and operations (5»25)-*-only the l a t -ter of whioh i s involved i n building up a l l other propositions as truth-functions of elementary propositions (5»3) or i n building up anything l i k e the Russellian hierarchy of orders or types ( 5 . 2 5 2 )—is to do away with the need for higher-order&functions e n t i r e l y . The only propositional functions required by the T are those whioh take p a r t i c u l a r s as arguments and whioh, when completed, constitute elementary propositions.1 If I am r i g h t about t h i s , i t i s hard to see what Wittgenstein would have i n mind i n suggesting that the axiom of r e d u o i b i l i t y was a contingent proposition. Even i f we ignore t h i s d i f f i c u l t y , however, and grant that the axiom of r e d u o i b i l i t y i s indeed a ^Granted, examples are given i n 3.333 of functions whioh have other funotions as arguments, but Wittgenstein's p o i n t—v i z . , that we do not need a theory of types to t e l l us that a function cannot be i t s own argument, since this i m p o s s i b i l i t y i s shown by the form of the function i t s e l f—nee d not be taken as a concession that funotions can have other funotions as arguments. He might simply be adopting Russell's viewpoint for the sake of argument and then saying that even i f functions could serve as arguments, we would not i n any case need a theory of types to t e l l us where they could so serve. I l l contingent proposition which asserts that a certain predicative funotion e x i s t s , Copi i s s t i l l mistaken i s supposing that the f a l s i t y of th i s p r o p o s i t i o n , and hence the non-existence of a predicative function of the r e q u i s i t e character as specified by the axiom of r e d u o i b i l i t y , shows that the existence of suoh funotions, or properties, i s contingent. For Wittgenstein, the contingency of a proposition concerns only the configuration of objects and not objeots themselves. Whether the objeots named i n a proposition are or are not i n a oertain configuration i s contingent, but the existence of these objeots themselves i s guaranteed by the very meanlngfulness of the i r names and so can-not be contingent. Even when, by making an e x i s t e n t i a l general-i z a t i o n , we assert the existence of an object without naming i t and thus without presupposing i t s existence, the existence of the object, i n this oase as w e l l , i s not contingent. That i s , we are only claiming that there i s an object answering a certain description; i f our assertion i s f a l s e , this again shows only that a oertain configuration does not e x i s t . For example, the f a l s i t y of ' (3x) x i s red* would show only that no p a r t i c u l a r happens to be configured with redness, but not that the number of p a r t i c u l a r s themselves i s less than i t otherwise might have been.1 S i m i l a r l y , the f a l s i t y of a proposition suoh as ' (3jrf)jrfa', -^It might be supposed that 5*55» which says that "• . , we are not able to give the number of names with d i f f e r e n t meanings • • • " i s evidenoe that Wittgenstein does regard the number of p a r t i c u l a r s , or of objeots generally, to be contingent. I sus-pect, however, that our i n a b i l i t y to give the number of objeots i s thought by him to be a matter only of our having not yet suooeeded i n reaching the ultimate analysis of propositions, the r e s u l t s of whioh, though presently unknown to us, are nevertheless not contingent. In any case, Copi would not p r o f i t by arguing 112 where the function i s generalized, would show only that the objeot a was not configured with; any properties (whioh would be the case, e.g., i f a were to be found only as a term i n various r e l a t i o n s ) , and not that those properties which do exist are oontingent. These same considerations would apply to the e x i s t e n t i a l olaim put forth i n the axiom of r e d u o i b i l i t y , were this axiom conoeded to be a oontingent proposition. The f a l s i t y of t h i s proposition would show only that no predicative function happened to be configured with other objects i n the manner speoified i n the proposition; i t would not show, as Copi olaims, that properties themselves are contingent. Therefore, t h i s argument of Copi's can be dismissed. In his second argument, Copi observes that 6.3751 shows that color-predications are not elementary propositions, since two d i f f e r e n t color-predications can be mutually incompatible, whereas elementary propositions are always compatible. He assumes that i f any properties were simple objeots, colors would be, and since 6.3751 shows that colors are not simple objects, therefore no properties are such objects.1 Edwin A l l a i r e , another defender of a r e a l i s t i c i n t e r p r e t a t i o n of the T, examines this objection at length,! so I w i l l confine myself to a few remarks. Copi's conclusion does not follow, because his assumption i s not against my int e r p r e t a t i o n of 5*55* since, i f the number of p a r t i c -ulars were contingent, then the contingency of properties would not count against t h e i r status as objects e i t h e r . I c o p i , p. 183. 2Edwiri B. A l l a i r e , "The ,T r a o t a t u s,« Nominalistio or Re a l i s t l o ? " i n gT, p. 338ff. 113 shared by Wittgenstein, To be sure, the T treats colors as i f they were simple objeots just as i t treats s p a t i a l points as i f they were objeots. At the same time, Wittgenstein's view that elementary propositions are mutually independent requires that colors be regarded as complex, and the common-sense view that colors are simple must therefore give way to this requirement. We should note that Wittgenstein had already acknowledged the p o s s i b i l i t y of incompatible color-predications i n the patently r e a l i s t i o NB, where he speaks of an inoompatibility i n the structures of the two objects ( i . e . , complex objects) red and green (16.8.16), and where the complexity of colors does not count against properties being objects, but rather counts against colors being (simple) properties (11.9.16). We can easily see, therefore, that 6.3751 provides no evidenoe against the view that Wittgenstein regards properties as objeots. F i n a l l y , I s h a l l consider four arguments advanced by Keyt. The f i r s t argument ooncerns the metaphor of l o g i c a l spaoe: In t h i s metaphor an existent atomlo fact i s the analogue of a material point; a possible atomic f a c t , of a geometrioal point; and an object, of the reference of a co-ordinate of a p o i n t . In a n a l y t i c geometry a point may be indicated by writing three nu-merals i n a oertain r e l a t i o n , for example, by writing "(3* 2, 1 ) " . Si m i l a r l y an atomic faot may be ind1oated by writing names i n a oertain r e l a t i o n . . . . Now i f Wittgenstein held that i n every atomic f a c t one object must be a quality or a r e l a t i o n , then i n the metaphor a point should be indicated not by three co-ordinates but by two co-ordinates and the sign of a function . . . .1 As should be obvious from my previous discussion of l o g i o a l spaoe, the d i f f i c u l t y which Keyt poses for realism arises only because he has the analogues a l l wrong. Thus, a material p o i n t , or •"-David Keyt, "Wittgenstein's Notion of an Object," i n JST, pp. 290-291. 114 s p a t i a l objeot, has as i t s analogue, not an existent state of a f f a i r s , but an objeot i n l o g i o a l space. A geometrical, or s p a t i a l , point has as i t s analogue, not a possible state of a f f a i r s , but any of those objects the t o t a l i t y of which con-s t i t u t e s the l o g i c a l space of the objeot i n t h i s space, just ass space proper i s constituted by the t o t a l i t y of s p a t i a l points within whioh the s p a t i a l object i s sit u a t e d . As far as states of a f f a i r s are concerned, there i s no question of an analogue between space proper and l o g i o a l space; i n either case, states of a f f a i r s i n the usual sense r e s u l t when the objeot i n space i s configured with one of the objects whioh together constitute that spaoe. Sinoe, on this account, objects i n l o g i o a l space are not the analogue of co-ordinates, Keyt's objection that a function-sign i s the wrong kind of sign to be a co-ordinate loses i t s f o r c e . Keyt's other three arguments concern the 'chain* simile of 2.03. The f i r s t argument i s that the notion that " . . . objeots hang together l i k e the l i n k s of a chain • • ." (2.03), suggests that objects are a l l fundamentally a l i k e . Therefore, objects cannot include both universals and p a r t i c u l a r s , sinoe these are fundamentally unlike one another.1 The answer to this argument i s that, for Wittgenstein, p a r t i c u l a r s and universals are funda-mentally a l i k e , as I have shown, and 2.03 can therefore be no evidence against realism on this account. Keyt's second argument Is that a r e l a t i o n Is of a higher 1Keyt, "Wittgenstein's Notion," p. 290. See also P i t c h e r , p. 114, and Mueller, p. 84. 115 type than i t s terms, so that a r e l a t i o n cannot be an object i n •1 the chain, since a l l such objects are surely of the same type, This argument f a i l s to see that, not only i n the T, but even i n the e a r l i e r t r a n s i t i o n from (1) to (2), one of Wittgenstein's main objectives has been to render the theory of types super-f l u o u s . There i s no need i n the T for a type-distinotion to insure that we do not try to give r e l a t i o n s a syntaotio r o l e appropriate only to p a r t i c u l a r s , sinoe the proper syntactic r o l e for eaoh kind of object i s 'written i n t o ' the object i t s e l f , i n vi r t u e of the object's l o g i o a l form. Hence, p a r t i c u l a r s , prop-e r t i e s , and r e l a t i o n s , d i f f e r i n g as they do both i n form and i n content, nevertheless stand on equal footing as objeots. Keyt's t h i r d argument i s that i f we include r e l a t i o n s among objeots, the chain-simile breaks down, because a three-termed r e l a t i o n i s more l i k e a key-ring that holds three keys than i t i s l i k e the l i n k i n a chain, which i s connected at best with only two other l i n k s .2 This argument, to be sure, i s e v i -denoe for the truism that s i m i l e s , when pressed, tend to break down, but there i s no reason to suppose that Wittgenstein him-s e l f intends the simile to extend t h i s f a r . Surely the point i s that, just l i k e the l i n k s of a chain, objeots i n a state of a f f a i r s f i t into one another d i r e c t l y , without the need of some connecting 'cement' suoh as the oopula whioh Wittgenstein had o r i g i n a l l y postulated i n p o s i t i o n (1). The fact that a ohain-•'•Keyt, "Wittgenstein's Pioture Theory of language," i n ET, P. 338. """* 2Keyt, "Picture Theory," pp. 383-384. 116 l i n k connects with two l i n k s while a r e l a t i o n may connect I t s e l f with three, or twenty, terms i s , for the point of the s i m i l e , i r r e l e v a n t . •2.03, then, i s i n no way incompatible with realism; on the contrary, t h i s passage i s quite l i k e l y intended to support r e a l -sim, p a r t i c u l a r l y against the BradleyIan charge that the admission of r e l a t i o n a l e n t i t i e s leads to an i n f i n i t e regress. Whether or not Wittgenstein ever read Bradley, we can presume that he was acquainted with at l e a s t that famous paragraph from Appearance  and Reality (sinoe i t i s quoted by Russell i n OKEW. whioh Witt-genstein did read) i n whioh Bradley argues that i f a r e l a t i o n i s to be anything to i t s terms, i t must i t s e l f bear a r e l a t i o n to these terms, with the r e s u l t that . . . we are hurried o f f into the eddy of a hopeless pro-cess, since we are forced to go on finding new r e l a t i o n s without end. The l i n k s are united by a l i n k , and thi s bond of union i s a l i n k whioh also has two ends; and these require each a fresh l i n k to connect them with,the o l d .1 Bradley assumes that the l i n k i n g of a r e l a t i o n with one of i t s terms r e q u i r e s , per Impossible, an i n f i n i t e number of i n t e r -mediate l i n k s . This conception of the way a r e l a t i o n manages to l i n k i t s e l f to a term treats the l i n k as something external to the nature and function of the r e l a t i o n i t s e l f . I f we take i n -stead the Fregean approach and allow r e l a t i o n s themselves to perform t h i s l i n k i n g r o l e by supposing that, apart from t h e i r actual performanoe of t h i s r o l e , r e l a t i o n s are merely incomplete, unsaturated e n t i t i e s , we eliminate the need for postulating 1?. H. Bradley. Appearance and R e a l i t y , pp. 32-33, quoted i n OKEW. p. 14. 117 intermediate l i n k s . In other words, on Bradley's view, for a r e l a t i o n to be something to i t s terms, i t must be more than i t i s i n i t s e l f apart from connection with i t s terms, and t h i s 'something extra' oan only be another, intermediate r e l a t i o n . For Frege, on the other hand, a r e l a t i o n i s only f u l l y what i t i s i n the act of being saturated by i t s terms, thereby r e l a t i n g them. Apart from th i s state of saturation, the r e l a t i o n retains i t s i d e n t i t y as a d i s t i n c t i v e e n t i t y , but only i n a kind of p r i -vative status, so that when the saturation does come about, i t i s more as though a p o t e n t i a l r e l a t i o n were actualized than that something extra i s added to what was already a s e l f - s u f f i c i e n t , independent e n t i t y . Wittgenstein, as we have seen, adopts the Fregean approach, though modifying i t to include not only r e l a t i o n s , but th e i r terms as w e l l , among unsaturated e n t i t l e s . So, while i n Frege's conception, a r e l a t i o n might be appropriately oompared to a board with a number of holes into whioh f i t corresponding pegs, the r e l a t i n g of the pegs being accomplished by th e i r being plugged into the holes, the chain-simile of 2.03 i s more appro-p r i a t e for Wittgenstein, sinoe each l i n k must be entered i n t o , or 'saturated* as i t were, by another l i n k i n order to be part of the ohain (whereas f o r Frege the status of the pegs, at l e a s t , would seem to remain e s s e n t i a l l y the same whether or not they were plugged i n ) . I would contend, then, that the chain-si m i l e i s c a r e f u l l y chosen by Wittgenstein and i s probably meant to i l l u s t r a t e his modified Fregean response to Bradley.1 •'•An in t e r e s t i n g comparison oan be made here with a passage 118 That t h i s should be the nature of Wittgenstein's answer to Bradley w i l l no doubt disappoint those who suppose that one of the most ingenious accomplishments of the T l i e s i n getting r i d of r e l a t i o n s altogether—or at l e a s t i n r e t a i n i n g them i n a way whioh avoids Bradley's criti c i s m s—by the doctrine that r e l a t i o n s are only shown, and not s a i d , by language i n i t s properly analyzed form* Of course, i f Wittgenstein's view were that we avoid trouble by employing a symbolism whioh completely excludes relation-words, thereby depriving Bradley of a way of formulating h i s objection to r e l a t i o n s , this would s t i l l leave unclear the ontologioal status of r e l a t i o n s . I have called suoh an i n t e r -p retation of the T 'nominalistio' f o r the sake of convenience, but suoh a view i s not nominalistio i n any ordinary sense. I t i s not that r e l a t i o n s themselves are held to be a ' l i n g u i s t i c f l o t i o n ' , on t h i s i n t e r p r e t a t i o n ; ordinary, unanalyzed language misleads us only i n supposing that we represent r e l a t i o n s i n the same way that we do p a r t i c u l a r s— v i z . , according to the name-object treatment—but that we dj> somehow represent r e l a t i o n s— v i z . , by their being shown i n a perspicuous language—is not denied. So the question remains as to what i t i s that Is being represented. I f a r e l a t i o n i s not an objeot, does i t nevertheless have a kind of oontent? Or i s i t perhaps a kind of oontentless, s t r u c t u r a l entity? And i n either oase, to what extent has a r e a l i s t ontology of some sort thereby been avoided, i f at a l l ? i n PB whioh resembles 2.03 of the T i n i t s general purport and, i n doing so, gives a r e a l i s t i c treatment to colors: "It i s clear that there i s no r e l a t i o n of 'finding o n e s e l f whioh exists be-tween a color and a place i n which the color 'finds i t s e l f . There i s no middle-member between color and space. Color and space saturate one another"Q(PB, p. 257). 119 There i s the further problem, again supposing that the view that r e l a t i o n s are. shown rather than said were indeed Witt-genstein's, that whatever gains might be had i n being able to answer Bradley i n a way that.avoids treating r e l a t i o n s as ob-jects would seem to be at l e a s t p a r t i a l l y o f f s e t by the incr e d -i b l e complexity of any supposedly perspicuous symbolism which attempted to carry out the programme of showing d i f f e r e n t r e l a t i o n s by means of the various s p a t i a l arrangements of names of p a r t i c u l a r s . Now, i t : i s no doubt true that, a symbolism whioh renders 'aRb' as *ab', *aSb* as 'ba', and so on, i s i n some sense as perspicuous as the symbolism i t replaces, but we need l i t t l e imagination to see that the obstacles i n developing suoh a sym-bolism to handle more than a few two- or three-plaoe r e l a t i o n s would be formidable—the prospect i s even ludic r o u s . I oonoede, however, that the bizarreness of the programme of 'showing' r e l a t i o n s by scattering symbols a l l over a page does not by i t s e l f show that Wittgenstein could not have held t h i s view; one could argue, I suppose, that Wittgenstein's concern i s solely with the p o s s i b i l i t y , and. not the p r a c t i c a l i t y , of thi s sort of symbolism, since his programme stems entirely from a p r i o r i con-siderations which demand t h i s 'showing* capability of the sym-bolism as a l o g i c a l requirement. Fortunately, there i s no need to pursue these matters, at l e a s t as they bear on the views of the T i t s e l f , f or we already have ample evidence that the doc-t r i n e of showing^is by no means intended to do away with r e l a t i o n s or properties as named objeots. Wittgenstein i s not, with his dootrine of showing, trying to avoid the d i f f i c u l t i e s posed by 120 Bradley, and therefore the p e c u l i a r i t i e s of t h i s nominalist version of the doctrine need not concern us f u r t h e r . Having said t h i s , I bring to an end my survey of the various arguments i n favor of a nominalistio interpretation of the A l l such arguments have been shown to be based upon misunder-standings, and none poses a serious challenge to my own view, I submit, therefore, that the realism of the T has been con-c l u s i v e l y established. There i s one other prominent c r i t i c of a r e a l i s t i o i n t e r -p retation of the T—James G r i f f i n , i n Wittgenstein's Logioal  Atomism—whose viewpoint I have not s p e c i f i c a l l y examined, as his arguments overlap those already considered. I t might be worth pointing out that one of G r i f f i n ' s main theses, whioh he attempts to support by his nominalist int e r p r e t a t i o n of the Tf« v i z . , that Wittgenstein's notion of analysis i n the T i s very much d i f f e r e n t from R u s s e l l ' s—G r i f f i n ' s contention being that analysis f o r Russell oonsists i n s h i f t i n g descriptions from the subject to the predicate p o s i t i o n , whereas for Wittgenstein descriptions are broken up into sub-descriptions whioh remain i n the subject p o s i t i o n ( G r i f f i n , p, 48), this process con-tinuing u n t i l a l l general terms are analyzed away—is not only oontradioted by my contention that general terms are not ana-lyzed away, but also by Wittgenstein's remark i n £Gs ''I have myself i n e a r l i e r times spoken of a 'complete a n a l y s i s ' , with the thought that philosophy had to dissect a l l propositions com-p l e t e l y i n order to c l a r i f y a l l connections and to eliminate every p o s s i b i l i t y of misunderstanding. As i f there were a c a l -culus i n which th i s dissection were p o s s i b l e . I imagined some-thing l i k e the kind of d e f i n i t i o n whioh Russell had given for the d e f i n i t e a r t i c l e " (PG., p. 211). This shows that Wittgen-s t e i n oonoeived his own programme of analysis along Russelllan l i n e s rather than according to the a l t e r n a t i v e theory of des-cri p t i o n s whioh G r i f f i n proposes. (Further evidence that the early Wittgenstein i s b a s i c a l l y i n sympathy with Russell's theory of descriptions i s to be found i n NB. p. 128, and i n £ B , pp. 200-201.) PART TWO THE LATER WITTGENSTEIN 122 INTRODUCTION TO PART TWO The f i r s t part of this work has shown that the T, far from being nomlnalistio or even leaning i n that d i r e c t i o n , i s on the contrary an example of P l a t o n i s t1 metaphysics par excellence. Granted, t h i s i s a Platonism tempered by the influence of R u s s e l l , i n whioh the programme of analysis allows us to r e f r a i n from pop-u l a t i n g the universe with universals such as horseness or table-ness, the presumption being that, e.g., the expression 'table* oan be eliminated i n favor of those simpler expressions which define i t , so that only those elements whioh remain as the end-products of analysis are the true u n i v e r s a l s . S t i l l , i f the excesses of Platonism have been curbed, these simple elements, at any r a t e , have the same f u l l - f l e d g e d ontologioal status as the p a r t i c u l a r s with which they are configured i n states of a f f a i r s . I t i s against the background of thi s endorsement of Platonism that we must view Wittgenstein's subsequent attaok upon i t . Wittgenstein's attack on Platonism i n his l a t e r philosophy i s well known. As he says i n the BB, we have a p r i m i t i v e , too simple idea of the structure of language when we suppose that • • ••• Properties are ingredients of the things whioh have the properties; e.g. that beauty i s an ingredient of a l l beau-I mean by 'Platonism' any view which accords universals at lea s t equal status with p a r t i c u l a r s . For Plato himself, of course, the immutability of universals makes them even more r e a l than p a r t i c u l a r s , and the T, i n seeking to minimize the d i f f e r -enoes between the two, i s obviously not Pl a t o n i s t i n t h i s sense. 123 t i f u l things as alcohol i s of beer and wine, and that we there-fore oould have pure beauty, unadulterated by anything that i s beauti f u l (BB, p. !?)• What i s not commonly appreciated, of oourse, i s that the view of properties as ingredients, capable of existenoe apart from the things i n which they are 'contained*, i s prec i s e l y Wittgen-stein's own e a r l i e r p o s i t i o n i n the T. The importance of appre-c i a t i n g t h i s i s that i f we do not, and instead assume that this attack on Platonism i s just part of a generally a n t i - r e a l i s t a t t i t u d e prevalent i n the T as w e l l , we w i l l overlook the pos-s i b i l i t y that Wittgenstein's attack i s not against realism as such, but only against a Platonic r e a l i s m , i n whioh properties are held to exist separately from things which have these prop-e r t i e s . There i s , of oourse, another version of realism, put forward most notably by A r i s t o t l e , i n which properties r e a l l y e x i s t , though only i n things (In Re) as th e i r common q u a l i t i e s and not apart from these things (Ante Rem). Now, there can be no doubt that a major thrust of Wittgenstein's l a t e r philosophy i s not just against Platonic realism, but against an u n c r i t i c a l realism as suoh. The c r i t i c i s m of "The tendency to look for something In oommon to a i l the e n t i t i e s whioh we commonly sub-sume under a general term . . . " (Bg, p. 17) applies both to Platonic and A r i s t o t e l i a n r ealism. The issue, however, i s not whether Wittgenstein rejeots the A r i s t o t e l i a n account of the use of most general terms—for he c l e a r l y does—but whether he rejeots suoh an aooount for a l l general terms. My contention i s that Wittgenstein aoes i n faot acknowledge that some general words are names of non-particular objects, i . e . , u n i v e r s a l s , and 124 that he Is therefore an A r i s t o t e l i a n r e a l i s t to at l e a s t a limi t e d extent. This contention requires two q u a l i f i c a t i o n s . F i r s t , I am not maintaining that the l a t e r Wittgenstein thinks of himself as an advocate of a limited A r i s t o t e l i a n realism; indeed, i t seems d e a r that he thinks he has originated a revolutionary method of philosophizing whioh obviates the need for taking sides i n t r a d i t i o n a l disputes, Including ontologioal disputes. My contention i s that, whatever he may think, his method of philosophizing i s not ontologioally n e u t r a l , and that his stand-point i s one of a limited r ealism. Secondly, even though Witt-genstein i s a r e a l i s t , his ontologioal p o s i t i o n d i f f e r s r a d i c a l l y from A r i s t o t l e ' s , or from other t r a d i t i o n a l p o s i t i o n s , i n being held to be determined by and dependent upon language, a p o s i t i o n which i s the reverse of the more orthodox view that language i s a r e f l e c t i o n of ontology. So, whatever ontologioal p o s i t i o n Wittgenstein can be said to hold, i n his view i t i s to language, rather than to an independent r e a l i t y , that we must ultimately address ourselves. Both of these q u a l i f i c a t i o n s w i l l be d i s -cussed i n greater d e t a i l i n the l a t t e r part of Chapter V and i n Chapter VII, r e s p e c t i v e l y . I s h a l l begin, however, by examining my own i n t e r p r e t a t i o n of Wittgenstein i n l i g h t of other i n t e r -pretations which would deny my claim that he i s even i m p l i c i t l y a limited A r i s t o t e l i a n r e a l i s t . 125 CHAPTER V WITTGENSTEIN'S MODERATE REALISM A l l commentators, so f a r as I am aware, agree that the l a t e r Wittgenstein i s not a r e a l i s t , though there i s some d i s -agreement among them concerning the nature of his attaok upon re a l i s m . In t h i s chapter, I s h a l l examine the views of various commentators and defend my own interpretation against t h e i r s . I begin by considering the view of Renford Bambrough, who thinks that Wittgenstein i s opposed to any version of realism and that t h i s opposition i s based on his doctrine of family-resemblances. Bambrough c i t e s #66-6? of the P I—i n which Wittgenstein says that the general term 'game' i s applied to i t s various instances, not on the basis of some element common to them a l l , but rather on the basis of a network of s i m i l a r i t i e s whioh serve to form a family of oases—and then argues that what holds for the general term 'game' holds for a l l general terms. Thus, Bambrough contends against Ayer that Wittgenstein does not i n t h i s example mean to contrast a complicated, concept l i k e 'game*1 with a simple con-cept l i k e 'red*, because the observation that there i s no essen-t i a l element common to a l l instances of games applies equally well to a l l instances of red things. To show that t h i s i s Wltt-^I should point out that the l a t e r Wittgenstein no longer regards concepts as a kind of ontologioal item. As £1 #532 shows, he now thinks of a oonoept as a use of words; e.g., my concept of understanding i s constituted by the kinds of ways I use the word 'understanding*. Accordingly, my practide i n Part Two w i l l be to indicate the subject-matter of a concept by single quotes—e.g., I w i l l say "the concept *game*" rather than, as i n the s p i r i t of Part One, "the oonoept game" or "the concept gameness"—in order to indicate that t h i s subject-matter i s i t s e l f l i n g u i s t i c rather than ontologioal. 126 genstein's own view, Bambrough c i t e s two passages. The f i r s t i s fr,om the BRB. p. where Wittgenstein poses the r h e t o r i c a l questions "Could you t e l l me what i s i n common between a l i g h t red and a dark red?"—the implication being that there need be no!thing i n common. The second i s PI #73, where Wittgenstein again askss "Which shade i s the 'sample i n my mind' of the color green—the sample of what i s common to a l l shades of green?"