Wittgenstein Mind and Language

WITTGENSTEIN: MIND AND LANGUAGE SYNTHESE LIBRARY STUDIES IN EPISTEMOLOGY, LOGIC, METHODOLOGY, AND PHILOSOPHY OF SCIENCE Managing Editor: JAAKKO HINTIKKA, Boston University Editors: DIRK VANDALEN, University of Utrecht, The Netherlands DONALD DAVIDSON, University ofCalifornia, Berkeley THEO A.F. KUIPERS, University ofGroningen, The Netherlands PATRICK SUPPES, Stanford University, California JAN WOLENSKI, Jagiellonian University, Krak6w, Poland VOLUME245 Ludwig Wittgenstein, Bleistijtzeichnung von Drobil, © Frau Katarina Eisenburger, reproduced here by kind permission of Frau Katarina Eisenburger. WITTGENSTEIN: MIND AND LANGUAGE Edited by ROSARlA EGIDI Department 0/ Philosophy Third University 0/ Rome, Italy SPRINGER-SCIENCE+BUSINESS MEDIA, B.V. Library of Congress Cataloging-in-Publication Data mind and language / edited ＠ Rosaria Egld'. -- <Synthese ＱＧ＠ ; v. 245) "Revised vers'on ｛Ｌ｝＠ the papers Ｌ＠ the W'ttgenstein conference held in November 1989 Ａ＠ the Un'verslty Ｌ＠ ＠ La Saplenza"--Pre'. Includes indexes. ISBN 978-90-481-4475-4 ISBN 978-94-017-3691-6 (eBook) DOI 10.1007/978-94-017-3691-6 1. Wlttgensteln. Ludwlg. 1889-1951--Contrlbutlons Ｑ＠ ph,losophy Ｌ＠ mind--Congresses. 2. W'ttgenste'n. Ludwig. 1889-1951--Contrlbutl0ns in ph'losophy Ｌ＠ language--Congresses. 3. Ph,losophy Ｌ＠ m'nd-Congresses. 4. Language and languages--Ph,losophy--Congresses. I. Egld'. Rosar'a. II. Ser'es. B3376.W564W552 1995 192--dc20 94-37362 Wittgenstein Ｎ＠ Ｎ＠ ISBN 978-90-481-4475-4 ＠ Rights Reserved. © 1995 Springer Science+Business Media Dordrecht Academic Publishers in 1995 Originally published ＠ ｬｵｷ･ｲ＠ No part of the material protected ＠ this copyright notice ＠ ＠ reproduced or uti1ized in any form or ＠ any means, electronic or mechanical, including photocopying, recording or ＠ any information storage and retrieval system, without permission from the copyright owner. In memory ofNorman Malcolm ( 1911-1990) TABLE OF CONTENTS Preface xi GEORG HENRIK VON WRIGHT / Wittgenstein and the Twentieth Century 1 PART 1. SCIENCE, MIND, AND METAPHYSICS DAVID PEARS / Wittgenstein on Philosophy and Science ANTHONY I.P. KENNY / Wittgenstein on Mind and Metaphysics JOHANNES L. BRANDL / Wittgenstein's Alleged Metaphysics of 23 37 47 Mind PART H. IN THE DOMAIN OF POSSIBILITIES: RULES, FORMAL AND INNER LANGUAGES IACQUES BOUVERESSE / Le reel et son ombre: la theorie wittgensteinienne de la possibilite CARLO CELLUCCI / Wittgenstein on the Meaning of Logical Symbols AVRUM STROLL / On Following a Rule DIE GO MARCONI / Fodor and Wittgenstein on Private Language MASSIMO DELL'UTRI / Mentalesians and Wittgenstein's Private Language 59 83 93 107 117 PART III. INTENTIONALITY AND THE GRAMMAR OF MENTAL LIFE ALDO G. GARGANI / Intentionality in Wittgenstein's Works HERBERT HOCHBERG / Intentional Reference as a Logical Relation: A Variation on a Theme in Moore, Russell, Wittgenstein, and Bergmann ROSARIA EGIDI / Wittgenstein Between Philosophical Grammar and Psychology ROBERTO CASATI / Notes on Phenomenology and Visual Space ix 127 137 171 185 CONTENTS x PART IV. BELIEF, DOUBT, AND RELATIVISM NORMAN MALCOLM / Disentangling Moore's Paradox EVA PICARDI / Maleolm on Moore's Paradox WOLFGANG CARL / Wittgenstein's Refutation on Scepticism in 'On Certainty' RUDOLF HALLER / PART V. 195 207 211 223 Was Wittgenstein a Relativist? TOPICS IN THE PHILOSOPHY OF PSYCHOLOGY S. STEPHEN HILMY / Wittgenstein on Language, Mind and Myth- ology JOACHIM SCHULTE / 235 Emotion: Remarks on Wittgenstein and William James GUIDO FRONGIA / Wittgenstein and Memory FRITZ G. WALLNER / How Ludwig Wittgenstein Reacted to Recent Changes in Psychology 249 263 Would Have 279 INDEX OF NAMES 289 INDEX OF SUBJECTS 293 PREFACE The essays collected in this volume represent, in a revised version, the papers of the Wittgenstein Conference held in November 1989 at the University ofRome 'La Sapienza' to celebrate the centenary ofhis birth. They offer a systematic account ofWittgenstein's philosophy ofmind and contribute to illuminate his later conception of perceptive, emotional and cognitive language. Some of the reasons why it seemed the right time to promote an ampIe confrontation ofideas on Wittgenstein's mature perspective are sufficiently c1ear as they derive from the need to sum up the state of research based on the availability of the Nachlaß and the publication in the last decade of a conspicuous quantity ofwritings dedicated to philosophical psychology; other reasons are more complex as they depend on the already noticed tendency in the recent epistemological debate to interpret Wittgenstein's provocative and controversial theses in a "perverse" way, in a way which has been used as a banner for epistemic relativism, subjectivism, and irrationalism. The intention of this collection of essays is to construct an image of Wittgenstein's thought, which is as faithful as possible to his philosophy of mind and language from both a theoretical and exegetical point of view. The book also strives to assess the continuity and internal coherence of the theses developed throughout the different phases of his research. Attention is focused on the fundamental topics of philosophical grammar investigated by Wittgenstein starting from the Thirties: they are intentionality, analysis of belief and Moore's paradox, certainty and doubt, will, memory, sensations and emotions, the theory of aspects and of private language in its relationship with scepticism and psychologism. All these topics are part of a great plan for the treatment of non-phenomenologicallanguage to which, according to Wittgenstein's late perspective, belong both psychological, as weIl as mathematical concepts and whose criteria of meaning are not of an empirical nature but have the "generality ofa supreme rule ofthe game admitting ofno appeal". The purpose ofthe division into thematic areas is to provide, without claiming them to be exhaustive, some important points of reference and, in a way, to create a framework for Wittgenstein's grammar of inner experience with the aim to exhibit the connections between his philosophy ofmind and general xi R. Egidi (ed.), Wittgenstein: Mind and Language, xi-xii. © 1995 Kluwer Academic Publishers. xii PREFACE views on metaphysics, logic, and psychology. In comparison to the recent specialized literature dedicated to the themes of philosophie al psychology, some of the papers presented in this volume dedicate special attention to the historical and critical context of Wittgenstein's theses on the subject and to the examination of their roots in the European thought of his time. I would like at this point to express my gratitude to the contributors to this volume, to the Universita di Roma 'La Sapienza', to the Consiglio Nazionale delle Ricerche and to the Istituto Italiano per gli Studi Filosofici for their substantial support at the Conference. I also wish to thank my colleagues and collaborators at the Dipartimento di Filosofia e Teoria delle Scienze Umane. Last but not least, to Brian McGuinness and Rudolf Haller go my special thanks for their help and encouraging cooperation. Rome, September 1992 ROSARlA EGID! GEORG HENRIK VON WRIGHT WITTGENSTEIN AND THE TWENTIETH CENTURY I have called my paper "Wittgenstein and the Twentieth Century". It is not my intention, however, to discuss how Wittgenstein has impressed and changed thinking in our century; nor shall I discuss the influences and impulses which he may have received from other philosophers. What Iwanted to do is something which is more hazardous and indefinite - and in the opinion of many perhaps not very important either. Iwanted to relate Wittgenstein to a prevailing climate of opinion or cultural situation, to something which mayaIso be called the "moods" or Stimmung of a time. I have made an effort in this direction once before.' Here 1 wanted to follow it up with thoughts on how Wittgenstein' s rejection of the civilization of contemporary Western society reflects a basic attitude of his to life and on how this attitude carries both his earlier and his later work in philosophy. I. A dominant feature in the spiritual physiognomy of the twentieth century is Modernity. It has become recognized under that name largely in retrospect and in contrast to tendencies whieh either are critieal of it or champion a new, "post-modern" mood of the time. Modernity, thus conceived, is our legacy of the Enlightenment and the French Revolution. It is the Age of Reason matured to become an age of science and technology, of an industrial mode of production, and of democratic forms of government. In origin it was an optimistic mood. It cherished avision of linear and unlimited perfection and progress towards a regnum hominis of free and equal men. The yoke of superstitious beliefs being lifted, also that of despotie government would never again be allowed to oppress man. This original mood was partly reflected in but partly also reinforced by the ideas of evolution which were characteristie ofl9th century scientific thought. They range from the historicallinguistics of aRasmus Rask and J acob Grimm, through Charles Lyell's new biography ofthe earth, to Darwin's theory ofthe origin of species and descent of man. These scientific achievements encouraged the beliefs and sentiments, epitomized in the positivist philosophy of Auguste Comte and his followers, the evolutionist progressivism of Herbert Spencer, and also, though with more ambiguous and sophisticated overtones, in Hegel's phenomenology ofthe spirit. R. Egidi (ed.), Wittgenstein: Mind and Language, 1-19. © 1995 Kluwer Academic Publishers. 2 GEORG HENRIK VON WRIGHT This climate of opinion, here briefly characterized, prevailed not least in late Victorian and Edwardian England. It enjoyed an exceptionally intense "Indian Summer" of creative intellectual talent in pre-first-war Cambridge where the young Wittgenstein came to study logic with Bertrand Russell. The First World War gave a shock to this mood but it by no means crushed it. The war could also be viewed as a great convulsion needed for breaking the fetters of unreason in which reactionary forces of the past had tried to hold back humanity on its way to Modernity. To many the revolution in Russia seemed the continuation and final breakthrough of the spiritual forces first let loose in France more than a century earlier. The so-called modernistic movements in architecture, art and literature testifY to this renewed optimism of the post-first-war period. And so do the efforts to create a global international organization to secure peace and progress for a world made - as said a slogan ofthe time - "safe for democracy". This rejuvenated Modernity had one of its most consequential and serene reflections in the philosophic trend known as logical positivism. I say "consequential", thinking ofthe repercussions which this trend has had on philosophy throughout the century. And I say "serene", thinking of the rationalist ethos and consciousness of a message which animated many of its early protagonists. This feeling of unisonity with a tune of the times is perhaps nowhere reflected as movingly as in Carnap's Foreword to Der logische Aufbau der Welt. I quote from it here in English: We feel that there is an inner kinship between the attitude on which our philosophical work is founded and the intellectual attitude which presently manifests itselfin entirely different walks oflife; we feel this orientation in artistic movements, especia1ly in architecture, and in movements which strive for meaningful forms ofpersonal and collective life, of education, and of external organization in general. We feel all around us the same basic orientation, the same style of thinking and doing. - Our work is carried by the faith that this attitude will win the future. 2 I vividly remember how these words moved me as a young student in Helsinki in the mid-1930s - and I think we were fortunate in my country to have received OUT inspiration in philosophy from a charismatic teacher, Eino Kaila, who saw hirnself in the vanguard of a radically new way of thinking. The optimistic mood apart, what was new about this thinking were above all two things. One was the strong emphasis on logic as the core and center of philosophy. The other was the alignment with science, the feeling that philosophy too had, at long last, attained "den sicheren Gang der Wissenschaften".3 "Wissenschaft" should then be understood in the traditional English sense of the word, comprising mathematics and the natural sciences, rather than in the broader German sense. WITTGENSTEIN AND THE TWENTIETH CENTURY 3 It is worth noting here that the idea of the natural sciences setting the pattern also for philosophie thinking has deeper roots in Austrian than in German or even in French or English philosophy. Schlick was successor to the chair once held by Mach in Vienna, and the Wiener Kreis could be called an inner circ1e of the Ernst Mach Society which introduced the Kreis to the world in its manifesto Wissenschaftliche Weltauffassung. But al ready long before that, Brentano had proc1aimed in his Habilitationsschrift of 1866 that "vera philosophiae methodus nulla alia ni si scientiae naturalis est".4 The key position of Brentano and his several pupils in the development of philosophy in our century has with time become increasingly obvious when the ties ofthe "new philosophy" to the phenomenalistic sensualism ofMach have gradually loosened, and the current originating as logical positivism has broadened into the mighty stream of analytic philosophy. In spite of the many tributaries which have, in the course of the years, emptied their waters into this river, I think it is right and illuminating to call analytic philosophy the mainstream of philosophie thinking in this century. In all its heterogeneity it retains the two features which I al ready mentioned as typical of its origin: the emphasis on logic and the alignment with science. It is, in short, the philosophy most characteristic of a culture dominated by scientific rationality. As is well attested, the influence ofWittgenstein's Tractatus on the Vienna Circ1e was profound. The authors - Carnap, Hahn, and Neurath - of the "manifesto" hailed Wittgenstein, together with Einstein and Russell, as the leading representative of a "scientific view of the world" or, with its more telling German name, "eine wissenschaftliche Weltauffassung".5 That this characterization is profoundly untrue of Wittgenstein's Denkart is obvious to every serious student ofhis philosophy. Yet it is even today quite common to label Wittgenstein, if not a "logical positivist" so at least an "analytic philosopher". There are deeper reasons for this misunderstanding than just the difficulties of getting rid of a label once attached. If Wittgenstein is not an analytic philosopher, what kind of philosopher is he then? This question certainly cannot be answered in the terms of current c1assifications. He is not a phenomenologist or hermeneuticist, nor an existentialist or hegelian, least of all is he a marxist. It is, moreover, not too difficult to reinterpret Wittgenstein's thinking so as to fit the analytic mould. He said in the Tractatus that "Philosophy is not one of the natural sciences" (4.111). But nor would the logical positivists have c1aimed that to be the case. He also said that philosophy is "something which stands above or below, but not beside the natural sciences" (ibid.). How does 4 GEORG HENRIK VON WRIGHT this differ from Carnap's view that philosophy is the logical syntax of the language of science?6 Is not the logic of science exactly something which stands "above" or "below" but not "beside" the sciences themselves? And Wittgenstein said many more things about what philosophy is and what it is not which can without too much ado be reconciled with the opinions both of earlier and later so-called analytical philosophers. Yet there are, 1 think, profound differences. 2. There bad existed throughout the 19th century undercurrents which did not share the optimistic belief in progress through modernity and did not vest hopes for the future of man in the further development of science and technology. Kierkegaard's opposition to Hegel and system-philosophy, particularly in his Concluding Non-Scientific Postscript, struck an early note of discord. Dostoyevsky' s Notes from the Underground dug deep into those dark layers ofthe human soul which might one day erupt in violent protest against a rationally organized, progressive society. The climax of 19th century critique of civilization is, of course, the "Umwertung aller Werte" attempted by Nietzsche. Of Nietzsche Wittgenstein once said that he had perhaps touched on "problems of the intellectual world of the West" which no other philosopher had "tackled and wrestled with" and which could only be written about "in the language of prophesy, comprehensible to the fewest". 7 The mood ofthese writers is not necessarily pessimistic. But it is a sombre mood of se1f-reflexion and questioning of dominant currents of their time, as they saw them. And these writers, we know, were more congenial to Wittgenstein than any 19th century philosopher of the established style. From his early years he distanced hirnself from and condemned modernity in all its philistine manifestations. It is but natural that the catac1ysm of the first War should have nourished this mood and added to it apocalyptic overtones just - as it is also understandable that the same disasters were hailed by others as having created a tabula rasa for the groundwork of a brave new world. The doomsday prophet par excellence is, of course, Spengler. Wittgenstein has mentioned Spengler as one of those who had influenced hirn. He says he took over from Spengler a line of thinking and seized on it with enthusiasm for his own work. 8 This, and the fact that Spengler's name occurs on the list - beginning with Boltzmann and ending with Sraffa - of persons, whom Wittgenstein recognized as influences, does not mean, however, that Spengler had deeply influenced the mood in which Wittgenstein viewed his times. It means, in the first place, that he had received from WITTGENSTEIN AND THE TWENTIETH CENTURY 5 Spengler's Untergangthe germ of one ofthe pervasive ideas ofhis later philosophical thinking. This is the notion of conceptual family resemblances. It is quite another thing that Wittgenstein also shared the apocalyptic view of Spengler's. With the years, this even deepened to a hatred of our decaying civilizationand a wish for its destruction. "Do you really think that Europe needs another great war?", I once asked hirn in the twilight of prewar sentiments in 1939. "Not one, but two or three", was his reply. This shook his young interlocutor deeply and seemed to hirn then unintelligible. The apocalyptic views of Wittgenstein are best reflected in some of the "general remarks" he wrote in the late 1940s. But the best and fullest articulation of his attitude to Modernity is of much earlier date. 1 am thinking, in particular, ofthe sketches ofa Foreword written in 1930 for a projected work. A version was printed with the Philosophische Bemerkungen, the earliest of Wittgenstein's prepared typescripts after his return to philosophy in the late 20s, published in 1964. The book for which it was written cannot, however, have been it. More likely, the Foreword was meant for the great and externally more accomplished work to which Wittgenstein's literary executors used to refer as the "Big Typescript" and which - most regrettably - to this day has not been published in its entirety. In one of the sketches he characterizes the spirit in which his work was written as follows: This is not ... the spirit ofthe main current ofEuropean and American civilization. The spirit of this civilization makes itself manifest in the industry, architecture and music of our time, in its fascism and socialism, and it is alien and uncongenial to the author ... Our civilization is characterized by the word 'progress'. Progress is its form ... It is occupied with building an ever more complicated structure ... I am not interested in constructing a building, so much as in having a perspicuous view ofthe foundation ofpossible buildings ... So I am not aiming at the same target as the scientists and my way ofthinking is different from theirs. 9 Wittgenstein's early sketches for aPreface should be read in juxtaposition to Carnap's Preface to Aufbau. They afford an impressive and nice illustration to the contrast between the protagonists and the critics of the spirit of Modernity. Although we have no documentary evidence for it, the thought is elose at hand that Wittgenstein wrote his words in reply to Carnap's. The way the two prefaces match seems to me too good to be a result of sheer coincidence. 3. The Hungarian philosopher and historian of ideas Christoph Nyiri sees in Wittgenstein a representative of the neo-conservative trend which rose to prominence in the 1920s, chiefly in the German-speaking world. 1O It continues an earlier conservatism which had taken a critical view of the rationalist 6 GEORG HENRIK VON WRIGHT spirit of the Enlightenment and of the progressive democratization and secularization of 19th century society. An early exponent of this type of conservative thinking is the great Austrian poet Grillparzer. Nyiri points to the deep impression Grillparzer had made on Wittgenstein and which anybody knowledgeable in literature cannot fail to notice throughout his writings also in places where there is no mention of Grillparzer's name. I have great respect for Nyiri's attempt to locate Wittgenstein on the cultural map. 1 think it in many ways illuminating of Wittgenstein' s attitudes and thoughts. But 1 also have reservations. The label "conservatism" does not seem to me appropriate - pace Nyiri's own observations on its ambiguitiesY Wittgenstein was much more anxious to combat and distance hirnself from a prevailing climate of opinion than to work for the restoration of one which was already fading. He is as little nostalgie in his thinking as are Dostoyevsky or Nietzsche. Moreover, the philosopher who wrote "I destroy, I destroy, 1 destroy"12 was not alien to the thought that something new could be built once the heap of rubble of a decaying culture had been cleared away. Not unlike some radicals of the left he appears to have seen something hopeful in the drastic sweeping away of an obsolete social order which had taken place in Russia. His plans of settling in the Soviet Union can be viewed in this light too. Among his Cambridge friends in the 1930s many had a pronouncedly marxist orientation. The only periodical which 1 have seen hirn reading and not frowning upon was The New Statesman and Nation - much more in tune with the tastes of left intellectuals than with those of apolitical conservatives. Just as Wittgenstein's philosophy defies classification in relation to movements and trends in 20th century thought, so also is his attitude to modemism and its critics far from univocal. It may even appear to us loaded with paradox and contradiction. 