—with the same implication as before.1 In fairness to Ayer, i t should be said that i f Bambrough's conclusion were co r r e c t , Ayer would be no less oorreot i n sup-posing that 'game' and 'red' d i f f e r markedly i n that the former term does not, while the l a t t e r does,; mark 'a simple and s t r a i g h t -i forward resemblance* between the things to which the word i s ap-p i l e d , because Ayer*s p o s i t i o n i s i t s e l f n o n - e s s e n t i a l i s t . That i s i , for Ayer the difference between *game* and 'red' i s not that the l a t t e r term names a common qu a l i t y of i t s instances while the former term does not, but only that the resemblance between red things i s more straightforward than the resemblance between games. So Bambrough's disagreement with Ayer i s not over the account of the difference between 'game' and 'red', but only oyer the question of whether the point of PI #66-67 i s to empha-si z e the contrast between complicated and simple conceptsj Bam-brough thinks i t i s not. ; Bambrough's argument i s r e a l l y d i r e c t e d , not against Ayer, but against a defender of the view that at l e a s t some general terms are applied to th e i r instances i n v i r t u e of the presence I ^ e n f o r d Bambrough, "Universals and Family Resemblances," i n WPI, p. 193. i • i 1 • 1 • 127 of a common element i n these instances. leaving aside for the moment the question of Wittgenstein's own intentions, oan we say that the two passages cited by Bambrough are s u f f i c i e n t to refute the defender of a moderate realism, provided he accepts t h e i r implication? C l e a r l y , I would contend, they are not. The r e a l i s t can r e a d i l y accept that d i f f e r e n t shades of red are a l l ca l l e d 'red' only because they resemble one another and not be-cause they possess a common q u a l i t y . I t i s not the general term 'red*, but the general term for a s p e c i f i c shade of re d , e.g., 'fire-engine red*, whioh provides a better example for the r e a l -i s t . D ifferent things whioh are fire-engine red do not merely resemble one another, as something l i g h t red resembles something dark red; r a t h e r , they each possess a quality common to them a l l . Surely, i f I out three hankderchiefs from a piece of u n i -formly dyed ol o t h , i t would be a strained use of language to say that the three handkerohiefs, resembled one another i n respect of t h e i r c o l o r , or that th e i r colors were quite s i m i l a r . Rather, we would say that they have the same co l o r . In f a c t , our l i n -g u i s t i c practices tend to lean i n the opposite d i r e c t i o n : i n most contexts we would c a l l d i f f e r e n t shades of a oolor the same color when, s t r i c t l y speaking, there would be only s i m i l a r c o l o r s , so that, i n the context under consideration, where the purpose i s to make the most precise discrimination possible with respect to observed differences of co l o r , we would say that the color of the handkerchiefs i s exactly or pr e c i s e l y the same, to d i s t i n -guish the oases where there was no observable difference of oolor at a l l from those oases where there were at lea s t s l i g h t 128 d i f f e r e n c e s .1 I t could be objected that I have assumed that the three handkerchiefs are cut from a uniformly dyed piece of olo t h , which insures that the observable differences between the hand-kerchiefs with respect to oolor w i l l be n i l , but that i n fa c t no general term, even one ostensibly standing for a s p e c i f i c shade of c o l o r , ever a c t u a l l y names just p r e c i s e l y one shade, because even i f the three handkerohiefs have p r e c i s e l y the same shade of oolor, i t i s nevertheless not i n v i r t u e of having t h i s s p e o i f i c c o l o r - q u a l i f y , and no other, that they are cal l e d ' f i r e -engine red* or whatever, since they might have varied s l i g h t l y i n color and s t i l l properly be cal l e d 'fire-engine red'. While th i s c r i t i c i s m i s , from a oertain perspective, a just one, i t i s not f a t a l l y damaging, for onoe we admit that two or more things can have pr e c i s e l y the same c o l o r , the way has been cleared i n p r i n o l p l e for the recognition of suoh common q u a l i t i e s , though i n p r a c t i c e we may not be equipped to make such precise d i s t i n c t i o n s with the general terms at our disposal where, at l e a s t i n the i r normal use, suoh a degree of p r e c i s i o n would be i r r e l e v a n t . Paint stores, on the other hand, might have occasion to use general terms for naming s p e c i f i c shades of color with the r e q u i s i t e degree of p r e c i s i o n , a feat they could Wittgenstein's argument i n PI #88 that there i s no i d e a l standard of exactness does not a f f e c t the point I am making here. I am not maintaining that there i s an absolute standard of exactness which renders the normal, wide use of 'same' i n -oorrect, but only that this wide use would be loose and mis-leading i n a context where the purpose i s to express one's deteotion of the s l i g h t e s t observable differences i n c o l o r . 129 e a s i l y manage by u t i l i z i n g a oolor-ohart ana stressing that, i n applying a color-name to a patch of p a i n t , no observable d i f f e r -ence from the sample i n the oolor-ohart corresponding to that name was to be permitted. Nothing i n p r i n c i p l e , then, prevents t h i s degree of rigour from being i n fo r c e , ana thi s i s a l l that we neea to oounter Bambrough*s claim that realism oannot possibly be true because there oannot be any general terms that name oom-mon elements. My argument i s that there could* and even very well may, be suoh general terms. Besides s p e c i f i c shades of co l o r , other l i k e l y candidates for the status of common elements nameable by general terms would be s p e c i f i c t a s t e s , smells, tex-tures, e t c .— i n short, s p e o i f i o sensible q u a l i t i e s . I f i t i s granted that Bambrough's a l l - o u t attack on realism i s at le a s t prima f a c i e unoonvincing, the question then becomes: i s Bambrough*s position,, nevertheless, also Wittgenstein's? The answer, I submit, i s that Wittgenstein does not try to dispose of a l l u n i v e r s a l s , i n Bambrough*s fashion, but rather that he i s content to maintain a limitea realism i n which at lea s t s p e c i f i c shades of color are regarded as common elements of the things they q u a l i f y . I w i l l attempt to make the argument only for c o l o r s , i f for no other reason than that those remarks of Witt-genstein's which can be cited as p o s i t i v e textual evidence f o r his realism are themselves invariably concerned with the topic of c o l o r . What other klnas of universals he would acknowledge, i f any, I am not prepared to say, though a good case could also be made for shapes, since the topios of oolor and shape are f r e -quently discussed together i n a way that suggests that Wittgen-130 s t e i n treats colors and shapes on a par as the most obvious examples of q u a l i t i e s common to many p a r t i c u l a r s . In any case, sinoe the common view i s that the l a t e r Wittgenstein i s , for one profound reason or another, opposed to realism i n p r i n c i p l e , and not just i n p r a c t i o e , my argument that he admits the e x i s t -enoe of at least one kind of universal w i l l be s u f f i c i e n t to show that the common view i s seriously mistaken. Before c i t i n g my p o s i t i v e evidence, I want f i r s t to rebut Bambrough's claim that Wittgenstein intends the family-resemblance treatment for a l l oonoepts. The i n v i t a t i o n i n PI #66 for us to 'look and see' i f there i s a common element, an essence, of a l l games i n v i r t u e of which they are called 'games' shows that the issue i s not to be determined by general, a p r i o r i considerations, as Bambrough supposes, but i s rather to be determined by seeing how, i n each case, the general term i n question actually oper-a t e s , the implication being that some general terms may very well name oommon elements. Furthermore, i n the course of arguing that there i s no common element of gamehood, Wittgenstein makes clear that he i s not attacking the notion of a oommon element as such, f o r , i n denying that there i s a oommon element i n a l l i n -stanoes of games i n v i r t u e of whioh they are games, he admits that various kinds of games might have oommon elements. Thuss "look f o r example at board-games . . . Now pass to card-games . • . many common features drop out, and others appear" (PI, #66). So, within the concept 'game', there are various sub-ooncepts such as 'board-game', 'oard-game', 'ball-game*, e t c . , and though the general term *game* i s not applied i n v i r t u e of the presence of a oommon element, we can by no means conclude that the general 131 terms •board-game*, 'ball-game', e t c . , follow s u i t . On the con-t r a r y , 'board-game* i s presumably applied to a l l i t s Instances i n v i r t u e of t h e i r having i n common that they are a l l played on a board, 'ball-game* i s applied to those games which involve the use of a b a l l , and so on. The emphasis of Wittgenstein's attack, then, i s not against as e s s e n t i a l i s t aooount of the use of concepts generally, but rather i s directed against the s i m p l i s t i c a t t i t u d e which advo-cates the e s s e n t i a l i s t account not only for a sub-concept of some more i n c l u s i v e oonoept (e.g., a species-oonoept, such as 'board-game', of the genus-conoept 'game'), where such an acoount might be p e r f e c t l y proper, but also extends t h i s e s s e n t i a l i s t aooount to the more inolusive oonoept as w e l l . Thus, the term 'board-game' may be applied i n v i r t u e of i t s instances having a common element, but the mistake i s i n supposing that what holds for 'board-game' also holds for 'game'. An exact p a r a l l e l can be drawn i n this respect between Wittgenstein's discussion of the concept 'game* and the concept •number*. The general term 'number*, we are told i n PI #6?, i s applied to a l l i t s instances, not i n v i r t u e of t h e i r possessing a common element, but rather i n v i r t u e of a l l these instances having various resemblances. At the same time, the sub-concepts •cardinal number*, 'rational number', 'real number', et c . , admit of an e s s e n t i a l i s t aooount, as th e i r genus-conoept 'number' does not. Wittgenstein implies t h i s i n his response to the hypothet-i c a l i n terlocutor who t r i e s to make a case for the e s s e n t i a l i s t acoount of the genus-oonoept 'number' by saying that what i s 131 oommon to the various kinds of numbers i n vi r t u e of whioh the general term 'number' applies to them a l l i s ". • • the d i s -junction of their common properties (Gemeinsamkei ten) . . . " (PI,#67). Wittgenstein's reply here i s that the interlocutor i s only " . . . playing with words. One might as well say: 'Some-thing runs through the whole thread—namely the continuous over-lapping of those f i b r e s ' . " Now, suoh a reply would be completely inappropriate i f Wittgenstein were attacking the assumption that oommon elements existed even within various kinds of numbers. Surely he i s admitting that suoh common elements do exist i n connection with the sub-oonoepts, his point being that i t i s a sop h i s t i c t r i o k which t r i e s to do for the genus-concept 'number* what oan be genuinely and.helpfully done for the species-concepts •cardinal number', 'rational number', etc., v i z . , to give an ao-oount of the use of the term by reference to an essential feature or group of features belonging to a l l of the instances to which the term i s applied. Obviously, then, Wittgenstein intends the family-resemblance acoount only for some concepts and not for others; suoh an account cannot, therefore, even i m p l i c i t l y be intended to have any bearing on any general solution of the •problem of universals', as Bambrough claims, and certainly can-not be used to show that sensible q u a l i t i e s suoh as s p e c i f i c shades of color are not universals. P o s i t i v e evidence that Wittgenstein admits that d i f f e r e n t p a r t i c u l a r things can have a color as their oommon quality i s found i n PI #72 (a simi l a r disoussion also occurring on pp. 131-132 of the MB). I w i l l f i r s t put this passage i n i t s proper 133 s e t t i n g so that i t s f u l l import may be understood. In #?1, Witt-genstein i s defending the appropriateness—indeed, the necessity— of explaining cert a i n concepts such as 'game* by giving various examples of games rather than by formulating a general d e f i n i t i o n whioh w i l l apply to a l l examples. In the case of the concept 'game1, the absenoe of any element common to a l l games i n vi r t u e of whioh they are called 'games* makes any general d e f i n i t i o n impossible, or at l e a s t unwieldy and useless, unless we concede-as Wittgenstein apparently does not—that there i s a s u f f i c i e n t l y small core of properties to allow, as something more substantial than an ad_ hoo exercise, a general disjunctive d e f i n i t i o n of 'game' along the l i n e s of that suggested i n PI #6? for 'number*. Also, and perhaps the more decisive reason for our being confined here to giving examples rather than a general d e f i n i t i o n , the concept •game* has 'blurred edges* or, to use Waismann*s expression, i s * open-textured*—i.e., we do not give the concept *game* a s t r i c t boundary which provides us with a procedure for deciding of every-thing whether or not i t i s a game—and this shows, not only that we must be oontent with explaining the concept 'game* by giving examples, but furthermore that there i s not even suoh a thing here as enumerating a l l the examples. We must be oontent to describe various games and to adds "This and si m i l a r things are called 'games'" (PI, #69), where the phrase 'and sim i l a r things' cannot be more precise l y f i l l e d i n . Thus, when I point out to someone examples of games i n the hope of bringing him to understand what games are, I am not using an i n d i r e c t method whioh could be replaced by the d i r e c t method of i s o l a t i n g the 1 3 ! essential element of gamehood and direoting his attention to that alone, i n abstraction from the p a r t i c u l a r examples. There i s not even a p o s s i b i l i t y of enumerating a l l the relevant ex-amples and thereby achieving a comprehensive, i f not an essen-t i a l i s t , account of the ooncept 'game', for this concept has 'blurred edges'. Rather, i t i s the nature of the case that I can only give p a r t i c u l a r examples and hope that they are under-stood i n the relevant way. That th i s kind of explanation oan be misunderstood does not make i t i n f e r i o r to an e s s e n t i a l i s t account, sinoe ". • • any general explanation oan be misunder-stood too" (PI, #71). The upshot of #71. then, i s that certain oonoepts are legitimately and solely explained by the giving of p a r t i c u l a r examples. Now, the implication of #71, and the point that #72 proceeds to discuss, i s that there i s more than one way that examples oould be involved i n the explanation of a concept. That i s , #71 says that, i n explaining to someone what a game i s by giving examples, I do not . . . mean by this that he i s supposed to see i n those examples that oommon thing whioh I—for some reason—was unable to express . . . Here giving examples i s not an i n d i r e c t means of explaining—in default of a bet t e r . . . . The point i s that t h i s i s how we play the game (I mean the language-game with the word "game".) This suggests t h a t , while giving examples i s as direot a means as one could have for explaining the ooncept 'game', there are other ooncepts i n whioh the giving of examples would be merely an i n d i r e c t means of explanation, capable of being replaced by a more d i r e c t one, presumably one i n which the oommon element i n the examples could be d i r e c t l y pointed out. That Wittgenstein 135 believes there can be both i n d i r e c t and d i r e c t explanations of oonoepts by means of giving examples i s oonfirmed by #72, sinoe the purpose of th i s passage i s to i l l u s t r a t e these d i f f e r e n t kinds of explanations. The passage I t s e l f Is as follows t Seeing what Is oommon. Suppose I show someone various multi-coloured p i o t u r e s , and say: "The colour you see i n a l l these i s oalled 'yellow ochre'".—This i s a d e f i n i t i o n , and the other w i l l get to understand i t by looking for and seeing what Is oommon to the p i c t u r e s . Then he can look a t , oan point t o , the common thing. Compare with t h i s a case i n whioh I show him figures of d i f f e r e n t shapes a l l painted the same colour, and say: "What these have i n oommon i s oalled 'yellow ochre'". And compare t h i s case: I show him samples of d i f f e r e n t shades of blue and say: "The colour that i s oommon to a l l these i s what I o a l l 'blue'". Three d i f f e r e n t oases are being considered here. The f i r s t oase i s obviously intended as an i l l u s t r a t i o n of an i n d i r e c t ex-planation whioh oould be replaced by a d i r e c t one. Instead of explaining the conoept 'yellow ochre' by saying that i t i s the color oommon to a l l the pi o t u r e s , the instructor oould d i r e c t l y display t h i s quality oommon to a l l the examples by using a color-chart or by pointing to the color as i t i s found on one or more of the p i c t u r e s . The seoohd oase i s also an i l l u s t r a t i o n of an in d i r e c t explanation, s l i g h t l y d i f f e r e n t from the f i r s t i n that the instructee Is not told what kind of oommon element he i s looking f o r , and he must determine by a process of elimination that i t i s the oolor and not, e.g., the shape, whereas i n the f i r s t oase the instruotee knows what kind of oommon quality he i s looking f o r , and the prooess of elimination i s oonfined to the task of ploking out the correct oolor among others. Sinoe the same concept i s being explained i n the second case as i n the f i r s t , the second kind of i n d i r e c t explanation oould be replaced by the same d i r e c t explanation that could replaoe the f i r s t explanation. The third case i s s i g n i f i c a n t l y d i f f e r e n t from the other two. Here, I take the point to be that various samples of d i f f e r e n t shades of blue could not be dispensed with i n favor of a d i r e c t display of thei r common quality of blueness, because there i s no suoh oommon quali t y to be found. We o a l l these various samples 'blue*, not i n . v i r t u e of a quality they commonly share, but i n v i r t u e of resemblances which they bear to one another. The ooncept 'blue' i s , i n thi s respect, l i k e the oon-oept 'game*, both suoh oonoepts being contrasted with a oonoept l i k e 'yellow ochre', where the general term stands for a oommon qua l i t y that oan be d i r e c t l y displayed.1 In #73 Wittgenstein does go on to say that even a concept l i k e 'green' or 'blue* could be explained by a single color-sample, but suoh a oolor-sample would not be a d i r e c t display of the color-quality blue or green, as would a oolor-sample of yellow-ochre (assuming that 'yellow ochre* i s here regarded as the name of a s p e c i f i c shade of o o l o r ) , because the former sort of sample would have a d i f f e r e n t kind of app l i c a t i o n than the l a t t e r , representing a whole range of c o l o r - q u a l i t i e s rather than one s p e c i f i c c o l o r - q u a l i t y . The notion that i t i s the use to which a sample i s put that determines whether i t i s a s p e c i f i c •^Admittedly, i n normal use even 'yellow oohre', l i k e 'blue', i s applied to more than one shade of oolor, ranging from a pale yellow to an orangish or reddish yellow; s t i l l , i t i s clear from the context of #72 that 'yellow oohre' i s here being used, not to cover a whole range of shades of a oolor, but rather to r e f e r to a s p e c i f i c shade of oolor which i s the oommon quality of a number of things. sample of pure green or a general sample representing many d i f f e r e n t shades of green, rather than some i n t r i n s i c differenoe i n the two samples—since the same sample could he used for either purpose—shows that even a s p e c i f i c sample aots as a d i r e c t display of a cert a i n color only i n the oontext i n whioh we treat the sample accordingly. Apart from th i s oontext of human a c t i v i t y , the sample i s 'dead', i . e . , i t has no i n t r i n s i c force to compel us to see i t as a s p e c i f i c sample, just as the act of pointing has no i n t r i n s i c power to establish whioh objeot i s being singled out. S t i l l , we do successfully single out ob-jeo t s , or oommon q u a l i t i e s , by p o i n t i n g , just as we use s p e c i f i c color-samples, and so t h e i d i s t i n o t i o n between concepts l i k e 'blue* and 'green8, on the one hand, and 'yellow ochre', on the other, remains a v a l i d and—for bur purposes—a c r u c i a l one. We may conclude our examination of #72 and neighboring passages, then, with the assurance that there Is clear evidence that Wittgenstein acknowledges one oase of a quality oommon to many instances, v i z . , a s p e c i f i c shade of c o l o r , and to t h i s extent, at l e a s t , he gives an I m p l i c i t endorsement to A r i s t o t e l i a n realism. There i s , however, another interpretation of Wittgenstein, advanced by W. E. Kennick, which also purports to show that Wittgenstein i s not a r e a l i s t , but for reasons quite d i f f e r e n t •'•This reference to 'pure c o l o r s ' shows that Wittgenstein thinks that even terms l i k e 'blue*, 'green', 'red', etc., can be used to name s p e o i f i o o o l o r - q u a l i t i e s ; as we s h a l l l a t e r see (especially i n Chapter VT), there are many Instances when, presumably to avoid making cumbersome and Irrelevant q u a l i f i -cations, Wittgenstein uses 'red' as just suoh a name. 'Red', i n f a c t , i s his f a v o r i t e example of the name of a oolor. 138 than those proposed by Bambrough, I s h a l l now show that Ken-nick's arguments succeed no better than Bambrough*s. Kenniok*s p o s i t i o n oan be summarized as follows: The passages i n the PI on family-resemblances are not an attack on realism, but on essentialism. These two positions should be kept d i s t i n c t , but they are confused because there i s a r e a l i s t as well as an e s s e n t i a l i s t interpretation of the doc-t r i n e that * i f two or more things are univooally called by the same name, they must have something i n common*. On the essen-t i a l i s t i n t e r p r e t a t i o n , this doctrine i s simply an i m p l i c i t d e f i n i t i o n of *univocity*—not an argument for universals—and i t oan be r e j e c t e d , as Wittgenstein does i n the passages on family-resemblanoes, by arguing that a word oan be univooally applied to things that merely resemble one another i n oertain ways. In order, however, for Wittgenstein's r e j e c t i o n of t h i s doctrine to be considered as a r e j e c t i o n of realism, we must add to t h i s doctrine the r e a l i s t assumption that , , . a general name, univooally applied , . , names, denotes, stands f o r , or designates one thing, the same and Ident i c a l i n eaoh case, which i s other than the thing of which i t i s predicated,1 Now, sinoe one can be an e s s e n t i a l i s t without granting this a d d i t i o n a l assumption of realism—(Kennick thinks that Socrates, i f we accept A r i s t o t l e ' s testimony, was such an e s s e n t i a l i s t ) ~ there i s no reason to assume that i n the passages on family-resemblances Wittgenstein i s doing more than attaoking essen-t i a l i s m . Admittedly, he challenges t h i s a dditional assumption iw. E. Kenniok, "Philosophy as Grammar and the Reality of Universals," i n ludwig Wittgenstein. Philosophy and language, ed. by A l i c e Ambrose and Morris Lazerowitz, p. 17&. 139 of the r e a l i s t as w e l l , hut he does so i n the oourse of his attack on the name-object theory of meaning, which occurs i n a quite separate context from the discussion of family-resemblances; i n short, although Wittgenstein attacks both essentialism and r e a l i s m , his two attacks ought to be kept d i s t i n c t . The ammu-n i t i o n for the former attack i s the notion of family-resemblances, but the ammunition for the l a t t e r attack consists i n various strategies designed to "undermine that picture of language i n which the meaning of a word i s regarded as detachable from the word i n the way that the bearer of a proper name i s detachable from his name. These strategies include, f i r s t , the denial that the bearer of a name Is the same as the meaning of the name (PI, #40); second, the denial that properties are ingredients of things (BB, p. 17); t h i r d , the admonition against looking for a substance answering to every substantive (BB, p. 1 and 5); f o u r t h , and perhaps most importantly, the Injunction to speak of the use of a word, rather than i t s meaning, wherever we can (PI. #43). If we think of meaning as use, . . . we may no longer be inclined to say that every general term (primitive predioate) must denote something. For we are not incl i n e d to think of the use of a word as an entity corre-sponding to the word, any more than we are incl i n e d to think of the use of a hammer as an entity corresponding to the hammer. What, then, do Wittgenstein's reminders add up to? Appar-ently to the claim that being the meaning of a general word or p r i m i t i v e predioate i s not being something corresponding to the word, and hence something the word stands for or denotes i n ad-d i t i o n to the things to whioh i t may be applied or of whioh i t may be predicated.1 In response to Kenniok, I would say, f i r s t , that i t i s 1Kennick, pp. I82-I83. doubtful that a clear-cut d i a t i n o t i o n between essentlalism and realism oan be maintained. Presumably even Kenniok would admit that realism implies essentlalism—-i.e., that the universal named by a general term i s the oommon element, or essence, of those Instances to whioh, and i n v i r t u e of whioh, the general term i s applied?—so the only question i s whether essentlalism implies realism. Kenniok c i t e s Socrates as an example of a non-r e a l i s t e s s e n t i a l i s t , i . e . , as one who holds that every general term must be definable i n terms of a conjunction of properties whioh o o l l e o t i v e l y constitute the general term's essence, but who does not suppose that t h i s essence i s a thing which i s named, denoted, e t c , by the general term. Now, I do not understand on what basis Kenniok oan claim that Socrates denied that there are suoh things as v i r t u e and knowledge, other than that Socrates may not have held that these things ( i . e . , properties) e x i s t apart from t h e i r instances i n a separate realm of forms. No doubt Kenniok i s r i g h t In supposing that essen-t l a l i s m does not imply a Platonic realism, but the issue ought to be whether Soorates, or anyone e l s e , oan be an e s s e n t i a l i s t without being to any extent any kind of a r e a l i s t . Even i f some sophisticated argument could be produoed showing that essentlalism does not imply r e a l i s m , the prima f a c i e connection between the two, at any r a t e , i s s u f f i c i e n t l y intimate that the burden of proof i s on Kenniok to show that Wittgenstein does not regard an attack on one as an attack on the other. As i t i s , I oan see no reason to deny that Wittgenstein, i n the passages on family-resemblanoes, oonoelves himself to be attacking both p o s i t i o n s , 141 to the extent that he i s attacking either one, and Bambrough can-not be f a u l t e d , therefore, for regarding these passages as having a bearing on the problem of u n i v e r s a l s . Bambrough's mistake, as we have seen, i s to assume that the family-resemblance account i s intended for a l l general terms, but at l e a s t those general terms for which the account i s correct are thereby shown not to be names of un i v e r s a l s , and to th i s extent the family-resemblance doctrine does help to undermine an un o r i t i o a l . realism. I f there i s any truth to Kenniok's claim that the target of the dootrine of family-resemblanoes i s essentlalism rather than realism, i t i s that Wittgenstein undoubtedly recognizes that an u n c r i t i c a l e s s e n t i a l i s t i s not necessarily an u n c r i t i c a l r e a l i s t .1 For example, an e s s e n t i a l i s t could, and probably would, give an account of the concept 'game' without supposing that 'game' i s i t s e l f the name of a u n i v e r s a l . Even the essential properties of a game would not necessarily be thought of as u n i v e r s a l s , provided that they oouia be defined by means of s t i l l simpler p r o p e r t i e s , and so on. Only at the point where this analysis oould go no further and we arrived at p r i m i t i v e general terms would the e s s e n t i a l i s t have to acknowledge universals corre-sponding to these terms. So, while not agreeing with Kennick that the doctrine of family-resemblances does not by i t s e l f bear on the problem of universals at a l l , I oonoede that the dootrine i s much more concerned to oombat an u n c r i t i c a l essen-t i a l i s m than an u n o r i t i o a l realism. •'•Indeed, t h i s accurately desorlbes Wittgenstein's own e a r l i e r p o s i t i o n i n the T. L 1412 . Leaving aside the question of Kenniok's correctness i n playing down the relevance of the family-resemblance dootrine to the problem of un i v e r s a l s , the more important issue for our purposes i s his claim that Wittgenstein produces other grounds for r e j e c t i n g r e a l i s m . I f Kenniok i s r i g h t , then, no matter what we conclude about the family-resemblance issue, my i n t e r -p retation of Wittgenstein i s wrong. In f a c t , however, the com-p a t i b i l i t y of Wittgenstein's attack on the name-objeot theory of meaning with his endorsement of a moderate realism i s eas i l y shown* Of the four strategies ci*ed by Kenniok as being intended to d i s c r e d i t the view that a word i s related to i t s meaning as a proper name i s related to i t s bearer, the second—viz., that properties are not ingredients capable of existence apart from the things they qualify—and the t h i r d— v i z . , that there i s not a substanoe corresponding to every substantive—are p l a i n l y com-p a t i b l e with a limited A r i s t o t e l i a n realism. The second strategy attacks, only Platonio r e a l i s m , and the thir d strategy i s com-p a t i b l e , not only with a moderate A r i s t o t e l i a n i s m , but even with a moderate Platonism. The two remaining s t r a t e g i e s , whioh con-s t i t u t e an e x p l i c i t attack on the name-object theory of meaning, f i r s t , by denying that the meaning of a name i s the bearer of the name, and secondly, by urging that we think of the meaning of a word as the word's use, are of no p a r t i c u l a r help to Kenniok eit h e r ; i f the i d e n t i f i c a t i o n of meaning with use i s s u f f i c i e n t to j u s t i f y our doing away with universale, sinoe they do not constitute the meaning of general terms, then, by pa r i t y of reasoning, we also ought to be able to do away with the bearers 143 of proper names, since they do not constitute the meaning of proper names. It i s presumably to mitigate the force of t h i s p a r a l l e l that Kenniok says, a l l u d i n g to PI #43, that • '. the meaning of a word can sometimes be explained by pointing to i t s bearer • • • but not always • .. • , "1 the implication being that the bearers of proper names can be relevant to the meanings of these names, and that the meanings of general terms, on the contrary, Involve no such referenoe to bearers or objeots. But t h i s i s simply to beg the question. Even i f the injunction to think of a word's meaning as i t s use, rather than as an objeot corresponding to the word, is.intended to eliminate the postu-l a t i o n of a l l sorts of superfluous e n t i t i e s In some realm of meaning, nothing i n t h i s Injunction runs counter to the p o s i t i o n that the meanings of some general terms are explained by reference to u n i v e r s a l s , and, as we have seen, the evidence c l e a r l y i n d i -cates that Wittgenstein regards at l e a s t the names of s p e c i f i c shades of color to be explained i n just ,this way. There i s no b a s i s , then, for regarding the meaning-is-use dootrine as an attack against realism i n general. There i s one other l i n e of argument against my interpretation of Wittgenstein to whioh Kenniok b r i e f l y alludes i n a footnote,2 only to dismiss i t as being uncharacteristic of Wittgenstein's usual strategy, but whioh I believe deserves more serious con-s i d e r a t i o n than the arguments preferred by Kenniok himself. This i s the argument that Wittgenstein's r e j e c t i o n of realism i s shown k e n n i o k , p. 182. 2Kenniok, p. 179. by h i s insistence on the difference i n funotion between general words and proper names. Kenniok ci t e s only one page reference i n the BRB as evidence for t h i s kind of approach, but there are other passages which could have been cited as w e l l , not only i n the BRB, but also i n the P I . That words have various functions i s indeed an important and recurring theme i n Wittgenstein's l a t e r philosophy, and i f there i s any reason at a l l to-believe that the sounding of this theme i s intended to d i s p e l the notion that amy general words are names of un i v e r s a l s , then my i n t e r -pretation faoes a serious, even insurmountable, d i f f i c u l t y . I t i s , therefore, c r u c i a l for me to show that there i s no basis for supposing that Wittgenstein regards th i s theme as incompatible with r e a l i s m , and I think that t h i s can be shown, even though a s u p e r f i c i a l reading of cert a i n passages might suggest otherwise, l e t us f i r s t consider the passage to whioh Kenniok r e f e r s : Our use of expressions l i k e "names of numbers", "names of colours", "names of materials", "names of nations" may spring from two d i f f e r e n t sources. One i s that we might imagine the funotions of proper names, numerals, words for colours, etc., to be much more a l i k e than they actually are. If we do so we are tempted to think that the funotion of every word i s more or le s s l i k e the funotion of a proper name of a person, or such generic names as "table", " c h a i r " , "door", e t c . The second source i s t h i s , that i f we see how fundamentally d i f f e r e n t the functions of such words as "table", "chair", etc., are from those of proper names, and how d i f f e r e n t from either the func-tions o f , say, the names of colours, we see no reason why we shouldn't speak of names of numbers or names of directions e i t h e r , not by way of saying some suoh thing as "numbers and directions are just d i f f e r e n t forms of objects", but rather by way of stressing the analogy whioh l i e s i n the lack of analogy between the functions of the words "ohalr" and "Jack" on the one hand, and "east" and "Jack" on the other hand (BRB. p. 82). An i n i t i a l l y p l a u s i b l e reading of t h i s passage i s that i f we are to speak of c o l o r s , e t c . , as having names, such 'names', at any r a t e , do not have objeots corresponding to them. That i s , the reason i t i s unenlightening to say that numbers and direotions are d i f f e r e n t forms of objeots than ohairs and tables i s because the former are not objeots at a l l . I t seems to me, on the other hand, that Wittgenstein i s not concerned i n t h i s passage to deny that numbers and directions are objeots; he may indeed want to deny t h i s , but his point here i s only that i t i s , i n any case, an unhelpful explanation of the difference i n function of various names to say merely that the names r e f e r to d i f f e r e n t kinds of objects, as i f such an explan-a t i o n oould be s u f f i c i e n t by i t s e l f . In other words, he i s trying to get away from the s i m p l i s t i c doctrine, c h a r a c t e r i s t i c of his own e a r l i e r philosophy, whioh says that to know the function of a name i s simply to be acquainted with the objeot named. Even i f numbers and direotions were objeots, the difference between the function of names of numbers or of directions from the function of names of pieces of furniture would not be explained by saying that numbers, d i r e c t i o n s , e t c . , are just d i f f e r e n t forms of objeots, for we would also have to know how these ob-jects were involved i n the use of these names, and the nature of th i s involvement can only be discovered by looking beyond the objeots to the l i n g u i s t i c p r a c t i c e I t s e l f . Let us apply both interpretations—the a n t i - r e a l i s t one as well as my own—to another passage from the BRB just a few pages p r i o r to the one we have been examining, where Wittgenstein i s discussing the differenoe between the demonstrative teaohing of names of kinds of building blocks, of names of numerals, and of proper names. He says: This difference does not l i e . . . i n the aot of pointing and pronouncing the word or, i n any mental act (meaning?) ac-companying i t , but i n the r o l e which the demonstration (pointing and pronouncing) plays i n the whole tr a i n i n g and i n the use which i s made of i t i n the pr a c t i c e of communication by means of th i s language (BRB, p. 80). He then goes on to say that One might think that the difference could be described by saying that i n the d i f f e r e n t cases we point to d i f f e r e n t kinds of objects. But suppose I point with my hand to a blue jersey. How does pointing to i t s colour d i f f e r from pointing to i t s shape? (BRB, p. 80) C l e a r l y , he fe e l s that the answer to t h i s l a s t question w i l l go some way towards showing why the. description i n terms of ' d i f -ferent kinds of objects' i s unsatisfactory; but what, we may ask, could be the purpose of t h i s question i f he were trying to d i s -c r e d i t realism? Is he r h e t o r i c a l l y suggesting that since shapes and colors are not objeots, we oannot r e a l l y point to one or the other, and that the description of the demonstrative teaching of these other names i n terms of pointing to d i f f e r e n t kinds of objects f a i l s for the same reason, v i z . , because there are no objects to be pointed to i n these other cases as well? Surely, his argument i s not that we oannot point to colors or shapes of things—for we c e r t a i n l y can—but rather that there i s no char-a c t e r i s t i c mental act of meaning the color or the shape whioh must accompany thi s act of pointing to one or the other. The difference between pointing to the oolor and the shape, . . . one might say, does not l i e i n the aot of demonstra-t i o n , but rather i n the surrounding of that aot i n the use of the language (BRB. p. 80). Obviously, then, colors and shapes are not chosen as examples because they have an especially dubious ontologioal status, but rather because the l m p l a u s i b i l i t y of regarding an ostensive d e f i n i t i o n as the pr i m i t i v e l i n g u i s t i c datum from whioh a l l subsequent explanations concerning the use of a word must flow, i s i n the i r case espe c i a l l y evident. That i s , we might suppose that pointing to a oolor or a shape would have to be a mental a c t , since a physical act of pointing with one's finger would not by i t s e l f serve to draw attention to the shape of a thing as opposed to i t s c o l o r , or vioe versa, but since there i s no c h a r a c t e r i s t i c mental aot of pointing e i t h e r , we are led to acknowledge the importance of the l i n g u i s t i c praotioes which surround the aot of demonstratively teaching the names of colors and shapes. Thus, i n the case of explaining the difference i n function between names of colors and names of shapes, the i n -adequacy of the bare act of pointing to a thing, apart from a wider context of l i n g u i s t i c p r a c t i c e , i s apparent, and i t i s th i s inadequacy that Wittgenstein means to expose i n his c r i t -ioism of the proposed description of the differences i n other sorts of demonstrative teaching. There i s one other passage i n the BRB that might be i n t e r -preted as going along with an a n t i - r e a l i s t interpretation of the passage on p. 80 of the BRB. This i s the passage on p. 176, i n whioh Wittgenstein says that . . . we can draw a d i s t i n c t i o n between obseEving, attending t o , the shape of the sample and attending to i t s colour. But, attending to the colour can't be described as looking at a thing whioh i s oonneoted with the sample, r a t h e r , as looking at the sample i n a peculiar way. Wittgenstein might be thought to be saying here that the color of an object i s not I t s e l f an object connected with the object which i t q u a l i f i e s , and t h i s — i f the phrase ' i s connected with' 148 i s taken widely enough to include the 'connection' of objeots with their inherent q u a l i t i e s—co u l d be taken as an attack on both the Platonic and A r i s t o t e l i a n versions of r e a l i s m . But i t i s p l a i n from the oontext that the 'thing' whioh i s oalled into question i s not the color of the sample, but rather the sense-impression whioh i s supposed to be d i r e c t l y seen i n the course of attending to the oolor. Thus, Wittgenstein continues: When we obey the order, "Observe the colour...", what we do i s to open our eyes to oolour. "Observe the colour..." doesn't mean "See the colour you see". The order, "look at so and so", Is of the kind, "Turn your head i n this d i r e c t i o n " ; what you w i l l see when you do so does not enter t h i s order. By attending, looking, you produce the impression; you oan't look at the Impression (BRB, p. 176), His point i s that attending to an object's color and then to i t s shape supposedly r e s u l t i n d i f f e r e n t v i s u a l experiences, but because these experiences are themselves brought about by the process of attending, they cannot r e a l l y be 'things' whioh we point to with our a t t e n t i o n . As he says on p. 174, i f I suppose myself to be attending to an inner ' f e e l i n g ' , . . . I don't point to the f e e l i n g by attending to i t . Rather, attending to the f e e l i n g means producing or modifying i t . (On the other hand, observing a chair does not mean pro-ducing or modifying the chair.) That i t i s the status of one's supposedly private sense-impression of oolor, and not the status of the oolor i t s e l f , that i s the source of Wittgenstein's concern i s seen by his treatment of the same subject i n PI #275-276, where colors are regarded as ob-j e c t i v e , p ublic q u a l i t i e s , and only the status of color-impressions i s oalled into question. Wittgenstein's mature p o s i t i o n on the r e l a t i o n between colors and color-impressions i t s e l f r a i s e s many questions which we need not go into here; i t i s s u f f i c i e n t for 149 our purposes to show that he does want to preserve the p u b l i c i t y of c o l o r s , and thus to c l a s s i f y them together with c h a i r s , rather than with f e e l i n g s , with respect to th e i r being possible objects of a t t e n t i o n . I conclude that the passage on p. 176 of the BRB i s not intended to impugn the ontologioal status of colors them-selves, and so cannot be taken as supporting an a n t i - r e a l i s t i n t e r p r e t a t i o n of the passage on p. 82 of the BRB. My strategy so f a r has been to contend that those passages i n the BRB which seem to be explaining the difference i n function of various words by denying that cert a i n words do name objeots are not r e a l l y making any such d e n i a l . S t i l l , the text of the PI—whioh i s , a f t e r a l l , our main concern—leaves no doubt that Wittgenstein does regard the name-object acoount to be gratuitous and misleading as an explanation of the funotion of many words. To c i t e the most famous (or notorious) example, words expressing sensations are not names of inner, p r i v a t e objects; r a t h e r , i n t h e i r second- and third-person uses, suoh words allude to a com-bination of behavior'and circumstances, while i n their f i r s t -person use, suoh words do.not name or describe anything at a l l , and are i n t h i s respect l i k e 'ouch* or 'hurrah*. Again, words l i k e 'expecting*, 'wishing*, 'hoping', 'believing', e t c . , are not names of inner objeots, i . e . , s p e c i f i c mental s t a t e s , but are family-resemblance words which characterize a complex blend of a wide variety of behavior and dispositions to behave, as manifested i n many d i f f e r e n t kinds of circumstanoes. Indeed, from the very f i r s t paragraph of the PI, Wittgenstein sets out to combat what he takes to be the Augustinian view of language, In whioh a l l words are treated as names of objeots, and I have yet to determine the extent of this attack. Is he denying that any general words are names of objects, or i s his attack con-s i s t e n t with the admission that some general words name univer-sals? With a view towards answering this question, I s h a l l now examine the main d r i f t of the argument i n the f i r s t few pages of the PI. So that we may from the beginning avoid confusion, i t should be clear that the question i s not whether Wittgenstein admits that any words other than proper names are names of objeots, sinoe he obviously does admit t h i s . Thus, i n PI, #1 and #2, gen-e r a l terms l i k e " c h a i r1, 'slab', e t c . , are given as the prime examples of words for whioh the name-object account i s appro-p r i a t e , but the objects i n this case are not universals—'chair* i s not, as i t were, a quasi-proper name designating the universal ohairness, but i s a generic name designating nothing more extra-ordinary than p a r t i c u l a r c h a i r s . That i s , the difference between generic names such as these and proper names l i e s , not i n the kind of objeot they name, but rather i n the d i f f e r e n t ways they name what are i n each instance p a r t i c u l a r objeots. The question I am concerned to answer i s whether Wittgenstein allows f o r the existenoe of general words which name non-particular objects, i . e . , u n i v e r s a l s . Evidence that he does, make such an allowance can be found at the very beginning of the PI, i n #1. Here, he t r i e s to com-bat the tendency to characterize a l l words as having the same kind of function by giving us an example of a language-game where 151 the d i f f e r e n t funotions of various words i s obvious. Thus: I send someone shopping. I give him a s l i p marked " f i v e red apples". He takes the s l i p to the shopkeeper, who opens the drawer marked "apples"; then he looks up the word "red" i n a table and finds a colour sample opposite i t ; then he says the series of cardinal numbers--I assume that he knows them by heart—up to the word " f i v e " and for eaoh number he takes an apple of the same colour as the sample out of the drawer . — I t i s i n this and sim-i l a r ways that one operates with words. —"Bu t how does he know where and how he i s to look up the word *red* and what he i s to do with the word ' g i v e ' ? "—We l l , I assume that he acts as I have described. Explanations oome to an end somewhere.—But what i s the meaning of the word "five"?—No such thing was i n question here, only how the word " f i v e " i s used. Now, i t might be thought that Wittgenstein's explanation of the funotions of these three words, by stressing what we do with the words and how we use them as we use instruments or tools for various jobs, i s i n the oase of each word meant to undercut the idea that the word names an object. But t h i s cannot be the gen-e r a l moral that he intends to draw, since 'apple* (or 'apples') i s the kind of word for whioh the name-object account i s correct, i . e . , i t i s a generic name of p a r t i c u l a r objects, on a par with 'chair' or 'slab'. Wittgenstein i s thus making the more subtle point that even where a word names an object, the meaning of the word i s not i d e n t i f i e d with the named object, but i s instead the use to whioh the name i s put. Even i f ' f i v e ' , 'red*, and 'apples* were a l l names of objects, this would not a f f e c t the main p o i n t , which i s that the funotions of the three words d i f f e r markedly. That he does not, however, regard a l l three words as names of objects i s indicated by the question: "But what i s the meaning of the word * f i v e * ? " and the answer: "No such thing was i n ques-t i o n here, only how the word *five* i s used." Wittgenstein i s not here saying that he has not given the meaning of the word 152 • f i v e1 by describing ,its use, but only that he has not s a t i s f i e d the requirement of a p r i m i t i v e theory of meaning, i . e . , his ex-planation of the meaning has not alluded to any object designated by the word ' f i v e * . So, while a l l u s i o n to an object may be r e l e -vant to the explanation of the meaning of a word, i t need not be; such an a l l u s i o n i s evidently relevant i n explaining the meaning of 'apples', but ir r e l e v a n t i n explaining the meaning of ' f i v e * . We come now to the c r u c i a l question: Is 'red', i n thi s respeot, l i k e 'apples' or l i k e 'five'? Wittgenstein provides the answer by r e s t r i c t i n g the r h e t o r i c a l charge—viz., that his description of the use of these words has f a i l e d to give the meaning—to the oase of the word ' f i v e ' . That i s , the hypotheti-c a l Interlocutor who complains that Wittgenstein's account has not given the meaning of ' f i v e ' makes no comparable complaint about 'apples* or *red*, the implication being that at le a s t the explanation of the use of these two words Involves an a l l u s i o n to the appropriate objects, v i z . , the apples i n the drawer and the color of the color-sample, whioh are named by 'apples* and 'red', r e s p e c t i v e l y . Judging by the question that the i n t e r -locutor i s allowed to pose concerning the use of the word 'red' — v i z . , where and how are we to look up the word?—Wittgenstein apparently f e e l s that his explanation of the use of 'red' pre-sents a problem to the t r a d i t i o n a l i s t , not i n f a l l i n g to allude to an object, but only i n supposing that the connection between name and object i s made i n this case by the u t i l i z a t i o n of a color-chart, where the shopkeeper's a b i l i t y to use t h i s chart i s l e f t unexplained. Had Wittgenstein f e l t that an additional 153? problem existed as to whether the c o l o r - q u a l i t i e s shown by the color-chart were legitimate objeots, he l i k e l y would have found a way to r a i s e r h e t o r i c a l doubts on this score as w e l l , so that he could have the opportunity of sweeping aside suoh troubles by giving the same response f o r 'red' as he does for ' f i v e ' , v i z . , that we need not suppose that the word names an objeot i n order for i t to have meaning. As he does not a v a i l himself of this opportunity, we have good reason to suppose that he has no qualms about regarding c o l o r - q u a l i t i e s as objeots. The next few pages of the PI are mainly concerned with s p e l l i n g out more e x p l i c i t l y the nature of Wittgenstein's de-parture from a s i m p l i s t i c name-object theory of meaning. The language-game described i n #2, consisting i n the four words 'block', ' p i l l a r ' , 'slab', and 'beam', whioh are names of four d i f f e r e n t kinds of building-stones, i s offered as an example of a language " . . . f o r which the description given by Augustine i s r i g h t " (PI, #2).1 Shortly thereafter, Wittgenstein describes an expanded version of language-game (2) i n order to show that the Augustinian aooount, whioh i s appropriate enough for the simpler language-game, ,no longer applies here, where words with r a d i c a l l y d i f f e r e n t functions than 'slab* and 'briok* have been introduced. • Thus: Iwittgensteln means that Augustine's description i s r i g h t only to the extent that the Individual words of thi s language are a l l names of objeots. Augustine's other assumption, v i z . , that learning a language requires the making of inferences (presumably i n one's own pr i v a t e language) concerning the inner, mental states of language-users on the basis of the i r outward behavior, i s repudiated by Wittgenstein without q u a l i f i c a t i o n . 15,4 Besides the four words "block", " p i l l a r " , e t c . l e t i t con-t a i n a series of words used as the shopkeeper i n (1) used the numerals ( i t oan be the series of l e t t e r s of the alphabet); f u r -ther, l e t there be two words, whioh may as well be "there" and " t h i s " (because t h i s roughly indicates t h e i r purpose), that are used i n oonnexion with a pointing gesture; and f i n a l l y a number of oolour samples, A gives an order l i k e : "d—slab—there". At the same time he shews the assistant a colour sample, and when he says "there" he points to a place on the building s i t e . From the stock of slabs B takes one for each l e t t e r of the alpha-bet up to "d", of the same colour as the sample, and brings them to the place indicated by A.—On other occasions A gives the order " t h i s—th e r e " . At " t h i s " he points to a building stone. And so on (£1, #8). Then, to the Augustiniah's -question: "What do the words of t h i s language s i g n i f y ? " Wittgenstein r e p l i e s : "What i s supposed to show what they s i g n i f y i f not the kind of use they have? And we have already described that" (PI. #10). In saying that what the word s i g n i f i e s i s shown by i t s use, Wittgenstein i s stressing two points: f i r s t , even where a word does straightforwardly name an object, i t could s t i l l be misleading to say that the word s i g -n i f i e s the object, i f t h i s were taken to mean that the funotion of the name i s thereby given. As he says: Of oourse, one can reduoe the description of the use of the word "slab" to the statement that t h i s word s i g n i f i e s t h i s ob-j e c t . This w i l l be done when, for example, i t i s merely a matter of removing the mistaken idea that the word "slab" r e f e r s to the type of building-stone that we i n f a c t c a l l a "block"—but the kind of ' r e f e r r i n g ' t h i s i s , that i s to say the use of these words for the r e s t , i s already known (PI, #10). Secondly, by seeing that the meaning of a word amounts to no more and no less than i t s use, we f e e l no need to deny dogmatically that signs l i k e *a', 'b', e t c , s i g n i f y numbers. As Wittgenstein shows, there are contexts i n whioh saying some such thing would be p e r f e c t l y proper, as . . . when, for example t h i s removes the mistaken idea that "a", "b", "o", play the part actually played i n language by 155 "block", "slab", " p i l l a r " . And one can also say that "o" means this number and not that one; when for example this serves to explain that the l e t t e r s are to be used i n the order a, b, c, d, e t c , and not i n the order a, b, d, c (PI, #10). We can comfortably admit that 'a', 'b*, etc., can be said to s i g n i f y numbers, and know that we are not thereby committing ourselves to an ontology of numbers i n the same way that we acknowledge the existenoe of slabs and b r i c k s , because our con-cession f i n a l l y amounts to no more than that these various signs are used thus and so, and while these uses involve objects such as slabs and b r i o k s , they do not involve objeots such as numbers. So, while Wittgenstein does not dogmatically deny the correctness of the assertion that number-words s i g n i f y numbers, he does sug-gest that a s s i m i l a t i n g the description of the use of number-words to that of words l i k e 'brick* oan be seriously misleading, because the words themselves are used i n t o t a l l y d i f f e r e n t ways. Wittgen-stein's answer, then, to the question: "Do the signs *a', 'b', e t c . , s i g n i f y numbers?" oan best be given by a remark made i n another oontext: "Say what you choose, so long as i t does not prevent you from seeing the f a c t s . (And when you see them there i s a good deal that you w i l l not say.)" (PI, #79) What can we conclude from a l l this about the ontologioal status of colors? As f a r as Wittgenstein's discussion of language-game (8) i s concerned, we oan conclude nothing at a l l d i r e o t l y , for t h i s language-game does riot even include color-words. He i s anxious to make the p o i n t , for reasons which w i l l become clear in the next chapter, that color-samples themselves can be regarded as instruments of language, i . e . , as the means of representation rather than that which i s represented (PI. #50), and so he uses the occasion of his description of language-game (8) to make polor-samples do the job of color-words. S t i l l , there i s i n this discussion at l e a s t some evidence supporting my e a r l i e r contention that a word whose function d i f f e r s from that of a word l i k e 'brick', which can be straightforwardly regarded as the name of an object, i s not on the basis of this difference necessarily assumed by Wittgenstein not to be the name of an object. As the following example suggests, there i s also the p o s s i b i l i t y that t h i s word names an object as w e l l , and that the word d i f f e r s from 'briok' only i n the kind of name i t i s , i . e . , the way i t r e l a t e s to i t s object. Thus, a f t e r suggesting that a number-word i s not r e a l l y the name of an object—or, at any r a t e , that to speak of numbers as objects i s not to commit ourselves to the existenoe of curious entities—Wittgenstein i l l u s t r a t e s his point by comparing d i f f e r e n t kinds of words to d i f f e r e n t kinds of t o o l s , and says: "The funotions of words are as diverse as the funotions of these objeots. (And i n both cases there are s i m i l a r i t i e s . ) " (PI, #11) The remark about ' s i m i l a r -i t i e s ' i s meant to indicate that the differences between the funotions of various words i s not always so extreme as that between aiaword l i k e 'brick' and a number-word, whioh suggests. that the scope of the ap p l i c a t i o n of the name-object account of language i s not to be set t l e d by studying two kinds of examples. Even i f a number-word i s not the name of an object, other kinds of words' may well be such names, even though they, too, d i f f e r from words l i k e 'brick'. That there are d i f f e r e n t kinds1 of s i g n i f i c a t i v e funotions among words, i n addition to words whose function i s not to s i g -n i f y any object at a l l , i s again implied i n PI #14: Imagine someone's saying: " A l l tools serve to modify some-thing. Thus the hammer modifies the p o s i t i o n of the n a i l , the saw the shape of the board, and so on."—And what i s modified by the r u l e , the glue-pot, the nails?—"Our knowledge of a thing's length, the temperature of the glue, and the s o l i d i t y of the box. "—Would anything be gained by t h i s a s s i m i l a t i o n of expressions?