4. Allan Janik, co-author ofthe book Wittgenstein's Vienna and another writer who has tried to place Wittgenstein in a broader cultural setting, has criticized Nyiri's diagnosis. He objects above all to the way in which Nyiri exploits for his purposes two key notions in Wittgenstein' s later philosophy. These are the notions of rule-following and form of life. Nyiri tri es to link them to political and social conservatism. This does not seem to me right. Janik's critical points against Nyiri I find weIl taken here. Janik suggests that it is characteristic of Wittgenstein that "he wished to separate his philosophizing from his personal beliefs" and that, accordingly, he "took great pains to separate his cultural pessimism from, say, his view of WITTGENSTEIN AND THE TWENTIETH CENTURY 7 meaning and thinking."13 If this is right, it would be futile to look for a connection between Wittgenstein's philosophy and the mood in which he viewed his times. One could go one step further and say that also the modernism and rationalist optimism, characteristic of many of the leading logical positivists, bears only a contingent relationship to their philosophy. It is true that neither Wittgenstein nor the positivists (with the sole exception of Neurath) elaborated their views of culture, history, politics, and society in the discursive form of philosophical or scientific writings. Their "Weltanschauung" shines forth only in various obiter dicta; in the case of the positivists the Vienna Circle manifesto or Carnap's Preface to Aufbau; in the case of Wittgenstein some "aphorisms" in the Tractatus, the prefaces to works not published in his lifetime, and a good many "general remarks" scattered throughout his writings. (Hither mayaiso be counted the notes on Frazer's The Golden Bough). The truth of these observations notwithstanding, I disagree with Janik's suggestions above. Fichte's famous words "Was für eine Philosophie man wählt, hängt davon ab, was für ein Mensch man ist", may not be interestingly applicable to the average, mediocre, academic philosopher. But for the great ones it is, I think, profoundly true. Their philosophy reflects their personality, and vice versa. And if personalities differ profoundly, so will the philosophies. Therefore it is not futile to look for the way in which Wittgenstein's thought can be said to reflect his view of life. 5. A different attempt from Janik's to resolve the puzzles here is made in Stephen Hilmy's book The Later Wittgenstein. It is another one of those not too many writings which discuss Wittgenstein's thought in relation to intellectual currents of the time. (A virtue of Hilmy's work is that it bases its argumentation directly on Wittgenstein's Nachlaß and does not confine it to the printed sources only.) Hilmy's view is that Wittgenstein's later philosophical deliberations should be seen "as a struggle against dominant intellectual trends in our modern civilization"14 and also "against this trend as it manifested itself in his own earlier philosophy". 15 Thus, according to Hilmy, Wittgenstein of the Tractatus was himself"caught up in the scientific current ofthe times"16 and had "succumbed to the scientific intellectual trends of the day".n Simplifying a little, one could then say that the logical positivists were right when they saw in the Tractatus an ally to their own philosophy, and that it is only the later thinking ofWittgenstein which reflects his Abstandnahme from that intellectual mood ofthe times to which I have here alluded with the term "modernity". 8 GEORG HENRIK VON WRIGHT There is certainly some truth in Hilmy's view. At least two arguments can be given in support of it. One is so to speak "biographieal". It points to changes in Wittgenstein's life. The Wittgenstein ofpre-war Cambridge, a student of Frege's and Russell's logic, was in many ways a different man from the Wittgenstein who returned to Cambridge in the late 1920s. His experiences from the years of the war had deeply affected his sentiments. After return from captivity he renounced his fortune and withdrew to a life of Tolstoyan simplicity in remote countryside villages. His life later at Cambridge and in his hut in Norway is stamped by the same austere frugality. The professor of philosophy refused to let himselfbe integrated in the academic establishment of one ofthe world's most distinguished universities. This was Wittgenstein's mood when he met the philosophers of the Vienna Cirele, - but it was a different mood from the one in which he had first conceived the ideas for the work which meant so much to them. The Wittgenstein who had radically changed his ways of life was now also on the road to a new philosophy which was to become less and less congenial to most of his former admirers. The other argument in support of Hilmy's view trades on the difference between the philosophy of the Tractatus and that of the Philosophische Untersuchungen, between the philosophies sometimes referred to simply with labels "Wittgenstein I" and "Wittgenstein 11". The first belongs in the atmosphere of logic and the science-oriented thinking of the positivists and later the analytic philosophers. The second is deeply infected by doubts about the influence of scientific rationalism on our thought and our lives. What 1 should like to do now is the following: Against Janik I shall try to show that there is a elose correspondence between Wittgenstein's thought and that which I have called his "mood". To this end we must look towards features of his personality which lay deeper than those habitually referred to as his "cultural pessimism" and which remained substantially ,unaffected by external changes in his way of life. Against Hilmy 1 shall try to show that there is a continuity in Wittgenstein's thought which links his "two philosophies" in the very way which makes them correspond to a fundamentally unchanged attitude to life. This connecting link is found, not surprisingly, in Wittgenstein's conception ofphilosophy and his attitude to language. 6. Wittgenstein's intellectual and moral personality must be understood against the background of the society in which he grew up. He was a son of the late Habsburg Empire. This was a very peculiar socio-political construct. The multinational and multi-lingual state was in many ways an obsolete phenomenon in a Europe then on its road towards democracy and industrlaliza- WITTGENSTEIN AND THE TWENTIETH CENTURY 9 tion in the frames of consolidated national states. It was a reactionary bulwark against progressing modernization. At the same time it appears to us today strangely modern, a forecast of what may come in a Europe now in a process of integration with national borders breaking down and a new mixing of languages and nationalities in the offing. Of this "kaiserliche und königliche" construct - for which the author of Der Mann ohne Eigenschaften invented the name with ironic Greek overtones "Kakanien" - was characteristic a hypocrisy and doubleness of morality which, though certainly not unique in history, was perhaps unique in the Europe of the 19th century. Or how else shall one understand that in this atmosphere of conventional half-truth and insincerity there arose such a strong reaction against it, a violent passion for truthfulness and purity, unsparing efforts to debunk the illusions and lay bare the underground of the human soul? We witness this reaction in the puristic architecturallanguage of Adolf Loos, in the stern atonality of Schönberg's music, in the searching cultural criticism of authors like Hennann Broch and Robert Musil, and in the apocalyptic irony ofKarl Kraus. In the same cirele also belongs the greatest debunker of all, Sigmund Freud. It is against this background in Austrian Geistesgeschichte that we have to understand Wittgenstein: his passion for truth and sincerity and his longing for pure and simple fonns of life. Wittgenstein had a rare sense for detecting or, to use a favourite word of his, "smelling" even the slightest trace of conventionallie, untruthfulness, artificiality and pretense in the people whom he met. This made relations with hirn strained, in some cases even unbearably difficult, for those who had the privilege of coming elose to hirn personally. The statement holds true equally of the Wittgenstein who came to pre-war Cambridge to attend Russell' s lectures and of the one who returned there in 1929 after ten years ofvoluntary withdrawal from the world. "Are you thinking oflogic or ofyour sins" Russell asked hirn in a memorable conversation. 18 "Ofboth" was Wittgenstein's reply. Paraphrasing a word by Ibsen -certainly well known to Wittgenstein - one could say that philosophical thinking was to hirn a perpetual holding of doomsday with hirnself. To the best of my knowledge, Wittgenstein is the only one among those twilight figures mentioned of Musil' s "Kakanien", whose craving for honesty prompted hirn to radical efforts to change his (outward) life. Here the influence ofTolstoy on Wittgenstein must be rated high. Tolstoy was at horne in a society at least as morally depraved as "Kakanien" but perhaps with greater resources ofunconsumed vitality than the latter. Tolstoy's mature life was a relentless fight to free hirnself from the nets of bigotry and conventionality. 10 GEORG HENRIK VON WRIGHT He lured, moreover, a number of men and women all over the world to try to "return" to a pure and simple life in truth. He was hirnself tragically incapable of this return; the last effort he made was a suicide of its kind. Also Wittgenstein's withdrawal to be a schoolmaster for peasant children was a failure. I think that both Wittgenstein and the great Russian were partly victims of an illusion about the actual existence of a country "Erehwon" uncontaminated by the moral ills of the society we live in and where we can go to start a new life. Tolstoy went to Caucasus and to the Bashkirs in search of it. Wittgenstein sought it in remote villages in lower Austria, in the wilderness ofthe fjords ofNorway or on the shore ofthe Ocean in Galway. His abortive plans of settling in the Soviet Union is also a chapter in this story. But the land he sought is really the land "Nowhere". 7. In a social atmosphere of bigotry and insincerity language also tends to become corrupted. It is infiltrated by euphemisms. Things are no longer called by their blunt and simple names, but are wrapped up in evasive circumscriptions or referred to by invented technical terminology. Style becomes unperspicuous, meaning unclear. In modern society this distortion of language has assumed grotesque proportions in the jargon of public administration and the media. Austria under the late Habsburgs may have been the place where this "illness ofthe times", now universal, first flared up. To fight it was the set task of that generation of Austrian "purifiers" of which Wittgenstein was a member. Fritz Mauthner' s Sprachkritik was an early sampIe of this reaction to language. The fact that Wittgenstein in the Tractatus said (4.0031) that all philosophy is critique of language, though not in Mauthner' s sense, has, I am afraid, long distracted modern philosophers from paying due attention to Mauthner's work. It belongs in the same tradition of which - not counting Wittgenstein - Karl Kraus is the supreme and most influential example. Wittgenstein admired the work ofKraus. He counted Kraus and Loos among those who had influenced hirn. It is an interesting but perhaps unanswerable question, whether the Krausian moral attitude to language, which Wittgenstein shared, was perhaps a force which made hirn broaden his philosophy of logic of the early Cambridge years to the philosophy of language embodied in the Tractatus. It is sometimes held that Wittgenstein's concerns in the Tractatus were the conditions which a logically ideal or perfect language has to fulfil. Thus, for example, Russell in his Introduction. This has also been disputed. I think those who dispute it are right, if the ideallanguage was to be a perfection of WITTGENSTEIN AND THE TWENTIETH CENTURY 11 the Begriffsschrift first propounded by Frege and then further developed by the authors of Principia Mathematica. Wittgenstein's contribution to their project was in the first place the tabular theory of truth-functions and the notion of tautology. But although logic no doubt was the gate through which Wittgenstein entered philosophy, his work in the tradition ofFrege and Russell was of short duration. It ended, 1 would say, when in the early part of the 1914-1918 war the thought of the proposition as a picture first dawned upon him and he wrote in his notebook in January 1915 "My whole task consists in explaining the nature of the proposition" .19 This had not been the "whoie task" of Frege or of Russell. Rather than saying that Wittgenstein in the Tractatus was concerned with the conditions of an ideal language we should, I think, say that he was in search of pure language or of language in its pure and uncorrupted form. The language whieh would depiet the world as it really is, absolutely "true to fact". It was to this end that, at a later stage of his work on the book, he conceived of the ontology of immutable and indestructible things to whieh there correspond names the configurations of which in language picture the contingent configurations of things in reality. There are some indieations in the notebooks that Wittgenstein's position was initially wavering on the questi on wh ether such a language could actually be found, examples given of things and corresponding names. But in the end he realized that the form and content ofthe "pure" language is ineffable. The language ofwhieh he was in search is, to use Max Black's happy phrase, a "never-never language".2o It is as remote from the language we speak as is the nowhere land of pure and simple life from the contaminated societies in which we are destined to live. Because ofthe way in which the pure language ofthe Tractatus is supposed to reach up to reality, one can also say that the world it is supposed to picture is a "never-never-world", a postulated ideal which is nowhere to be found. It has occurred to me that the traits of relativism and opposition to realism which some interpreters see in Wittgenstein' s later work have their root in the impasse of the Tractatus approach to language. Just as there is no "pure" language there is also no "pure" reality for language to depict. 8. 17th and 18th century philosophers had entertained the fiction of "man in astate of nature". This fiction was thought useful in their search for the raison d 'etre of the state and the legitimation of societal institutions. Some fancied it to have been an existing state of affairs in the "uncorrupted" infancy of prehistorical man. Even if this was an illusion, it does not necessarily ruin the philosophie value of the fiction. 12 GEORG HENRIK VON WRIGHT In a somewhat similar sense Wittgenstein may be said to have been in search, throughout his reflective life, of a "natural state of language". In the Tractatus the thought it was hidden under the disfiguring veils of ordinary speech. It existed somewhere deep under the surface of language as used. But he did not find it there. He came to the conc1usion that it was not there to be found either. He had been looking for it in the wrong direction. It was not to be attained by diving under a surface but by looking at the surface itself. There was also to be found the uncorrupted language of which he was in search. The difficulty was only to see it. "Nothing is hidden." lt is often said that one ofhis early ideas whieh Wittgenstein never gave up is his conception of philosophy. Basically this is true, I think. But it is not a truth without modifications. It holds for his conception of philosophy as an activity and not a doctrine (Lehre), and ofphilosophy as critique oflanguage. But it does not hold for his specific view that philosophieal problems are due to linguistic confusions, to what he later called the "bewitchment" (Verhexung) of our thinking by the means of language. There is no statement to this effect in the early writings. There he speaks ofthe philosophie problems resting on a misunderstanding of the ("true") logic of our language, on a confusion of the grammatical with the logical form of thought. 21 The problems which arise through a "linguistic bewitchment" of our thought are not questions in search of an answer, but "burnps that the understanding has got by running its head up against the limits of language". 22 A main reason why we get these "bumps" is "that we do not command a clear view of the use of our words".23 We miss the übersichtliche Darstellung ("perspicuous presentation") which will expose to us the undistorted use of language and "bring words back from" what Wittgenstein calls "their metaphysical to their everyday use".24 The clarity which the perspieuous presentation gives is an absolute. It makes the philosophieal problem disappear completely.25 The disquietude is gone and the philosopher attains his aim which is "Friede in den Gedanken".26 Philosophy which follows these lines is strictly descriptive. It does not explain anything - for example how it is possible for signs to mean. Nor does it answerquestions of essence, for example what thinking, or truth, or logical necessity are. Unlike most philosophy after Kant it is neither transcendentalist nor essentialistY It puts the "how possible?"- and "what is?"-questions aside by directing our attention to the role which the problematic words play in actual communication. In this sense "the work of the philosopher" can be said to consist "in assembling reminders for a particular purpose".28 A philosophy which does not look for answers to questions, does not ex- WITTGENSTEIN AND THE TWENTIETH CENTURY 13 plain or theorize about the things which attract the philosopher's curiosity, and does not try to provide the foundations for our beliefs, is not a philosophy for which scientific thinking sets the pattern. It, on the contrary, fights the infiltration of this thinking into philosophy and makes it responsible for the confusions from which the philosopher tries to rid himself. It is not, need not be, hostile to science as such. But it may be said to take a critical or even ho stile attitude to the influence of science outside its proper domain - and in particular on philosophic thought. In this it runs counter to an intellectual mainstream of the century. 9. There are a few points on which one might wish to challenge the consistency of Wittgenstein's (later) view of philosophy. On the one hand Wittgenstein says that the philosopher must "in no way interfere with the actual use of language" and that philosophy "leaves everything as it iS".29 But he also says that he is "engaged in a struggle with language".3o In a passage in the unpublished part of the "Big Typescript" he describes this struggle as follows - I have here translated it into English: Human beings are deeply embedded in philosophical, i.e. grammatical confusions. Freeing them from these presupposes tearing them away from the enormous number of connecting links that hold them fast. A sort ofrearrangement ofthe whole oftheir language is needed. (Man muß sozusagen ihre ganze Sprache umgruppieren.) - But of course that language has developed the way it has because some human beings felt-and still feel-inclined to think that way. So the tearing away will succeed only with those in whose life there already is an instinctive revolt against the language in question and not with those whose instinct is for the very herd which created that language as its proper expression. 31 This sounds almost like a cry for a "cultural revolution" needed for putting the philosophic worries to rest. Because is not the "Kampf mit der Sprache" also an interference with it? Particularly ifit requires a "rearrangement ofthe whole of language"? And a similar question can be raised about the attempt to "bring back" (in German zurückführen) the words from their "metaphysical" to their "everyday" use. Wittgenstein's talk about philosophy suggests to me the following pieture: There are in the large garden of language tidy plots of land where uncorrupted language-games are being played by human users. But this garden is also partly overgrown by metaphysical weeds which hide the plots from sight and blur their borders and thereby confuse those who play with language. The good philosopher should be a gardener who by clearing away the weeds displays the plots oflinguistic ground in their purity and thereby helps communication to go on unimpeded by the metaphysical-philosophical confusions. 14 GEORG HENRIK VON WRIGHT But is this picture not another vision ofthe pure and uncorrupted language of which Wittgenstein had been in search in the Tractatus? He could not find it then because he sought it under the surface of everyday language rather than on the surface itself. But if this surface is partly veiled, hidden, overgrown by metaphysical weed which must be removed before we command a perspicuous view of what things are like, may it then not be the case that this uncorrupted ordinary (use ot) language too is a never-never thing and that the search for it must end in frustration? Does belief that this is not so perhaps rest on an illusory idealization of ordinary language?32 My impression sometimes is that Wittgenstein in the Tractatus had attained a position which, as it stood, was unassailable. As we know, he himself assailed it later. He even talked of "grave errors" in his earlier writing. 33 His thinking took what may be called a "philosophic V-turn": from the surface to the hidden depths, then from the depths back to the surface. 34 But I know from what Wittgenstein told me that there were moments when he doubted wh ether this was a turn in direction to final c1arity or whether it too would end in an impasse of obscurity. This is the radical, "existential" doubt of one who cannot resolve the question whether all his efforts have been a failure. No outsider can resolve it for hirn either. 