— The implication i s not only that i t i s misleading to say of c e r t a i n tools that they modify something, but also that at l e a s t two d i f f e r e n t kinds of tools can appropriately be said to modify something. Since the analogy i s between a t o o l modifying something and a word s i g n i f y i n g something, we can conclude that the s i g n i f i c a t i v e funotion, l i k e w i s e , i s not the exclusive pre-rogative of one kind of word. Applying t h i s analogy of tools to the words of the language-game described i n #1, we oan say that •apples' and 'red' are l i k e the hammer and the saw, while 'fi v e * i s l i k e , e.g., the glue-pot. We oan see from Wittgenstein's discussions of these various language-games that there i s no basis for concluding that no gen-e r a l words serve to name non-particular objects. Considering the emphasis placed on the m u l t i p l i c i t y of functions of words, we are on precarious ground i n a t t r i b u t i n g to Wittgenstein any p o s i t i o n characterizing general words as a c l a s s . I t i s only as a protest against t h i s tendency to over-simplify, and not as the advocacy of some s i m p l i s t i c view of his own, that we should take Wittgen-1But see PI #18, which warns us against a s i m p l i s t i c view of what •kindsT~are. stein's comment i n _____ #26: One thinks that learning language consists i n giving names to objects. V i z , to human beings, to shapes, to colours, to pains, to moods, to numbers, etc. To repeat—naming i s some-thing l i k e attaching a la b e l to a thing. One can say that t h i s i s preparatory to the use of a word. But what i s i t a prepa-r a t i o n for? And the answer to this question w i l l not be determined by some neat, pre-conoeived idea such as that only p u b l i c l y observable p a r t i c u l a r s q u a l i f y as objeots. Bather, each case w i l l be judged on i t s own merits, and i t i f turns out that there are no objects such as pains (as Wittgenstein does subsequently contend), then t h i s conclusion i s determined, not i n advance, but by the same procedure of ca r e f u l l y weighing l i n g u i s t i c f a c t s1 which shows, e.g., that human beings are indeed properly considered to be objeots named by the generic term 'human being' (Mensoh), a procedure whioh could also be used to show that a good case can be made for saying that colors are non-particular objects. And, as I have contended on other grounds, Wittgenstein d e f i n i t e l y goes some way towards making this case. To sum up the r e s u l t s of our examination, we have found nothing i n Wittgenstein's l a t e r writings incompatible with a limited A r i s t o t e l i a n realism, and some passages which can r e a -sonably be taken to imply that t h i s i s his p o s i t i o n . S t i l l , i t must be admitted that at no place i n the l a t e r writings does he ^•It could, of course, be argued that on this question of the status of sensations, Wittgenstein's examination of l i n -g u i s t i c practices i s not as careful as i t should be, but my present point i s only that he certainl y intends that his judg-ment should be determined on this basis, and not on some pre-conception of the way language works, or ought to work. 1591 go out of his way to claim this p o s i t i o n as his own. There are, I think, c h i e f l y two reasons for t h i s . The remainder of this chapter w i l l concern i t s e l f with the f i r s t reason, and the second reason w i l l be disoussed i n the following chapter. The f i r s t reason for the lack of an open proclamation of realism i s that Wittgenstein oonceives himself to have devel-oped a method of treating philosophical problems which t r a n -scends the t r a d i t i o n a l approach of affirming or denying spe-c i f i c p o s i t i o n s . The long-standing disputes between metaphy-sic i a n s of d i f f e r e n t persuasions, whioh are viewed as disagree-ments about the nature of r e a l i t y and are thought to be decid-able by q u a s i - s c i e n t i f i c methods, are viewed by Wittgenstein as no substantive disagreements at a l l , but rather as confusions a r i s i n g from mistaken views concerning the functions of language. He does not consider himself to be taking sides with any of the t r a d i t i o n a l p a r t i e s to these disputes, and hence would hot want to opt for any of the positions i n the realist-oonoeptualist-nominalist controversy. Indeed, at least one commentator, A l i c e Ambrose, has been so impressed by Wittgenstein's claim to have avoided taking sides i n t r a d i t i o n a l disputes that she considers th i s feature of his philosophical method--viz., that whioh pur-ports to show that ostensibly substantive disagreements are r e a l l y grammatical confusions—as providing the key to Wittgen-stein's 'solution' to the 'problem' of u n i v e r s a l s . Thus, she notes that philosophers other than Wittgenstein have also s t r i v e n to 'command a clear view of the use of our words', but that, whereas they wanted to use this knowledge to gain an insight 16c? into the nature of r e a l i t y , The philosophical investigations of the l a t e r Wittgenstein were aimed neither at obtaining information about the world nor at r e f u t i n g the positions of philosophers who claimed to have succeeded i n t h i s . l I t i s a mistake, she says, to suppose that Wittgenstein holds that philosophers have been misled by confusions about language into pronouncing f a l s e views, or that he i s denying what these philosophers assert to be t r u e ,2 He • • • i s not concerned to question a matter of fact and i s not to be construed as contesting the claimed truth value of any philosophical p o s i t i o n , although he oan be interpreted as con-cerned with removing the temptation to adopt one and defend i t . - * Now, I do not doubt that Ambrose has accurately characterized what Wittgenstein conceives himself to be doing, and i f this i s i n fact what he is . doing, then my claim that he i s a limited A r i s t o t e l i a n r e a l i s t — o r , for that matter, that he adheres to any ontologioal p o s i t i o n—i s i n c o r r e c t . But I think that t h i s claim of Wittgenstein's (and of Ambrose on his behalf) to have succeeded i n maintaining ontologioal n e u t r a l i t y , as i t were, i s simply naive. This i s not the place to attempt an extensive examination of Wittgenstein's philosophical method and of the discrepancy between the way he characterizes i t and the way i t actually works in p r a c t i c e , and I must therefore be content to ^Alioe Ambrose, "Wittgenstein on Universals," i n Ludwip;  Wittgenstein; The Man and His Philosophy, ed. by K. T. Fann, P. 336. 2Ambrose, p. 339» This i s also Kenniok's professed view of what Wittgenstein i s doing, but i n practice Kenniok treats Witt-genstein's arguments as having the effect of denying the truth of realism. ^Ambrose, p. 352. 161 make a few general observations, Wittgenstein's conviction that philosophical positions can-not take a stand on any fa c t u a l issue i s , f i r s t of a l l , just a matter of the way he chooses to define 'philosophy'. Insofar as fa c t u a l issues bear on any p o s i t i o n , he regards t h i s p o s i t i o n as a ' s c i e n t i f i c ' one, to be v e r i f i e d or f a l s i f i e d by empirical 1 methods. Philosophy, on the other hand, i s an a c t i v i t y of con-ceptual c l a r i f i c a t i o n brought about by a consideration, from a non-empirioal perspective, of the workings of language, i . e . , language i s investigated, not qua anthropological phenomenon i n whioh the causes of i t being the way i t i s are to be ascertained, but rather qua rule-governed a c t i v i t y , where we look at the way words are used, i n the sense of looking at the rules (to the extent and i n the manner which rules operate at a l l ) aocording to which the various language-games are played. On this view, there can be no d i s t i n c t i v e l y philosophical propositions, and therefore no formulation of d i s t i n c t i v e l y philosophical p o s i t i o n s . (In this respect, Wittgenstein never departs from the views of the T.) A l l that philosophy oan properly do i s to draw attention to the actual uses of words and thereby remove the i l l u s i o n s of t r a d i t i o n a l philosophers, whose misconceptions about the functions of words r e s u l t s i n the mistaken conception of philosophy as an area of substantive inquiry d i s t i n c t from science. Leaving aside the claims that the disputes of t r a d i t i o n a l Isome metaphysicians l i k e Whitehead, on the other hand, would say that philosophy i s simply the most general l e v e l of empirical i n q u i r y . See Process and R e a l i t y , p. 5. 162 philosophers concern no matters of fact and a r i s e solely from a mistaken view of the funotions of language, both of whioh are cer t a i n l y open to question, Wittgenstein's own claim to be i n -vestigating language as i t i s actu a l l y used does not i n any case absolve him from the charge of engaging i n ontologioal i n q u i r y . We can, I suppose, respect Wittgenstein's wishes and say that insofar as his remarks about the funotions of certain words com-mit him, e.g., to the existenoe of p a r t i c u l a r things l i k e b r i c k s , apples, or people, he i s acting i n his capacity as a s c i e n t i s t rather than as a philosopher, but thi s somewhat eccentric char-a c t e r i z a t i o n of the business of science does not change the Important f a c t , v i z . , that he i s oommitted at least to an ontology of these p a r t i c u l a r things.1 In the same way, i f his remarks show him to be oommitted to the existence of c o l o r - q u a l i t i e s , we may then include universals within his ontology as w e l l , and need only to be oareful not to c a l l his ( i m p l i c i t ) affirmation of the existenoe of universals the affirmation of a philosophical p o s i t i o n . Another p e c u l i a r i t y of Wittgenstein's own conception of his philosophical method concerns his assumption that as long as c r i t i c i s m i s directed against a philosophical p o s i t i o n on 'gram-he e d l e s s to say, the way our use of words commits us to an ontology i s not a simple matter to express. While i t seems clear that the users of language (2) are committed to an ontology of brioks , s l a b s , e t c . , i n that the existence of suoh objects i s required i f the language-game i s to be played at a l l , people who play the language-game, i f we can c a l l i t that, of t e l l i n g f a i r y -t a l e s are not, on the other hand, committed to an ontology of unicorns or elves, even though 'unicorn' and ' e l f * are, i n some sense, used as names of objects.. Nor, to take a more mundane ex-ample, are we oommitted to Napoleon's existence when we use his proper name. We oan say that the existenoe of the named object may or may riot be presupposed, depending on the language-game, but this whole question admittedly r a i s e s oomplex issues whioh I do not pretend to be able to deal with here. 163 matioal' rather than factual grounds, the c r i t i c i s m oannot be said to bear on an ontologioal dispute. The clearest example of t h i s assumption at work i s i n Wittgenstein's attack on the notion of a private objeot, which he feels can be carried out without f a l l i n g into the behaviorists' error of denying that private objeots e x i s t . In PI #304-308 Wittgenstein assures us that his c r i t i c i s m of the. view that a mental process i s a private experience does not mean that he wants to deny that we have mental processes. Behaviorists speak of mental processes as f i c t i o n s , but t h i s makes the dispute between behaviorists and introspeo-t i o n l s t s appear to be over a matter of f a c t , l i k e the dispute about the existence of the loch Ness monster, whereas Wittgen-s t e i n does not want to appear to be taking one side of a factual dispute, and therefore he speaks only of the i n t r o s p e c t i o n i s t s ' account as involving a grammatloal f i c t i o n , v i z . , a ' f i c t i o n a l ' ( i . e . , wrong) account of the function of language as i t i s used to t a l k about inner experience.1 As he says i n the "Notes on Privacy," The 'private experience' i s a degenerate construction of our grammar (comparable i n a sense to tautology and contra-d i c t i o n ) . And t h i s grammatical monster now fools us; when we wish to do away with i t , i t seems as though we had denied the existence of an experience, say, toothache.2 Ambrose c i t e s this strategy, i n which Wittgenstein i s claim-ing to correct only misconceptions about the grammar of expressions •'•Wittgenstein's notion of a 'grammatical f i c t i o n ' almost assuredly derives from pp. 98-99 of Ggden and Richards' The  Meaning of Meaning. For an acoount of the influence of t h i s work on the l a t e r Wittgenstein, see Appendix B. Wittgenstein, "Notes on Privacy," i n Philosophical Review 77 (1968), p. 314. 164 concerning inner experience, as being a prime example of how he does not attempt to refute philosophers' claims about the nature of the world. I t seems patently obvious to me, however, that a philosopher who believes that the word 'pain* i s a generic name for a certai n kind of private objeot, and who includes such ob-jects i n his ontology, has every r i g h t to f e e l that Wittgenstein i s attempting to refute him i n arguing that *pain* i s not the name of an objeot at a l l , and that we are only fooled by the s u p e r f i c i a l grammatical p a r a l l e l s between expressions l i k e *I have a pineapple* and *I have a pain* into thinking that i n both oases I am talking about an object that I have. Granted, Wittgenstein i s not denying that there are objeots answering to the name *pain*, i n the sense that there are no objeots answering to the name 'unicorn*, because, as he contends, 'pain' i s not used as the name of an object at a l l .1 Nevertheless, this i s xThe extent to which commentators have f a i l e d to grasp the point of this r e l a t i v e l y straightforward 'grammatical' c r i t i c i s m i s evidenced by the frequency with which a passage l i k e PI #304 has been misunderstood. When Wittgenstein says that a sensation ". . . i s not a something, but not a nothing either J" he i s not contending with Specht that a sensation i s indeed a something, only not a private something (E. K. Specht, The Foundations of  Wittgenstein's Late Philosophy, p. 93), nor with Pitoher that a sensation i s a pr i v a t e something, only that i t plays no part i n language-games and i s therefore 'as nothing' i n such Janguage-games (Pitcher, p. 300)» nor with Donagan that a sensation i s a priv a t e something that does play a part i n pub l i c language by being referred t o , though i t i s s t i l l i r r e l e v a n t to the use of names t e l l i n g us what i s referred to, since these names require p u b l i c c r i t e r i a of application (Alan Donagan, "Wittgenstein on Sensation," i n WPI, pp. 345-346). Against Pitcher and Donagan, Wittgenstein cannot be oonceding that a sensation i s a private something, because he has argued that the very notion of a p r i -vate object i s u n i n t e l l i g i b l e . It i s equally gratuitous, i n . . l i g h t of Wittgenstein's insistence that a sensation i s not a something, to suppose as Specht does that pain i s a public some-thing. Obviously, the point of #304 i s to correct that mis-conception of the grammar of sensation-words whioh interprets 165 small consolation to the philosopher who wants to include p r i -vate objeots i n his ontology and who, i f he accepts Wittgenstein's arguments, w i l l have to consider his former p o s i t i o n to have been refute d , just as surely as i f i t had been denied on the basis of empirical f a c t . So, whatever Wittgenstein thinks he i s doing, his actual praotioe i s not one of preserving n e u t r a l i t y on ontologioal issues; and when I say that he i s a r e a l i s t , I am r e f e r r i n g only to the ontologioal p o s i t i o n implied by what he says i n the course of doing philosophy, and not to the p o s i t i o n—or lack of one— that he conceives himself as having when he self-consciously t r i e s to characterize what he i s doing. I r o n i c a l l y , Ambrose—who thinks she i s defending Wittgenstein's claim of n e u t r a l i t y— i n effect concedes my interpretation of Wittgenstein as a moderate r e a l i s t when she says: Now i t i s quite true that many substantives do have a naming function, e.g. proper names, l i k e "John", names of s p e c i f i c c o l -o r s, l i k e "oerise", general names, l i k e "chair" (though amongst these are important d i f f e r e n c e s ) . But not a l l words stand for things that can be pointed to..... • . suoh words as names of objects,^ the only two alternatives being that the -objects designated by such names do e x i s t , just as there r e a l l y are objects named by 'horse', or that they do not e x i s t , just as no objects answer to the name 'unicorn'. Wittgenstein i s saying that neither of these alternatives hold i n connection with a word l i k e 'pain' any more than they hold for a word l i k e 'sake'. I f someone contends that sakes e x i s t , the proper answer i s not to deny that sakes e x i s t , but rather to show him that he has misunderstood the funotion of the word 'sake*, whose use i s not to designate an object at a l l . This i s precisely the approach taken by Wittgenstein in.hi s discussion of sensations. For a f u r -ther discussion of the status of private objeots, see Appendix A. ^Ambrose, pp. 346-347. She might also have added that not everything that, i n some sense, can be pointed to i s r e a l l y an objeot; e.g., we can ostensively explain what the number two i s by pointing to a group of two nuts. 16.6; Aside from Ambrose's general conviction that nothing Witt-genstein says commits him to an ontologioal p o s i t i o n , there i s an additional reason for her f a i l u r e to see how his admission that there are s p e c i f i c c o l o r - q u a l i t i e s bears on the question about u n i v e r s a l s , and this i s that she poses the question ex-c l u s i v e l y i n terms of a P l a t o n i s t account of u n i v e r s a l s . Thus, the metaphysical p o s i t i o n that there are universals i s i d e n t i f i e d with the p o s i t i o n that there are • • refined object[s] being apprehended by an inner v i s i o n • . .1 , 1 and existing i n their own • . shadow world . . . . "2 We can only suppose that she i s so conoerned to exhibit the ' l i n g u i s t i c obsessions', as she c a l l s them, manifest i n Platonism, that the p o s s i b i l i t y of examining Wittgenstein's remarks from the perspective of another version of realism simply does not occur to her. This leads us to the second reason for Wittgenstein's own lack of an overt proclamation of his realism: l i k e Ambrose, Wittgenstein himself i s preoccupied with Platonism, and that he does not t o t a l l y r e j e c t realism as such i n the course of his repudiation of Platonism i s a matter which he considers to be of only i n c i d e n t a l importance, i f he considers i t at a l l . To show that his in t e r e s t i n the problem of universals does indeed center almost exclusively around the issue of Platonism, and that his opting for A r i s t o t e l i a n realism occurs merely as a casual by-product of his repudiation of Platonism, w i l l be the task of the next chapter, where I study Wittgenstein's views i n the context of t h e i r h i s t o r i c a l development. ^-Ambrose, p. 338. ^Ambrose, p. 3^9. 16? CHAPTER VI FROM PLATO TO ARISTOTLE: THE TRANSITION TO THE LATER PHILOSOPHY What Is the oonneotion between language ana r e a l i t y ? A fundamental concern of Wittgenstein i n the T i s to provide an answer to thi s question. The connection must be suoh as to allow the expression of propositions whose sense i s independent of their truth or f a l s i t y . This separation of sense from truth-value i s aohieved by the picture theory, i n which a proposition has sense by being a picture of a (possible) f a c t , the arrange-ment of elements i n the proposition corresponding to the (pos-s i b l y existing) arrangement of objects i n r e a l i t y . Whether the arrangement of objects pictured by the proposition actually oc-curs i n r e a l i t y i s an empirical issue; i n either case, the prop-o s i t i o n suooeeds i n being a p i c t u r e . As opposed to f a c t s , whose existence i s contingent, objects must necessarily e x i s t . If the existence of objeots were a con-tingent matter, then.we would again have f a i l e d to separate questions of sense from questions of t r u t h , sinoe, e.g., the sense of a proposition containing the name 'A' would depend upon the truth of another prop o s i t i o n , v i z . , that proposition which s a i d , i n e f f e c t , that A e x i s t s . I t i s for this reason that we cannot sensibly assert of objects that they exist or do not ex i s t ; t h e i r existence i s already guaranteed by the very meaning-fulness of thei r names. At the l e v e l of propositions, the oonneotion between language 168 and r e a l i t y consists i n propositions having the same form, or l o g i o a l m u l t i p l i c i t y , as (possible) f a c t s . Since, however, there need be no actual fact corresponding to a proposition, the connection between language and r e a l i t y i s ultimately made at the l e v e l of names and their corresponding objects, for i t i s the connection at thi s juncture that i s presupposed even for propositions to make sense.1 The manner i n which this dootrine of the necessary existence, or • i n d e s t r u c t i b i l i t y ' , of elements i s , i n the T, bound up with Platonism i s simply that some of these elements are univ e r s a l s . An A r i s t o t e l i a n version of realism i s precluded by the co n f i g -uration of universals and p a r t i c u l a r s being a contingent matter, so that the existence of universals can never depend upon any such configuration. For example, i t must be possible for me sensibly to assert of some p a r t i c u l a r that i t i s red ( i . e . , q u a l i f i e d by the universal redness), even i f i n fa c t i t i s not, or even i f there are no red things i n the world at a l l . Redness, then, must e x i s t , whether i t happens to be configured with any 2 p a r t i c u l a r s or not. •"•The primacy of the name-object oonneotion i s i m p l i c i t l y acknowledged i n T 3*13* "A proposition, therefore, does not actually oontain~its sense, but contains the p o s s i b i l i t y of ex-pressing i t . . . . A proposition contains the form, but not the content, of i t s sense." That i s , the elements of the proposi-t i o n a l sign have the same p o s s i b i l i t i e s of arrangement as objects i n r e a l i t y , and so to this extent the sense of a proposition need draw only on the resources of language i t s e l f . But i n order to t e l l what the arranged elements i n a propositional sign mean, i . e . , i n order to give the sense of a proposition a oontent as well as a form, we must look beyond language to the objects of r e a l i t y . Names must be correlated with objects. 2T h i s assumes that the.T would consider colors to be u n i -v e r s a l s , which of course i t does not, though Wittgenstein c l e a r l y 169 I t i s these key features of the picture theory—viz., that propositions have the same p o s s i b i l i t i e s of arranging th e i r e l e -ments as there are arrangements of elements i n possible f a o t s , thereby giving language and r e a l i t y a oommon l o g i o a l form, and also that the elements corresponding, to names must exist—whose repudiation provides a dominant theme i n Wittgenstein's l a t e r work, especially i n the t r a n s i t i o n a l period, beginning with his paper of 1929, "Some Remarks on l o g i o a l Form," and ending with PG, where the essentials of his mature p o s i t i o n are established. I s h a l l now trace the development of his views on this Issue of the connection between language and r e a l i t y , with the purpose of showing that, to the extent that Wittgenstein i s concerned i n his l a t e r philosophy with the question of universals at a l l , i t i s almost exclusively within the context of trying to elucidate the nature of the connection between language and r e a l i t y while avoiding the Platonism of the T. The process which eventually leads to the t o t a l collapse of the picture theory begins innocently enough, when Wittgenstein observes, i n his paper of 1929, that the forms of elementary propositions cannot be foreseen. I t might be thought that he already says the same thing i n the T, 5.55 to 5.5571, but his point there i s that, although the general form of an elementary proposition i s known a p r i o r i , the existenoe of propositions of any s p e c i f i c form—e.g., whether there are elementary propositions would l i k e to think of colors as simple objeots and i s only pre-vented from doing so by his requirement that elementary propo-s i t i o n s be independent. 170 of a twenty-place r e l a t i o n a l form, or even of a two-plaoe r e l a -t i o n a l form or a subject-predicate form—remains to be somehow discovered by a n a l y s i s ,1 S t i l l , i f Wittgenstein r e f r a i n s i n theory from making any dogmatic committment, there i s no doubt that he assumes i n practice that there probably are, e.g., e l e -mentary propositions of the subject-predicate form and of the two-place r e l a t i o n a l form (as i s shown by T 4.24), and i t i s thi s comfortable assumption that i s attacked i n "Some Remarks on Logioal Form." The most we oan say about the subject-predicate and r e l a t i o n a l forms i s that These forms are the norms of our p a r t i c u l a r language into whioh we project i n ever so many d i f f e r e n t l o g i o a l forms.2 To show how d i f f e r e n t the actual l o g i o a l forme of propo-s i t i o n s are from these norms of ordinary language, Wittgenstein c i t e s as his prime example the f a c t that ". . . for their repre-sentation numbers (rational and i r r a t i o n a l ) must enter into the structure of the atomic propositions themselves."^ Thus, a description of our v i s u a l f i e l d i n whioh the shape and p o s i t i o n In December of 1931, Wittgenstein comments on his e a r l i e r view as follows: "I wrote Q.n the T| s 'One cannot give a des-c r i p t i o n of elementary propositions1", and that was en t i r e l y c o r r e c t . I t was clear to me that here there are, at any r a t e , no hypotheses, and that one cannot approach this question l i k e Carnap, assuming i n advance that elementary propositions should consist of two-place r e l a t i o n s , e t c . But I s t i l l thought that at a l a t e r date the elementary propositions would be able to be described. Only i n reoent years have I freed myself from th i s e r r o r . . . . One oannot dlsoover anything i n philosophy. I s t i l l had not under-stood that c l e a r l y enough myself, however, and violated the very p r i n c i p l e " (WJTK, pp. I82-I83). See also PG, pp. 210-212. 2"Some Remarks on Logioal Form," i n ET, p. 33. 3"Remarks," p. 33. 171 of every patoh of oolor i s given with respect to a co-ordinate system • • by statements of numbers whioh have the i r s i g n i f -icance r e l a t i v e to the system of co-ordinates • • .1 , 1 i s the only kind of description whioh w i l l have the r i g h t l o g i o a l mul-t i p l i c i t y . A l s o , numbers w i l l have to enter those propositional forms • • • when—as we should say i n ordinary language—we are dealing with properties which admit of gradation, i . e . proper-t i e s as the length of an i n t e r v a l , the p i t c h of a tone, the brightness or redness of a shade of colour, e t c .2 Although? Wittgenstein does not i n this paper go on to s p e l l out the consequences of these observations for the doctrines of the T, he makes c l e a r , i n remarks recorded a few months l a t e r by Waismann, that the very p o s s i b i l i t y of forms of description other than that of subject-predicate, etc., renders gratuitous the aooount of o b j e c t s—imp l l o i t i n the T—which characterizes them with respect to t h e i r d i f f e r e n t grammatical funotions. As he says: "Now the question has no sense: Are objeots t h i n g - l i k e , standing as i t were i n the subject-place, or p r o p e r t y - l i k e , or are they r e l a t i o n s and so on?" And he oonoludes: "We speak of objects simply where we have equal (glelohbereohtigte) elements of representation" (WWK, p. 43). The example used by Wittgenstein to i l l u s t r a t e what sort of thing he means by an 'element of representation' i s that i f we were to desoribe the surface of a room a n a l y t i c a l l y by an equation and describe the d i s t r i b u t i o n of colors on these sur-faces, and i f we were to give an analysis of these colors them-1"Remarks," p. 33. 2"Remarks," p. 34. 172 selves by saying how they were to be produced by the four primary c o l o r s ,1 then the primary colors themselves, v i z . , red, yellow, blue, and green, would be the elements of representation. Pre-sumably, then, the primary colors would be the only simple ob-jeots involved i n descriptions of v i s u a l spaoe, and the phrase 'element of representation' i s merely intended to remind us not to try to c l a s s i f y these objects according to a grammatical funotion. In PB Wittgenstein speaks of 'elements of knowledge (Erkenntnls)' rather than 'elements of representation', but the two expressions are probably intended to mark out the same prov-i n c e , since he gives as examples of elements of knowledge H. • • the four primary c o l o r s , spaoe, time, and such else that i s given" (PB, p. 169)• These elements of knowledge are character-ized as the "things themselves" (PB, p. 169), the "simple objeots" (PB. p. 169), and are contrasted with the "things of physics" (PB. p. 168), and also with "i n e s s e n t i a l (uneigentliohen) objeots" such as p a r t i c u l a r sense data (PB, p. 136). I t might be thought that what we have here i s s t i l l Platonism, though admittedly of an austere v a r i e t y , having l i t t l e else i n i t s ontology than primary c o l o r s . But to claim even th i s much would be a mistake, for PB c l e a r l y indioates that colors ought not to be supposed to have an independent ontologioal status outside of t h e i r s p e c i f i c s p a t i a l and temporal manifestations. Thus: In what way i s endless time a p o s s i b i l i t y and not a r e a l i t y ? For one could object against me that time must be a r e a l i t y just •••For reasons whioh esoape me, Wittgenstein here and e l s e -where (e.g., Z #331) includes green among the primary c o l o r s . Since nothing""important hinges on t h i s , I w i l l , for the purposes of this discussion, adopt his c l a s s i f i c a t i o n . 173 as much as oolor. But i s not oolor by i t s e l f also just a p o s s i -b i l i t y so long as i t does not exist at a s p e c i f i c time and i n a s p e c i f i c place? (PB, p. 163) In this off-hand remark, made en t i r e l y without fan-fare, we find the clear implication that Wittgenstein has abandoned the P l a t o -nism of the T. The way i n which the doctrine of * elements of representation', or 'elements of knowledge', contributes to the erosion of Platonism i s that, i n a r r i v i n g at this doctrine, Witt-genstein establishes the t a c t i c of diminishing the r o l e of r e a l -i t y and increasing that of language i n explaining their connec-t i o n . That i s , whereas i n the T various kinds of objects, v i z . , p a r t i c u l a r s , p r o p e r t i e s , and r e l a t i o n s , were thought to be needed to provide the r e q u i s i t e oontent for the sense of propositions, these objeots are now rendered superfluous by co-ordinate sys-tems, e t c . , which are themselves part of language. And i t i s th i s same strategy of transferring what seems to belong to the province of r e a l i t y into the province of language that provides Wittgenstein with a way of escaping from the P l a t o n i s t assumptions of the T altogether. The PB i s a t r a n s i t i o n a l work which shows Wittgenstein's d i s s a t i s f a c t i o n with the T's version of the pioture theory and his e f f o r t to provide a new account, but the strategy to be employed i s s t i l l only hinted at here, and only i n the PG does i t emerge with f u l l f o r c e . I s h a l l now trace the oourse of this development from PB to PG. In the PB Wittgenstein wants, among other things, to defend the pioture theory of language against a oausal theory of the 17k connection between language and r e a l i t y .1 According to the p i c -ture theory, the truth of a thought (or a proposition) involves only the thought and the fact that makes the thought true, while the causal theory involves an additional faotor, v i z . , the f e e l i n g of s a t i s f a c t i o n whioh r e s u l t s when the fact f u l f i l l s the thought. On the causal account, the way I intend a proposition to be taken depends upon my subsequent reaction; e.g., i f I say 'It i s r a i n -i n g ' and—after looking out the window and seeing the drops of water f a l l i n g from the sky—do not experience the r e q u i s i t e f e e l i n g of s a t i s f a c t i o n , or whatever, then I must conolude that I could not have meant what i s normally meant by the words 'It i s r a i n i n g ' . But, as Wittgenstein oontends, the sense of a proposition oan never be decided by experience, and I must be able to know at the time I assert a proposition how I intend i t to be taken. As he says: "The intention expresses even now how I now compare the p i c t u r e with r e a l i t y " (PB, p. 65). likewise, to expect, wish f o r , or search for something presupposes that I know what i t i s that I am expecting, wishing f o r , or searching for (or at l e a s t t h i s i s so for one commonly used sense of 'ex-pecting*, e t c . ) , which cantabe summed up by saying that . . . the f a c t whioh s a t i s f i e s the expectation of p i s represented by the proposition p . . • Therefore, not by the description of an e n t i r e l y d i f f e r e n t event (PB. p. 66). Prom here, Wittgenstein i s led into considering more pre-c i s e l y what i t i s that thoughts, expectations, wishes, searches, ^ R u s s e l l , i n The Analysis of Mind, and Ogden and Richards, i n The Meaning of Meaning are the.causal theorists Wittgenstein has p a r t i c u l a r l y i n mind. See PB, p. 63. 175 e t c . , presuppose, and this brings us back to the issue of Platonism: Searching presupposes that I know what I am searching f o r , without what I am searohing for r e a l l y having to e x i s t . I ex-pressed t h i s e a r l i e r by saying that searching presupposes the . elements of the complex but not the combination for which I search. And t h i s i s not a bad comparison. l i n g u i s t i c a l l y , t h i s i s expressed by saying that the sense of a proposition presupposes only the grammatically correct application of cer-t a i n words (PB, p. 67). This sounds l i k e an endorsement of the T, but a few pages l a t e r , while r e - i t e r a t i n g the point that the expectation, thought, wish, etc., that p must have the same m u l t i p l i c i t y as that ex-pressed i n p i t s e l f , Wittgenstein makes an i n t e r e s t i n g q u a l i f i -c a t i o n . After c a l l i n g such prooesses of expectation, e t c . , a r t i c u l a t e , i n v i r t u e of t h e i r possession of the same l o g i c a l m u l t i p l i c i t y as the propositions with which they are involved, he says that these a r t i c u l a t e prooesses might also be oalled the in t e r p r e t a t i o n of signs (meaning by t h i s that the same in t e r p r e -t a t i o n of signs i s involved i n the expectation that p as i s i n -volved i n p I t s e l f ) , and then adds: Perhaps one must say that the expression 'interpretation of signs' i s misleading, and instead of this one should say 'the use of signs'. For ' i n t e r p r e t a t i o n ' sounds as i f one now correlates with the word 'red* the oolor red (even when i t i s not present) and so f o r t h . And again the question a r i s e s : What i s the connection between sign and world? Could I search for something i f the space were not there i n whioh I oould search for i t ? Where does the sign connect to the world? (PB, p. 70) The problem ar i s e s here beoause Wittgenstein has rejected the s o l u t i o n of the T whioh t e l l s us that the oolor red i s always 'present' i n the world as a Platonic u n i v e r s a l . I f , as he now holds, the oolor red i s an a c t u a l i t y only i n i t s s p e c i f i c s p a t i a l and temporal manifestations, what does the word 'red' mean when 176 the oolor Is not being manifested? There i s a further problem involved i n the i d e n t i f i c a t i o n of the meaning o f , e.g., the word 'red' with a Platonic univer-s a l (or with any one object, for that matter). In "Some Remarks on Logioal Form" Wittgenstein had shown that, contrary to the views of the T, i t i s possible for elementary propositions to be mutually exclusive, a view whioh he now expresses i n the PB by sayings "I do not apply the proposition as a measuring r u l e to r e a l i t y , but rather the system of propositions" (PB, p. 110). Thus, i n applying the proposition "This i s red' to r e a l i t y , I am simultaneously applying the propositions 'This i s not green', 'This i s not blue', e t c , since, i f the f i r s t proposition i s true, a l l these other propositions w i l l be true as w e l l . This admission i s by i t s e l f a s i g n i f i c a n t a l t e r a t i o n of the picture theory, as i t removes the one-to-one correspondence between propositions and possible situations ( i . e . , i f the possible s i t u a t i o n i s determined to be an actual f a c t , t h i s makes true a whole system of propositions, and not just one pr o p o s i t i o n ) . But the T's account of the oonneotion between language and r e a l i t y i s thereby a f f e c t e d , not just at the l e v e l of propo-s i t i o n and f a c t , but also at the l e v e l of name and objeot, be-cause, e.g., the measuring r u l e for colors must be involved i n the very meanings of the color-words themselves. I oannot under-stand the meaning of 'red' just by seeing something red; i n ad d i t i o n , I must understand the operation of the whole measuring r u l e of whioh red i s but one graduation mark, which amounts to knowing what oolors there are besides red and how they are a l l 177 grammatically related to one another.1 Obviously, the T*s account of what i t i s for a name to have meaning i s no longer tenable. Nor does the PB provide an ent i r e l y s a t i s f a c t o r y a l t e r n a t i v e . I t does, however, move i n the r i g h t d i r e c t i o n by holding that the measuring r u l e for the representation of colors i s i t s e l f an instrument (Mittel) of language (PB. p. 79)» and thus stressing that d i f f i c u l t i e s i n understanding how propositions have sense, i . e . , how language ultimately connectsVup with r e a l i t y , are to be resolved by seeking an explanation i n the operations of language i t s e l f rather than i n the e x t r a - l i n g u i s t i c r e a l i t y to which language i s a p p lied. S t i l l , merely to diminish the r o l e of objects i n determining the meanings of names does not solve the problem of the ontologioal status of suoh objeots. Only i n the PG does the strategy for eliminating Platonism receive i t s d e f i n i t i v e formu-l a t i o n , and so i t i s to t h i s work that we now turn. As we would expect from the t i t l e , the underlying theme of PG i s to show how ostensibly 'philosophical' problems are r e a l l y nothing but confusions about the grammar of language. In p r i n -c i p l e , of oourse, Wittgenstein had even i n the T rejected the view that philosophy had.its own d i s t i n c t i v e subject-matter, since the contingent status of a l l propositions, whioh placed ^A rough idea of what i s meant by the grammatical r e l a t i o n between colors oan be given i n statements l i k e : 'Purple i s more closely related to blue than to green', or 'Orange i s a mixture of red and yellow, but yellow i s not a mixture of red and green', which are not meant as empirical statements about oolor pigments, or about l i g h t waves, or about processes i n our r e t i n a s , but are instead conceptual truths concerning the colors themselves i n their purely phenomenologioal aspect. For further d e t a i l s , see PB, p. 273ff. 178 them a l l In the province of 'science' (this term being used i n a very extended sense), l e f t no room for philosophy other than as an a o t l v i t y of making clear the aotual l o g i c a l structure of these propositions. In p r a c t i c e , however, his e f f o r t to explain the connection between language and r e a l i t y , which led him to p o s i t simple objects, I.e., i n d e s t r u c t i b l e substance, as an a p r i o r i requirement of the p i c t u r e theory, placed him firmly i n the ranks of t r a d i t i o n a l metaphysicians. Even when, i n the PB, t h i s doctrine of i n d e s t r u c t i b l e substance became a source of embarrassment, Wittgenstein had no ready solution for avoiding entanglement i n t h i s apparently substantive metaphysical i s s u e . But by the time of the writing of the PG, the way out has become cl e a r : the ostensibly metaphysical issue involved i n explaining the connection between language and r e a l i t y i s not r e a l l y a sub-stantive issue, for "Like everything metaphysical, the harmony between thought and r e a l i t y i s found i n the grammar of language" (PG, p. 1 6 2 ) . 1 Simply put, Wittgenstein's contention i s that insofar as an objeot i s relevant to determining the meaning of the name corre-sponding to i t , as i n the case of ostensive d e f i n i t i o n , the object i s i n t h i s capacity part of language and not part of the r e a l i t y whioh language represents. Thus: The interpretation of written and spoken signs through ostensive explanations i s not the a p p l i c a t i o n of language, but part of the grammar (Sprachlehre) (PG. p. 88). 1A11 quotes are taken from the published version of PG, whioh was not written u n t i l 1933-34, but the e a r l i e r version, written during the period 1930-32, a r t i c u l a t e s e s s e n t i a l l y the same p o s i t i o n . See especially #13, 31, 54, 173, and 441. 179 The connection between language and r e a l i t y i s made through the explanation of words—whioh belongs to the grammar, so that language remains i n i t s e l f closed, autonomous (PG, p. 97)• The c o r r e l a t i o n of name and object i s produced i n no other way than through a ta b l e , an ostenslve gesture and a pronouncing of the word, e t c . It i s part of the symbolism (PG, p. 97)• In short, there i s no longer a problem of explaining the conneotion between language and r e a l i t y i n a way that avoids Platonism, beoause there i s no such conneotion at a l l . This absence of any connection i s not i t s e l f known through a meta-physical discovery whose a r t i c u l a t i o n competes with the olaims of Platonism, but i s rather the r e s u l t of a l i n g u i s t i c decision to include i n the realm of 'grammar' ". . • a l l conditions (methods) of comparing a proposition with r e a l i t y . . . " (PG, p. 88), which by d e f i n i t i o n places objeots themselves, insofar as they are relevant to the ostensive explanation of words, within the realm of language. In the PB, Wittgenstein comes very close to t h i s dootrine of l i n g u i s t i c autonomy1 when, i n trying to make the point that colors we imagine have the same 'grammatical behavior' as colors we see, he notes that our ordinary language does not have the means of describing preoise shades of oolor, so that . . . when I want to communicate to someone which color some thing should have, I send him a sample, and obviously this sample belongs to language; and the memory-image or other image of the c o l o r , whioh I produoe by a word, belongs to language i n just the same way . . . (PB, p. 73), • • • S t i l l , as close as he comes, he does not, i n the PB, take the f i n a l step. Thus, a f t e r oonoeding the d i f f i c u l t i e s involved i n connecting language to r e a l i t y , he nevertheless continues to hold that "What i s c h a r a c t e r i s t i c of propositions of the kind 'This i s . . . ' i s only, that i n the symbol there somehow enters the r e a l i t y outside of the so-oalled sign-system" (PB, p. 120). 180 the point being that the grammatical p o s s i b i l i t i e s of oolor-samples we actually see are no d i f f e r e n t than those of colors we imagine (e.g., we oan no more imagine reddish-green than we oan see a sample of i t ) . The view that colors are at le a s t sometimes instruments of, rather than the object of, communi-c a t i o n—v i z . , i n the i r status as samples—is but a short way from the view that colors always function i n this paradigmatic capacity when they are being ostensively defined, and although Wittgenstein extends th i s grammatical status to a l l objeots that oan enter into ostensive d e f i n i t i o n s , e.g., to people, a c t i v i t i e s , e t c . ,1 colors are probably the most natural and obvious candidates for thi s dual- or shifting-status treatment, and there i s l i t t l e question that i n making th i s d i s t i n c t i o n between the funotion of objeots as part of grammar and as part of r e a l i t y , Wittgenstein has i n mind c h i e f l y the example of c o l o r s . This pre-oooupation with c o l o r s , whioh Issues i n the development of a strategy p a r t i c u l a r l y appropriate for them2 i s also explainable by the fact' that oolors are a prime example of that seemingly r e c a l c i t r a n t , e x t r a - l i n g u i s t i c residue whioh had been c o l l e c t i v e l y referred to i n PB as the 'elements of know-ledge'. I f oolors can be disposed o f , Wittgenstein no doubt 1Thus, p. 143 of PG says: • • the wish 'he should come' i s the wish that r e a l l y he should r e a l l y come. And i f one wanted a further explanation of thi s a f f i r m a t i o n , we would say 'and by "he" I understand th i s here, and by "come" I mean the ac t i o n . . . ' . But those are explanations of grammar, whioh we create through language. Everything i s carried out within the language." 2I t i s much more natural to think of a patch of color i n a oolor-chart as an instrument of language than, say, the E i f f e l Tower, though on Wittgenstein's aooount, both are part of the grammar i n the context of being ostensively defined. 181 f e e l s that the r e s t w i l l take oare of i t s e l f . Accordingly, re-uses the example of colors i n formulating the general problem* . . . the thought that p i s the oase does not presuppose that i t i s the oase. . . . on the other hand something must be a presupposition for the thought i t s e l f (I oannot think that something i s r e d , i f the oolor red doesn't exist) (PG, p. 142). And then i t gives the solutions The proposition *I could not think that something i s red i f red didn't e x i s t ' r e a l l y means the image of something red, or the existenoe of a red sample, as part of our language. But, of oourse, one cannot say that our language must contain suoh a sample. I f the language does not contain the sample, i t i s just another language. But one can say and emphasize that the languag does contain; the, sample (PG, p. 143). It i s t h i s n o t i o n—vi z . , that, i n the name-object r e l a t i o n , the object i s relevant only as a component of a grammatical r u l e— which removes the l a s t excuse for equating the meaning of a word with the objeot for which i t stands, and clears the way for the idea that the meaning of a word i n every case Just is. i t s use i n the language, no more and no l e s s . For this reason, the dootrine of the autonomy of language has a good claim to be regarded as the decisive step i n Wittgenstein's t r a n s i t i o n to his l a t e r philosophy. The implications of t h i s doctrine of l i n g u i s t i c autonomy w i l l become clearer i n the next chapter, but for now I wish to note only that the absence of any e x p l i c i t mention of the 'auton-omy' of language i n the PI marks, not any change i n Wittgenstein' basic p o s i t i o n , but only a r e a l i z a t i o n that t h i s p o s i t i o n oan be expressed i n a way that i s not needlessly provocative. Thus, once the point has been made that the meaning of a word i s i t s use, and not the objeot, i f any, that corresponds to the word, then whether or not we oontinue to i n s i s t that the objeot, qua 182 component of an ostensive d e f i n i t i o n , i s part of language i s a matter of only secondary importance, since, i n any case, there i s no longer thought to be one fundamental semantic l i n k which we can c a l l the name-object r e l a t i o n ; there are many suoh r e l a t i o n s , constituted by various l i n g u i s t i c praotioes, and this i s enough ^ o repudiate the myopic outlook of the T without i n -s i s t i n g that, e.g., i n the proposition 'That i s called the E i f f e l Tower' (said while one points to the E i f f e l Tower), the E i f f e l Tower i s part of the grammar. I am not saying that i n the PI Wittgenstein would have denied that i t i s part of the grammar, but only that he there uses more muted tones to make i n effeot the same po i n t . Thus, i t i s enough to say that i n making the connection between the E i f f e l Tower and i t s name, we have not yet made a move i n any language-game, but have only made the preparation for a move. And since the kind of prepa-r a t i o n we make varies from one language-game to another, and does not consist i n establishing some simple, invariable name-objeot r e l a t i o n , MWe may says nothing has so far been done, when a thing has been named. I t has not even got a name except i n the language-gameM (PI, # 4 9 ) . For our purposes, anyway, the important point i s that i n the PI .Wittgenstein does e x p l i c i t l y r e t a i n expressions oharaoter-i s t i o of the doctrine of U n g u i s t i o autonomy i n discussing the status of those ostensibly simple objects with reference to which the T's account of the connection between language and r e a l i t y i s conoeived i n the f i r s t plaoe.1 So, e.g., i n describing ^His discussion of the status of simple objects i s compli-183 language-games i n whioh oolor-samples are employee), he c a l l s such samples 'instruments of language' (PI, #46) and a means of representation rather than something that i s represented (PI, #50). And even i n language-games whioh involve no d i r e c t appeal to oolor-samples as paradigms, so that the color s i g n i f i e d by a word has simply to be borne i n mind, the status of these imagined oolors i s e s s e n t i a l l y no d i f f e r e n t than that of external para-digms. One might suppose otherwise, i n that, while external paradigms can be destroyed, we always r e t a i n the a b i l i t y to bring fort h Images of these c o l o r s , whioh shows that the oolors themselves, though not t h e i r external paradigms, are indestruct-i b l e . But, against the view that i t must always be possible to remember, e.g., what oolor i s called 'red', Wittgenstein argues that the c r i t e r i a for judging memories to be correct cannot be provided exclusively by these, acts of memory themselves. Some-times, i t i s true, we judge on the basis of memory that a sample has changed c o l o r , but sometimes we appeal to the sample i n judging that our memory-image i t s e l f has changed. As he had said e a r l i e r i n the PG: When the oolor-sample appears darker to me than I remember i t from yesterday as being, I.need not, and sometimes do not, suppose my memory i s co r r e c t . And I could very well speak of a darkening of my memory*5 (PG, p. 95) • oated by his claim that there are no absolutely simple o b j e c t s -s i m p l i c i t y being r e l a t i v e to the language-game—but for now we can ignore this q u a l i f i c a t i o n , since his chief argument against regarding c o l o r s , eto., as examples of in d e s t r u c t i b l e substance i s independent of the claim that no objeots are absolutely simple. As we s h a l l see i n Chapter VTI, the q u a l i f i c a t i o n that no objects are absolutely simple i s only part of a much more important q u a l i f i c a t i o n about ontologioal o l a s s i f i o a t i o n as suoh. 184 But i f memories cannot, by some i n t r i n s i c feature, insure t h e i r own correctness, and are themselves sometimes judged by external paradigms, then, were a l l suoh paradigms, and red things gener-a l l y , destroyed, the s i t u a t i o n oould oome about that we no longer remembered whioh oolor our image was an image of. And When we forget whioh oolor i t i s that has thi s name, i t loses i t s meaning for us; that i s , we are no longer able to play a p a r t i c u l a r language-game with i t . And the s i t u a t i o n then i s comparable with that i n which we have l o s t a paradigm whioh was an instrument of our language (PI_, #57 )• So, even though the meaning of the word 'red1 i s independent of the existenoe of red things qua objects i n r e a l i t y , i f we assume the destruction of red things qua paradigms of language—whioh, Wittgenstein seems to imply, would r e s u l t i n the absenoe of an external check for the correctness of our memory-images, and would thus remove" them from relevance to our language-games just as e f f e c t i v e l y as i f they had been destroyed i n the manner of external paradigms—then there i s a sense i n whioh we could very well say that the color 'red1 had been destroyed. That i s , with the destruction of the relevant paradigms, we oould no longer play our acoustomed language-game with the word 'red*. What a l l t h i s shows i s that the so-oalled ' i n d e s t r u c t i b i l i t y * of red i s not a matter of the existenoe of a Platonic u n i v e r s a l , but merely of the existenoe of a paradigm as part of the language. To assign the r e s p o n s i b i l i t y for providing the meanings of words to l i n g u i s t i c paradigms i s not, however, to dispose of the onto-l o g i o a l issue e n t i r e l y , because we must now consider the ontolog-i o a l status of the paradigms themselves. Granted, things i n the i r funotion as paradigms do not have an ontologioal status. But a paradigm, a f t e r a l l , just i s a thing l i k e a piece of colored paper, and i s exempt from ontologioal status only i n respect of i t s s p e c i a l r o l e i n the language-game. Indeed, Witt-genstein admits that the same oolored thing could function simul-taneously as a paradigm and as an object i n r e a l i t y , depending on the perspective from which i t i s regarded. Thus, i n d i s -cussing how the words 'That i s blue* can function either as a statement about the objeot pointed t o , or as an explanation of the word 'blue*, he adds* "It i s also possible for someone to get an explanation of the words out of what was intended as a piece of information" (PI, p. 18). In this case, what funotions for the speaker as a r e a l objeot, functions for the hearer as a paradigm. So the question s t i l l remains as to the ontologioal status of paradigms apart from t h e i r function as paradigms. I f oolored things can serve as paradigms for c o l o r , we must s t i l l consider the ontologioal status of oolored things themselves. Now, although Wittgenstein's main ontologioal concern—as I have here t r i e d to show—is to repudiate Platonism, and not to argue for any of the remaining alternatives—whether i t be nominalism, conoeptualism, or A r i s t o t e l i a n realism—his remarks c l e a r l y show that, regarding the ontologioal status of oolored things, he does indeed opt for one of these a l t e r n a t i v e s , v i z . , A r i s t o t e l i a n r e a l i s m . We have already seen, on p. 63 of PB, how Wittgenstein i s tending towards t h i s version of realism (though his view i n PJB i s complicated by the subsequently aban-doned doctrine of 'elements of knowledge', from the standpoint of which oolors—and only primary c o l o r s , at that—are the 186 things themselves and are thus not amenable to an A r i s t o t e l i a n treatment i n whioh they are regarded as properties of p a r t i c u l a r s ) , and the PG makes his p o s i t i o n even c l e a r e r . Thus, i n contrast to nominalism and oonoeptuallsm, both of which deny that a color r e a l l y exists i n objects themselves as their common q u a l i t y , Wittgenstein expresses his p o s i t i o n as follows: This i s a possible d e f i n i t i o n : "to point to a color" means to point to the object which has the color (PG, p. 63)• . . . "The color brown ex i s t s " means nothing at a l l , other than that i t i s present here, or there as the oolor of an object • . • (PG, P. 137). . . . we calculate with the word 'red* by describing the place where the color i s found, whioh shape or si z e the patch (Fleck) or the body has whioh bears the oolor, whether i t i s pure or mixed, darker or l i g h t e r , changes or stays the same, e t c . . . . (PG, p. 67). That th i s remains his view throughout the l a t e r PI as well i s evident, not only from PI #72, whioh, as I have already a r -gued, implies that i t i s possible for two or more objeots to have p r e c i s e l y the same color as thei r oommon q u a l i t y , but also hy #58» i n whioh, a f t e r explaining the sorts of considerations which would lead us to deny that we oould sensibly assert that red exists (or that red does not e x i s t ) , he adds: In r e a l i t y , however, we quite r e a d i l y say that a p a r t i c u l a r colour e x i s t s ; and that i s as much as to say that something ex-i s t s that has that oolour. And the f i r s t expression i s no less acourate than the second; p a r t i c u l a r l y where 'what has the c o l -our* i s not a physical object. This passage shows how w i l l i n g Wittgenstein Is to embrace an A r i s t o t e l i a n account of the status of c o l o r s— i . e . , that they exi s t only as properties actually Instantiated i n things—even when the 'things' i n question are only mental images,1 which, •'•It i s evident from the oontext that these are the kind of 187 i f not for the generally adopted strategy of eliminating P l a t o -nism by i n s i s t i n g that universals exist only as q u a l i f y i n g aotual p a r t i c u l a r s , might not otherwise be so re a d i l y assumed to be property-bearing things at a l l .1 So, what was only an i m p l i c i t suggestion i n PI #1, v i z . , that 'red', l i k e 'apple', i s the name of an objeot, i s f i n a l l y given an e x p l i c i t confirmation i n PI #58: the oolor red i s , indeed, an objeot, whose mode of existence i s that of being the property of existing red things. Prom the foregoing, i t i s evident that the 'grammatical' so l u t i o n of the problem of the oonneotion between language and r e a l i t y does not go as far as Wittgenstein might have hoped towards removing him from the ranks of metaphysicians, sinoe the strategy of invoking paradigms—while allowing him to dispose of Platonism—involves an ontologioal oommlttment of i t s own. S t i l l , although Wittgenstein does not succeed i n t o t a l l y expunging onto-l o g i o a l implications from his l a t e r views, the next chapter w i l l show that, from the perspective of this l a t e r philosophy, the nature of ontologioal committment as suoh i s r a d i c a l l y trans-formed, so that an important q u a l i f i c a t i o n must be made i n c a l l i n g him a moderate A r i s t o t e l i a n r e a l i s t . I t i s to the task of making t h i s q u a l i f i c a t i o n that I now proceed. non-physical objeots Wittgenstein has i n mind. •^ Of course, the question of the status of mental images, after-images, and the l i k e , i s further complicated by consider-ations advanced elsewhere i n the PI against the p o s s i b i l i t y of naming or describing private objeots, processes, or stat e s , either i n a private language or a public language. In Appendix A, I w i l l show how the admission that mental Images, et c . , are indeed private objects could have been made with an untroubled conscience by Wittgenstein himself, had he taken more seriously one of his own l i n e s of argument. • 188 CHAPTER VII THE PRIMACY OF LANGUAGE A B opposed to the view that universals are f i c t i o n s of language (nominalism) or mental creations (conceptualism), realism—whether of the Platonic or A r i s t o t e l i a n variety—holds that suoh items are actually to be found i n that e x t r a - l i n g u i s t i c extra-mental, objective realm which, for short, we o a l l • r e a l i t y ' Now, although Wittgenstein obviously i s a r e a l i s t , i n the sense that he regards colors as r e a l objeots, just as much as the things they q u a l i f y , h i s ontologioal p o s i t i o n must be d i s t i n -guished from t r a d i t i o n a l realism, or from any t r a d i t i o n a l onto-l o g i o a l p o s i t i o n , for that matter, with regard to the nature of r e a l i t y i t s e l f and i t s r e l a t i o n to language. In other words, the q u a l i f i c a t i o n I am about to make i s not especially directed to Wittgenstein qua r e a l i s t , but to Wittgenstein qua ontologist, so that the q u a l i f i c a t i o n whioh applies to his acknowledgment of the existenoe of universals equally applies to his acknowledgment of the existenoe of p a r t i c u l a r things. Wittgenstein's departure from the more t r a d i t i o n a l approaoh to ontology—including that of Frege and Russell as well as that ^ o r a f u l l e r discussion of the d i s t i n c t i o n between these various sorts of p o s i t i o n s , see R. I. Aaron's The Theory of Uni-versals . which c r i t i c i z e s nominalism and conoeptualism from the standpoint of the same sort of moderate A r i s t o t e l i a n realism that I have argued—subject to the q u a l i f i c a t i o n of t h i s chapter—is the p o s i t i o n of the l a t e r Wittgenstein. I r o n i c a l l y , i n his i n t e r p r e t a t i o n of Wittgenstein, Aaron makes Bambrough's mistake of pl a c i n g the whole burden on the family-resemblance strategy, thus turning Wittgenstein into a kind of resemblance-theory nominalist despite the a f f i n i t y of Wittgenstein's p o s i t i o n with Aaron's own. 189 of the T—concerns the issue of the j u s t i f i c a t i o n of language, i . e . , the grounding of language i n r e a l i t y . B r i e f l y , the usual view i s that such a j u s t i f i c a t i o n oan, and indeed must, be had i f there i s to be such a thing as f a c t - s t a t i n g language at a l l , whereas the view of the l a t e r Wittgenstein i s that language i s not, nor oan i t possibly be, grounded i n r e a l i t y . In his view, language i s a f r e e - f l o a t i n g structure, having no ontologioal basis as i t s anchor and foundation. Not only can we say that for Wittgenstein language i s not answerable to r e a l i t y , but we can even a f f i r m the opposite: r e a l i t y , and therefore ontology, i s answerable to language. So we cannot say that the nature of e x t r a - l i n g u i s t i c r e a l i t y i s suoh as to force us to include ex-pressions i n our language which designate this or that u n i v e r s a l , or any universal at a l l , on pain of our language being somehow incomplete, i . e . , incapable, because of the lack of these ex-pressions, of saying a l l that i t ought to be able to say i f i t purports to deal thoroughly with the r e a l i t y with whioh i t i s ostensibly concerned. We oan no more contend that a language lacking expressions for universals i s incomplete than we can say that checkers i s an incomplete game because i t contains no b i s h -ops or knights. Checkers i s simply a d i f f e r e n t game than chess, and i n the same way, a language lacking names for universals i s just d i f f e r e n t from a language containing such names, and i s not an incomplete version of the l a t t e r . This comparison between languages and games i s not my own, but Wittgenstein's; so apt and illuminating does he consider t h i s comparison, i n f a c t , that he incorporates i t into the 190 concept of language i t s e l f . Language, for the l a t e r Wittgen-s t e i n , beoomes the language-game.. Granted, the notion of a language-game cannot be said to have as i t s ra t i o n a l e only one consideration. As Wittgenstein himself indicates i n the PI, the expression 'language-game* serves p a r t l y to remind us that language i s pri m a r i l y an a c t i v i t y , whose function must be under-stood against a s p e c i f i c background of the customs and practices of a c u l t u r e . The expression *language-game* i s also p a r t i c -u l a r l y appropriate for characterizing the more pr i m i t i v e and simple l i n g u i s t i c praotioes of savages and infants which serve as perspicuous objeots of comparison with our own more complex l i n g u i s t i c practices (whioh are,©nevertheless, also called *language-games'). But the one fundamental consideration which gave r i s e to the notion of a language-game i n the f i r s t place i s undoubtedly the one I have suggested," v i z . , that language, l i k e a game, i s not answerable to r e a l i t y . Thus, Wittgenstein holds that a game i s defined by i t s r u l e s , and not by some purpose l y i n g outside of the rules whioh serves to j u s t i f y the rules depending upon how well they f a c i l i t a t e the achievement of thi s purpose; i f the r u l e s are changed, we do not have a better or worse version of the same game (judged by the increased or de-creased effectiveness of the game i n s a t i s f y i n g some purpose outside the game), but merely a d i f f e r e n t game. In the same way, language i s not to be characterized i n terms of the achieve-ment of some purpose outside of i t s e l f which i t i s supposed to f a c i l i t a t e , such as the purpose of describing r e a l i t y ; r a t h e r , language i s defined by Its grammatical r u l e s , whioh are no less 191 a r b i t r a r y than the rules ef a game. Two languages having d i f -ferent grammatical rules are simply d i f f e r e n t language-games, and nothing more can be s a i d—ce r t a i n l y not that one language i s more adequately grounded i n r e a l i t y than the other. In order to show that i t i s indeed t h i s concern—viz., to emphasize the ab-sence of an ontologioal j u s t i f i c a t i o n for grammatical r u l e s— whioh Is the fundamental consideration underlying Wittgenstein's notion of a language-game, I s h a l l once again examine his views i n the context of t h e i r h i s t o r i c a l development, from which we oan r e a d i l y see that the notion of a language-game i s es s e n t i a l l y a reformulation and recasting of the formalist strategy as i t i s described and o r i t i c i z e d by Frege. Part Two of the Grunflgesetze der Arlthmetlk contains a lengthy discussion of the views of two mathematicians, E. Heine and J . Thomae, whom Frege groups together as advocates of what he c a l l s formal arithmetic. In contrast to meaningful (inhaeltlloh ) arithmetic, whose subject-matter i s actual numbers, i . e . , those e n t i t i e s which the signs for numbers designate, formalism1 i s concerned only with the manipulation of number-signs themselves and regards the meaningful arithmetician's concern with the designata of number-signs as i r r e l e v a n t for the purposes of arithmetic and l i a b l e to lead to 'metaphysical d i f f i c u l t i e s ' , e.g., when we wonder what sort of entity corresponds to a sign for an i r r a t i o n a l number. Heine's i s the cruder version of -'-I w i l l be using the expression 'formalism' to r e f e r only to the p o s i t i o n discussed by Frege, and not to the work of sub-sequent mathematicians suoh as H i l b e r t , who are also called ' f o r m a lists'. 192 formalism—he apparently believes that numbers just are the tangible ink marks on pieces of paper, or chalk marks on black-boards, etc.—and Frege has no trouble disposing of t h i s v ersion. The more in t e r e s t i n g version i s that of Thomae, who does not i d e n t i f y numbers with tangible signs, but instead characterizes numbers i n terms of the rules according to whioh number-signs are manipulated. The following quote, cited by Frege, gives the crux of Thomae*s p o s i t i o n : For the f o r m a l i s t , arithmetic i s a game with signs, which are called empty. That means they have no other content (in the c a l c u l a t i n g game) than they are assigned by th e i r behaviour with respect to c e r t a i n rules of combination (rules of the game). The chess player makes si m i l a r use of his piecesj he assigns them certain properties determining th e i r behaviour i n the game, and the pieces are only the external signs of t h i s behaviour. To be sure, there i s an Important difference between arithmetic and chess. The rules of chess are a r b i t r a r y , the system of rules for arithmetic i s such that by means of simple axioms the numbers can be referred to peroeptual manifolds and oan thus make im-portant contribution to our knowledge of nature.2 The fact that the numbers i n the oaloulating game can be 're-ferred to perceptual manifolds' i s supposed to mitigate the oharge of a r b i t r a r i n e s s , but Frege correctly points out that, before we consider any possible applioation of the ca l c u l a t i n g game to our peroeptual manifolds, the calculating r u l e s , as they ooour within the domain of formal arithmetic i t s e l f , are no less a r b i t r a r y than the rules of chess. As Frege says: In formal arithmetic we need no basis for the rules of the game—we simply s t i p u l a t e them. . . . we absolve ourselves from accounting for one choioe of rules rather than another.3 ^Gottlob Frege, Grundgesetze der Arithmetik, i n Philosophical  Writings, ed. by Geaoh and Black, pp. 194-195. 2J . Thomae, Elementare Theorie der analytisohen Funotionen  einer oomplexen Veraenderlichen, quoted i n Frege, pp. 183-184. 3Frege, p. I85. 193 From the standpoint of meaningful arithmetic, on the other hand, ca l c u l a t i n g rules are saved from the accusation of a r b i -trariness i n that the number-signs designate e n t i t i e s whioh • . supply the grounds for the rules . • . m 1 by which these signs are manipulated. On Thomae's view, we oan appeal to no such e n t i t l e s . Frege expresses Thomae's p o s i t i o n thus: We do not derive these rules from the reference of the signs, but lay them down on our own authority, re t a i n i n g f u l l freedom and acknowledging no necessity to j u s t i f y the r u l e s ; though we exercise this freedom with an eye to possible a p p l i -cations, since otherwise arithmetic would be a game and nothing more.2 Even this concession—viz., that the rules of formal arithmetic, though applicable to r e a l i t y , are i n themselves arbitrary—would not s a t i s f y Frege, since, he argues, a prerequisite for the a p p l i c a b i l i t y of a caloulus just i s that i t i s grounded In r e a l -i t y . Thus, considered merely as ink marks on a piece of paper, a formula oan no more be applied to r e a l i t y than oan a c o n f i g -uration of chess pieces on a ohess board; what makes the formula, but not the configuration of ohess pieces, applicable to r e a l i t y i s that the ink marks of the formula express a sense, whioh they can only do because number-signs purport to stand for e n t i t i e s . And once i t i s admitted that arithmetical formulas express a sense, we have an additional reason, Frege thinks, for denying that the ca l c u l a t i n g rules can be a r b i t r a r y : the rules must be so chosen as to insure that, from formulas expressing true prop-o s i t i o n s , there, can only be derived other formulas whioh l i k e -wise express true propositions. So, i n order for the gulf to l-Frege, p. I85. 2Frege, p. 189. 194 be bridged between arithmetical formulas and their a p p l i c a t i o n s , • • i t i s necessary that formulas express a sense and that the rules be grounded in. the reference of the signs . • • ,M l and these two requirements are Incapable of f u l f i l l m e n t i n formal arithmetic. l e s t we suppose that Thomae i s misrepresented i n being said to hold that numerical signs designate nothing at a l l , since he does say that suoh signs do have a kind of oontent i n v i r t u e of t h e i r behavior with respect to the rules of the c a l -culating game, Frege assures us that Thomae's apparent a s c r i p t i o n of a kind of content to numerical signs or chess pieces . . . i s due to an inaccurate formulation, prompted perhaps by a c e r t a i n repulsion from empty signs. . • • I am aware that the chess pieces are given, likewise that rules for their manip-u l a t i o n have been established, but I know nothing of any content. I t oan surely not be said that the black king, i n consequence of these r u l e s , designates something, as, say, the name ' S i r i u s ' designates a cert a i n fixed s t a r . On the contrary, the appro-p r i a t e way of speaking i s to say that the rules of chess treat of the black k i n g .2 Frege adds that i t i s misleading to speak of the behavior of signs with respect to r u l e s , since i t i s the player or c a l c u -l a t o r who, by obeying or disobeying the r u l e s , behaves with respect to them, and also that i t i s eccentric to speak of a numerical sign or a ohess pieoe as an external sign of i t s be-havior, since the more natural way of speaking would be to say that the rules of the calculus or of ohess treat of the manip-u l a t i o n of the signs or pieces.3 i n short, Frege altogether dismisses the attempt to give numerical signs a kind of content •'•Frege, p. 188. 2Frege, p. 190. 3Frege, p. 191. 195 constituted by r u l e s rather than by designated e n t i t i e s , and he continues to see the contrast between meaningful arithmetic and formal a r i t h m e t i c—in Thomae's as well as i n Heine's version as that between arithmetic whose formulas have a oontent, i n v i r tue of t h e i r ontologioal ground, and arithmetic whose formulas are composed of empty signs. In t h i s dispute between Frege and Thomae, as presented i n the Grundgesetze. we find a large part of the background against which Wittgenstein works out his own views concerning the neces-s i t y of grounding l i n g u i s t i c signs i n r e a l i t y . We can roughly characterize the evolution of Wittgenstein's p o s i t i o n by saying that the early Wittgenstein sides with Frege and that the l a t e r Wittgenstein sides with Thomae. Thus, i n the T, Wittgenstein goes along with Frege's view "that the use of a l l p r i m i t i v e signs must be j u s t i f i e d by the existence of e n t i t i e s corresponding to these signs and providing t h e i r ontologioal ground, since the only p r i m i t i v e signs recognized by the T are names of simple objeots, and i t i s of course these objects themselves, i n v i r t u e of t h e i r form, whioh provide the ground for the s y n t a c t i c a l l y correct use of t h e i r names i n propositions. Even i n r e j e c t i n g the Fregean approach to arithmetic, as i s done i n the T, Witt-genstein shows his respeot for Frege's p r i n c i p l e s by holding that numerical signs are merely exponents of operations (6.021) and not p r i m i t i v e signs, implying that i f they were p r i m i t i v e , then there would have to exist e n t i t i e s as t h e i r ontologioal grounds, and, presumably, mathematical equations would then be expressions of thoughts, as Frege supposes but as the T denies 196 (6.21). Then, beginning with PB, we see Wittgenstein breaking away from th i s Pregean p o s i t i o n and adopting Thomae*s p o s i t i o n , not only for arithmetic, but for language as suoh. So, whereas Thomae i m p l i c i t l y contrasts the c a l c u l a t i n g game of arithmetic, where signs have only a rule-constituted meaning, with language proper, where words and their s y n t a c t i c a l behavior presumably are ontologioally grounded i n e n t i t l e s designated by these words, we find Wittgenstein saying i n PB that the syntax of language generally, and not just of the signs of the arithmetical c a l c u l u s , cannot and need not be J u s t i f i e d by r e a l i t y (PJ3, p. 53); we also fin d that Thomae*s analogy between chess pieces and numerioal signs has been replaced by the broader analogy between chess pieces and words generally (PB, p. 61), though there s t i l l seems to be at l e a s t i m p l i c i t l y a contrast between those kinds of words which have merely a rule-constituted meaning and those whioh have an object-designating meaning. These points are r e - i t e r a t e d and amplified i n the notes recorded by Waismann i n WWK. where Wittgenstein again takes up the formalist strategy by contending that mathematios deals with ink marks on paper i n the same sense that chess has to do with figures of wood. That i s , we are not concerned with chess pieces qua physical p r o p e r t i e s , but r a t h e r , so to speak, qua l o g i o a l p r o p e r t i e s , whioh these pieces have i n v i r t u e of the rules of ohess. So, when I say that the knight can only move over three squares, the oastle only i n a s t r a i g h t l i n e , e tc., this *oan* means a grammatical, and not a p h y s i c a l , p o s s i b i l i t y ; i . e . , I am saying 197 something about the syntax, or grammar, of. ohess, ana not enun-c i a t i n g a law of physios.1 That pawns, kings, e t c . , are not to be characterizea merely as pieces of wood of oertain shapes, nor are they to be regarded as functioning as representatives of other objects, shows that Frege i s wrong i n objecting to the f o r m a l i s t s ' notion of rule-constituted meaning (and here, on p. 105 of WWK. Wittgenstein e x p l i c i t l y mentions Frege i n t h i s regard). The moral which Wittgenstein i s anxious to draw from thi s i s that ". • . not only the axioms of mathematics, but a l l syntax i s a r b i t r a r y . . . " (WWK. p. 103), and that . . apart from Its a p p l i c a t i o n s , considered for i t s e l f alone, i t [ i . e . , syntax] i s a game, exactly as chess i s " (WWK, p. 105). Here, we can p l a i n l y see that Wittgenstein's i n t e r e s t i n games i s from the perspective of the formalists games are the most obvious example of rule-governed a c t i v i t i e s i n which the rules require no ontologioal ground. Wittgenstein's advance beyond the f o r -malist conception i s to regard, not only mathematics, but a l l of language as a game i n this respect. With the advent i n PG of the doctrine of complete l i n g u i s t i c autonomy, the formalist strategy i s extended s t i l l f u r t h e r . Rule-oonstltuted meaning i s no longer contrasted with object-designating meaning, for even objects are now considered to be components of grammatical rules with respect to th e i r r o l e i n •^This i s c l e a r l y directed at Frege, who writes as i f the only sense i n whioh we can treat of the p o s s i b i l i t y of movement of a ohess piece i n the context of a chess game i s qua physical p o s s i b i l i t y of moving a p a r t i c u l a r piece of wood around on a ohess board. See Frege, pp. 190-191. 198 ostensive explanations. Wittgenstein is now able to say, quite generally, that the meaning of a word is i t s use, whioh, from the context in which this slogan f i r s t appears, can be seen to be largely a response in the formalist vein to the view that the meaning of a word is an objeot whioh provides the ontologioal ground for the word's use, the latter being thought of as distinct from, and a consequence of, i t s meaning. Thus, Wittgenstein be-gins by saying that the rules governing the use of the negation-sign seem to follow from the nature of negation, which is an independently deacribable process (PG, p. 53), and generally that a sign, in virtue of having a meaning by standing for some-thing, seems to contain within i t the whole grammar of i t s use, in the way that a box contains a string of beads which are wait-ing to be taken out (PG, p. 55 )• But this way of thinking of the relation of meaning and use rests on a primitive conception of meaning, which sees everything as being accomplished in the simple act of naming an object (PG, p. 56)• Aotually, words have many different kinds of uses (PG, p. 58), and an apt com-parison can be made between the different kinds of words and the different kinds of ohess pieces (PG, p. 59). The appropriateness of this analogy gives us a more helpful way of viewing the relation between meaning and use, v i z . , by identifying them. Thus, the meaning of a word is i t s place in the grammar (PG, p. 59) or i t s use in the language (PG, p. 60). Immediately after stating the meaning-is-use doctrine, Wittgenstein adds: Grammar describes the use of words In language. It is thus related to language similarly to the way that the description of a game, such as is given in the rules of the game, is related to the game (PG, p. 60). 199 The influence of the formalist outlook on Wittgenstein's formu-l a t i o n of the meaning-is-use doctrine i s , therefore, r e a d i l y apparent.1 Lest we overlook the implication whioh this oomparison of grammatical rules with rules of a game has for the r e l a t i o n of language to r e a l i t y , Wittgenstein t e l l s us a few pages l a t e r that . . . a name has meaning, a proposition has sense, i n the calculus to which i t belongs. This (jjaloulusp i s , so to speak, autonomous.—Language must speak for i t s e l f (PG, p. 63). And to say that language i s autonomous i s to say that, while we j u s t i f y moves within the calculus of language by appealing to the grammatical rules which sanction, these moves, the rules themselves have and need no j u s t i f i c a t i o n . In other words, Grammar i s not accountable to r e a l i t y . Meaning i s f i r s t determined (constituted) by grammatical r u l e s , and therefore they are not answerable to any meaning and are i n this respect ar b i t r a r y (PG, p. 184). The rules of grammar are arb i t r a r y i n a way that the rules of cooking are not, because the.concept 'cooking' i s defined by the purpose of cooking, while the concept 'language' i s not defined by some purpose outside of i t s e l f . Thus: Whoever cooks by other than the correct rules cooks badly; but whoever follows rules d i f f e r e n t than those of ohess plays another game; and whoever follows other grammatical rules than the usual ones does not speak f a l s e l y , but just speaks of some-thing else (PG, p. 185). A l l of thi s goes to show that the expression 'language-game' has been c a r e f u l l y chosen with the view towards empha-s i z i n g that the formalist contention, v i z . , that the arithmetical calculus i s l i k e a game i n having and needing no ontologioal •^Also, i n the BB the meaning-is-use doctrine i s introduced as a response to FregeTs" c r i t i c i s m of the fo r m a l i s t s . See BB, p. 4. 200 ground for i t s r u l e s , should be extended to cover language i n general; as suoh, the notion of a language-game marks an ex-tremely important step i n Wittgenstein's philosophical develop-ment. C l e a r l y , standard aooounts suoh as Quinton's of the rati o n a l e for the expression 'language-game'—viz., that various l i n g u i s t i c practices are a l l called 'language', not because they possess some oommon element, but because they have certain family-resemblances with one another, and that since l i n g u i s t i o practices are i n thi s respect l i k e games, the two terms of the simile are run together to emphasize that both 'language' and 'game' have a m u l t i p l i c i t y of uses1—are s u p e r f i c i a l at best. Perhaps the widespread f a i l u r e to recognize the influence of formalism i n Wittgenstein's notion of a language-game i s to some extent his own f a u l t , since i n the PI the formalist strategy i s greatly de-emphasized, though why this i s so I cannot say. S t i l l , I assume that i n the PI Wittgenstein does continue to hold that language has no ontologioal ground, and I base this assumption c h i e f l y on the evidence of PI #561-564. In #561, Wittgenstein asks why we want to say, as we do, that our use of the word ' i s ' both for the copula and for the sign of equality constitutes two di f f e r e n t meanings for the word ' i s ' rather than one meaning whose use oonsists, as i t were, i n both of these oomponent uses. We want to say that the use of the same word i n both oases i s i n e s s e n t i a l , and that two d i f f e r e n t words could serve Just as w e l l . He then poses, i n #562, some important questions: •^-Anthony Quinton, "Contemporary B r i t i s h Philosophy," i n WPI. p. 12. 201 But how can I decide what i s an e s s e n t i a l , and what an i n e s s e n t i a l , a c c i d e n t a l , feature of the notation? Is there some r e a l i t y l y i n g behind the notation, whioh shapes i t s grammar? If he were now inc l i n e d to answer this l a s t question i n the af f i r m a t i v e , thereby r e j e c t i n g his e a r l i e r a ssimilation of grammatical rules to rules of a game, we should expect as a follow-up to the question some kind of remark to this e f f e c t , considering the si g n i f i c a n c e of suoh a s h i f t i n p o s i t i o n . What he says instead, i s t h i s : Let us think of a simi l a r case i n a game: i n draughts a king i s marked by putting one piece on top of another. Now won't one say i t i s i n e s s e n t i a l to the game for a king to oon-s i s t of two pieces? (PI, #562) In #563 he then considers a simi l a r example pertaining to the use of the king i n ohess to determine whioh player moves f i r s t , depending on the oolor of the king one player s e l e c t s from the closed f i s t of the other; he asks whether we want to count this as part of the r o l e of the king i n ohess, and, implying that he does not, concludes i n #564: So I am inclin e d to dis t i n g u i s h between the essential and ine s s e n t i a l i n a game too. The game, one would l i k e to say, has not only rules but also a p o i n t . That i s , even though a game i s constituted by i t s rules—and not by the way and the extent to whioh these rules happen to be grounded i n some external reality—we nevertheless are able to dis t i n g u i s h e s s e n t i a l from i n e s s e n t i a l r u l e s , and the point of making this observation where he does can only be to argue that grammatical r u l e s , too, can be distinguished into those that are esse n t i a l and those that are i n e s s e n t i a l for a p a r t i c u l a r language-game, without our having to suppose that this d i s t i n c t i o n requires that the language-game be grounded i n r e a l i t y . 202 There i s s t i l l other evidence that Wittgenstein continues to characterize language from the formalist perspective. In PI #492 he says: To invent a language could mean to invent an instrument for a p a r t i c u l a r purpose on the basis of the laws of nature (or consistently with them); but i t also has the other sense, analogous to that i n whioh we speak of the invention of a game. This other sense of 'invent1 means simply the act of s t i p u l a t i n g a system of rules so as to constitute a language-game; i n thi s sense, we are looking at language, not as an instrument for bringing about some e f f e c t , but as a grammatical system. Thus: Grammar does not t e l l us how language must be constructed i n order to f u l f i l l i t s purpose, i n order to have such-and-such an effect on human beings (PI., #496). The rules of grammar may be called " a r b i t r a r y " , i f that i s to mean that the aim of the grammar i s nothing but that of the language (PI, #497). Of course, there i s also a sense i n whioh our grammatioal rules are not a r b i t r a r y , v i z . , i n the sense i n whioh our grammar gets i t s character from our form of l i f e , i . e . , our c u l t u r a l back-ground, customs, education, i n s t i n c t i v e behavior, e t c . But these factors provide, not an ontologioal ground, or j u s t i f i c a t i o n , for our grammatical r u l e s , but rather a genetic, or causal, ex-planation of why we have them. Likewise, though our form of l i f e depends i n turn upon facts of nature being as they are, suoh facts themselves are relevant to our grammatical rules only as causes and not as grounds. To i n s i s t that grammatical rules must have, not only causes, but grounds as well i s to display a misconception about the j u s t i f i c a t i o n a l process, against whioh Wittgenstein f i g h t s no less e x p l i c i t l y i n the PI, e.g., i n #217, than he does i n the PG when he says that "A reason (Grund) oan 203 be given only inside a game. The chain of reasons oomes to an end, and indeed does so at the boundary of the game. (Reason and oause.)" (PG, p. 97) Even i n his l a s t philosophical work, 0C, Wittgenstein oontinues to maintain th i s l i n e of thought by saying that a language-game ". . • i s not based on grounds. I t i s not reasonable (or unreasonable). I t i s t h e r e—lik e our l i f e " (OC, #559). S t i l l other evidenoe that Wittgenstein never renounoes the view that grammar has no ontologioal ground i s to be found i n a number of remarks preserved i n Z, including two of the more ex-p l i c i t remarks to thi s e f f e c t , closely resembling versions of which had o r i g i n a l l y appeared i n P_G? These include the passage comparing rules of grammar and rules of cooking—to whioh i s added, i n the version i n Z_: "That i s why the use of language i s i n a oertain sense autonomous, as oooking and washing are not . . . " (Z, #320)—and also Z #331 (resembling the passage on pp. 185-186 of PG), which begins: One i s tempted to j u s t i f y rules of grammar by sentences l i k e "But there r e a l l y are four primary c o l o r s " . And the saying that the rules of grammar are ar b i t r a r y i s directed against the pos-s i b i l i t y of t h i s j u s t i f i c a t i o n , whioh i s constructed on the model of j u s t i f y i n g a sentence by pointing to what v e r i f i e s i t . And the passage oonoludes by suggesting that, although we do i n faot o l a s s l f y r e d , green, yellow, and blue, as primary oolors, i t i s not as though the nature of objective r e a l i t y provides grounds for thi s c l a s s i f i c a t i o n as opposed to some other one. We could, i f we so chose (or i f our form of l i f e so d i c t a t e d ) , group r e d , green, and c i r c u l a r together; such a c l a s s i f i c a t i o n would be no les s legitimate than any other, since the notion of 204 an i n t r i n s i c a l l y legitimate c l a s s i f i c a t i o n , grounded i n the true nature of r e a l i t y , simply has no p l a c e , I conclude, therefore, that Wittgenstein's considered, mature view continues to he that language i s not grounded i n r e a l i t y . How this conclusion bears on Wittgenstein's status as a r e a l i s t can be seen, e.g., by examining the language-game des-cribed i n PI #48. The words of this language are 'R', 'G*, 'W*, and 'B', whioh name red, green, white, and black squares, respec-t i v e l y . Now, i f Wittgenstein were a t r a d i t i o n a l r e a l i s t with respect to shades of oolor, he would have to maintain that a oolored square i s , i n a fundamental sense, undeniably composite, since i t contains as a component a color which i s a (universal) element i n i t s own r i g h t . What he actually says, however, i s that the oolored squares are the primary elements. Then, to the questions "But are these simple?" he r e p l i e s s I do not know what else you would have me o a l l "the simples", what would be more natural i n this language-game. But under other circumstances I should c a l l a monochrome square "composite", con-s i s t i n g perhaps of two rectangles, or of the elements colour and shape (PI, #48). He does not specify what.these 'other circumstances' might be, but one suoh circumstance i n whioh we would be inclined to treat color and shape as distinguishable components would be where the elements were d i f f e r e n t shapes as well as of d i f f e r e n t oolors. As i t i s , with every element dealt with i n language (48) being a square, the d i s t i n c t i o n between color and shape beoomes i r r e l -evant. If we i n s i s t that, relevant or not, the d i s t i n c t i o n i s s t i l l there, waiting to be made, thi s i s to miss the point that the only d i s t i n c t i o n s whioh exist are provided for i n the language-205 game i t s e l f . In language (48) the d i s t i n c t i o n between color and shape simply does not e x i s t . That ontologioal d i s t i n c t i o n s are r e l a t i v e to, and dependent on, the language-game being played receives further i l l u s t r a t i o n from PI #64s l e t us imagine language game (48) altered so that names si g n i f y not monochrome squares but rectangles each consisting of two suoh squares. Let such a rectangle, which i s half red half green, be called "U"; a half green half white one, "V"; and so on. Could we not imagine people who had names for such combina-tions of colour, but not for the ind i v i d u a l colours? . . . In what sense do the symbols of this language-game stand i n need of analysis? How far i s i t even possible to replaoe this language-game by (48)?