10. The bewitchment of our thinking through language happens when words take on what Wittgenstein calls a "metaphysical" use. So one can say that Wittgenstein's fight with language was a fight with metaphysics. This was precisely what the logieal positivists had been engaged in. "Überwindung der Metaphysik durch logische Analyse der Sprache" was the titte of one of Carnap's artic1es in Erkenntnis. So what difference, if any, was there in their respective attitudes to metaphysics? For the positivists, questions were "metaphysical" when they could not be decided by the methods of natural science or by logieal deduction from scientifically acceptable premisses. As amply testified in the documents from the heyday of positivism, metaphysics was seen by the champions of the "new philosophy" as an outgrowth and residue of religious beliefs of apre-modem society, as a rationalized disguise of at bottom irrational attitudes. Metaphysies was philosophy which had helped reactionary forces in society to block and retard the progress of emancipated, rational, and secularized man. Wittgenstein's fight against metaphysics was something very different. By a metaphysical use of language Wittgenstein means the "free-wheeling" of language which occurs when words get detached from their actual use in the language-games of communicative discourse and are being used for construct- WITTGENSTEIN AND THE TWENTIETH CENTURY 15 ing what Wittgenstein calls Luftgebäude ("houses of cards") in the linguistic isolation of the philosopher's mind. 35 In past centuries of European history the thinking of metaphysicians was to a large extent nourished by the linguistic rituals of a religious culture. This was a culture in which language-games with words like "God", and "sin", "grace" and "doom" and "redemption" had an established everyday use. In a similar manner, the thinking Wittgenstein calls metaphysical is stamped by the linguistic patterns and thought habits of a predominantly scientific civilization. The metaphysics which Wittgenstein is fighting is thus not one rooted in theology but one rooted in science. He is fighting the obscuring influence on thinking, not of the relicts of a dead but of the habits of a living culture. Of this he gives c1ear warning in the 'BIue Book' where he wrote: Philosophers constantly see the methods of science before their eyes, and are irresistibly tempted to answer questions in the way science does. This tendency is the real source of metaphysics, and leads the philosopher into complete darkness. 36 He immediately gives examples: The craving for general theories, contrasted with what he ealls "the eontemptuous attitude to the partieular ease"37; the tendeney to explain the eoneept of number, to reduee the infinite to the finite, mathematies to logic, intentional behaviour to bodily movement. The most vulgar examples of these tendencies we find, it seems to me, in contemporary philosophy of mind, be it in the form of the physiealists' "identifieation" of so-called mental states with brain proeesses or the eliminative materialists' rejeetion of our common sense psychologieal eoneepts - what they eall "folk-psyehology" - as a "radically false theory" eventually to be replaeed by a perfected neuroseience. Parther into the jungle ofmetaphysics, as Wittgenstein saw it, philosophy can hardly lose itself than in these latter day phenomena of a philosophie eulture gone "scientist". Thus one ean say - I am here quoting Stephen Hilmy's book on the later Wittgenstein - that Wittgenstein's "eonception of metaphysies was sueh that it eneompassed not only traditional metaphysies, but also, and especially, the dominant mode of reflection of our own epoch; and it is primarily the latter, that constituted the intellectual current against whieh he was struggling".38 It is ironie that the metaphysies which Wittgenstein was fighting was exactly the one in the eobwebs of whieh the logical positivists and a good many of their followers among analytieal philosophers according to him had been eaught. It is worth asking here how Wittgenstein himself viewed the great efforts of past thinkers who had tried to eneompass the world intelleetually in systems 16 GEORG HENRIK VON WRIGHT we eall "metaphysieal". The thing to say in response to the question is that he had no need to eombat these systems and that he therefore also was, or would have been, able to bestow on them the respeet and awe (Ehrfurcht) which he was disposed to feel towards great aehievement in all spheres of human life. But he was also acuteiy aware that he was not continuing a tradition. What he did was as different from what a Leibniz or a Spinoza had been doing as is life in our times different from what it was in theirs. Yet there is also enough resemblance to make his philosophy, as he said hirnself, "a legitimate heir" to those intellectual pursuits of the past which traditionally go under the name "philosophy" .39 I hope to have made it clear in whieh sense Wittgenstein's efforts in philosophy were a fight against a dominant climate of opinion in our eentury. I have called this climate that of Modernity. It was in origin coupled with a euphorie belief in progress thanks to the managerial uses of reason in industrialized democratic societies. This optimism has largely faded in the wo es of two global wars and under the load which man has imposed on nature, threatening the biosphere with breakdown. But also amidst the sombre or even apoealyptic mood of scared humanity, this rake's progress continues. "Science and industry", Wittgenstein wrote a few years before he died, "might turn out to be the most enduring thing in the modern world".40 But he also wrote that there is nothing absurd in the belief "that the age of science and technology is the beginning of the end of humanity" and that mankind in seeking to steer its course towards the future relying on scientific rationality "is falling into a trap" Y Wittgenstein's philosophy and also his life was a protest against these trends and an effort to set an example of "a changed mode of thought and life" which, if followed might provide a eure for what he thought of as a siek time. But no rule can, I think, be laid down for how to imitate his path. Academy ofFinland Helsinld NOTES von Wright (1978). Carnap (1928), p. v f.: "Wir spüren eine innere Verwandtschaft der Haltung, die unserer philosophischen Arbeit zugrunde liegt, mit der geistigen Haltung, die sich gegenwärtig auf ganz anderen Lebensgebieten auswirkt; wir spüren diese Haltung in Strömungen der Kunst, besonders der Architektur, und in den Bewegungen, die sich um eine sinnvolle Gestaltung des menschlichen Lebens bemühen: des persönlichen und gemeinschaftlichen Lebens, der Erziehung, der äusseren I 2 WITTGENSTEIN AND THE TWENTIETH CENTURY 17 Ordnungen im Grossen. Hier überall spüren wir dieselbe Grundhaltung, denselben Stil des Denkens und Schaffens ... Der Glaube, daß dieser Gesinnung die Zukunft gehört, trägt unsere Arbeit." 3 As early forceful expression ofthis view of"the new philosophy" is Russell (1914). It is interesting to read it in conjunction with the writings from the years ofthe flowering ofthe Vienna Circ1e. 4 Quoted from Haller (1981). 5 Hahn (1929), p. 54: "Einstein, Russell und Wittgenstein seien hier als diejenigen unter den führenden Denkern der Gegenwart genannt, die die wissenschaftliche Weltauffassung am wirkungsvollsten in die Öffentlichkeit vertreten und auch stärksten Einfluß auf den Wiener Kreis ausüben." 6 Camap (1934), pp. iii-iv: "Philosophie wird durch Wissenschaftslogik ersetzt; Wissenschaftslogik ist nichts anderes als logische Syntax der Wissenschaftssprache. " 7 See Wittgenstein (1980), p. 9. 8 Ibid., p. 19. In German: "Ich habe sie (sc. Gedankenbewegung) nur sogleich leidenschaftlich zu meinem Klärungswerk aufgegriffen." 9 Ibid., pp. 6-7. In German: "Dieser Geist ist ... ein anderer als der des großen Stromes der europäischen und amerikanischen Zivilisation. Der Geist dieser Zivilisation, dessen Ausdruck die Industrie, Architektur, Musik, der Faschismus und Sozialismus unserer Zeit ist, ist dem Verfasser fremd und umsympatisch ... Unsere Zivilisation ist durch das Wort 'Fortschritt' charakterisiert. Der Fortschritt is ihre Form ... Ihre Tätigkeit ist es, ein immer komplizierteres Gebilde zu konstruieren ... Es interessiert mich nicht, ein Gebäude aufzuführen, sondern die Grundlagen der möglichen Gebäude durchsichtig vor mir zu haben. Mein Ziel ist also ein anderes als das der Wissenschaftler,und meine Denkbewegung von der ihrigen verschieden." \0 See in particular his essay 'Konservative Anthropologie: Der Sohn Wittgenstein', in Nyiri (1988). The essay opens with the words (p. 91): "Ludwig Wittgenstein begann um die Wende der zwanziger/dreissiger Jahre des 20. Jahrhunderts jenen Ideenkreis auszuarbeiten, den man heute, rückblickend, wahrscheinlich als die tiefste und umfassendste Grundlegung vom konservativen Bild des Menschen bezeichnen kann." 11 Nyiri (1988), p. 104 ff. 12 Wittgenstein (1980), p. 21. 13 Janik (1985) p. 130. 14 Hilmy (1987), p. 190. IS Ibid., p. 192. 16 Ibid., p. 194. 17 Ibid., p. 210. 18 Told and retold by Russell in many places. See, for example, Russell (1968), p. 99. 19 Wittgenstein (1979), p. 39: "Meine ganze Aufgabe besteht darin, das Wesen des Satzes zu erklären." 20 Black (1964), p. 11. 21 See, for example, Tractatus, the Preface (Vorwort) and ibid., 4.002, 4.003, and 4.0031. 22 Wittgenstein (1953), § 119. 23 Ibid., § 122. 24 Ibid., § 116. 25 Ibid., § 133. 26 Wittgenstein (1980), p. 43. 27 Both attributes, in my opinion, apply to the Tractatus. There Wittgenstein is in search of the essence oflanguage (the proposition) and ofthe trascendental apriori conditions ofthe possibility of language. (The resemblance with Kant has often been noted). The later philosophy ofWittgenstein is opposed to both tendencies. There is a profound change in this regard as compared with the Tractatus. Some attempts to "transcendentalize" Wittgenstein's later philosophy of language 18 GEORG HENRIK VON WRIGHT therefore seem to me substantially mistaken. (Which does not prevent them from having some independent interest in themselves.) 28 Wittgenstein (1953), § 127. 29 Ibid., § 124. 30 Wittgenstein (1980), p. 11. "Wir kämpfen mit der Sprache. Wir stehen im Kampf mit der Sprache." 31 MS 213, p. 423. 32 The phrase "idealization of ordinary language" was suggested to me in conversation by Heikki Nyman. 33 Wittgenstein (1953), Vorwort, p. x. 34 See Wittgenstein (1953), § 108: "Die Betrachtung muß gedreht werden, aber um unser eigentliches Bedürfnis als Angelpunkt." 35Ibid.,§ 118. 36 Wittgenstein (1958), p. 18. 37 Ibid. 38 Hilmy (1987), p. 225. 39 Wittgenstein (t 958), p. 62. The fullest statement by Wittgenstein hirnself ofhis views ofphilosophy is the chapter "Philosophy" in TS 213 (the "Big Typescript"). This chapter has now been published (Wittgenstein 1989.) - Of great interest are the pronouncements made by Wittgenstein on his philosophy recounted by G. E. Moore in Moore (1954-1955) (particularly the third article). In the early 1930s Wittgenstein too, somewhat in the style ofthe Vienna positivists, seems to have thought of his way of doing philosophy as a great break with tradition. Moore III, p. 26: "He said that what he was doing was a 'new subject', and not merely a philosophy, a 'kink' in the 'development of human thought', comparable to that which occured when Galileo and his contemporaries invented dynamics; that a 'new method' had been discovered, as had happened when 'chemistry was developed out of alchemy'." lt is perhaps significant that statements with this colouring are not found in Wittgenstein's own written work. 40 Wittgenstein (1980), p. 63. 41 Ibid., p. 56. "Es ist z.B. nicht unsinnig, zu glauben, daß das wissenschaftliche und technische Zeitalter der Anfang vom Ende der Menschheit ist; daß die Idee vom großen Fortschritt eine Verblendung ist, wie auch von der endlichen Erkenntnis der Wahrheit; daß an der wissenschaftlichen Erkenntnis nichts Gutes oder Wünschenswertes ist und daß die Menschheit, die nach ihr strebt, in eine Falle läuft. Es ist durchaus nicht klar, daß dies nicht so ist." REFERENCES Black M. (1964):A Companion to Wittgenstein 's Tractatus, Cambridge University Press, London. Carnap R. (1928): Der logische Aufbau der Welt, Weltkreis Verlag, Berlin-Schlachtensee. Hahn H. (1929): Wissenschaftliche Weltauffassung, Der Wiener Kreis, hrsg. vom Verein Ernst Mach, Artur Wolf Verlag, Wien. Haller R (1981): 'Wittgenstein and Austrian Philosophy', in Austrian Philosophy: Studies and Texts, ed. by J. C. Nyiri, Philosophia Verlag, München. Hilmy S. (1987): The Later Wittgenstein, B. Blackwell, Oxford. Janik A. (1985): Essays on Wittgenstein und Weininger, Rodopi B. V., Amsterdam. Moore G. E. (1954-55): 'Wittgenstein's Lectures in 1930--33', I-III, Mind, 63, pp. 1-:-15; 289-315; 64, pp. 1-27. Reprinted in G. E. Moore,Philosophical Papers, GeorgAllen & Unwin, London 1959. WITTGENSTEIN AND THE TWENTIETH CENTURY 19 Nyiri J. C. (1988): Am Rande Europas. Studien zur österreichisch-ungarischen Philosophie geschichte, Akademiai Kiado, Budapest. RusseIl B. (1914): Our Knowledge ofthe External World as a Fieldfor Scientific Method in Philosophy,AIlen & Unwin, London. RusseIl B. (1968): The Autobiography ofBertrand RusselI, vol. II, AIlen & Unwin, London. Wittgenstein L. (1933 2 ): Tractatus logico-philosophicus, German text with facing translation by C. K. Ogden & F. P. Ramsey, Kegan Paul, Trench, Trubner & Co., London. Wittgenstein L. (1953): Philosophische Untersuchungen. Philosophical Investigations, ed. by G. E. M.Anscombe & R. Rhees. Trans. by G. E. M.Anscombe, B. BlackweIl, Oxford. Wittgenstein L. (1958): The Blue and Brown Books, ed. and with aPreface by R. Rhees, B. Blackwell, Oxford. Wittgenstein L. (1961): Notebooks /9/4-/916, ed. by G. H. von Wright & G. E. M. Anscombe. Trans. by G. E. M. Anscombe, B. BlackweIl, Oxford (1979 2). Wittgenstein L. (1969): On Certainty. Über Gewissheit, ed. by G. E. M. Anscombe & G. H. von Wright. Trans. by D. Paul & G. E. M.Anscombe, B. Blackwell, Oxford. Wittgenstein L. (1975): Philosophical Remarks. Philosophische Bemerkungen, ed. by R. Rhees. Trans. by R. Hargreaves & R. White, B. BlackweIl, Oxford (1964 1). Wittgenstein L. (1979): 'Bemerkungen über Frazers The Golden Bough. Remarks on Frazer's The Golden Bough ',ed. by R. Rhees, Synthese, 17, pp. 233-53. Trans. by A. C. Miles, Humanities Press, Atlantic Highlands, N.J. Wittgenstein L. (1980): Culture and Value. Vermischte Bemerkungen, ed. by G. H. von Wright in coIlaboration with H. Nyman. Trans. by P. Winch, B. BlackweIl, Oxford. Wittgenstein L. (1989): 'Philosophie', hrsg. von H. Nyman, Revue Internationale de Philosophie, 43, pp. 175-203. Wright von G. H. (1978): 'Wittgenstein in Relation to His Times' ,in Wittgenstein and His Impact on Contemporary Thought, ed. by E. Leinfellner et al., Hölder-Pichler-Tempsky, Wien, pp. 73-78. PART I SCIENCE, MIND, AND METAPHYSICS DAVID PEARS WITTGENSTEIN ON PHILOSOPHY AND SCIENCE In this paper I shall discuss the question, "How did Wittgenstein see the relation between philosophy and science?" This is not a question about his philosophy 0/ science. It is not, for example, direcdy concerned with the brief account of scientific theories that he gave in the Tractatus. I My inquiry is a more fundamental one. 1 want to uncover his view of the relation between science and philosophy. If he showed scientific discourse, and, more generally, factual discourse on the same map as philosophical discourse, how would he place them? And at how many points did he see a risk of our unwittingly crossing the frontier which divides them? It is, as usual, worth looking at the Tractatus first, because although it only supplies sketchy answers to my question, they do set the scene for later developments. They are, perhaps, rather disappointing answers, because Wittgenstein merely gives us a diagram in which all kinds of discourse are related to the single line dividing what can be said from what can only be shown. Everything that can be said belongs to factual discourse, all of which is sometimes counted as "scientific".2 Everything else, inc1uding philosophy, is placed on the other side ofthe line. It is all done on such a small scale that many important distinctions are not c1early represented. What we are told about the differences between stating facts and explaining them seems to make science almost an offshoot ofphilosophy.3 Also, not enough is said about the differences between philosophical discourse and other modes of showing. The flow of thought is always from the general to the particular and so from the theory of language to its application to various disciplines. Even important differences are treated as points to be accommodated within the theory without being given a chance to alter its structure. The result is an excessively simple demarcation of the boundary between philosophy and science. Still, it is an interesting picture, both in itself and because it helps us to understand what came later. Philosophy is assigned the task of analyzing language in use,4 and what it tries to say can only be shown, because it is not concerned with the mere sounds and shapes of words, but takes them with their meanings. The argument is simple: if what philosophy tries to say were true, it could not have been false and so it would not be true in the ordinary, contingent way, and, therefore, could not properly be expressed in factual 23 R. Egidi (ed.), Wittgenstein: Mind and Language, 23-36. © 1995 Kluwer Academic Publishers. 24 DAVID PEARS language. Its message, if it can be called one, is that, though the structure of factuallanguage is largely optional, its fundamental features are not optional but forced on it by the nature ofreality. Science5 exploits some ofthe options and invents systems of description which are more perspicuous than our ordinary ones and yield factual generalizations ofwider scope. 6 However, this is not a method which makes it possible to explain everything. 7 There are points in this simple diagram which Wittgenstein elaborated after 1929, often without any changes in his basic ideas. But let me start with a development which combines continuity with radical change. In the Tractatus he maintains that much scientific theorizing has an apriori character. Instead of simply recording our discoveries, it imposes on them new systems of description which are more perspicuous and more widely applicable than our ordinary ones. This, as already remarked, brings science elose to philosophy, because philosophy too is driven by the same need for a panoptic view. There are, at this stage in the development of Wittgenstein's thought, two differences between the disciplines: one is that the philosophical theory which meets this need, 10gical atomism, is, unlike scientific theories, total, and therefore, does not belong to factuallanguage and can only be shown: the other is that it is a theory which is completely determined by the nature ofthings,8 whereas scientific theories are under-determined by the nature of things which, in their case, leaves scope for options. After 1929 this map undergoes a radical change. The relative positions of science and philosophy remain more or less the same, but the distance between them is increased. Philosophy still go es further than science in its search for a commanding view, but what it investigates is no longer an independent reality but only our own linguistic reactions to whatever reality there may be. The whole idea of an independent world imposing its structure on language has gone, not because Wittgenstein has become a conventionalist instead of a realist, but because he has seen that the difference between these two views vanishes at the point where theorizing becomes total. So in his later philosophy he is directly concerned with language in its place in our lives. What he looks for is no longer any kind oftheory because the special features ofphilosophical results, which previously made hirn treat them as theories which can not be stated but only shown, now make hirn say that they are not theories at all. The commanding view that he seeks is no longer a small-scale map - the whole terrain seen from a high altitude - but a complete grasp of the relationship between the various moves made on the ground by people using language. There are several ways of describing these developments in Wittgenstein's philosophy. One, which he himse1f often used, was to compare his new work WITTGENSTEIN ON PHILOSOPHY AND SCIENCE 25 with psycho-therapy. In this paper a different point ofview is adopted and an attempt will be made to relate the changes to the line which divides philosophy from science. Suppose we agree with the popular view that the task of science is to go behind the phenomena in order to explain them. Then a plausible way of assimilating philosophy to science might seem to be to try to transcend the limit of language and to theorize about possibilities9 in the way that science theorizes about facts. That is to say,just as a scientist explains why a possibility is or is not realized as a fact, so too a philosopher might try to explain why a certain candidate for possibility does or does not achieve that status. This way of assimilating philosophy to science is rejected in the Tractatus, but there is only a very schematic demonstration of what is wrong with it: facts are surrounded by identifiable possibilities, but possibilities are not surrounded by a further ring of identifiable candidates for possibility. That puts in a nutshell an idea, whose range and power only become apparent in Wittgenstein's later work, where it supplies hirn with a strategy that can be used against many metaphysical theories. A weIl developed example ofthis strategy is provided by his later treatment of solipsism. The solipsist uses the sentence "I do not see what you see" not in the ordinary way, to say something which sometimes happens to be true, but rather to express a universal truth about the human predicament. 