—It i s just another language-game; even though i t i s related to (48). In other words, we cannot say that language (48) allows for a more accurate representation of r e a l i t y by providing for d i s -t i n c t i o n s which are slurred over by language (64), and this i s because there i s nothing that can be called ' r e a l i t y * which functions as a standard external to a l l language-games and against whioh they oan be compared. Insofar as we talk about r e a l i t y at a l l , we oan only be understood within the framework of a p a r t i c -ular language-game. In a manner reminiscent of Kant's categories and forms of i n t u i t i o n , the language-game has been plaoed as a necessary intermediary between us and noumenal r e a l i t y , as i t were—even to the poi n t , as could be contended i s also the case for Kant, that the very i n t e l l i g i b i l i t y of the notion of a r e a l i t y standing on the other side of thi s medium i s i n question. 'Reality* just i s • r e a l i t y as defined by the language-game'. Therefore, when I say that Wittgenstein makes statements implying that he i s a r e a l i s t with respect to s p e c i f i c shades of oolor, or whatever, 206 what I r e a l l y mean i s that he makes statements whioh show that he i s playing a language-game appropriate to realism, or at l e a s t describing r e a l i s t language-games and implying that i n - o u r form of l i f e we do sometimes play them. So, e.g., when he says that colors e x i s t , i . e . , that things exist which have c o l -ors as p r o p e r t i e s , we should, on his behalf, contend at most that he i s committing himself to a r e a l i s t ontology only i n the sense that he happens to be playing or describing a language-game from which this ontology issues. To go further than t h i s , saying some suoh thing as that r e a l i t y compels us, upon pain of inac-curacy or incompleteness, to acknowledge the existence of univer-sals such as c o l o r s , and to provide for the p o s s i b i l i t y of naming or describing them i n language, would be inadmissible. I f this appears to be a rather d r a s t i c q u a l i f i c a t i o n to be made about Wittgenstein's status as a r e a l i s t , I should again point out that the q u a l i f i c a t i o n applies generally to any ontologioal p o s i t i o n to whioh one might want to say that Wittgenstein i s committed. He i s , at any r a t e , no less oommitted to the e x i s t -ence of colors as properties of things than he i s to the e x i s t -ence of such p a r t i c u l a r things themselves. With th i s oonoluslon, I am now i n a p o s i t i o n to examine s t i l l one more argument—advanced th i s time by Pitcher—purporting to show that Wittgenstein provides the basis for disposing of the errors of realism (and which, of oourse, takes for granted that Wittgenstein himself i s not a r e a l i s t ) . Pitcher claims that part of Augustine's over-simplified pioture of language-learning attaoked by Wittgenstein i s the view that we can see the universal 20? In the p a r t i c u l a r , o r—les s crudely put—that we can come to possess a oonoept by abstracting the relevant common features from many p a r t i c u l a r s . Claiming that Wittgenstein's dootrine of family-resemblances provides a powerful argument against the Augustinian p i c t u r e of concept formation as regards some concepts, Pitoher concedes that the Augustinian pioture has at least a prima faole p l a u s i b i l i t y for r e l a t i v e l y simple oonoepts suoh as those of oolors. Surely we want to say that a l l red things have the oolor red i n oommon (I am again assuming, to avoid i r r e l e v a n t complications, that 'red' i s being used as the name of a s p e c i f i c shade of c o l o r ) , and that to apply the oonoept 'red' to future things I may encounter, I need only note whether they, too, pos-sess this oommon feature. My contention has been that t h i s i s , i n f a c t , Wittgenstein's view, v i z . , that while some oonoepts, even the vast majority, treat only of family-resemblanoes, other oonoepts are indeed applied to various p a r t i c u l a r s i n v i r t u e of common features whioh the p a r t i c u l a r s share, the most obvious example, and the one usually given by Wittgenstein himself, being that of c o l o r s . But Pitcher contends that Wittgenstein has a l i n e of argument whioh can be used to show that even the modest claim that there are any common-feature concepts at a l l i s mis-taken. To say that many p a r t i c u l a r s have a common feature, whioh i s to say that they have the same feature, assumes, according to Pi t c h e r , that sameness "• . . i s a wholly n a t u r a l , as opposed to a conventional, r e l a t i o n .1 , 1 Pitoher oontends, to the contrary, 1George P i t c h e r , "About the Same," i n Ludwlg Wittgenstein.  Philosophy and Language. p. 122. 208 that '•• • • there i s an element of convention—indeed, a profound element of convention—in a l l sameness, m1 T O support this claim as regards oolor, he argues that we can imagine a society em-bodying a quite d i f f e r e n t form of l i f e than our own whioh would see nothing i n oommon between two c o l o r s , such as l i g h t blue and dark blue, whioh we o l a s s i f y as being two shades of the same c o l o r . His conclusion of this part of the argument i s that These examples show that there i s no answer to the question whether or not two colours are d i f f e r e n t shades of the same colour apart from the conventions embodied i n the colour-concepts of thi s or that language—conoepts determined by the 'form of l i f e ' of the language users and by certain general facts of nature, . . , There i s no sense to the claim that colours 1 and 2 just are, i n and of themselves, shades of the same colour: according to our colour-concepts, they are indeed shades of the same colour—namely, of blue; but according to the colour-concepts of our mythical t r i b e s -men, they are not.2 This part of the argument, however, does not yet undercut the r e a l i s t ' s claim that two or more things oan have as their common element p r e c i s e l y the same shade of co l o r , so Pitcher t r i e s to show that even the judgment that two things have the same oolor contains an element of conventionality. We oan imagine a t r i b e , he says, who l i v e . • . i n a jungle illuminated by an eerie l i g h t called jungle l i g h t . Most things are unohanged i n colour when they are brought from sunlight into jungle l i g h t , but certain kinds of materials x change s l i g h t l y i n colour when this happens to them. Hence i t i s quite possible that we should judge two pieces of oloth—one of kind x and another of kind not-x—to have exactly the same colour when the jungle tribesmen would judge them to be of d i f f e r e n t colours.3 This argument, though ingenious, w i l l not do what Pitcher requires of i t ; i t shows at most that the oolor an objeot i s Judged actually I p i t c h e r , p. 123. 2 P i t o h e r , pp. 131-132. 3pitoher, p. 133. 209 to have—as opposed to the oolor i t merely appears to have—is based partly on the convention of what we accept as standard l i g h t i n g conditions. But i t does not show that we are mistaken i n supposing that a s p e c i f i c shade of color oan and does q u a l i f y many d i f f e r e n t p a r t i c u l a r s as the i r common element. The r e a l i s t need not deny that what we c a l l a thing's ' r e a l ' oolor i s p a r t l y based on convention; the c r u c i a l p o i n t l i s that, even though two cultures may disagree as to which objects have, say, fire-engine red as t h e i r ' r e a l ' c o l o r , both cultures w i l l presumably agree that t h i s same shade of red q u a l i f i e s more than one thing. Nevertheless, leaving aside this objection to Pitcher's argument, the more Important objection to his whole approach i s not that he over-estimates, but rather that he under-estimates, the element of conventionality whioh Wittgenstein would .-say i s involved i n our oonoept formation. That i s , i n saying that the r e l a t i o n of sameness i s at le a s t i n part conventional (where 'conventional* oan be taken to mean 'arising from our form of l i f e rather than based on the way r e a l i t y i s i n i t s e l f ' ) , he implies that the r e l a t i o n i s i n part not conventional, so that when I judge of two things that they are i n some respect the same, my judgment i s j u s t i f i e d at l e a s t i n part by the way r e a l i t y i s . For Wittgenstein, however, the process of j u s t i f y i n g a move i n a language-game f i n a l l y terminates i n appealing to the rules of the language-game, and these rules can never be based on the nature of an e x t r a - l i n g u i s t i c r e a l i t y .1 Again, should i t •*-I say ' f i n a l l y ' , for c l e a r l y 'the way r e a l i t y i s ' can be cited as a j u s t i f i c a t i o n for a move within a language-game, though only as an intermediate step. Thus, e.g., I could jus-210 appear that I am making Pitcher's case against Wittgenstein's purported realism stronger than Pitcher makes i t himself, I can only repeat that Wittgenstein's ontology does oontain u n i v e r s a l s , no less than i t contains p a r t i c u l a r things, whatever we may think about c a l l i n g 'ontology' that whioh has i t s source i n language. t i f y my claim that the cat i s behind the couch by saying that I saw i t go behind the couch a minute ago and that I now hear a purring noise coming from that d i r e o t i o n . But i f I then were asked what j u s t i f i c a t i o n I had for regarding these observations as a j u s t i f i c a t i o n for my o r i g i n a l claim, I could only appeal to the r u l e s— t a c i t l y accepted and for the most part unarticulated— by which we play the language-game of giving evidence and which j u s t i f y me—if, indeed, we are w i l l i n g to suppose that the jus-t i f i c a t i o n a l demand could properly be taken even this f a r — i n supposing that my evidence i s strong enough to j u s t i f y my o r i g -i n a l claim. And, of course, these r u l e s , whioh form the frame-work of the language-game, are not themselves ontologically grounded. To approach the same example from a d i f f e r e n t angle, i f I had given my evidence and were then asked how I knew that what I heard was (or even seemed to be) a purring noise, I could only appeal to the rules for the use of the words 'purring noise' to j u s t i f y my description as being c o r r e c t , and these rules them-selves have no ontologioal ground. Therefore, the process of j u s t i f i c a t i o n , i n whichever direotion we take i t , ultimately terminates, not i n the way r e a l i t y i s , but i n grammatical r u l e s . (See, however, OC, i n whioh Wittgenstein argues that some empir-i c a l propositions are as certain as any grammatical rules whioh could be cited i n support of them—e.g., i n QC #306-307—and therefore themselves constitute part of the j u s t i f i o a t i o n a l framework of language-games. Whether or not this admission of content-expressing propositions to the j u s t i f i o a t i o n a l framework requires a fundamental modification of the notion of a language-game i s a question that I oannot here explore f u r t h e r , although I should add that Wittgenstein himself recognizes the problem and t r i e s to solve i t , i f not altogether convincingly, by t r e a t -ing such basic empirical propositions as descriptive norms and thus ass i m i l a t i n g them to grammatical r u l e s . See QjC, #95, 98, 308, and 321.) 211 AFTERWORD In retrospect, perhaps the most surprising r e s u l t of our examination i s that, considered apart from the philosophical systems i n whioh they are embedded, Wittgenstein's views on universals are rather staid and unexciting. The t r a n s i t i o n from Plato to A r i s t o t l e hardly has the makings of a philosoph-i c a l r e v o l u t i o n . Granted, the early Wittgenstein i s very far from P l a t o , and the l a t e r Wittgenstein Just as far from A r i s t o t l e , oonoerning whioh universals he acknowledges to e x i s t , but even those handy devices whioh he uses to avoid the excesses to whioh r e a l i s t s are l i a b l e , v i z . , the tool of analysis and the family-resemblance strategy, are not of his own making. The programme of a n a l y s i s , of course, i s taken over from R u s s e l l , and the family-resemblance strategy—as I s h a l l try to show i n Appendix B—is l i f t e d s t r a i ght out of the pages of Ogden and Richards1  The Meaning of Meaning. What i s exciting and o r i g i n a l i s the way Wittgenstein i s able to gather up old and borrowed views suoh as these and synthesize them into a powerful and compelling philosophical system d i s t i n c t l y his own. Even more remarkably, he i s able i n his l i f e t i m e to do this twioe. 212 BIBLIOGRAPHY Aaron, R. I. The Theory of Universale. 2nd ed. London: Clarendon Press, 1967. Allaire?, Edwin B. "The 'Traotatus': Nominalistic or Realistic?" Essays on Wittgenstein's Traotatus. Edited by I. M. Copi and R. W. Beard. New York: Macmillan, I966. Ambrose, Alice. "Wittgenstein on Universals." Ludwlg Wittgen-stein, The Man and His Philosophy. Edited by K. T. Fann. New York: D e l l , 19^77 Ambrose, Alice, and Lazerowitz, Morris, eds. Ludwig Wittgenstein.  Philosophy and Language. London: Allen & Unwin, 1972. Ansoombe, G. E. M. An Introduction to Wittgenstein's Traotatus. Harper Torchbooks. New York: Harper & Row, 1965. Ansoombe, G. E. M. "Mr Copi on Objects, Properties, and Relations in the 'Traotatus'." Essays on Wittgenstein's Traotatus. Edited by I. M. Copi and R. W. Beard. New York: Macmillan, 1966. Bambrough, Renford. "Universals and Family Resemblances." Witt- genstein: The Philosophical Investigations. Edited by George Pitcher. New York: Doubleday, 1966. Bernstein, Richard. "Wittgenstein's Three Languages." Essays on Wittgenstein's Traotatus. Edited by I. M. Copi and R. W. Beard. New York: Macmillan, 1966. Black, Max. A Companion to Wittgenstein's 'Traotatus'. Ithaca, N. Y.: Cornell University Press, 1964. Cook, John. "Wittgenstein on Privacy." Wittgenstein: The P h i l - osophical Investigations. Edited by George Pitcher. New York: Doubleday, 1966. Copi, I. M. "Objects, Properties, and Relations in the 'Trao-tatus'." Essays on Wittgenstein's Traotatus. Edited by I. M. Copi and R. W.-Beard. New York: Macmillan, 1966. Copi, I. M., and Beard, R. W. Essays on Wittgenstein's Traotatus. New York: Macmillan, 1966. Donagan, Alan. "Wittgenstein on Sensation." Wittgenstein: The  Philosophical Investigations. Edited by George Pitoher. New York: Doubleday, 1966. 213 Engelmann, Paul. Letters from Ludwlg Wittgenstein with a Memoir. Edited by B. F. MoGuinness. . Translated by L. Furtmueller. Oxfords Blaokwell, 196?. Evans, E. "About •aRb1." Essays on Wittgenstein's Traotatus. Edited by I. M. Copi and R. W. Beard. New Yorks Maomillan, 1966. Evans, E. "Traotatus 3.1432." Essays on Wittgenstein's Trao-tatus. Edited by I. M. Copi and R. W. Beard. New York: Maomillan, 1966. Fann, K. T., ed. Ludwlg Wittgenstein. The Man and His Philosophy. A Delta Book. New York: D e l l , 1967. Favrholdt, David. An Interpretation and Critique of Wittgenstein's  Traotatus. New York: Humanities Press, 1966. Feyerabend, Paul. "Wittgenstein's Philosophical Investigations." Wittgensteins The Philosophical Investigatlons. Edited by George P i t c h e r . New York: Doubleday, 1966. Frege, Gottlob. The Foundations of Arithmetic. Translated by J . L. Aust i n . Oxford: Blaokwell, 1953. Frege, Gottlob. Grundgesetze der Arithmetic (Vol. I I , #86-137). Translated by Max Black. Philosophical Writings of Gottlob  Frege. Edited by Peter Geach and Max Black. Oxford: Blaokwell, 1966. Frege, Gottlob. "On Conoept and Object." Translated by Peter Geach. Phllosophloal Writings of Gottlob Frege. Edited by Peter Geaoh and Max Black. Oxford! Blaokwell, 1966. Frege, Gottlob. "On Sense and Reference." Translated by Max Black. Phllosophloal Writings of Gottlob Frege. Edited by Peter Geaoh and Max Black. Oxford: Blaokwell, 1966. Geach, Peter, and Black, Max, eds. Phllosophloal Writings of Gottlob Frege. 2nd ed. Oxford! Blaokwell, 1966. G r i f f i n , James. Wittgenstein's Logioal Atomism. London: Oxford University Press, 1964. " Hartnaok, Justus. Wittgenstein and Modern Philosophy. Anchor Books. Translated by Maurice Cranston. New York: Double-day, 1965. Kenniok, W. E. "Philosophy as Grammar and the Reality of Uni-ver s a l s . " Ludwig Wittgenstein. Philosophy and Language. Edited by A l i c e Ambrose and Morris Lazerowitz. London: A l l e n & Unwin, 1972. 214 Kenny, Anthony, "Cartesian Privacy," Wittgenstein: The P h i l o - sophical Investigations. Edited by George Pitoher. New York: Doubleday, 1966. Keyt, David. "A New Interpretation of the Traotatus Examined." Philosophical Review 74 (1965). Keyt, David. "Wittgenstein's Notion of an Objeot." Essays on  Wittgenstein's Traotatus. Edited by I. M. Copi and R. W. Beard. New York: Macmillan, 1966. Keyt, David. "Wittgenstein's Pioture Theory of Language." Essays  on Wittgenstein's Traotatus. Edited by I. M. Copi and R. W. Beard. New York: Macmillan, 1966. Khatchadourian, Haig. "Common Names and Family Resemblances." Wittgenstein: The Philosophical Investigations. Edited by George P i t c h e r . New York: Doubleday, 1966. Klemke, E. D., ed. Essays on Frege. Urbana: University of I l l i n o i s Press, 1968. Klemke, E. D., ed. Essays on Wittgenstein. Urbana: University of I l l i n o i s Press, 1975T Malcolm, Norman. Ludwig Wittgenstein: A Memoir. London: Oxford University Press, 1958. ' ~ Malcolm, Norman. "Wittgenstein's Philosophical Investigations." Wittgenstein: The Philosophical Investigations. Edited by George Pitoher. New York: Doubleday, 1966. Marsh, Robert C , ed. Logic and Knowledge: Essays. 1901-1950. London: A l l e n & Unwin, 1956. MoGuinness, B. F. "Pictures and Forms i n Wittgenstein's'Trao-tatus'." Essays on Wittgenstein's Traotatus. Edited by I. M. Copi and R. W. Beard. New York: Macmillan, 1966. Moore, G. E. "Wittgenstein's Lectures i n 1930-33." Philosoph-i c a l Papers. London: A l l e n & Unwin, 1959. Mueller, Anselm. Qntologle i n Wittgensteins 'Traotatus'. Bonn: Bouvier, 1967. Ogden, C. K., and Richards, I. A. The Meaning of Meaning. 8th ed. New York: Harcourt, Brace and Company, 1946. Pears, David. Bertrand Russell and the B r i t i s h Tradition i n Philosophy. London: C o l l i n s Press, 1967. Pears, David. Ludwig Wittgenstein. New York: Viking Press, 1970. 215 P i t c h e r , George, "About the Same." Ludwlg Wittgenstein. P h i l - osophy and Language. London: A l l e n & Unwin, 1972. P i t c h e r , George. The Philosophy of Wittgenstein. Englewood C l i f f s , N. J . : P r e n t i c e - H a l l , 1964. P i t c h e r , George, ed. Wittgenstein: The Philosophical I n v e s t i - gations. Anchor Books. New York: Doubleday, 196ZI Quinton, Anthony. "Contemporary B r i t i s h Philosophy." Wittgen-s t e i n : The Philosophical Investigations. Edited by George P i t c h e r . New York: Doubleday, 1966. Ramsey, Frank. " C r i t i c a l Notice of the 'Traotatus1." Essays on  Wittgenstein's Traotatus. Edited by I. M. Copi and R. W. Beard. New York: Maomillan, 1966. R u s s e l l , Bertrand. The Analysis of Mind. London: A l l e n & Unwin, 1921. R u s s e l l , Bertrand. Introduction to Mathematical Philosophy. New York: Maomillan, 1919. R u s s e l l , Bertrand. Mysticism and Logic. Anchor Books. Garden C i t y , N. Y.: Doubleday, 1957. R u s s e l l , Bertrand. "On S c i e n t i f i c Method i n Philosophy." Mys- ticism and Loglo. Garden.City, N. Y.: Doubleday, 1957. R u s s e l l , Bertrand. "On the Relation of Universals and P a r t i c -u l a r s . " Loglo and Knowledge: Essays, 1901-1950. Edited by Robert C. Marsh. London: A l l e n & Unwin, 1956. R u s s e l l , Bertrand. Our Knowledge of the External World. A Mentor Book. New York: The New American L i b r a r y , I960. R u s s e l l , Bertrand. Philosophical Essays. London: A l l e n & Unwin, 1966. R u s s e l l , Bertrand. "Philosophy of Logioal Atomism." Loglo and  Knowledge: Essays. 1901-1950. Edited by Robert C. Marsh. London: A l l e n & Unwin, 195&. R u s s e l l , Bertrand. The P r i n c i p l e s of Mathematics, 2nd ed. New York: W. W, Norton & Company, 19^1 R u s s e l l , Bertrand, The Problems of Philosophy. A Galaxy Book. Oxford: Oxford University Press, 1959. S e l l a r s , W i l f r i d . "Naming and Saying." Essays on Wittgenstein's  Traotatus. Edited by I. M. Copi and R. W. Beard. New York: Maomillan, 1966. 216 Specht, Ernst Konrad. The Foundations of Wittgenstein's Late  Philosophy. Translated by D. E. Walford. Manchester: Manchester University Press, 1969. Stenius, E r i k . Wittgenstein's 'Traotatus1. Oxford: Blaokwell, I960. Strawson, Peter. "Review of Wittgenstein's Philosophical I n v e s t i -gations. " Wittgenstein: The Philosophical Investigations. Edited by George P i t c h e r . New York: Doubleday, 1966. Stroud, Barry. "Wittgenstein and Logioal Necessity." Wittgen- s t e i n : The Phllosophloal Investigations. Edited by George P i t c h e r . New York: Doubleday, 1966. Thomson, Judith J a r v i s . "Professor Stenius on the 'Traotatus'." Essays on Wittgenstein's Traotatus. Edited by I. M. Copi and R. W. Beard. New York: Maomillan, 1966. Waismann, F r i e d f i o h . Ludwig Wittgenstein und der Wiener K r e l s . Edited by B. F. McGuinness. Oxfords Blaokwell, 1967. Waismann, F r i e d r i o h . The P r i n c i p l e s of L i n g u i s t i c Philosophy. Londons Maomillan, I965I Walker, 'Jeremy. A Study of Frege. Ithacas Cornell University Press, 1965."" Weinberg, J u l i u s Rudolph. An Examination of Logioal P o s i t i v i s m . Paterson, N. J.s L i t t l e f i e l d , Adams & Co., I960. Whitehead, A l f r e d North, and R u s s e l l , Bertrand. P r l n o i p i a Math- ematloa to *56. Cambridge: Cambridge University Press, 1962. Whitehead, Alfred North. Process and R e a l i t y . Harper Torohbooks. New York: Harper & Bros., I96Q. Winch, Peter, ed. Studies i n the Philosophy of Wittgenstein. Londons Routledge & Kegan Paul, 1969. Wittgenstein, Ludwig. The Blue and Brown Books. Harper Torch-books. New Yorks Harper & Row, 1965. Wittgenstein, Ludwig. Lectures and Conversations on Aesthetics.  Psychology, and Religious B e l i e f . Edited by C y r i l B a rrett. Oxford: Blaokwell, 1966. Wittgenstein, Ludwig. Notebooks. 1914-1916. Harper Torohbooks. Edited by G. H. von Wright and G. E. M. Ansoombe. .Trans-lated by G. E. M. Ansoombe. New York: Harper & Row, 1969. 21? Wittgenstein, Luawig. "Notes Dictated to G. E. Moore i n Norway, A p r i l 1914." Notebooks. 1914-1916. Eaitea by<-G. H. von Wright ana G. E. M. Ansoombe. Translatea by G. E. M. Ans-oombe. New York: Harper & Row, 1969. Wittgenstein, luawig. "Notes on l o g i c , September 1913." Note- books, 1914-1916. Eaitea by G. H. von Wright ana G. E. M. Anscombe. Translatea by G. E. M. Ansoombe. New York: Harper & Row, 1969. Wittgenstein, luawig. "Notes on Privacy." Philosophical Review 77 (1968). Wittgenstein, Luawig. Cn Certainty. Eaitea by G. H. von Wright and G. E. M. Anscombe. Translatea by Denis Paul ana G. E. M. Ansoombe. Oxfora: Blaokwell, 1969. Wittgenstein, Luawig. Philosophical Investigations. 2nd ea. Eaitea by G. E. M. Ansoombe ana Rush Rhees. Translated, by G. E. M. Ansoombe. Oxford: Blaokwell, 1967. Wittgenstein, Ludwig. Philosophisohe Bemerkungen. Editea by Rush Rhees. Oxfora: Blaokwell, 1964. Wittgenstein, Luawig. Philosophisohe GrammatIk. Eaitea by Rush Rhees. Oxfora: Blaokwell, 1969. Wittgenstein, Ludwig. Proto-Traotatus. Edited by B. F. MoGuinness, T. Nyberg, ana- G. H. von Wright. Translatea by D. F. Pears and B. F. MoGuinness. Lonaon: Routleage & Kegan Paul, 1971. Wittgenstein, Luawig. Remarks on 'the Founaatlons of Mathematics. Eaitea by G. H. von Wright, Rush Rhees, ana G. E. M. Ans--;oombe. Translatea by G. E. M. Ansoombe. Oxfora: Black-w e l l , 1956. Wittgenstein, Luawig. "Some Remarks on Logical Form." Essays on Wittgenstein's Traotatus. Eaitea by I. M. Copi ana R. W. Beara. New York: Macmillan, 1966. Wittgenstein, Luawig. Traotatus Logloi-Phllosophious. Trans-latea by D. F. Pears ana B. F. MoGuinness. Lonaon: Rout-leage & Kegan Paul, 1961. Wittgenstein, Luawig. Z e t t e l . Eaitea by G. E. M. Ansoombe ana G. H. von Wright. Translatea by G. E. M. Ansoombe. Berke-le y : University of C a l i f o r n i a Press, 1967. 218 . APPENDIX A NONJUSTIPICATIONALISM AND THE PROBLEM OF PRIVATE OBJECTS In Chapter VII I contended that, whatever else i s packed into Wittgenstein's notion of a language-game, his fundamental r a t i o n a l e for comparing language to a game i s to extend the formalist strategy to language i n general. Just as the rules of a game are not answerable to r e a l i t y , l ikewise, the grammatical ru l e s of language have no ontologioal ground. J u s t i f y i n g a move i n a language-game i s ultimately a matter of appealing to the rules of the game, and beyond t h i s , the j u s t i f i c a t i o n a l pro-cess cannot go. I f we try to push the process back any f u r t h e r , we can only r e f e r to the form of l i f e we actually have. This a l l u s i o n to our form of l i f e , however, i s not r e a l l y part of the process of J u s t i f i c a t i o n , but i s at best a causal expla-nation of why we have the language-games we do and why we f o l -low the rules of the language-games i n one way rather than an-other. The moral to be drawn from t h i s , whioh appears as a theme throughout Wittgenstein's l a t e r works, i s that j u s t i f i -cation comes to an end, i . e . , we must always reach a point where the request for a j u s t i f i c a t i o n no longer makes any sense. Of oourse, the notion that the process of j u s t i f i c a t i o n i s f i n i t e i s hardly o r i g i n a l with Wittgenstein. What i s d i s t i n c t i v e about his approach i s his willingness to stop short of the place where others have tr i e d to push the J u s t i f I c a t i o n a l process. To give but one example, i n Chapter IV of PP Russell postulates a law of induction whioh, i n conjunction with our past experience, 219 i s supposed to j u s t i f y our b e l i e f that, say, f i r e w i l l burn us— this law of induction i t s e l f constituting a j u s t i f i o a t i o n a l t e r -minus, sinoe we oannot i n turn j u s t i f y i t . Wittgenstein prefers to omit the inductive p r i n c i p l e from the j u s t i f i o a t i o n a l process altogether; i f the process has to stop anyway, why not l e t i t stop at those p r i m i t i v e , fundamental b e l i e f s suoh as that i f we put our hand i n the f i r e , we w i l l be burned? No appeal to an inductive law could make these kinds of b e l i e f s any more c e r t a i n , psychologically or epistemologically, than they are already, and thi s further appeal can therefore be eliminated as superfluous.1 This desire of Wittgenstein's to avoid any gratuitous ex-tension of the j u s t i f i o a t i o n a l process, which makes him c r i t i c a l of the places at which many philosophers have t r a d i t i o n a l l y drawn the j u s t i f i o a t i o n a l boundaries (e.g., at i n t u i t i v e l y ' s e l f -evident' propositions, a.view whioh i s c r i t i c i z e d i n OC, #144 and 204), forms that part of his outlook whioh, for want of a better word, I c a l l his n o h j u s t i f i o a t l o n a l i s m . Although many oommenta-tors have noted the presence of this theme of n o n j u s t i f i o a t l o n -alism i n Wittgenstein's' work, I can think of no one who has given i t the emphasis i t deserves as a theme l y i n g at the very heart of Wittgenstein's l a t e r philosophy. Even Speoht, who, i n emphasiz-1See PI, #324-325. That p r i m i t i v e b e l i e f s , with an osten-s i b l y empirical content, oan form part of the j u s t i f i o a t i o n a l framework of language-games, apparently goes against my claim just previously made that the j u s t i f i o a t i o n a l process terminates i n the grammatical rules of the language-game. However, as I indicated i n the footnote on p. 210, the question of whether to include empirical propositions as part of the j u s t i f i o a t i o n a l framework of language-games was a source of d i f f i c u l t y for Witt-genstein himself and never adequately resolved. 