1O Wittgenstein makes two moves against this position. First, he demonstrates that, if the remark, made with the solipsist's intention, has asense, it does not have its ordinary sense. Then he argues that the solipsist is not really using existing language to express a new truth but, rather, proposing a linguistic innovation which, he hopes, will give us a better map of the possibilities than existing language. That, according to Wittgenstein, is his fundamental mistake. He is assimilating philosophy to science and modelling the relation which is supposed to make a language adequate to express the possibilities on the relation which produces true reports of the facts. These two moves can be described with a different emphasis. The point of the first one is to convince the solipsist that his remark does not start with the sense that is required for the universal truth that he wants to express with it. Contrary to what he thinks, we do not immediately recognize the possibility, seeing what others see, which it is our misfortune never to find realized. It is important to insist on this, because the solipsist nearly always begins by assuming that he has already identified the possibility which, he claims, necessarily remains unrealized. Sut there he is mi staken. Ifhe were using the words in the old way, his remark would be literally false, because according to the 26 DAVID PEARS old way of speaking, the possibility is frequently realized - we do often see the same things as other people. So he must go back and actually do the work of linguistic innovation. The point of the second move is that this is not the kind of work that he takes it to be. The relation between language and the world that yields sense is not at alltike the relation between language and the world that yields truth. He must not assurne that he can simply see the possibility that he has in mind, and that all that he needs to do is to express it in existing language, which a sympathetic audience will immediately appreciate, because they will understand its rules to have been changed in the appropriate way. But appropriate to what? We cannot know, until the possibility has actually been identified. It is worth looking more c10sely at the impact ofWittgenstein's two moves against solipsism. The strategy is a general one and he uses it against other metaphysical theories, but its use against solipsism is particularly c1ear and fully developed. It is, for example, obvious that he does not share the positivist view that such theories are soft-headed folly. On the contrary, he believes that the tendency to assimilate philosophical problems to scientific on es is natural and inevitable, and that it can only be held in check, but never eradicated. 11 Solipsism is not a gratuitous eccentricity but the understandable consequence of reflection on two connected facts which can easily make scepticism about other people's sensations seem inescapable: I can only feel my own sensations and only I can feel them. Ifwe focus exc1usively on Wittgenstein's first move against theories ofthis kind, we will be likely to find his philosophy more like Moore's or Austin's than it really iso But the first move is not intended to be a complete refutation of the solipsist. It is a preliminary skirmish designed to uncover his adversary's central position: he is really innovating, rather than using words in the old way to formulate a new truth about our predicament. The point of his innovation is that existing language pays too littIe heed to the facts of sensory isolation and attaches too much importance to our need to communicate, and to feel that we are communicating, with each other. The result is that the criteria adopted for everyday truth are excessively relaxed and the deeper truth about our predicament has been suppressed by a conspiracy not of silence, but of inexpressibility. Wittgenstein's second move is to challenge the solipsist's reason for his linguistic innovation. The reason offered by hirn is that his new way ofspeaking matches the material better than existing Ianguage: but the match that he has in mind cannot be truth, because any question of truth must wait until the rules for the use of the words have been fixed, and after they have been fixed, WITTGENSTEIN ON PHILOSOPHY AND SCIENCE 27 though different ways of speaking will all aim at truth, they will not compete for it. The vaunted match can only be between the solipsist's new way of speaking and the possibilities. But what possibilities? And by what roles will this more fundamental kind of match be judged? The solipsist thinks that he does not need to do any more work in order to answer these questions. But that is an illusion based on his unexamined assumption that the relation with the world that confers sense on language is like the relation with the world that confers troth on it. In fact, he has to do more in order to identify the possibility which, according to hirn, is never realized, and his further task, which is to introduce his new way of speaking, can only be carried out in existing language. The result will then necessarily be disappointing, because the suggestion of a real deprivation, originally presented in existing language, will have been cancelled by the explicit shift to the new language. When the objection, that match can only be determined by roles, is made against solipsism, it is, of course, assumed that the target of match is identified in a general way by the function of our existing sensation-Ianguage. That is the material which the solipsist is taken to have in mind, and his question is, what language would do justice to it. So when Wittgenstein objects that the question is unanswerable, because there are no roles for judging such a match, he is criticizing a philosophical theory which relies on part of existing language to identify the target of match. The controversy then turns on what shapes the detailed construction of sensation-Ianguage after its general function has been settled. The solipsist's claim is that, once we are committed to a language for reporting our sensations, we are forced by their nature to model it on our language for physical objects. Wittgenstein has two answers to this. One, which has not been discussed here, is the well-known anti-privacy argument, that the solipsist's claim is incoherent. The other answer, which has occupied the centre of this discussion, is that, even if it were coherent, his preferred language would not be forced on us by the material, in the way in which facts sometimes force us to acknowledge truths. We can still retain our existing language, which does justice in its own way to the facts cited by the solipsist. It certainly seems more natural to do so, and at least, it cannot be wrong. The idea, that we can find a viewpoint from which one language can be seen to be better adapted to the possibilities than another one, is rejected in the Tractatus: theories about possibilities cannot be modelIed on theories about facts. However, it would be an exaggeration to say that all the implications of this difference between scientific and philosophical theories were clear to 28 DAVID PEARS Wittgenstein when he was writing the Tractatus. If they had already been developed in his mind, he would have disallowed all philosophical theories from the start, even those that "can only be shown". But the Tractatus does offer a theory about the deep structure common to all languages which is imposed on them by the nature of reality.12 This structure allows us certain options in the rules adopted for our languages, but it sets a limit to the variations, because, once we have made our choices, it forces their consequences on uso This view, which leaves a certain lingering similarity between philosophy and science, is clearly expressed in the Tractatus,13 but its rejection is the central point of his later philosophy of language. So the idea, that theories about possibilities cannot be modelled on theories about facts, though implicit in the Tractatus, was not immediately taken to its logical conclusion, and the development of his critique of solipsism, described above, takes the rejection of that kind of metaphysics further than he took it in his early work. Ifthe idea, that our linguistic choices do not force any consequences on us, is the central point of his later philosophy of language, it will be worth looking at the way in which he developes it in his later work. For it ought to give us a clearer picture ofhis view ofthe line dividing philosophy from science. It is too brief and uninformative to say that the differences between the two disciplines struck hirn as so great that he disallowed all theorizing in philosophy. We need to appreciate the details and the best place to see them is in his treatment of the central problem of his later philosophy of language, rulefollowing. However, before we look at that problem, there is another part of his later philosophy which needs to be examined, his views about the relation between mind and brain. These views are sometimes interpreted in a way that runs counter to his central point, that the first attachment of a language to the world never forces any consequences on us, because what we do next will always depend on what we find it natural to do and not on any fixed structure ofreality. So the remainder of this paper will be concerned with two matters, Wittgenstein' s views about the relation between mind and brain, and his views about rule-following. In each case the goal will be a clearer picture ofhis line of demarcation between philosophy and science. The first ofthese two inquiries may be launched by a familiar question. If the philosophy of mind investigates mental concepts, what is its relation with neurology? Even ifthe further development of a language is not dictated by the metaphysical structure of the world, it might weIl be influenced by its physical structure. For example, there seems to be nothing in Wittgenstein' s later philosophy which would WITTGENSTEIN ON PHILOSOPHY AND SCIENCE 29 lead one to expect hirn to reject the suggestion that the concept of memory might develope in this way, or even to reject the suggestion that memory al ready involves general stipulations about the physical connection between the original perception and its later representation in the subject's mind. Even if there is no metaphysical force governing the development of mental concepts, they are surely sensitive to scientific discoveries. It is, therefore, surprising to find passages in which these natural suggestions seem to be rejected. In one ofthem he maintains that, when I remember someone's name, my brain may not be in astate caused by my being told his name in the past, and the only difference between me and someone else who had forgotten it might be that I would produce it if asked, but he would not. 14 In another passage he says that there may weIl be "no process in the brain connected with associating or thinking; so that it would be impossible to read off thoughts from brain-processes. I mean this: if I talk or write, there is, I assurne, a system of impulses going out from my brain and corre1ated with my spoken or written thoughts. But why should the system continue further in the direction of the centre? Why should this order not proceed, so to speak, out of chaos?"15 It is not clear how far Wittgenstein wants to go in these passages. It is possible that he is just speculating that the central operations ofthe brain may be chaotic. This would hardly be a contribution to neuro-physiology, but it would show something about the concepts of "associating" and "thinking", as we now have them: their use is compatible with this speculation, and they do not, at present, have the more complex structure of physical dispositions. But, of course, the speculation may turn out to be false, because the central operations of the brain may be discovered to be governed by causal laws. What will happen then? According to hirn it is not a philosopher's business to answer this question. Science constructs conceptual suburbs around the old city of ordinary language,16 but it is science itself and not philosophy that predicts such developments. All that a philosopher can do is to ask hirnself which deve10pment he would find natural if a particular realized speculation turned out to be false. E.g. if Wittgenstein's neurological speculation turned out to be false, his answer might well be that he would find it natural to assimilate mental concepts to physical dispositions. Others have found more exciting implications in Wittgenstein' s remarks about brain-processes - for example, that any such conceptual fusion of psychology and philosophy is impossible. But how could he have supported such a claim? Some have taken his supporting argument to have been that, when a mental concept seemed to be including its physical basis, what would reaIly 30 DAVID PEARS be happening would be its replacement by a different concept. However, that would only affect the description of the change and not the change itself. On the other hand, it must be admitted that the tone of his writing and one of the illustrations that he uses suggest that he is going beyond the point marked by my conservative interpretation. The illustration is taken from botany: according to hirn, the seeds oftwo species oftree might differ only in their origins and their later developments, and it is just a prejudice to insist that there must be an actual physical difference between them, derived from their divergent origins and the cause of their divergent developments. It is hardly necessary to observe that a geneticist would find this view entirely natural and Wittgenstein's view a prejudice. But really these speculations have little importance and the question that a philosopher must answer is, "What reaction would come most naturally to us if an underlying physical mechanism were discovered?" This is one of those cases where Wittgenstein's thought has acquired a momentum which seems to take hirn further than he can legitimate1y go. He cannot legitimately claim that there is any metaphysical force dictating the answer to the question. The momentum which carries hirn perilously near to making that claim comes from his successful polemic against the postulation of mental causes to explain such achievements as understanding the meaning ofa word. It looks as ifthe same strategy ought to be equally effective against the suggestion that the physical causes of mental performances might be discovered. This ho pe is encouraged by the obvious fact that no explanation extracted from physical causes could possibly yield the kind ofinsight that he sought as a philosopher. According to hirn, a typical philosophical investigation starts when people get lost in the intricacies of their own thoughts; not, of course, when they are actually thinking them, but when they pause to reflect on them. Their feeling then is that, though they experience no difficulty in thinking the thoughts, they cannot see how they are related to one another. They know their way around but they cannot see the relations between their own well-practised movements. The discovery ofthe physical basis ofmental phenomena will obviously do nothing to dispel this kind of perplexity. These considerations may explain the over-run of Wittgenstein' s thought or, more accurate1y, of its apparent implication - but they cannot justify it. Consider, first, his argument against the postulation of amental cause to explain such achievements as understanding the meaning of a word. It is, he argues, no good postulating a picture or a formula in a person' s mi nd to explain his grasp of a word's application. His point is not that such things never occur in the mind, or that they are no he1p when they do occur, but only WITTGENSTEIN ON PHILOSOPHY AND SCIENCE 31 that they cannot provide a complete explanation ofthe speaker's application of the word. The reason for the shortfall is simply that the picture, or the words in the formula themselves, have to be applied. 17 So it is amistake to suppose that amental cause of this kind could automatize a speaker's practice and eliminate all his difficulties of application. The mi stake is the offering of an explanans which itself re-introduces the explanandum, and so starts an infinite regress. However, this argument cannot be adapted for use against the suggestion that the neural basis ofa speaker's grasp ofthe application ofa word might be discovered, and even inc1uded in the mental concept. In this new area there is no point of contact between the cause and the application of the word as it is for the speaker when he makes it. In the other area there is a point of contact, because the mental picture can serve hirn as an aid, and that is the point of departure for Wittgenstein's argument, that such an aid could never become a complete guide for hirn, altogether replacing his basic discernment of the correct applications. But in this new area there never is any possibility of his using the physical cause as a guide. So there is no point of departure here for any parallel argument. To put the point in another way: a neural explanation ofthis kind would have no tendency to show that the speaker's proficiency was not basic. This is important, because his insistence that the speaker's reasons for his applications run out l8 is surely right. If this interpretation of Wittgenstein's remarks on the relation between mind and brain is correct, there is no need to suppose that he put mental language under any metaphysical constraint, or to infer that he deviated at this point from his settled, later view about the line dividing philosophy from science. But if we really want to trace that view to its origin, we must seek it in his later account ofrule-following. That is the next and last topic of this paper. In Philosophical Investigations Wittgenstein places the examination ofrulefollowing immediately after the critique ofhis early theory ofmeaning. There is an obvious point in this collocation: the system of the Tractatus is a target of the reductio which is the centre of the negative part of his later treatment of rule-following. It is only one ofthe targets, because other forms ofPlatonism are also under attack. For in his later book he repudiates any theory which appeals to an unspecifiable metaphysical paradigm to explain the constraint felt by rule-followers. All such theories merely project this feeling onto the world and do nothing to explain it. In fact, if his reductio is convincing, they offer the rule-follower a criterion of correctness which he could not possibly use. 32 DAVID PEARS "Whence comes the idea that the beginning of aseries is a visible section of rails laid to infinity? Weil, we might imagine raUs instead of a rule. And infinitely long rails correspond to the unlimited application of a rule." "All the steps are already taken" means: I no longer have any choice. The rule, once stamped with a particular meaning traces the line along which it is to be followed through the whole of logical space - But if something ofthis sort really were the case, how would it help?19 The objection is that this is a picture of guidance without any usable exemplification. Before we ask for the argument with which he supports this objection, it is worth taking one more look at the Tractatus. The point was made above20 that he there took the view that the world imposes a fixed structure on any language, so that, once we have made a linguistic choice, its consequences are forced on uso However, that was aremark made about the general structure of our language. If we ask whether he would have said the same thing about the application of a particular name to a particular object, it is not so clear what his answer would have been. Certainly, the point of the picture theory is that a name picks up from the object to which it is attached certain constraints on its subsequent use. But no account is given ofthe way the name is supposed to do this, nor is there any discussion of its re-applications or of the criterion of identity on which they would depend. So the later attack on Platonism, which does deal with the repeated applications of descriptive words, makes a somewhat oblique impact on the Tractatus. The objection to Platonism which is developed in Philosophical Investigations exploits the fact that any aid in the rule-follower's mind, such as a picture or an analytical formula, will itself have to be applied, and so will raise again the same problem that was raised by the original word. 21 So ifthe Platonist appeals to such aids - and he must point to something in the speaker' s mi nd which will reflect the paradigm outside his mind - he will never be able to explain the distinction between correct and incorrect applications of the word. The naturalness of certain applications is an essential part of the basis of the distinction. But the Platonist tries to dispense with it, and so in his theory the difference between obeying and disobeying a rule disappears. He will not allow people's natural tendencies to make any contribution to what counts as using the word correctly, but he cannot find any effective substitute for them. What we have found so far in Wittgenstein's later treatment of rule-following is the rejection of an assimilation of philosophy and science which exhibits a now familiar pattern. The task of philosophy is to investigate language in use, and, when we use a language, we feel ourselves under the constraint of WITTGENSTEIN ON PHILOSOPHY AND SCIENCE 33 the rules governing its words. So we theorize about the source of this constraint, modelling our theory on science. However, Platonism makes its source inaccessible - a point made by Aristotle against Plato himself - and Wittgenstein's achievement in the negative part of his treatment of rule-following was to demonstrate that, even if the Platonic paradigm had been accessible, it would have been no help. He also had a positive contribution to make to the investigation of rulefollowing, a contribution which takes his demarcation of science and philosophy one stage further. If we ask why anyone would postulate Platonic paradigms to explain the human practice of using words according to rules, we immediately open up a new approach to Wittgenstein's demarcation of philosophy and science. The philosophical investigation of rule-following is likely to start as a search for an empirically accessible criterion of correct usage which will confine itself to material available to speakers of a language. This search immediately runs into a difficulty: when people speak a language, they cannot always rely on specifiable criteria of correctness and must, in the end, trust their own competence to apply the words in its vocabulary. But this can easily seem to be a lamentable weakness in the foundations of language. So the Platonist steps in and offers to satisfy the demand for security by drawing on the resources of another world. However, that is not the whole story. For the really fundamental mistake made in this area is to demand a single criterion of correctness in any world. For example many commentators believe that Wittgenstein himselfproposed a single criterion in this world, agreement in judgements with other members of the community speaking the language. But in fact, he never goes quite so far in that direction, and the reasons why he stops short of such a theory are interesting. First, it would be a theory which took no account of the constraints imposed on the community by the world with which their language was concerned. But the correction of the mistake would not lead to a better theory about the ultimate touchstone of correct usage recognized by speakers of a language. For the only way to correct the mi stake would be to go beyond the speaker's point ofview and try to see what they are doing from the outside, in the wider context of their lives. So instead oftheorizing about rule-following, Wittgenstein simply describes the phenomenon. He presents language as a system of practices exhibiting regularities which cannot be tested against any single touchstone. That does not mean that any move, however bizarre, is permissible in a language game. For it is essential that the regularities be a viable part of human life in the world. That requires them to be sensitive both to the constraints imposed by 34 DAVID PEARS the world and to the constraints imposed by other people: they must promote our activities and the satisfaction of our needs and they must make communication possible. However, these general requirements cannot be captured in any single criterion of correctness in this world or in Plato's. They preserve the regularity of linguistic practices without giving the practitioners any single point of orientation. What this account of rule-following offers us is a way of avoiding the excessive intellectualization ofrule-following which can be found in most other treatments of it. Even the Tractatus had offered a theory about the fixed structure of the world to explain our feeling that when we use language we are under a completely external constraint. That theory, it is true, was only concerned with the general structure of the world, but its extension to particular features, such as the system of colours, is a step that is easily taken. The idea is that the task confronting the speaker of any language is a purely intellectual one: everything that keeps his language on the rails must be picked out by hirn and given a place on his list of criteria. Wittgenstein's account of rule-following offers a very different picture: an essential part ofthe basis of the regularities exhibited by language is human nature. If that seems too subjective, we should ask ourselves why we assume that everything objective needs to be included in the speaker's picture ofthe world as he finds it, and nothing in his finding it so. This is not the place to develop an interpretation of Wittgenstein's idea that language is based on apre-existent system of perceptions and actions the idea expressed in his remark, "In the beginning was the deed".22 The point that needs to be made here is only that he is rejecting other philosophers' excessive intellectualization of the problem of rule-following. If they confine themselves to the point of view of people actually using language, they will never be able to give a complete explanation of the regularity of their practices. At this point it will be objected that, if this is what Wittgenstein is doing, he must have erased the line separating philosophy from science. For this part of Philosophical Investigations seems merely to contain a mixture of psychology and anthropology. It is, of course, true that his observations about the place of language in our lives are factual. However, their aim is not the construction of any theory. They are offered as an antidote to the intellectualist prejudice, that a complete account ofrule-following can be given by a philosopher who confines himself to the point of view of a speaker following a rute. That, of course, was the prejudice of analytic philosophy. Wittgenstein's strategy is to suggest that we WITTGENSTEIN ON PHILOSOPHY AND SCIENCE 35 should not restrict ourselves to the rule-follower's inside story but should put it in a wider context which includes things which do not always, and, in some cases, could not ever, serve hirn as criteria of correct usage. These suggestions are not a contribution to any science, but of course, it is easy to cross over into science at this point without noticing the transition. For there are many tempting empirical questions just off. the narrow path trodden by Wittgenstein. Could someone solitary from birth set up a language for his own use? How much idiosyncrasy is allowed by the constraint of communication? How fast can the meaning of a word change? Interpretations of his later philosophy often stray into these scientific investigations, but it is surely possible to see how his observations could be factual but not steps towards the construction of any theory, and yet not in the least mysterious in their import and purpose. They are directed against the common philosophical tendency to over-intellectualize the practice of rule-following, and, if Wittgenstein is right, it did not take a theory to correct that mistake. Christ Church Oxjord NOTES 1 See Tractatus logico-philosophicus (TLP), ed. by D. F. Pears & B. F. McGuinness, Routledge & Kegan Paul, London 1961, 6.3-Q.372. 2 TLP, 4.11. 3 TLP, 6.32--{5.36. 4 TLP, 3.262 and 3.326-3.27 5 I.e. science in the narrow sense, in which it does not include the whole of factual discourse. 6 TLP, 6.342. 7 TLP, 6.372. 8 TLP, 6.124. 9 Which are identified with the senses of sentences in TLP. See 2.202-2.221. 10 See Wittgenstein, 'Blue Book', pp. 63ff., in The Blue and Brown Books, ed. and with aPreface by R. Rhees, B. Blackwell, Oxford 1958. 11 See his remarks on the nature of philosophy in Vermischte Bemerkungen. Culture and Value (CV), ed. by G. H. von Wright in collaboration with H. Nyman. Trans. by P. Winch, B. Blackwell, Oxford 1980, e.g. pp. 14--17. 12 TLP, 6.124. 13 TLP, 3.342. 14 Zettel (Z), ed. by G. E. M.Anscombe & G. H. von Wright, B. Blackwell, Oxford 1967, § 610. See Bemerkungen über die Philosophie der Psychologie. Remarks on the Philosophy ofPsychology, Vol. I (RPP I), ed. by G. E. M. Anscombe & G. H. von Wright. Trans. by G. E. M. Anscombe, B. Blackwell, Oxford 1980, § 905. 36 DAVID PEARS Z, § 608. See RPP I, § 903. See Philosophische Untersuchungen. Philosophical Investigations (PI), ed. by G. E. M. Anscombe& R. Rhees. Trans. byG. E. M.Anscombe,B. Blackwell, Oxford 1953, I, § 18. 17 PI I, §§ 139-14l. 18 PI I, §§ 217; 485. 19 PI I, §§ 218-219, I st paragraph. 20 See above, p. 24. 21 See above, p. 32. 22 CV, p. 31. 15 16 ANTHONY LP. KENNY WITTGENSTEIN ON MIND AND METAPHYSICS Wittgenstein is often regarded as being both positivist and behaviorist: positivist in rejecting all metaphysics, and behaviorist in denying inner human life. So far as concerns philosophy of mind, this view is based on a misunderstanding of Wittgenstein's work. He did indeed attack one particular metaphysical theory of mind: the Cartesian theory. Cartesianism is metaphysical in the sense of isolating statements about mental1ife from any possibility of verification or falsification in the public world. But much of Wittgenstein's work in philosophy of mind is devoted to showing the importance of distinctions between different kinds of potentiality and actuality. These distinctions were one of the major concerns of the work of Aristotle which was the first book to bear the name Metaphysics, and were a main target of c1assical antimetaphysicians. In this sense Wittgenstein himselfhad a metaphysics of mind; and the metaphysical sensitivity which he shared with Aristotle was what enabled hirn to reject Cartesianism without falling into behaviorism. In this paper I will try to illustrate different forms of metaphysics, and sketch Wittgenstein' s attitude to each. In his mature writings, Wittgenstein does not often speak of metaphysics. The word "metaphysical" occurs only twice in the Philosophische Untersuchungen.! In each case it has a pejorative sense. And Wittgenstein can sum up his method as being a corrective to metaphysics: Wenn die Philosophen ein Wort gebrauchen - "Wissen", "Sein", "Gegenstand", "Ich", "Satz", "Name" - und das Wesen des Dings zu erfassen trachten, muss man sich immer fragen: Wird denn dieses Wort in der Sprache, in der es seine Heimat hat,je tatsächlich so gebraucht? - Wir führen die Wörter von ihrer metaphysischen, wieder auf ihre alltägliche Verwendung zurück. (PI I, § 116). Metaphysics here seems to be identified with the search for essences. But there can be a legitimate attempt to understand essences, on which Wittgenstein himse1f is engaged: [W]ir auch in unsem Untersuchungen das Wesen der Sprache - ihre Funktion, ihren Bau - zu verstehen trachten ... (PI I, § 92). What is wrong is to consider the essence, not as something which lies open to view and must merely be given a perspicuous description, but something 37 R. Egidi (ed.), Wittgenstein: Mind and Language, 37-46. © 1995 Kluwer Academic Publishers. 38 ANTHONY J.P. KENNY interior and hidden: a kind of metaphysical clockwork or hardware which explains the functioning of mind and language. The kind of metaphysics which Wittgenstein systematically attacks is the metaphysics which consists of grammar masquerading as science. One source of metaphysics is the philosopher's temptation to mimic the claims and methods of science. Metaphysics, in this sense, is a quasi physics, an imaginary physics elevated into something sublime and mysterious. The tendency to create this kind of metaphysics is weH described in PI I, § 36, where Wittgenstein is talking about ostensive definition. Und wir tun hier, was wir in tausend ähnlichen Fällen tun: Weil wir nicht eine körperliche Handlung angeben können, die wir das Zeigen auf die Fonn (im Gegensatz z.B. zur Farbe) nennen, so sagen wir, es entspreche diesen Worten eine geistige Tätigkeit. Wo unsere Sprache uns einen Körper vennuten lässt, und kein Körper ist, dort, möchten wir sagen, sei ein Geist. In allusion to this passage, we might call this kind of metaphysics spiritualistic metaphysics. The passage al ready cited in which Wittgenstein describes the task of the philosopher as being to bring back words from their metaphysical usage itself comes in a section where Wittgenstein has been talking about the tendency to think ofthe proposition as something sublime, to put a halo around thought, to think of logic as a structure of crystalline purity. (PI I, §§ 74-108) We think ofthe mind as a mysterious medium, different from a physical medium, where the strict laws of logic operate. Let us look at some examples of spiritualistic metaphysics ofthe kind which was Wittgenstein's target. The metaphysical impulse may lead us to postulate spiritual substances, or spiritual processes. In each case we are misled by grammar: where it makes us expect a physical substance, but there is not one, we invent a metaphysical substance; where it makes us expect an empirical process, but cannot find one, we postulate an incorporeal process. (PI I, § 339). First, metaphysical substances. One ofthe most bizarre, as weB as the most ubiquitous, misunderstandings of the nature of the mind is the picture of mind's relation to body as that between a little man or homunculus on the one hand and a tool or instrument on the other. We smile when medieval painters represent the death of the Virgin Mary by showing a small scale model virgin emerging from her mouth: but basically the same idea can be found in the most unlikely places. Descartes, when first he reported the occurrence of retinal images, warned us not to be misled by the resemblance between images and their objects into thinking that when we saw the object we had another pair of eyes, inside the WITTGENSTEIN ON MIND AND METAPHYSICS 39 brain, to see the images. But he himself believed that seeing was to be explained by saying that the soul encountered an image in the pineal gland. This was a particularly striking version of what has been nicknamed "the homunculus fallacy": the attempt to explain human experience and behaviour by postulating a little man within an ordinary man. We humans are always inclined to explain things we only imperfectly understand in terms of the most advanced technology of the age we live in. As time passes and technology progresses the tool or instrument which the manikin is fancied to control gets more and more sophisticated. Thus Plato thought that the soul in its relation to the body could be compared with a sailor in a boat or acharioteer holding the reins. Many centuries later Coleridge said that what poets meant by the soul was "a being inhabiting our body and playing upon it, like a musician enc10sed in an organ whose keys were placed inwards". (Letters, i 278). More recently the mind has been compared to a signalman pulling the signals in his signal box, or the telephone operator dealing with the incoming and outgoing calls in the brain. Most recently the boat, the chariot, the railroad and the telephone exchange have given way to the computer, so that the relation ofthe soul to the body is envisaged as that of the programmer who writes the software to the hardware which executes the program. What is wrong with the homunculus fallacy? In itself there is nothing misguided in speaking of images in the brain, if one means patterns in the brain which can be mapped onto features of the sensory environment. There is nothing philosophically objectionable in the suggestion that these schematic images may be observable to the neurophysiologist investigating the brain. What is misleading is to say that these images are visible to the soul, and that seeing consists in the soul's perception ofthese images. The misleading aspect is that such an account pretends to explain seein"g, but the explanation reproduces exactly the puzzling features which it was supposed to explain. For it is only if we think of the relation between a soul and an image in the pineal gland as being just like the relation between a human being and pictures seen in the environment that we will think that talk of an encounter between the soul and the image has any illuminating power at aH. As a metaphor, manikin talk may be no more than a harmless necessary fancy; but as an element in a theory a manikin bedevils understanding. For whatever needs explaining in the behaviour of the man turns up, grinning and unexplained, in the shape of the manikin. From an example of a metaphysical substance, let us turn to considering metaphysical processes. Wittgenstein discusses the question: Is understanding 40 ANTHONY J.P. KENNY amental process? (pI I, § 151ft). Some philosophers have thought that understanding was a psychological process in the same sense as we might call "a psychological process" the reciting of a poem in one's head. But reflection soon shows that this is not so. If meaning was amental process accompanying the utterance of a sentence, it should be possible for the process of meaning to take place without the sentence being uttered at all. Can one, in fact, perform the act of meaning without uttering the sentence? If you try to do so, you are likely to find yourself reciting the sentence itself under your breath' But of course it would be absurd to suggest that simultaneously with every public utterance of a sentence there is a private one too: it would surely take great skill to ensure that the two processes were exactly synchronized with each other! And how terrible if the two came slightly out of synchrony, so that the meaning of one word got mistakenly attached to the next one! Moreover, the question whether somebody understands a sentence, and whether she really means it, can be raised about sentences uttered in the privacy of the imagination no less than about sentences uttered before a public audience. Infuriated by a curmudgeonly relation, I may mutter to myself "I wish he would drop dead!" Luckily, I don't mean it. I hum in my mind a Russian folk-song, enchanted by the sound of the words. But I haven't the faintest idea what they mean. If understanding and meaning were processes, they would have to accompany private utterances. So ifthe processes involved were some kind of inner utterance, we would be set off on an endless quest for the real understanding. Some philosophers have thought that understanding was amental process in rather a different sense. They have conceived the mind as a hypothetical mechanism postulated to explain the observable intelligent behaviour of human beings. If one conceives the mind in this way one thinks of amental process, not as something comparable to reciting the ABC in one's head, but as a process occurring in the special mental machinery. The process on this view is amental process because it takes pi ace in a medium which is not physical; the machinery operates according to its own mysterious laws, within a structure which is not material but spiritual; it is not accessible to empirical investigation, and could not be discovered, say, by opening up the skull ofa thinker. Such processes need not, on this view, be accessible either to the inner eye of introspection: the mental mechanism may operate too swiftly for us to be able to follow all its movements, like the pistons of a railway engine or the blades of a lawn mower. But we may feel that if only we could sharpen our faculty for introspection, or somehow get the mental machinery to run in WITTGENSTEIN ON MIND AND METAPHYSICS 41 slow motion, we might be able actually to observe the processes of meaning and understanding. According to one version of the mental-mechanism doctrine, understanding the meaning of a word consists in calling up an appropriate image in connection with it. In general, of course, we have no such experience when we use a word and in the case ofmany words (such as "the", "if', "impossible", "one million") it is difficult even to suggest what would count as an appropriate image. But let us waive these points, allow that perhaps we can have images in our mind without noticing that we do, and consider only the kind of word for which this account sounds most plausible, such as colour words. We may examine the suggestion that in order to understand the order "Bring me a red flower" one must have a red image in mind, and that it is by comparison with this image that one ascertains which flower to bring. This cannot be right: otherwise how could one obey the order "imagine a red patch?" Whatever problems there are about identifying the redness ofthe flower recur with identifying the redness of the patch. It is of course true that when we talk mental images often do pass through our minds. But it is not they which confer meanings on the words we use. It is rather the other way round: the images are like the pictures illustrating a text in a book. In general it is the text which tells us what the pictures are of, not the pictures which tell us what the words of the text mean. In fact, understanding cannot be thought of as a process at all. Understanding is a kind of ability, and therefore is astate rather than a process (PI I, p.59). In so far as the exercise ofunderstanding is an exercise ofintelligence, we may call understanding amental state. But it is important to guard against misunderstanding here. Understanding may be amental state, but it is not a psychological state like pain or depression or excitement. Such states last over periods, and can be continuous or interrupted; but one cannot know uninterruptedly what a word means. Wittgenstein's treatment of the many-faceted illusion that understanding is amental process is an example of his critique of spiritualistic metaphysics. He attacks metaphysics not by the blunt instrument of some positivistic verification principle, but by the careful drawing of distinctions which enable hirn to disentangle the mixture oftruism and nonsense in the metaphysician's concept of mind. Besides spiritualistic metaphysics, there is another kind of metaphysics to which Wittgenstein was implacably opposed. This is the view that there is a fundamental branch of philosophy which underlies and underpins the rest of philosophy and the rest ofthe sciences. We might call this kind of metaphysics 42 ANTHONY LP. KENNY foundationalist metaphysics. Descartes, who was the arch-exponent of spiritualistic metaphysics, can also be taken as a spokesman for foundationalist metaphysics. "The whole of philosophy - he wrote - is like a tree, whose roots are metaphysics, whose trunk is physics, and whose branches are all the other sciences". Not only Descartes, but many other thinkers have seen philosophy as an ordered system; a system which could perhaps be most perspicuously displayed by being cast into axiomatic form, as Spinoza tried to do. Wittgenstein's Tractatus has reminded many people of Spinoza; but his later philosophy was the very reverse of systematic. This does not mean that it lacked method, or rigour. It means rather that there was no part of philosophy which had primacy over any other part. One could start philosophising at any point, and leave off the treatment of one problem to take up the treatment of another. Philosophy had no foundations, and did not provide foundations for other disciplines. Philosophy was not a house, nor a tree, but a web. This is how we are to understand the famous passage PI § 133: Die eigentliche Entdeckung ist die, die mich fähig macht, das Philosophieren abzubrechen, wann ich will. - Die die Philosophie zur Ruhe bringt, so daß sie nicht mehr von Fragen gepeitscht wird, die sie selbst in Frage stellen. - Sondern es wird nun an Beispielen eine Methode gezeigt, und die Reihe dieser Beispiele kann man abbrechen. - Es werden Probleme gelöst (Schwierigkeiten beseitigt), nicht ein Problem. In refusing to countenance systematic or foundationalist metaphysics, Wittgenstein was distancing himself from many of the great philosophers of the past, inc1uding Aristotle. For Aristotle there was a philosophical discipline which deserved the title "First Philosophy"; and the attempt to delineate this is a central theme in the collection of treatises which we know as "The Metaphysics of Aristotle". Sometimes first philosophy is described as the discipline which studies being qua being; sometimes as the discipline which studies being qua divine. The two formulations probably are two ways of describing a single enterprise: one accounts for everything that is the case about everything there is by appealing ultimately to the divine movers unmoved, and "the study of being qua being" describes this investigation in terms of its explicandum, while "the study ofbeing qua divine" describes it in terms ofits explicans. But however one understands Aristotle's first philosophy it is c1ear that Wittgenstein would have accepted neither its methodological presuppositions nor its foundational role. However, a considerable part of Aristotle's Metaphysics (and ofhis other works which would nowadays be described as metaphysical) is devoted to a philosophical activity which resembles quite c10sely Wittgenstein's own WITTGENSTEIN ON MIND AND METAPHYSICS 43 method. The distinction between actuality and potentiality, and the c1assification of different kinds of potentiality, is universally recognized (by both friend and foe) as being one of Aristotle's most characteristic contributions to philosophy and in particular to the philosophy of mind. His distinctions were later systematised by scholastic philosophers in the middle ages. We might call the systematic study of actuality and potentiality dynamic metaphysics. Thus, according to Aristotle, active powers (e.g. the power to heat) differed from passive powers (e.g. the power to be heated). Natural powers (such as water's capacity to wet) were to be distinguished from rational powers (such as a pharmaeist's ability to prescribe). Natural powers needed certain preconditions for their exereise: fire will only bum wood ifthe wood is suffieiently dry. But if these conditions are met, then the power will infallibly be exercised. The ease is not the same with rational powers. A pharmacist may have the skill to preseribe, and may have the necessary pharmacopea; but he may fail to preseribe ifhis patient does not have suffieient funds. Natural powers, unlike rational powers, are also tendencies to aet in a certain mann er. The possession of rational powers, aceording to Aristotle, is peeuliar to human beings. Among the powers of humans there are some which are innate - the senses, for instanee - while others like the ability to play the flute, are acquired by praetiee. The liberal arts, and in general the skills whieh are the fruits of edueation, are powers of a partieular kind, namely dispositions. Dispositions are abilWes whose exercises are the relevant scientifie, artistie and eraft aetivities; but they are themselves actualisations ofthe eapacity to leam whieh is presupposed by edueation. They ean thus be ealled aetualisations as weH as potentialities. Medieval philosophers introdueed a teehnieal terminology here: the skills were first or primary aetualizations in contrast to the episodie employments of the skills whieh were secondary actualizations. Thus the ability to speak Greek is a first aetualization, while the actual utterance of a Greek statement or eommand, or the understanding of a particular Greek text on hearing it, is a secondary actualization. Wittgenstein undertook a prolonged investigation of the nature of potentiality in the 'Brown Book', where sections 58-67 are devoted to various language games with the word "ean'? The distinetions whieh he draws, in writing on the philosophy of mind, between proeesses and states, and between different kinds of states, eorrespond to the Aristotelian distinctions between kinesis, hexis, and energeia, and the criteria by which the distinctions are made often eoincide. The example which Wittgenstein diseusses at length, to discuss the 44 ANTHONY J.P. KENNY relationship between apower and its exercise, namely leaming to read (PI I, §§ 156ft) is the same as the standard Aristotelian example of amental hexis, namely, knowledge of grammar. In addition to the Aristotelian distinction between powers, their possessors, and their exercises, we may introduce the notion of the vehicle of a power or ability. The vehicle of an ability is the physical ingredient or structure in virtue of which the possessor of an ability possesses the ability and is able to exercise it. The distinction between abilities and their vehicles is not something which is peculiar to human beings and their abilities. Vodka has the power to intoxicate: the vehicle ofthe power ofvodka to intoxicate is the alcohol the vodka contains. A vehicle is something concrete, something which can be weighed and measured. An ability, on the other hand, has neither lenght nor breadth nor location. This does not mean that an ability is something ghostly: my front-door-key's ability to open my front door is not a concrete object, but it is not a spirit either. An important instance of the distinction between possessor, ability, and vehicle is the distinction between people, their minds, and their brains. Human beings are living bodies of a certain kind, with various abilities. The mind, as we have said, is the capacity to acquire or possess intellectual abilities. The vehicle of the human mind is, very likely, the human brain. Human beings and their brains are physieal objects; their minds are not, because they are capacities. Onee again, to say that the mind is not a physieal object is not to say that it is a ghostly spirit: denying that the mind has a length or breadth or loeation does not involve one in spiritualistic metaphysies. In every age sinee Aristotle, philosophers have been tempted to blur the distinetions he made. In philosophy there is a perennial temptation to reduee potentialities to aetualities. Some philosophers attempt to reduce powers to their exercises: thus, explieitIy, David Hume, who said the distinetion between apower and its exereise was frivolous. Some philosophers attempt to reduee powers to their vehicles: thus, implieitly, Deseartes, who wanted to identify the powers of bodies with their geometrieal properties. Philosophical errors about eapacities in general show up partieularly vividly when they oeeur in the philosophy of mind. Applied in this area, exercisereduetionism beeomes behaviourism: the attempt to identify mind with behaviour consists in treating the eomplex seeond-order capacity which is the mind as if it was identicaI with its particular exercises in behaviour. Applied in this area, vehicle-reduetionism beeomes materialism: the attempt to identify mind with brain eonsists in redueing my mental eapaeities to the parts and structures of my body in virtue of whieh I possess those eapacities. WITTGENSTEIN ON MIND AND METAPHYSICS 45 Materialism is a grosser philosophical error than behaviorism because the connection between a capacity and its exercise is in truth a more intimate one than the connection between a capacity and its vehicle. In the case of the mind, the connection between capacity and exercise is a conceptual connection: one could not understand what the mind was if one did not understand what kinds of thing constitute the exercise of mental capacity. The connection between capacity and vehicle, on the other hand, is a contingent one, discoverable by empirical science. Aristotle's grasp of the nature of mind will stand comparison with that of any subsequent philosopher; but he had a wildly erroneous idea ofthe relationship ofthe mind with the brain, which he believed to be an instrument to cool the blood. Wittgenstein rejected both behaviourism and materalism: in the area of philosophy of mind he is closer to Aristotle and his scholastic successors than he is to any ofthe more fashionable philosophies of our scientific and scientistic age. In one of his most characteristic, and most striking remarks, he goes so far as to entertain the possibility that some of our mental capacities may lack a vehicle altogether. Keine Annahme scheint mir natürlicher, als daß dem Assoziieren oder Denken kein Prozess im Gehirn zugeordnet ist; so zwar, daß es also unmöglich wäre, aus Gehirnprozessen Denkprozesse abzulesen. Ich meine das so: Wenn ich rede oder schreibe, so geht, nehme ich an, ein meinem gesprochenen oder geschriebeben Gedanken zugeordnetes System von Impulsen von meinem Gehirn aus. Aber warum sollte das System sich weiter in zentraler Richtung fortsetzen? .. Es ist also wohl möglich, daß gewisse psychologische Phänomene physiologisch nicht untersucht werden können, weil ihnen physiologisch nichts entspricht. ... Warum soll es keine psychologische Gesetzmässigkeit geben, der keine physiologische entspricht? Wenn das unsere Begriffe von der Kausalität umstößt, dann ist es Zeit, daß sie umgestoßen werden. 3 Here we see Wittgenstein both at his most antimetaphysical and at his most metaphysical. He is antimetaphysical in the sense that he is making a frontal attack on the scientism characteristic of our age: the assumption that there must be physical counterparts ofmental phenomena. Yet he is here also highly metaphysical. He is not embracing spiritualistic metaphysics: what does the associating, thinking, and remembering is a human being with a body, not aspiritual substance. But in the sense of dynamic metaphysics, Wittgenstein is here very c10se to Aristotle. For he is envisaging as a possibility a pure Aristotelian soul, or entelechy, which operates with no material vehic1e: a formal and final cause to which no mechanistic efficient cause corresponds. We can sum up Wittgenstein's attitude thus: he was hostile to spiritualistic metaphysics and to foundationalist metaphysics; but he was one of the most ANTHONY J.P. KENNY 46 consummate practitioners of the dynamic metaphysics which is one strand of the Aristotelian tradition. This can be brought out, finally, by pointing to one of the passages in which Wittgenstein most explicitly rejects spiritualistic metaphysics. In this very passage we find that, in the Aristotelian sense, he gives a surprisingly metaphysical formulation of the relationship between soul and body. Nur vom dem, was sich benimmt wie ein Mensch, kann man sagen das es Schmerzen hat. Denn man muß es von einem Körper sagen, oder, wenn du willst, von einer Seele, die ein Körper hat. Und wie kann ein Körper eine Seele haben? (PI I, § 283). How striking, that a body' s having a soul should seem more problematic than a soul' s having a body! Balliol College Oxford NOTES I Philosophical Investigations (PI), ed. by G. E. M. Anscombe & R. Rhees. Trans. by G. E. M. Anscombe, B. B1ackwell, Oxford 1953. 2 The Blue and Brown Books, ed. and with aPreface by R. Rhees, B. Blackwell, Oxford 1958. 3 Zettel, ed. by G. E. M.Anscombe & G. H. von Wright. Trans. by G. E. M.Anscombe, B. Blackwell, Oxford 1967, §§ 608-610. JOHANNES L. BRANDL WITTGENSTEIN'S ALLEGED METAPHYSICS OF MIND Anthony Kenny suggests in his stimulating paper Wittgenstein 's Metaphysies 0/ Mind that Wittgenstein's attack on Cartesianism is driven by a metaphysical view similar to Aristotle's theory ofpotentiality and actuality. He recommends using the threefold distinction between possessor, vehicle and exereise 0/ an ability for interpreting Wittgenstein' s view of mental events. As a heuristic device, this may prove useful, but I doubt that from this proposal there will emerge anything that may legitimately be called 'Wittgenstein's Metaphysics of Mind' . 1. THREE TYPES OF METAPHYSICAL CLAIMS What is a metaphysics of mind? That depends on what we mean by the term 'metaphysics'. I am ready to accept, with Kenny, the positivist criterion according to which a claim is metaphysical conceming some domain if it is neither verifiable nor falsifiable by observations within this domain. 1 Since the domain of possible observations can be more or less restricted, we should distinguish various types of metaphysical claims: 1) If a claim is neither verifiable nor falsifiable conceming any domain of possible observations, such a claim may be called purely speeulative. 2) If a claim is neither verifiable nor falsifiable by observations actually available at a certain stage of investigation, such a claim may be called investigation independent. 3) If a claim is neither verifiable nor falsifiable given the specific conditions that make observation possible for human beings, such a claim may be called transeendental. With this tripartite division in mind, let us see whether Wittgenstein says anything ab out the mental that could be classified in one of these categories. As concems category (l), it hardly needs arguing that Wittgenstein rejects claims ofthis sort as springing from a mythological, or quasi-scientific mindbody theory. It is typical of scepticism about other minds to raise doubts about facts whose presence is beyond any conceivable evidence. This form of scepticism takes mental states as something observable only by the inner eye of the subject. It thus assumes from the beginning that whatever I am directly 47 R. Egidi (ed.), Wittgenstein: Mind and Language, 47-56. © 1995 Kluwer Academic Publishers. 48 JOHANNES L. BRANDL aware of is inaccessible to anyone else. Conversely, I may have all sorts of beliefs about the minds of other persons, but I cannot find out whether these beliefs are correct. On the present assumption, the only way to do this would be to experience these mental states myself. But then they would no longer be the states of other minds, having become part of my own consciousness. Hence we can only speculate about what goes on in the mind of other people. Once the appropriate evidence for mental events has been restricted to the deliverances of an inner sense, claims about other minds turn out to be unverifiable in sense (1). The only facts which could tell in favour of their truth are facts about the way people act, and these have been excluded as the 'wrong' kind of evidence from the beginning. There can be little doubt that Wittgenstein has spotted the flaw in this kind of reasoning. Whenever he talks of'hidden phenomena' or 'hidden mechanisms' he wams against introducing mental entities which are accessible in no other way than by being aware ofthem immediately. For this reason he insists that instead of saying "There has just taken place in me the mental process of remembering..." we should rather say "I have just remembered... ".2 The point of this move is the following: When 1 talk of amental state (or process) in me I am not characterizing it in any particular way, but merely am hinting at what is typical of amental state in general. Ignoring this fact leads to confusion. Apparently my bodily states are 'in me' too; theyare my states after all. So the question arises: what is the contrast we draw by calling some states mental and others nonmental? It seems that this contrast will only emerge if 1 conceive of my bodily states not as states of me, but as states ofmy body (or processes in my body). Then 1 may conclude by parity ofreasoning that, strictly speaking, mental states are not my states either, but states of my mind (or processes in my mind). Thus I end up with the Cartesian distinction between two substances having different kinds ofproperties, mental and physical. The recommended translation should block this argument. Saying "I have just remembered... " should make it plain that 1 am talking about amental state ofme, not in me. The contrast between my bodily states and my mental states is not a contrast between states of my body and states of my mind. There is only one bearer for both kinds of states, namely me. The pronoun "I" in my memory report does not refer to amental entity, but to a person that has both mental and bodily states. Asked for the difference between these kinds of states 1 may claim some sort of privileged access to my mental, but not to my bodily states. This privilege does not exclude the possibility that amental state might make itself manifest in other ways too, besides occurring 'in my WITTGENSTEIN'S ALLEGED METAPHYSICS OF MIND 49 mind'. This is particularly clear in the case of feeling pains. Though this is a special form of cognitive awareness, one can know about pains also without being aware of them in this special way. If someone else cares for my pains, he knows about them, yet he does not share my feelings in pitying me. Ifwe observe Wittgenstein's advice not to talk ofprocesses in our minds, we may hope to escape scepticism about other minds. There simply is no domain of mental facts which are accessible only to the subject experiencing them. Since alt speculative metaphysics about the mind relies on such facts, there is a straightforward incompatibility between such a view and Wittgenstein's method. Whenever mental facts become the target of a sceptical argument, Wittgenstein advises us to change our way of describing these facts. Having redescribed them in the appropriate way, there is nothing left to concern a speculative metaphysician. Turning to the idea of metaphysics as telling us what things really are, I think it is equaIly clear that Wittgenstein has no such ambitions. The need to make investigation-independent claims about the mind arises once the accessibility of mental facts is admitted both from the first person and the third person perspective. Then we find ourselves in the position of Hume, expecting the sun to rise tomorrow morning, but not having fully sufficient reasons for a rigorous proof of this claim. In the same manner we expect other people to have pains when crying and to feel pleasure when they look pleased, but for aIl we know their feelings might be different. This form of scepticism is strong enough to turn anything we say ab out facts beyond our own consciousness into metaphysical claims of the investigation-independent sort. There is an ongoing debate over whether or not Wittgenstein endorses scepticism of a Humean variety. Unfortunately the arguments on both sides have become so complicated that it is impossible to rehearse them here. But I hope I can bring out one point on which aIl participants in that debate should be in agreement. This point suffices, I believe, to show that whether or not Wittgenstein was bothered by a sceptical paradox he certainly did not take re fuge in some sort of metaphysics in order to solve it. Consider the foIlowing situation. I am sitting in a bar waiting for the bill. I had two glasses of wine and I know that each of them is 2 pounds. So I expect the waiter to ask me for 4 pounds. Now a friend sitting next to me with his eye out for a quarrel raises the foIlowing question: "How in heaven's name can you know in advance what your bill will be?" "Since I have calculated it", I reply. "But that", my friend responds, "only tells you what you would charge yourself if you were the waiter. How can you predict the bill he will give you?" Already a little angry I answer: "WeIl, he will calculate it too, and since 50 JOHANNES L. BRANDL it is a very simple calculation it is quite improbable that he will make a mistake." At this point my friend comes up with Kripke's version of the argument ab out mle-following. "You expect the waiter to follow the same arithmetical mIes as you do. But for all you know, and for all anybody else may know (including the waiter hirnself) this may not be the case. He might be a quus-person, adding in the usual way under conditions C, but doubling the result under conditions C. On the assumption that conditions C' obtain you should get a bill of 8, not of 4 pounds. Now, tell me your reasons for thinking that the waiter is not a quus-person and that his way of doing addition is not conditional on anything you regard as irrelevant for doing arithmetic?" As 1 said, it is very debatable what the next move in this game should be according to Wittgenstein. Should 1 join my friend in being sceptical about what the waiter's calculation will be? Or should 1 argue from the fact that the waiter has been trained in our community and that there is sufficient mutual agreement between us to warrant my prediction? Or should 1 reply that no reasons are needed for excluding such queer cases as imagined by my friend in order to justify my belief about the bill? Whatever the recommended move may be, there are other moves which are definitely wrong. It would be a clear mistake to start an argument about understanding what the waiter says when he asks for the bill. 1 could not save my prediction by adding the hypotheses that the waiter might get the words wrong, that he would not really want 8 pounds even ifhe said so, that he certainly meant he should get only 4 pounds, but mistakenly uses the word "eight" to express his demand. 1fthat were so, 1 would expect hirn to be satisfied by getting 4 pounds because that is what he really asks me for. Naturally, 1 would also expect hirn to thank me for the 8 pounds 1 gave to hirn, because that is what he will call my four pounds according to the present hypothesis. This desperate move rests on a serious mistake. It presupposes that we can always drive a wedge between what somebody says with the words he uses and what he really means by them. Of course, this strategy allows me to excuse whatever the waiter says or does as a purely verbal mistake, as a surface phenomenon so to speak. The correct calculation may still be there at a deeper level, as a phenomenon in the mind of the waiter. But it will be there independently of anything he does or says. This is the price one has to pay for justifying an apparently mi staken calculation by invoking a linguistic error. Of course, sometimes we do have good reasons to doubt the linguistic competence of others, but the mere fact that their calculations are wrong is not a good reason. It is the arbitrariness in attributing a linguistic mistake to the waiter which leads to unacceptable consequences. WITTGENSTEIN'S ALLEGED METAPHYSICS OF MIND 51 If whatever seems to be wrong for a third person may be correct from the first-person perspective, it could just as well be the other way round. That is exactly how my friend in the bar wants to have it. If I allow the waiter to say "eight" and really mean "four", then he can allow the waiter to do an idiosyncratic calculation even though in the past he always came up with the expected results. That is to say, I might agree with all his previous uses of the plus-sign although in fact he always had a different calculation in mind. This difference might become apparentjust in presenting me the next bill. To avoid this, we must deny that saying and meaning can diverge in an arbitrary way. An argument which opens such a gap wi11leave us he1pless in convincing our sceptical friend that rule-following is not a matter of 'anything goes'. So far, I think, nothing controversial has been said ab out the sceptical underpinnings ofthe rule-following problem. My only point was that the problem becomes worse as soon as we concede that rules may have been obeyed in spite of all appearances to the contrary. That is exactly what happens when we divorce mental states from their role in explaining human behaviour. If the same behaviour may be effected by just any mental state whatever, then the truth about the mental is left completely moot. Mental states are now put beyond the reach of empirical investigation. Wittgenstein wanted to block this source of metaphysical theories about the mind with his rule-following argument. In view of it, I think, there is no space left for attributing to hirn claims about mental phenomena whose truth is independent of any empirical investigation about their causal effects. There is then left to consider, as a third possibility that Wittgenstein holds a transcendental metaphysics of the mind. This, again, is a big issue which should be discussed only on the basis of extensive textual evidence. But I think that something can be said in advance against this interpretation. As is wen known, Wittgenstein applies his considerations about rule-following to the case of a private language, i.e. to a language which cannot be shared by several speakers. He starts this discussion by asking wheter we could imagine a language in which a person records his inner experiences for his private use. 3 The argument then is intended to prove that no matter how we imagine such a language, we imagine something impossible. But why should anybody want to claim that there are languages which are in principle unsuited for communication? There are a number of prejudices ab out the mental which are liable to produce the myth of a private language. One such prejudice which is particularly