220 ing that for Wittgenstein " . . . the rules of a language-game cannot be read off from r e a l i t y . • .1 , 1 and that r e a l i t y i t s e l f " . . • i s only l i n g u i s t i c a l l y a c c e s s i b l e , i . e . , only given and attainable i n language-games • • • ," has grasped more cl e a r l y the f r u i t s of the formalist strategy than most, and perhaps a l l , other oommentators, nevertheless, does not r e a l i z e how deeply Wittgenstein i s committed to nonjustifioationalism as an i n t e g r a l part of t h i s strategy. Thus, Specht observes that language-games are supposed to be able to create ". . . a new organization of objeots into groups, according to certain points of view or paradigms . . . ,"3 but that, sinoe any suoh organization pre-supposes that we oan recognize these objeots as being relevantly the same, we have here the basis for a ori t i q u e of Wittgenstein's theory of language. This i s because the question of how t h i s knowledge of sameness i s possible i s ". • • one of the most d i f f i c u l t problems of philosophy . . ."^ and one to which Specht apparently f e e l s that Wittgenstein provides no sati s f a c t o r y so-l u t i o n , even though Specht's c i t a t i o n of two examples from p. 184 of the Remarks on the Foundations of Mathematics—the f i r s t of whioh says that we do not need a c r i t e r i o n for judging d i f f e r e n t things to be relevantly the same, that the word 'same' i s not used wrongfully just because i t i s used without j u s t i f i c a t i o n , and the second of which says that we neither need nor can we have a j u s t i f i c a t i o n for judging that a color agrees with the S p e c h t , p. 170. 2Speoht, p. 181. ^Specht, p. 182. ^Specht, p. 183. : 221 color sample—shows that he Is well aware of Wittgenstein's non-ius t i f i o a t i o n a l approach. Specht's evaluation of this approach i s that Wittgenstein wishes to make clear by means of these examples that there i s no j u s t i f i c a t i o n for the use of the word "same"; i n other words, that the question about knowledge of identity oannot be answered.1 To attempt an answer—Specht reads Wittgenstein as saying—would be to indulge i n metaphysical explanation, but of course Wittgen-s t e i n ". • • from the beginning denies his theory the p o s s i b i l i t y of giving metaphysical interpretations i n those oases where the problems oannot be solved along a n a l y t i c a l l i n e s . "2 Specht i s saying, i n other words,,that there i s a genuine problem here, which Wittgenstein cannot solve with the ©nly means at his d i s -p o s a l , sinoe he has rejected 'metaphysical explanation', so he i s forced by an ad hoo measure to appeal to the notion of a j u s t i f i o a t i o n a l terminus as a way of showing that this i s just one of those problems that cannot be solved, one of those ques-tions which admit of no answer. But t h i s i s to under-estimate seriously the force of the n o n j u s t i f l o a t i o n a l i s t approach. The f i n i t u d e of the j u s t i f i o a t i o n a l process does not mean that at some point we are relieved of the onus of solving oertain prob-lems or answering certain questions; i t i s rather that the point at whioh the demand for solutions or answers goes beyond the pro-Visions of the language-game i s also the point at whioh l e g i t i -mate problems or questions cease to e x i s t . The n o n j u s t i f l o a t i o n a l i s t theme crops up repeatedly i n ^•Specht, p. I83. 2Speoht, pp. 183-184. 222 Wittgenstein's work from PG onward, i n connection with such topics as knowing how to carry out an order, continuing a s e r i e s , following a r u l e , using words, etc., and i n his l a s t work, PC, the basic problems of epistemology are throughly explored from a n o n j u s t i f i o a t i o n a l i s t perspective.1 There can be no doubt, then, that the n o n j u s t i f i o a t i o n a l i s t theme assumes a central place i n Wittgenstein's l a t e r philosophy. I r o n i c a l l y , however, the most famous and controversial dootrine of the l a t e r Wittgenstein—viz., his denial of the p o s s i b i l i t y of a p r i v a t e language, whioh involves the concomitant and, i n my view, much more important dootrine of the u n i n t e l l i g i b i l i t y of the notion of a private something (where the same argument applies whether we regard t h i s 'something* as an objeot, an ex-perience, a process, or an event)—is a product of Wittgenstein's f a i l u r e to take his own nonjustifioationalism seriously enough. Wittgenstein's argument against the p o s s i b i l i t y of a l a n -guage being used to name or describe private objeots i s that the lack of p u b l i c access to private objeots prevents any intersub-j e c t i v e check to see i f these private objects are being correctly named ( i . e . , r e - i d e n t i f i e d ) and described; furthermore, that the notion of a private check, i n which one must r e l y on one's own memory to be assured that one i s applying language c o r r e c t l y , i s also vacuous, since there i s i n such a case no external check by whioh one can d i s t i n g u i s h correct from faulty memory; and there-1See, for example, QC #34, 77-78, 110, 130, 146, 148, 150, 166, 175, 192, 204, 206, 212, 307, 359, 429, 474, 499, 559, 563, 612, 625. 223 fore that we cannot make sense of the notion of a correct a p p l i -cation of language to private objects. "Whatever i s going to seem correot to me i s correot. And that only means that here we can't t a l k about correctness" (PI, #258). What this argument b o i l s down to i s that the application of language to private objeots cannot be j u s t i f i e d . Now, i t seems to me the:?; proper response to t h i s argument—a response i n the best Wittgensteinian tradition—would be: so what? I regard i t as per f e c t l y natural and untroublesome to suppose that when, e.g., I say I have a mental image of a farmer standing to the l e f t of a cow, I am describing a scene to which only I have d i r e c t access and to which anyone e l s e , through an examination of the state of my central nervous system or whatever, can have at best only an i n d i r e c t , i . e . , i n f e r e n t i a l , acoess; f u r t h e r , that no matter what happens to be the state of my b r a i n , I alone, through i n t r o s p e c t i o n ,1 am the ultimate judge of whether, say, the farmer i s standing to the l e f t of the cow rather than to the r i g h t of i t . On Wittgenstein's argument, i f I am not sub-jeot to some kind of external check, I cannot be sure that i n XI use the word 'introspect' with caution, since admittedly one usually does not determine what one i s imagining by inspect-ing one's image, as one might inspect a pioture to determine i t s contents. S t i l l , i f I were asked to imagine a farmer and a cow standing side by side and, aft e r I had done t h i s , I were then asked whether the farmer stood to the r i g h t or to the l e f t of the cow, there i s surely an admissible sense i n which I could be said to inspect my image to find out. Granted, my reported findings would be i n c o r r i g i b l e , but why should i n c o r r i g i b i l i t y count against saying that I had surveyed my mental image and had found out something about i t ? And, anyway, would a report concerning a comparable publlo picture be any more c o r r i g i b l e i n most circumstances? 224 my mental image the farmer r e a l l y i s to the l e f t of the cow, nor can I even be sure that what I mentally pioture i s a farmer or a cow at a l l , and t h i s r e a l i z a t i o n presumably helps to dispose of the existenoe of mental images e n t i r e l y , at least on the suppo-s i t i o n that suoh images, were they to e x i s t , would be e s s e n t i a l l y p r i vate rather than p u b l i c .1 I t seems to me, on the other hand, that we can r e a d i l y admit both (a) that mental images are p r i -vate (whioh i s not to deny that they may have public physiolog-i c a l c o r r e l a t e s , though these correlates are themselves estab-lished only on the basis of f i r s t - p e r s o n introspective testimony and may also be over-ridden by such testimony), and (b) that mental images are desoribable by public language, (a) and (b)^ are rendered compatible by my contention (c) that such descrip-tions are not amenable to j u s t i f i c a t i o n , and that none i s i n the i r oase required. common objection to the claim that Wittgenstein does not acknowledge the existence of mental images i s that he f r e -quently talks as though we do have such images. He admits, e.g., that a picture of a slab could come before a child's mind (PI, #6) or that one can describe what one imagines (PJL, #367)• But we must not be misled into supposing that Wittgenstein i s there-by conceding the existenoe of mental images. When he says: "One ought to say, not what images are or what happens when one Imagines anything, but how the word 'imagination' i s used. . . . th i s question jas to the nature of the imagination] i s not to be decided—neither for the person who does the imagining, nor for anyone else—by pointing; nor yet by a description of any pro-cess. The f i r s t question also asks for a word to be explained; but i t makes us expect a wrong kind of answer . . . " (PI, #370), I take t h i s to mean that the mistake i n asking what images are or what happens when we imagine anything l i e s i n imputing to 'image' or 'imagination' the wrong grammatical function. Witt-genstein no more denies that people have mental images than he denies that they have pains; but, i n either case, we create a •grammatical f i c t i o n ' i f we suppose that there are things suoh as images or pains whioh are designated by the words 'image' and 'pain'. 225 If I am able to play those language-games i n which my use of language i s subjeot to an inter-subjeotive check, then the fact that I am able to transfer these l i n g u i s t i c techniques to a description of my own private experience i s simply another related language-game, i n whioh the function of the j u s t l f i o a -t i o n a l demands i s correspondingly r e s t r i c t e d , or i n which such demands are even eliminated altogether. S i m i l a r l y , I use names of sensations to re f e r primarily to my own private experience, and then assume, not by a conscious, reasoned process of analogy, but by a spontaneous, i n s t i n c t i v e a t t i t u d e , that other people have the same kind of experience that I do when they make sen-sory contact with the same physical object as I do (or, gener-a l l y , with the same kind of environment)—e.g., we both b i t e into the apple and I therefore assume that you, too, have the priv a t e experience of apple-taste—or when, i n relevantly s i m i -l a r oircumstanoes, they manifest behavior si m i l a r to mine, e.g., when I have a pai n . I t i s part of our form of l i f e that we regard each other as private centers of experience, and a j u s t i - f i c a t i o n of t h i s a t t i t u d e , as i s attempted i n standard solutions to the other-minds problem, i s neither possible nor i s i t needed. This attitude towards others i s simply part of the background against which our language-games are played. On at le a s t one occasion Wittgenstein seems to be steering towards just such a n o n j u s t i f l o a t i o n a l i s t strategy for dealing with private experiences, when, i n a passage i n Z, he remarks; . . .when I say "It tastes exactly l i k e sugar", i n an important sense no remembering takes place. So I do not have  grounds for my judgment or my exclamation. If someone asks me 226 "What do you mean by 'sugar'?"--I s h a l l indeed try to shew him a lump of sugar. And i f someone asks "How do you know that sugar tastes l i k e that?" I s h a l l indeed answer him "I've eaten sugar thousands of times"—but that i s not a j u s t i f i c a t i o n that I give myself (£, #659). This suggests that Wittgenstein i s treating a taste as an exper-ience i d e n t i f i a b l e apart from the physioal s t u f f with which t h i s taste-experience i s associated, and that the i d e n t i f i c a t i o n of th i s experience as sugar-taste (along with i t s description i n p u b l i c language) oan legitimately take place without an appeal to memory—indeed, without the need of any j u s t i f i c a t i o n at a l l . His more usual approach, however, i s to suppose that l a n -guage, i n i t s f a c t - s t a t i n g capacity, always requires j u s t i f i -cation for i t s employment, and i f there are oases where no such j u s t i f i c a t i o n i s required, i n these oases language must be functioning i n some other capacity. Thus, i n PI #290 he admits that one does not i d e n t i f y one's sensations by c r i t e r i a , but he goes on to explain that this i s only because there i s no question here of i d e n t i f y i n g anything at a l l . The utterance of a sensation-word i s the beginning of the language-game, and not the outcome of the process of i d e n t i f y i n g an objeot (PI, #290). Granted, one talks about describing one's sensations, but Wittgenstein warns us that t h i s oannot be a description i n the sense i n whioh one describes one's room, or, generally, i n the sense of using language to state facts (PI, #292). So, although he does oon-oede that there are n o n j u s t i f i o a t i o n a l uses of language—"To use a word without j u s t i f i c a t i o n . • • ," as he says, ". • • does not mean to use i t without r i g h t . . . " (PI, #289)—this nonjus-t i f ioationalism i s made to complement the attack on private 227 experienoe rather than to render this attack unnecessary i n the way I maintain he ought to have done to avoid "being i n the un-tenable p o s i t i o n of having to explain away altogether the e x i s t -ence of p r i v a t e experience. As a way of tying i n my own remarks with the vast corpus of seoondary l i t e r a t u r e on the subject, I have decided to foous on two a r t i c l e s— t h e reviews of the PI by Strawson and Maloolm— whose disoussion of the key issues has remained unsurpassed, and usually even unapproached, by the flood of l i t e r a t u r e whioh has arisen sinoe these reviews made thei r appearance nearly twenty years ago. Strawson's p o s i t i o n i s e s s e n t i a l l y the one I have been advocating, the main difference between us being his contention that t h i s i s e s s e n t i a l l y Wittgenstein's p o s i t i o n as w e l l . That i s , Strawson believes that Wittgenstein wavers be-tween a stronger and a weaker t h e s i s , the stronger thesis being that no words name private experiences, and the weaker thesis being that M. • • oertain conditions must be s a t i s f i e d for the existenoe of a common language i n which sensations are ascribed to those who have them . . . ,1 , 1 where the sensations themselves are s t i l l regarded as private experiences. These conditions are that the names of sensations "• . . must always contain i n a more or le s s complex way, within t h e i r l o g i c , some a l l u s i o n to what i s not sensation, to what oan be seen and touched and heard."2 This i s because . . . a oommon language for describing or reporting requires •4?, P. Strawson, "Review of Wittgenstein's Philosophical  Investigations." i n WPI. p. 42. 2Strawson, p. 46. 228 general agreement i n Judgments. So for a (descriptive) word or phrase to belong to a oommon language, i t i s essential that the  occasions on which i t i s r i g h t to apply It should provide shared  experiences of a certain kind. the existence of which i s con-nected with the~rightness of applying the word.1 And sinoe sensations are p r i v a t e , i . e . , non-shareable exper-iences, we must look to people's observable behavior to provide the shareable experiences which constitute our c r i t e r i a for ascribing sensations to them. Because of t h i s need for behav-i o r a l c r i t e r i a for a oommon sensation-language, Strawson con-cludes that . . . i t i s necessarily empty and pointless (I w i l l not say meaningless) either (a) to speculate about the a s c r i p t i o n of pain to anything which does not exhibit behaviour compatible i n the relevant respects with human behaviour ( i . e . the behaviour of those who use the concept), or (b) to r a i s e generalized doubts about other people's experienoe of p a i n , or about one's own know-ledge of t h i s .2 He then adds: "It i s the above points whioh I take Wittgenstein e s s e n t i a l l y to be making."3 Now, while I disagree with Strawson's assumption that i n the PI Wittgenstein i s for the most part ad-vocating the weaker thesis—he i s too deeply oommitted to the stronger thesis for t h i s— I agree with Strawson that the weaker thesis i s the only defensible one, and the one Wittgenstein ought to have advooated. Although Strawson emphasizes the need for a J u s t i f i c a t i o n for our use of words i n a common language, he i s ultimately taking the same n o n j u s t i f i c a t i o n a l approach that I advocate, sinoe he c l e a r l y does not regard behavioral c r i t e r i a as l o g i c -a l l y guaranteeing, by d e f i n i t i o n , the existenoe of the. relevant sensations. If he were asked on what basis we suppose that •'•Strawson, p. 62. 2Strawson, p. 49. ?Strawson, p. 49. 229 another's pain-behavior shows the existenoe of a certain kind of p r i v a t e experience, rather than another kind, or even none at a l l , he could only reply that we have no j u s t i f i c a t i o n , but that suoh a skeptioal doubt, though not s t r i c t l y meaningless, would be empty and p o i n t l e s s . That i s , for Strawson i t i s not the assumption of the existence of another's private experience that i s empty, but rather the doubt of i t s existenoe. Furthermore, to the extent to which the a s c r i p t i o n of sensations does require j u s t i f i c a t i o n , the shareable experiences which are supposed to provide t h i s j u s t i f i c a t i o n cannot themselves be j u s t i f i e d i n t h e i r olaim to be shareable. Thus, Strawson says that we have a oommon impersonal language for describing colors because we can reach general agreement.in our oolor-judgments, but i f he were asked what j u s t i f i c a t i o n he has for supposing that we ever do reach suoh an agreement, e.g., that when we a l l say that some-thing i s red we do not a l l have d i f f e r e n t color-experiences whioh we simply c a l l by the same name, he could only dismiss this as an empty metaphysioal speculation—the p o s s i b i l i t y of j u s t i f i -cation here being out of the question—and say that our form of l i f e makes any suoh j u s t i f i c a t i o n a l attempt Irrelevant. Straw-son must admit, then, that the a s c r i p t i o n to others of shareable and non-shareable experiences a l i k e must ultimately be carried on without a j u s t i f i c a t i o n , and since he also e x p l i c i t l y says that one can recognize and Identify one's own experiences, both shareable and non-shareable, without the use of any c r i t e r i a ,1 we can readily see that the nonjustif1oational approach i s central •'•Strawson, p. 46 and 63. 230 to Strawson1B strategy. I turn now to Malcolm's review, which i s mainly concerned to elucidate Wittgenstein's aooount of how language i s related to inner experienoe, and to rebut Strawson's c r i t i c i s m of Witt-genstein. Malcolm takes Wittgenstein's chief insight to be that the connection between a sensation and i t s outward expression i s esse n t i a l rather than contingent. The presence of behavioral c r i t e r i a f or pain guarantees, by d e f i n i t i o n , the presence of pain; these c r i t e r i a constitute the pain (though Malcolm i s put i n the uncomfortable p o s i t i o n of having to admit that even for Wittgenstein these o r i t e r i a oan only be a l o g i o a l guarantee i n oertain circumstances). I f behavioral o r i t e r i a are regarded as anything short of l o g i c a l l y guaranteeing that for whioh they are o r i t e r i a , then t h e i r status as o r i t e r i a , even i n Strawson's weaker sense of being indications or signs of the inner exper-ience, comes into question. I f inner experiences are only con-tingently connected to th e i r outward expression, then one Is immediately faced) with the problem of showing how anything can be known about these inner experiences other than i n one's own case. Malcolm argues that the defender of the contingent-connection thesis i s foroed to admit that the pub l i c i n a c c e s s i -b i l i t y of inner experiences means that such experiences oan only be the subject-matter of one's private language, and then r a i s e s the standard d i f f i c u l t i e s about o r i t e r i a of correctness, etc., to show that even a private language for talking to oneself about one's inner experiences i s impossible. We are thus l e f t with only the essential-connection thesis—which denies that 231 so-called 'inner experiences' are r e a l l y private experiences or objects at a l l , separable from th e i r behavioral manifestation— as a viable p o s i t i o n . Without examining Malcolm's arguments i n d e t a i l , I want to show only how the seeming impregnability of his p o s i t i o n depends upon j u s t i f i o a t i o n a l i s t assumptions whioh the nonjustif1cational approaoh, i m p l i c i t l y advocated by Strawson, sets out to challenge. Strawson's view, we w i l l r e c a l l , i s that the p o s s i b i l i t y of nam-ing and describing non-shareable experiences i n our oommon l a n -guage requires that there be shareable experiences whioh are associated with, and thereby indicate the presenoe of, these non-shareable experiences. Were there no suoh association with shareable experiences, our non-shareable experiences could be the subject-matter only of each individual's private language, but as i t i s , they are the subject-matter, not of private l a n -guage, but of our oommon language. Malcolm, on the other hand, characterizes Strawson's view as ". . . that each of us not only can have but does have a private language of sensations • • ." and that this p r i v a t e language serves as the "• . • understruoture of the language we a l l understand."2 This must be Strawson's view, Malcolm supposes, because the claim that sensations are non-shareable experiences implies that there i s a feature of the use of sensation-words that cannot be taught, v i z . , t h e i r r e f e r -ft enoe. Only the possessor of the inner experience oould be ao-^orman Malcolm, "Wittgenstein's Philosophical I n v e s t i - gations ." i n WPI. p. 97. 2Maloolm, p. 98. 232 quainted with i t ; everyone else would have to be content to be acquainted with the shareable.experiences with which th i s non-shareable experience i s associated. So, on this view (whioh Malcolm imputes to Strawson), ". • . sensation-words w i l l have both a public and private meaning."1 Malcolm then adds: But i f my words, without these a l l u s i o n s , oan r e f e r to my sensations, then what i s alluded to i s only contingently related to the sensations. Adding the "allusions to what can be seen and touched" w i l l not help one l i t t l e b i t i n making us understand one another. For the behavior that i s , for me, contingently as-sociated with 'the sensation of pain' may be, for you, contin-gently associated with 'the sensation of t i c k l i n g ' ; the piece of matter that produces i n you what you c a l l 'a m e t a l l i c taste* may produce i n me what, i f you could experience i t , you would c a l l *the taste of onions*; my 'sensation of red' may be your 'sensation of blue'; we do not know and cannot know whether we are ta l k i n g about the same things; we cannot learn the essential thing about one another's use of sensation-words—namely, the i r reference.2 The phrase whioh Malcolm considers to be his knock-out punch, v i z . , 'we do not know and oannot know whether we are talk i n g about the same things *, i s a prime example of what Strawson c a l l s 'empty and pointless metaphysical speculation*. Malcolm's claim that there i s an epistemologioal gap here that demands to be f i l l e d by d e f i n i t i o n a l o r i t e r i a , whioh guarantee that the reference of a sensation-word i s a matter of public knowledge, oan be rejeoted with the observation that our form of l i f e makes no such demand. Granted that pains are non-shareable experiences, s t i l l , we a l l know what pains are, and the referenoe of the word 'pain* i s therefore p u b l i c , not p r i -vate. Malcolm's insistence that Strawson must believe that *pain' has both a pu b l i c and a private meaning stems from Malcolm's own assumption that the privacy of my inner experience Malcolm, p. 98. ^Malcolm, p. 98. 233 implies that no one else can know what I have. But what kind of knowledge does Malcolm suppose other people are deprived of here? Malcolm seems to want something concerning other people's experiences which would be as good i n every respect as anything they could have themselves, but i t i s p r e c i s e l y "because our status as i n d i v i d u a l centers of private experienoe makes suoh a wish u n f u l f i l l a b l e that our form of l i f e takes no notice of such a wish and therefore does not give r i s e to language-games i n which I oan be said to 'know* something about my own sen-sations that no one else can possibly know, even i f I t e l l him whatever I possibly can about my experienoe. Once we abandon t h i s use of 'know'—whioh, from the perspective of the language-games we actually play (in our non-philosophical moments) i s i l l e g i t i m a t e and i r r e l e v a n t—the d i s t i n c t i o n between a sensation-word's public and p r i v a t e meaning also loses i t s p o i n t . Malcolm's j u s t i f l o a t i o n a l i s t assumptions again come to the fore i n his discussion of how Strawson has f a i l e d to understand Wittgenstein on the topic of recognizing and i d e n t i f y i n g sen-sa t i o n s . According to Strawson, Wittgenstein correctly notes, f i r s t , that the expression of doubt has no place i n the language-game i n which one says 'I am i n pain*, and secondly, that one does not, when one says *I am i n pain*, Identify one's sensation by o r i t e r i a , but that Wittgenstein goes wrong i n supposing . . . that these facts can only be accommodated i f we regard 'I am i n p a i n ' as an expression or manifestation of pain . . . So regarded, 'pain' ceases to appear as the name or the description of a sensation. If we do not so regard i t , then we s h a l l require o r i t e r i a of i d e n t i t y for the sensation; and with these there would enter the p o s s i b i l i t y of error . . . What he has committed himself to i s the view that one oannot sensibly be said to reoog-234 nise or id e n t i f y anything, unless one uses o r i t e r i a ; and, as a consequence of t h i s , that one cannot recognise or id e n t i f y sensations.1 Strawson points out that we use no o r i t e r i a to ide n t i f y tastes or oolors, and that we also discriminate and recognize aohes and throbs, searing and jabbing pains, to whioh there do not obviously correspond d i f f e r e n t c h a r a c t e r i s t i c natural expressions of pai n . Malcolm's way of getting around these d i f f i c u l t i e s raised by Strawson, without r e a l l y facing them head-on, i s to argue that we cannot i n any straightforward sense be said to recognize or i d e n t i f y pains, i n the way we can i d e n t i f y , say, r a b b i t s , be-cause only i n the l a t t e r case i s there the p o s s i b i l i t y of making a mistake; and only where suoh a p o s s i b i l i t y exists can we speak of a correot i d e n t i f i c a t i o n at a l l . So, we oan, i f we l i k e , s t i l l speak of i d e n t i f y i n g sensations, as long as owe recognize that such ' i d e n t i f i c a t i o n s ' are i n c o r r i g i b l e . The fact that i t i s out of the question that one should make a mistake i n iden-t i f y i n g one's sensations shows that sensations and other inner experiences are not private objects, because i f they were such objeots, which I id e n t i f y and from which I derive t h e i r des-c r i p t i o n , ". . . why oannot my i d e n t i f i c a t i o n and description go wrong, and not just sometimes but always?"3 For a l l we know, our memory i s constantly deceiving us. Expressing Malcolm's conclusion i n terms of my own e a r l i e r example, the fact that i t does not make sense to ask me what my grounds are for supposing that, i n my mental image, the farmer i s standing to the l e f t of the cow i s supposed to show that my mental image oannot be a •'"Strawson, p. >5. 2Strawson, p. 46. ^Malcolm, p. 101. 235 p r i v a t e object. What i s i t , then, a p u b l i c objeot accessible i n p r i n c i p l e to anyone? Or i s i t just nothing at a l l , the words 'I now have a mental image i n whioh a farmer i s standing to the l e f t of a cow' being a non-indicative use of language on a par, i n this respect, with a groan or a ory? C l e a r l y , i t seems to me, an a l t e r n a t i v e preferable to either of these i s that mental images are indeed private objeots or experiences, and that the i n c o r r i g i b i l i t y of reports and descriptions of sensations, images, etc. , rather than showing that these things are not private ob-j e c t s , simply shows that the request for grounds concerning reports and descriptions of objeots of this status i s i r r e l e v a n t , and that Malcolm's challenge—viz., 'why oannot my i d e n t i f i c a t i o n and description go wrong'—is nothing less than a challenge to our form of l i f e . This p a r t i c u l a r 'why?' simply cannot, and need not, be answered. It would be giving the wrong impression of Malcolm to say that he i s t o t a l l y oblivious to the notion of an ungrounded form of l i f e . On the contrary, he i s well aware of this notion, but introduces i t into the discussion so as to lend support to the attack on private objects. Thus: As philosophers we must not attempt to j u s t i f y the forms of l i f e , to give reasons for them—to argue, for example, that we p i t y the injured man because we b e l i e v e , assume, presuppose, or know that i n addition to the groans and writhing, there i s p a i n . The faot i s , we p i t y himJl Unfortunately, Wittgenstein would probably have approved of the way Malcolm