suggestive takes every mental phenomenon to have a natural meaning, just as having pains is a natural indicator for some damage in the body. If this is the 52 JOHANNES L. BRANDL model for mental phenomena in general, then a belief or desire should be a natural indicator of the fact which obtains when the belief is true or the desire satisfied. When I leam to express my beliefs and desires in a conventional language, what happens, according to this picture, is that I leam for each sentence in the given language which mental states naturally indicate the fact described by this sentence. In doing so I will establish corre1ations of the following kind: Sentence S is true iff my belief that such-und-such is true. If the language has a productive vocabulary this requires an infinite list of such correlations. The only way in which I can handle such a list is by applying a procedure which generates infinitely many c1auses ofthe form: My beliefthat such-and-such is true iff such-and-such. But applying such a procedure means that lalready have some kind of language at my disposal. The upshot then seems to be that leaming the first conventionallanguage is only possible as a process of translating it into a more basic language, a language I somehow acquire together with my beliefs and desires even before I leam to express them verbally. This basic language must be a system ofnatural signs revealing their meaning simply by occurring in my mind; just as a certain feeling reveals itse1f to me without any need to leam in advance what it means. But now we have to observe that my pains, taken as a natural sign in my mind, reveal their meaning only to me. Therefore the basic language I start with must be a private one. How does Wittgenstein respond to this reasoning? I think his answer can be interpreted in two ways. First, Wittgenstein may be taken as offering a better argument in favour of the opposite conclusion. The defender of a private language starts from the vague characterisation of mental states as natural signs and ends up with the conclusion that for any language to be possible at all every speaker must be equipped with a language of his own, Le. with a system of signs that necessarily mean something only to hirn. The opposite view takes mental states to be necessarily linked to conventional manifestations, e.g. verbal utterances, and conc1udes from this that a private language is impossible since there are no conventional criteria for its application. The first argument amounts to a transcendental proof of the existence of a private language, the second offers an equivalent proof of the opposite. Reading Wittgenstein in this way means that he indulges in transcendental metaphysics. He does so by claiming that every mental state has to be linked with public criteria. But the conflict can be described in a different way too. On this second reading Wittgenstein is not aiming at any proof at all. Rather he wants to undermine the alleged proof either ofthe necessity or the impossibility of a private language: By revealing both conc1usions to be too strong he WITTGENSTEIN'S ALLEGED METAPHYSICS OF MIND 53 argues for the falsity of their premises. As I see it, Wittgenstein was prepared to mistrust any philosopher who sets out to prove how things must be or how they cannot be. In the present case this throws doubt on the characterisation of mental states as natural signs, as weIl as on the idea of naming sensations by conventional signs. The philosopher's task is not to establish transcendental arguments; rather he should dismantle them whenever they are forced upon hirn. Once we conceive of the private language argument not as an argument in and of itself, but as a strategic move against a philosophical prejudice, I think that turning Wittgenstein into a transcendental philosopher becomes very difficult. He induces us to mistrust the "hardness of the logical must"4 and to question the rigidity of explanations based on mere analogy. To work out "preconceived idea[s] to which reality must correspond" was exactly the opposite of what he wanted to achieve. He called it the "dogmatism into which we fall so easily in doing philosophy". 5 2. THREE GRADES OF MENTAL INVOLVEMENT Having listed three reasons for not counting Wittgenstein's philosophy of mind 'metaphysical', let me turn to some considerations of a more systematic kind. One thing Wittgenstein makes us aware of, I think, is that we should be very careful in defining the vocabulary with which we attribute cognitive states. Roderick Chisholm once suggested that the terms we use in attributing epistemic states may all be defined in terms ofthe single notion of 'preferring one proposition to another'.6 Thus we may focus on this single notion and distinguish three different ways of defining 'epistemic preference': Df1: S prefers p to q = df It is more likely that S will accept the proposition that p than the proposition that q (given that he considers these propositions as mutually exclusive alternatives). Df2 : S prefers p to q = df It is more likely that S will sincerely assent to an utterance saying that p than to an utterance saying that q (given that he understands these utterances as conflicting with each other). Df3 : S prefers p to q = df It is more likely that S will act successfully if it is the case that p than if it is the case that q given that his intentions depend on these matters of fact). 54 JOHANNES L. BRANDL Definition Dfl is the one that Chisholm would prefer. The idea behind it is that epistemic preference involves adecision among a given set of propositions. These are entities to which no one has a privileged access. Two persons, even when speaking different languages, may nonetheless grasp the very same proposition. That gives epistemic states an objective value. But there is an additional reason for having propositions in ones' ontological baggage. Propositions are semantically transparent entities. The only way to know a proposition is to know what it means. Linguistic entities, by contrast, may be encountered without grasping their meaning. In this respect propositions are like private experiences: They reveal to us their meaning as soon as they occur to uso But unlike private experiences they may occur to more than one individual. Therefore, defining cognitive states in terms of an awareness of propositions guards against the idea of a private language. And this may weIl elose the door to a transcendental metaphysics of the mind. Definition Df2 has a different advantage. If our cognitive states are defined in terms of our readiness to assent to utterances, that makes it plain why only persons (and perhaps some animals) can be in such states. Neither a Cartesian ego nor a mindless body could assent to an utterance. It requires more than taking an attitude towards a proposition and also more than a bodily movement like nodding ones' head. Assent is the sincere expression of agreement, which involves both mental and bodily states. Thus Dt; guards against theories which turn the domain of other minds into a domain of speculative metaphysics. Both ofthese advantages are shared by definition Df3 • A Cartesian ego may have mental states, but only a person with a body is able to perform actions. Since Df3 rests on this latter notion it secures the link between the mental and the physical. The other advantage, namely the objectivity of mental states, can be guaranteed not only by introducing abstract propositions, but also by relating a person's actions to what is actually the case. When two people act successfully in the same way under the same conditions, we have reason to assurne that their preferences are the same as welF Thus Df3 prevents mental states from being locked up in a private sphere. Besides sharing the advantages of the other definitions, Df3 has an advantage of its own. It shows that there is no more to the cognitive states of a person than what is manifested in her actions. This is not so elear from the other definitions. They require us to know which propositions a person considers as the relevant alternatives or to translate the utterances to which the person assents into our own language. In each case there may be more than one way to fulfill this requirement. Yet, judging from what the person does, there may be no difference between choosing one translation over the other or WITTGENSTEIN'S ALLEGED METAPHYSICS OF MIND 55 one set of propositions rather than another. According to definitions Df} and Df2 this indicates a difference in the cognitive states ofthe person. Any difference in the objects of the attitudes implies a difference in the mental state itself. In drawing such fine-grained distinctions we find that the definitions Df] and Df2 lead us to investigation-independent claims about our cognitive states. Definition Df3 avoids this danger by reporting mental states without relating a person to any specific object. The definition only reHes on what a person S achieves given that it is the case that p or the case that q. Here the variables "p" and "q" may be replaced by non-synonymous sentences without implying that a different mental state is attributed to S. In this way we are prevented from a conception of mental states that we cannot verify or falsify by even the best investigative procedures available. 8 3. MENTAL AND MODAL INVOLVEMENT Let me close this note with a comparison between the three definitions of epistemic preference and Quine's objections to modallogic. Quine recommends taking "necessity" as a semantic predicate to be attached to names of sentences. 9 This corresponds to the meta-linguistic definition Df2 • He objects to using "necessarily" as an operator on open sentences because this amounts to introducing essential properties. A similar criticism may be urged against definition Df}. According to this definition there ought to be propositions that a person prefers to any other alternative. Such a commitment to 'epistemically perfect' propositions is just as objectionable as a commitment to essentialism. But there is amiddie way in the case of modallogic and also in the case of defining epistemic preferences. The third option is to take "necessity" as an operator on closed sentences only. This is still objectionable from Quine's point of view because it sacrifices extensionality. But, as Arthur Prior once remarked, it would be mere stubbornness to deny that the sentence "S thinks that p " is a function of the sentence "p" only because the function has to be intensional. lO I think it would be equally stubborn to insist on adefinition of mental states in terms of assent or dissent, only to avoid the full blown mentalistic idiom of grasping propositions. There is a modest form of getting involved both in mental and in modal affairs. It was Wittgenstein's advice, I think, to find the right degree of involvement in the first case. That was his way to avoid dubious metaphysical claims about the mind. l1 Universität Salzburg 56 JOHANNES L. BRANDL NOTES 1 Kenny mentions this criterion also in his book The Metaphysics of Mind, Oxford University Press, Oxford 1989, p. ix. 2 L. Wittgenstein, Philosophische Untersuchungen. PhilosophicalInvestigations, ed. by G. E. M. Anscombe & R.Rhees. Trans. by G. E. M. Anscombe, B. Blackwell, Oxford 1953, § 306. 3 IMd., § 243. 4 Ibid., § 437. s Ibid., § 131. 6 Roderick M. Chisholm & R. Keim, 'A System ofEpistemic Logic' ,Ratio, 14, 1972, pp. 95-110. 7 CharlesTravis pointed out to me that this inference cannot be drawn simply from the fact that two persons achieve the same results under the same conditions, because that may be a matter of pure luck. The demand that they must 'act successfully in the same way' should rule this out, but it is not cJear to me how to explicate this phrase appropriately. 8 Most ofthe literature on individuating mental states focusses on the opposite problem, namely, that the criteria ofindividuation offered (for instance) by possible-worlds-semantics are too coarse grained. But there is pressure from both sides here to draw just as many distinctions as there are in psychic reality. 9 w.V.O. Quine, 'Three Grades ofModal Involvement', in: The Ways ofParadox and other Essays, Harvard University Press, Cambridge Mass., 1966, pp. 158-176. 10 Arthur Prior, Objects ofThought, Clarendon Press, Oxford 1971, p. 48. 11 I am indebted to Tony Anderson, Barry Smith and Charles Travis for their help with this paper. PART II IN THE DOMAIN OF POSSIBILITIES: RULES, FORMAL AND INNER LANGUAGES JACQUES BOUVERESSE LE REEL ET SON OMBRE: LA THEORIE WITTGENSTEINIENNE DE LA POSSIBILITE ABSTRACT Wittgenstein repeatedly criticised the idea of certain "shadowy beings" which we believe to be necessary to relate two radically different things which apparently cannot be put directly in contact with each other, like thought and what it represents, sentence and what verifies it, desire and what realizes it, expectation and what satisfies it, etc. Wittgenstein tried to show that we are brought no nearer to the goal by the interposition of intermediary entities of that kind. In the case of a dec1arative sentence, the shadowy being must be a possibility which is expressed by the sentence and can be or not be realized. If the sentence is true, the possibility is supplanted by the corresponding fact; and if it is false, the only reality we have is so to speak the shadow, the possibility itself. One ofthe most fundamental theses ofWittgenstein's philosophy is that all possibility (at least "logical" or "grammatical" possibility, as he calls it) must be contained in language, and not between language and reality, in and intermediary realm of possibilities, and that it is all of the same kind and on the same level. To say that something is possible is simply to say that an expression has sense; and to say that it is impossible is not to say that the expression represents an impossible sense, as if we had first to consider a possible sense and then to exc1ude it as unreal or absurd, but simply that it has no use and does not belong to language. An important consequence of that is that, for Wittgenstein, there are no degrees of impossibility, no impossibilities that could be deeper and more radical than others. We cannot distinguish between a nonsense which makes (or at least could make) sense and a nonsense which does not. As Wittgenstein says, "what we exc1ude has no semblance of sense", it is not thinkable and does not have to be thought in some way in order to be exc1uded. What we exclude is always a use and not, so to speak, a possibility which turns out to be impossible. I Wittgenstein a consacre une part importante de ses reflexions a la critique d' "etres fantömatiques" (shadowy beings), qui sont supposes servir 59 R. Egidi (ed.), Wittgenstein: Mind and Language, 59-81. © 1995 Kluwer Academic Publishers. 60 JACQUES BOUVERESSE d'intermediaires obliges entre deux choses qui ne paraissent pas pouvoir etre mises en relation l'une avec l'autre: la pensee et ce qu'elle represente, la phrase et ce qui la verifie, le desir et ce qui le realise, l'attente et ce qui la satisfait, etc. Puisque la chose que nous attendons peut ne pas se realiser, "ce que nous attendons n'est pas le fait, mais une ombre de ce fait; pour ainsi dire, la chose qui est presque le fait".C)1 (BLB, p. 36) De meme, puisque le fait qui rendrait notre pensee vraie, s'il existait, peut tres bien ne pas exister, nous sommes tentes de dire que l'objet de notre pensee n'est pas le fait luimeme, mais plutöt une ombre de ce fait. "11 y a, remarque Wittgenstein, differents noms pour cette ombre, par exemple 'proposition', 'sens de la phrase'." (ibM., p. 32) Puisqu'une pensee peut etre fausse, elle doit contenir, a defaut du fait lui-meme, qui, en l'occurrence, n'existe pas, du moins la possibilite consideree comme une sorte de substitut fantömatique ou de double annonciateur du fait qui la verifierait. Wittgenstein s'efforce de montrer que nous pourrions tout aussi bien faire l' economie de ces entites intermediaires qui peuplent le royaume des ombres, parce qu'elles ne font que repousser un peu plus loin le probleme qu'elles etaient supposees resoudre: Nous n'avons rien gagne en admettant qu'une ombre devait intervenir entre I'expression de notre pensee et la realite alaquelle se rapporte notre pensee. Nous avons dit que, si nous voulions une image de la realite, la phrase elle-meme est une image de ce genre (bien que ce ne soit pas une image par ressemblance). (ibid., p. 41) Le probleme est que nous demandons aux ombres que nous evoquons, de realiser une chose tout a fait impossible. Nous disons qu'un etat intentionnel comme le desir doit contenir une ombre de la chose desiree en voulant dire par la qu'il doit en quelque sorte contenir la chose elle-meme, mais a l'existence pres. C' est ce que nous pouvons dire de moins absurde, si nous ne voulons pas etre obliges de proferer une chose encore plus absurde, a savoir que la chose desiree elle-meme doit etre presente dans le desir. Mais, objecte Wittgenstein: Comment pouvons-nous desirer que ceci precisement arrive, si ceci precisement n'est pas present dans notre desir? Il est tout a fait vrai de dire: la simple ombre ne ferait pas l'affaire; car eHe s'arrete immediatement avant I'objet; et nous voulons que le desir contienne l'objet lui-meme. (ibid., p. 37) la Dans ses ｬ･ｾｯｮｳ＠ des annees 1930-1932, Wittgenstein souligne ce point de ｦ｡ｾｯｮ＠ suivante: Si j' attends que Mr Smith entre, mon attente contient quelque chose de semblable aMr Smith; si j'attends une tache rouge, mon attente ne peut pas contenir la tache rouge elle-meme, mais seulement quelque chose qui lui est semblable. Mais attendre quelque chose de 'semblable' LE REEL ET SON OMBRE 61 est deja attendre quelque chose de different. L'element 'semblable', I'ombre, dans I'attente, est differente de la realisation; vous n'etes pas plus pres de la nlalisation.2 En d'autres termes: Interpoler une ombre ou une chose du meme genre entre un symbole ou une proposition et un fait ne resoudra rien a moins que nous n'ayons une autre ombre pour servir d'intermediaire entre I' ombre et le fait; nous avons alors une regression a I'infini, nous ne sommes pas arrives plus pres. Une similitude n'explique pas I'ombre, elle n'explique pas non plus la relation au fait. (WLl930-32, p. 32) Dans le cas d'une proposition declarative, l' ombre mediatrice est la possibilite du fait concerne, c'est a dire le fait precis qui devrait etre realise pour que la proposition soit rendue vraie, moins la realisation. Et l'objection est evidemment la meme que precedemment. Pas plus que la realisation du desir, l'actualisation de la possibilite ne peut etre comprise en termes de concordance d'un fait ree1 avec une image ressemblante. Ou plus exactement, si I'expression du desir et la description du fait ne sont pas capables d'aller elles-memes jusqu'a la chose concernee, aucune image plus ressemblante ne le fera davantage. Dire qu'une proposition exprime une possibilite n'est donc pas dire que la proposition contient une image suffisamment ressemblante pour etre l'image de ce fait a l'exclusion de tout autre, mais simplement qu'elle a un sens; et dire qu'une proposition a un sens, ne veut pas dire qu'il lui correspond, dans un univers qui n'est pas celui des faits ree1s ni celui des expressions, une entite qu'on appelle son sens. Nous ne pouvons donc pas essayer d'expliquer ce qu'on veut dire en disant qu'une proposition pa un sens en ｲ･ｭｰｬ｡ｾｮｴ＠ "p a un sens" par "p represente une possibilite" et "p n'a pas de sens" par "p est une proposition a laquelle ne correspond aucune possibilite" ou, ce qui est pire encore, "a laquelle correspond une impossibilite". "p a un sens" et "p exprime un fait possible" doivent avoir exactement le meme sens; et, si I' on considere les choses dans Ia perspective du Tractatus/ I'une et l'autre sont des propositions impropres, qui essaient de dire que1que chose qui peut seulement se montrer. 11 va sans dire que l'on n'ameliorerait pas les choses en disant que "(La phrase) p a un sens" veut dire que "La proposition exprimee par pest possible", alors qu'une phrase denuee de sens exprime ou tente d'exprimer une proposition impossible. "p a un sens", "p exprime une proposition" et "p exprime une proposition possible" disent en fait rigoureusement la meme chose. Dans un passage dont Ia redaction (qui n'est pas due a Wittgenstein lui-meme) est assez confuse, mais qui est neanmoins suffisamment clair sur Ie point essentiel, Wittgenstein dit: 62 JACQUES BOUVERESSE Nous ne pouvons pas dire 'Une proposition (proposition) p estpossible'. Si p a un sens, celaen lui -meme montre qu' elle est possible. Si p n' etait pas possible, ce ne serait pas une proposition du tout. Tout ce que nous pourrions vouloir dire par 'p est possible' serait 'p a un sens'; il ne pourrait y avoir aucune indication concemant la question de savoir si ce que passerte est arrive ou pourrait arriver. De ce fait, 'p est possible' n'est pas une proposition legitime. (Possibilite veut dire ici possibilite logique). (WLl930--32, p. 9) C'est une des theses fondamentales de la philosophie de Wittgenstein que toute la possibilite (du moins la possibilite de l'espece qu'i! appelle "logique" ou "grammaticale") est contenue dans le langage. C'est le systeme de representation que nous adoptons qui determine les possibilites accordees a la realite concernee. Wittgenstein dit que "la possibilite est exprimee (contenue) dans le langage lui-meme". (WL1930--32, p. 8) Et il faut comprendre cela comme signifiant en particulier que la possibilite n'est pas contenue dans un uni vers de possibilites qui occuperait une position intermediaire entre le langage et la realite. Plus precisement, la possibilite est contenue dans la grammaire du langage, puisque c' est la grammaire qui determine ce qui a un sens et ce qui n'en a pas, c'est a dire ce qui est possible et ce qui ne l'est pas. "La grammaire, ecrit Wittgenstein, n'est pas l'expression de ce qui est le cas, mais de ce qui est possible; i1 y a donc un sens auquel la possibilite est la forme logique". (WL1930--32, p. 10) 11 resulte de ce1a que ce qui est possible et ce qui ne I' est pas est arbitraire dans un certain sens, le sens auquel, comme le dit et le repete Wittgenstein, la grammaire est arbitraire: "Ce qui est appele possible et ce qui n'est pas appele possible est en un certain sens arbitraire. Nous disons que, bien qu'aucun homme ne soit assis sur cette chaise, quelqu'un le pourrait". Cela signifie en gros que: La phrase 'Quelqu'un est assis sur cette chaise' a un sens, c'est ä. dire, il y a une possibilite logique que quelqu 'un y soit assis. Il est theoriquement possible, c 'est ä. dire possible dans une theorie, qu'un certain hydrogene ait six fois la valence normale. Cela pourrait etre possible dans certaines theories, mais peu pratique. Les systemes peu pratiques sont rejetes, et le rejet est traite comme si ce qui est rejete etait quelque chose de faux. Le rejet d'un systeme grammatical est comme le rejet d'un etalon de longueur, et l'acceptation d'une grammaire, d'un symbolisme, est comme l'acceptation d'un etalon de longueur. 4 On pourrait donc dire que "p est possible" veut dire "p a un sens" et "p est possible chimiquement" veut dire "p a un sens en chimie", c'est adire dans le cadre d'une certaine theorie chimique admise. Le choix d'un systeme de representation determine un espace de possibilites pour les faits de la categorie concemee; et inversement, ce qui est possible est toujours ce qui peut se dire de fa<;on douee de sens dans un certain systeme de representation. Wittgenstein LE REEL ET SON OMBRE 63 compare, du reste, a I' occasion, le cas de la possibilite logique a celui de la possibilite chimique. Dans les Recherehes Philosophiques S on trouve la suggestion suivante: Comparez 'logiquement possible' avec 'chimiquement possible' . On pourrait appeler par exemple chimiquement possible une combinaison pour laquelle il y a une formule de structure avec les bonnes valences (par exemple H-O-O-O-H). Une teIle combinaison n'existe evidemment pas forcement; mais meme a une formule HO z' il ne peut correspondre moins dans la realite qu'aucune combinaison. (PI I, § 521) Vue du point de vue de la realite, l'impossible devrait, semble-t-il, representer quelque chose de moins que la simple non-existence, une forme aggravee de la non-existence, tout comme la necessite ressemble aune forme accentuee de l'existence. Pourtant, aune formule qui represente une combinaison impossible correspond dans la realite exactement la meme chose, asavoir simplement aucune combinaison. La difference entre les deux est faite uniquement par le systeme de representation adopte. Quand Wittgenstein dit que la grammaire est arbitraire, il veut dire qu'il n' existe pas de faits auxquels nous pourrions essayer de nous referer pour etablir que les possibilites et les impossibilites qui ont ete decretees par la grammaire etaient bien des possibilites et des impossibilites reelles. L'argument essentiel qu'il utilise sous des formes diverses au debut des annees trente pour justifier cette impossibilite de soumettre la grammaire au verdict d'une realite quelconque revient toujours a dire que la description de possibilites et d'impossibilites qui sont supposees preceder la grammaire et rendre ses regles obligatoires, presupposerait elle-meme une grammaire et donc une distinction prealable, qui ne peut etre justifiee, entre comment pourrions-nous justifier les le sens et le non-sens. De toute ｦ｡ｾｯｮＬ＠ exc1usions qui ont ete operees par la grammaire sans etre contraints d'utiliser des expressions auxquelles la grammaire que nous cherchons ajustifier ne donne precisement aucun sens? Pouvons-nous donner une description qui justifiera les regles de la grammaire? Pouvonsnous dire pourquoi nous devons utiliser ces regles? Notre justification ne pourrait prendre que la forme consistant a dire: 'Etant donne que la realite est teIle ou teIle, les regles doivent etre teIles ou teIles'. Mais pour decrire une realite dans laqueIle la grammaire serait autrement, je devrais utiliser les combinaisons precises que la grammaire defend. Les regles de la grammaire distinguent le sens et le non-sens et, si j'utilise les combinaisons defendues, je protere du non-sens. (WLl930-32, p. 47) Dans la Grammaire Philosophique,6 Wittgenstein formule sa position ainsi: "J'aimerais dire: 'Je dois commencer avec la distinction du sens et du nonsens. Avant elle, rien n'est possible. Je ne peux pas lajustifier"'. (pG, § 81) 64 JACQUES BOUVERESSE La difficulte qui se presente ici tient au fait que, lorsque nous disons d'une chose qu'elle est impossible, il semble que nous ayons du d'abord en quelque sorte penser une possibilite et plus precisement nous la representer, pour pouvoir ensuite l'exclure. A I'epoque du Tractatus, Wittgenstein avait ete aux prises avec la question de savoir ce qui peut bien correspondre dans la realite ala proposition "non-p", lorsque celle-ci est vraie. Si I' on exclut, comme il le fait, que le fait negatif puisse etre con9u comme resultant du fait positif par l'adjonction d'un constituant supplementaire qui correspondrait a la negation (un fait negatif est la "non-existence d'etats de choses" (2.06) et non l'existence d'etats de choses negatifs), ce qui correspond en pareil cas a "nonp" ne peut etre que la non-realisation du fait decrit par "p". Dans les le90ns des annees 1930-1932, Wittgenstein formule les choses ainsi: On demande parfois: comment pouvons-nous nier une proposition? Car si non-p est vrai, il n'y arien pour lui correspondre; il n 'y a quelque chose pour correspondre ap que lorsque pest vrai. Qu'est-ce qui correspond a 'La porte n'est pas ouverte' quand elle est ouverte? Mais il y a ici une fausse analogie: p ne correspond pas aque/que chose. Et ce qui correspond anon-p est que p n 'est pas /e cas. (WL1930-32, p. 52) Mais cela suppose evidemment que le fait que p, dont la propostion negative asserte la non-realisation, soit represente - et cela signifie necessairement, represente comme une possibilite - dans l' expression de la proposition. Que se passe-t-il apresent, si nous essayons de nier une proposition qui dit, non pas que quelque chose est le cas, mais que quelque chose est possible, c'est a dire si nous essayons d'asserter que quelque chose n'est pas possible? La possibilite de la chose, qui est exprimee par la proposition positive, n'estelle pas du meme coup necessairement presupposee? Ou encore, pour pouvoir dire que p n'a pas de sens, ne faut-il pas avoir donne d'abord un sens a p ou en tout cas envisage une possibilite de sens pour p? Wittgenstein reconnait qu'il y a quelque chose de paradoxal dans le fait que nous puissions dire que quelque chose est impossible, puisque cela implique, semble-t-il, que nous puissions dire ce qu'est la chose dont nous sommes en train d'exclure la possibilite: Il est etrange que nous disions ce que c'est, qui est impossible, par exemple que le dessus de cheminee ne peut pas etre jaune et vert en meme temps. En parlant de ce qui est impossible, il nous semble que nous concevions I'inconcevable. Lorsque nous disons qu'une chose ne peut pas etre verte et jaune au meme moment, nous excluons quelque chose, mais quoi? Si nous devions decouvrir quelque chose que nous decririons comme vert et jaune, nous dirions immediatement que ce n' etait pas un cas exclu. Nous n' avons pas exclu du tout un cas quel qu 'il soit, mais plutöt l'usage d'une expression. Et ce que nous excluons n'a pas de semblance de sens. La plupart d'entre nous croient qu'il y a un non-sens qui a du sens, et un non-sens qui n'en n'a pas-que LE REEL ET SON OMBRE 65 c'est un non-sens qui n'est pas du meme genre que de dire 'Ceci est vert etjaune au meme moment' et de dire ' Ab sur Ah'. Mais ce sont des non-sens dans le meme sens, la seule difference etant dans la ｦ｡ｾｯｮ＠ dont sonnent les mots. (WL 1932-35, p. 64) Ce qui nous induit en erreur dans des cas de ce genre est que nous avons 1'impression que nous faisons un effort desespere pour reussir it penser ou it imaginer quelque chose et qu'une proposition comme "Aucun objet ne peut etre vert et jaune au meme moment", par exemple, enregistre simplement l'echec de cette tentative. Si j'essaie de soulever un homme avec une seule main, je peux m'apercevoir que j'en suis incapable et dire qu'il m'est impossible de soulever un homme avec une seule main. Mais si la possibilite dont il s'agit est la possibilite logique, le point important est que ce que je ne peux imaginer est egalement quelque chose que je ne peux pas essayer d'imaginer, parce que je ne sais tout simplement pas ce qu 'il pourrait etre question d' essayer d'imaginer. Rien, en effet, n'a ete decrit ou meme simplement suggere par l'expression dont on entend exc1ure l'usage: "Nous pourrions demander: it quoi ressemble une structure impossible? Peut-elle etre decrite? Lorsque nous appelons une chose que1conque impossible, il n'y arien que nous appelions sa description".7 (LFM, p. 148) La situation est evidemment un peu plus delicate lorsqu'une expression qui donnait l'impression d'avoir un sens apparatt comme etant en realite depourvue de sens au terme d'un processus de demonstration plus ou moins complique. Wittgenstein est revenu frequemment sur la question de savoir comment nous pouvons, en mathematiques, chercher une chose dont nous decouvrons finalement qu'elle est impossible, comme par exemple la trisection de l'angle ou la construction de l'heptagone regulier avec la regle et le compas, ce qui donne l'impression que la chose que l'on cherchait est une chose que I' on ne pouvait pas non plus reellement chercher. Dans ses le90ns sur les fondements des mathematiques de 1939, Wittgenstein dit que: Tout ce que la demonstration mathematique du fait qu'un heptagone regulier ne peut etre construit avec la regle et le compas effectue, est de nous donner de bonnes raisons d' exclure I' expression 'construction de I'heptagone' de notre notation. Par consequent, 'Smith a trace la construction de I'heptagone' n 'est pas faux, mais denue de sens. Il utilise une expression qui n 'est pas seulement une expression a laquelle on n'a pas donne de signification, mais qui a ete exc\ue. Elle a ete exclue pour des raisons d' experience, bien que I'enonce selon lequel c' est impossible ne soit pas un enonce d'experience. (LFM, p. 47) Wittgenstein ne va pas jusqu'it nier que 1'0n puisse faire une difference entre le non-sens immediat et le non-sens que l'on pourrait dire differe, si I' on entend par lit que certaines expressions peuvent etre exclues 66 JACQUES BOUVERESSE immediatement eomme denuees de sens, alors que d'autres ne le peuvent qu'au terme d'une suite d'operations dont le resultat peut etre, au moins du point de vue psyehologique, tout afait inattendu. Mais le non-sens est toujours le non-sens d'une expression exc1ue, et non d'un sens impossible, d'un sens qui, en quelque sorte, n'a pas de sens. Wittgenstein soutient qu'il n'y a pas de degres dans l'impossibilite ou le non-sens, que tout ee qui est impossible ou denue de sens l'est au meme degre et de la meme ｦ｡ｾｯｮＬ＠ meme s'il ne l'est pas immediatement: Les regles pour I'usage des mots peuvent exclure eertaines eombinaisons, et eela de deux fa..ons: (I) lorsque ee qui est exclu est reeonnu eomme non-sens des qu'il est entendu, (2) la Oll des operations sont requises pour nous permettre de le reeonnaitre eomme non-sens. Le fait que la negation d 'une tautologie eomplexe soit une eontradietion est deeouvert par les memes moyens par lesquels on trouve quex2 + 6x + 7 a deux raeines entieres est vrai -a I'aide d'operations. Nous pourrions eroire, par exemple, que 'S ax paires de ehaussures, aveex2 = 2' a un sens, paree que nous aurions la sensation que nous pourrions extraire un sens de eette phrase en resolvant I'equation quadratique x 2 =2. Le fait que nous ne voyions pas ee qU'est le resultat est une raison de eroire que nous pourrions appeler la phrase une espeee de non-sens differente de 'tables, ehaises, chaussures'. Le mot 'non-sens' est utilise pour exclure eertaines ehoses, et pour des raisons differentes. Mais i1 ne peut pas etre vrai qu'une expression est exclue et eependant pas tout a fait exclue - exelue paree qu'elle designe I'impossible, et pas tout a fait exelue paree qu'en l'excluant nous devons penser I'impossible. Nous exc1uons des phrases eomme 'C'est a la fois vert et jaune' paree que nous ne voulons pas les utiliser. Bien entendu, nous pourrions donner un sens aces phrases. J'ai dit anterieurement que ce qui est possible ou impossible est une question arbitraire. Nous pourrions instaurer eomme regle, par exemple, que 'Vert et jaune peuvent etre a la meme plaee au meme moment' devra avoir un sens. (WL1932-35, p. 64) Qu'une expression soit exc1ue direetement ou que, eomme e'est le eas en mathematiques, elle le soit a la suite d'un ealcul ou d'une demonstration, l'exclusion est done dans tous les eas aussi eompU:te et radicale et ne laisse rien subsister de ee qui pouvait ressembier au premier abord aune possibilite ou une potentialite de signifieation. Wittgenstein, comme on I' a vu, refuse de eonsiderer la possibilite eomme une forme affaiblie ou amoindrie de l'existenee, et done eomme eapable de preexister en un sens quelconque anotre decision d'adopter une eertaine norme d'usage pour les expressions eoneernees. Le "reel du possible", eomme on disait a l'epoque de Leibniz, ne peut resider ailleurs que dans le langage luimeme: Si vous voulez eomprendre ee qU'est la possibilite logique, une bonne analogie est la possibilite ehimique. Prenez une formule de strueture en ehimie: LE REEL ET SON OMBRE 0= H H- I o I 67 o I o I O-H 'H 20 4 est chimiquement possible bien qu'il n' existe pas' . Cela etant, qu' est-ce qui est possible apropos de HP4' s'il n'existe pas? Est-il dans ce cas un peu plus reel par le fait d'etre possible? Est-il plus proche de l'existence que H0 2? Vous pourriez dire: 'Ya-t-il des stades de l'existence?' (cf 'Cela n'est pas arrive, mais c'etait tout a fait possible, ce qui est deja quelque chose!). Quelle est l'utilite de cette idee que c'est une possibilite chimique? Y a-t-il ici aussi une realite qui est de la nature d'une ombre? Et Oll est l'ombre? - Les chimistes ne pourraient-ils pas croire qu'il y a un monde dans lequel il y a un H20 4, bien qu'il ne soit pas passe a la realite? N'y a-t-il pas ici une realite faite d'ombre - que vous avez reellement dessinee? Votre langage n'est-il pas la realite faite d'ombre? Lorsque vous dites que 'H 20 4 est possible', vous voulez dire simplement que c'est un signe dans votre systeme. Ce systeme de valences n'a pas ete choisi au hasard, mais parce qu'il s'accordait bien avec les faits. Mais une fois qu'il a ete choisi, ce qui est possible est ce dont il y a une image dans le langage des valences. Nous avons adopte un langage dans lequel cela a un sens de dire 'H 20 4' •• ' - ce n'est pas vrai, mais ce la a un sens. Exactement la meme chose est vraie de la possibilite geometrique. Dire qu'une ligne droite peut etre tracee, c' est dire que cela a un sens de parler de tracer une ligne droite. (LFM, p. 146) La metaphore de l' ombre interposee entre la realite et nous rend justice a notre idee que le possible precede et annonce dans certains cas le reet. Nous pouvons par exemple, dans certaines conditions, voir l'ombre d'un homme qui s'apprete a entrer avant de voir l'homme lui-meme. Mais si le possible n'est pas reel, c'est l'ombre qui devient en quelque sorte la realite, la realite demeuree al' etat de possible. Aussi determinantes que puissent etre les raisons que nous avons d'instaurer une possibilite ou une impossibilite grammaticales en adoptant un mode d'expression determine, Wittgenstein soutient que nous sommes radicalement empeches de les concevoir et de les presenter comme equivalant ala reconnaissance directe ou indirecte d'une possibilite ou d'une impossibilite qui existaient en dehors du langage et attendaient d'y etre enregistrees. Puisque les propositions mathematiques sont des propositions grammaticales, on ne peut evidemment pas non plus dire d'elles qu'elles se contentent de decrire fidelement des possibilites et des impossibilites objectives qui sont en l'occurrence reconnues essentiellement par la procedure de 68 JACQUES BOUVERESSE la demonstration. Comme toutes les propositions grammaticales, elles determinent des possibilites et des impossibilites qu'elles donnent l'impression de decouvrir dans une realite qu'elles sont supposees decrire, en l'occurrence la realite mathematique, I' exemple le plus parfait de ce que pourrait etre une realite purement grammaticale, si cela avait un sens de parler de quelque chose de ce genre. Pour Wittgenstein, les propositions de mathematiques pures operent au niveau grammatical de la determination de sens (Sinnbestimmung), et non au niveau descriptif de l'utilisation de sens (Sinn verwendung). Wittgenstein dit de la proposition mathematique qu'elle "doit en verite nous montrer ce que cela a un SENS de dire"8 et de la demonstration qu'elle "doit nous determiner a reconnaitre ceci comme sens et a ne pas reconnaitre ce1a comme sens." (ibid.) 11 en resulte immediatement que la possibilite en mathematiques n'est pas et n'a pas besoin d'etre le reflet d'une existence qu'elle revete et qui la fonde. Wittgenstein evoque ace propos, avec le melange de deference extreme et de distanciation radicale qu'il manifeste generalement en pareil cas, le point de vue tout a fait caracteristique de Frege: Frege, qui etait un grand penseur, a dit que, bien qu'i! soit dit dans Euclide qu 'une ligne droite peut etre tracee entre deux points quelconques, en fait la ligne existe deja, meme si personne ne l'a tracee. L'idee est qu'i! y a un univers de la geometrie, dans lequelles entites geometriques existent. Ce que dans le monde ordinaire nous appellerions une possibilite est dans le monde geometrique une realite. Dans le ciel euclidien, deux points sont deja relies I'un ill' autre. C' est une idee tres importante: I'idee de la possibilite comme etant une espece differente de realite; et nous pourrions l'appeler une ombre de la realite. (LFM, pp. 144-145) Wittgenstein fait allusion ici a un passage de Über die Zahlen des Herrn Schuber! dans lequel Frege ecrit que: [Les postulats d , Euclide] ne sont a proprement parler rien d' autre que des axiomes qui enoncent qu'i! y ades constructions-points, lignes, surfaces-constituees d'une certaine maniere. C'est ainsi que la proposition qui demande que I'on trace une ligne droite d'un point quelconque a un autre dit qu'i! y a, pour deux points pris de fat;:on quelconque, une ligne droite qui les joint. Lorsque nous tra..ons une ligne nous dirigeons notre attention sur elle, qui est iI proprement parler deja la. La possibilite objective de tracer une ligne est aproprement parler la meme chose que I'existence objective de cette ligne. 9 Selon une conception courante que l'on a de la distinction qui existe, dans le systeme d'Euclide, entre les axiomes et les postulats, on est enclin a dire que les axiomes enoncent que certaines choses sont ou ne sont pas le cas, alors que les postulats garantissent certaines possibilites, essentiellement la possibilite d'effectuer des constructions d'un certain type. Frege considere, bien entendu, cette distinction comme illusoire. Pour lui, asserter la possibilite LE REEL ET SON OMBRE 69 de tracer une certaine figure veut dire asserter I' existence de cette figure dans l'univers des entites geometriques. L'image qu'il y a derriere cette conception est, comme le remarque Wittgenstein (ibid.), qu'une droite est deja tracee entre deux points quelconques, mais une droite extremement mince et a peine perceptible, sur laquelle nous en ｲ･ｴ｡ｾｯｮｳ＠ en quelque sorte une plus epaisse et plus grossiere (nous la "materialisons" simplement). On peut dire que Wittgenstein propose, comme Frege, d'ignorer la distinction qui vient d'etre evoquee; mais, chez lui, la reunification s'effectue en sens inverse: les mathematiques construisent des possibilites et en eliminent d'autres, sans jamais se rMerer a une realite determinee et immuable qui serait presente a l' arriere-plan. Wittgenstein a toujours considere comme une source de confusion desastreuse l'idee que les mathematiques peuvent distinguer entre leur possibilite et leur realite et traiter en quelque sorte separement de 1'une et de l'autre. Dans les Philosophische Bemerkungen,JO il observe que: "Sans ces se on se laisse egarer par 1'idee: 'Mais peut-il y avoir une possibilite sans qu'il y ait une factualite qui lui corresponde?'" (p. 164, note). Dans le cas des mathematiques, la situation est la suivante: Le sentimentque I'on aest: en mathematiques, i1 ne peutpas y avoirrealite et possibilite. Tout est sur un seul plan. Et c' est adire, en un certain sens, ree!. Et c' est exact. Car ce que les mathematiques expriment avec leurs signes est tout sur un seul plan; c 'est adire: elles ne parlent pas tantöt de leur realite et tantöt de leur possibilite. Elles ne doivent, au contraire, pas du tout essayer de parler de leur possibilite. Mais i1 y a bien dans leurs signes une possibilite, asavoir la meme qui reside dans les propositions proprement dites, dans lesquelles les mathematiques sont appliquees. Et lorsqu'elles essayent (comme dans la theorie des ensembles), d' enoncer (aussprechen) leur possibilite, c'est adire !orsqu'elles la confondent avec leur realite, alors on a le droit de les renvoyer aI'interieur de leurs frontieres. (PB, pp. 164-165) La difference que les mathematiques cherchent a faire entre leur possibilite et leur realite se manifeste evidemment en priorite dans la distinction que l'on croit devoir instaurer entre l'infini potentiel et l'infini actuel. Cantor souligne que c'est a peine si l'infini potentiel peut encore etre appele "infini", puisque ce que l'on designe sous le nom d'infini potentiel est en realite simplement du "fini variable". En revanche, l'infini proprement dit, l'infini actuel, est quelque chose de tout a fait determine: "Le concept 00, par exemple, ne contient rien de fluctuant, rien de variable, rien de potentieI, il n'est pas unümapov, mais un acpooptcr/-levov, et la meme chose est vraie de tous les autres nombres transfinis".ll Si 1'0n se demande quelle est la relation exacte qui existe entre ces deux concepts si differents de l'infini, la reponse de Can- 70 JACQUES BOUVERESSE tor est tout a fait c1aire: " ... Au vrai, I'infini potentiel n'a qu'une ｮｾ｡ｬｩｴ･＠ empruntee, dans la mesure Oll il renvoie toujours a un infini actuel, qui seul le rend possible. D'Oll I'epithete accolee de ｦ｡ｾｯｮ＠ pertinente par les scholastiques a I'infini potentiei: cruVKO:tl1yopl1 Jlu'tn:coc;". (ibid., p. 404) Wittgenstein plaide, au contraire, pour ce qu'on pourrait appeler l'autonomie radicale de la possibilite. Il soutient que l'on ne peut pas la decrire en la transformant, comme le fait la theorie des ensembles, en une realite infinie qui echappe par essence a toute possibilite de symbolisation directe: La theorie des ensembles est fausse pour la raison qu'elle presuppose apparemment un symbolisme qu'il n'y a pas, au lieu de celui qu'il y a (qui est seul possible). Elle construit sur UD symbolisme fictif, donc sur le non-sens. (PB, p. 211) Wittgenstein soutient que I'infini ne peut apparaitre dans les mathematiques que sous la forme de la possibilite infinie, qui est refletee, et non pas decrite, dans les possibilites infinies du symbolisme. Un axiome qui asserte I' existence de l'infini ne pourrait donc qu'etre un non-sens: L'axiom 0/ injinity est dejA un non-sens pour la raison que la possibilite de l'enoncer presupposerait UD nombre infini de choses - donc ce qu 'il veut affirmer. Des concepts logiques, par exemple de !'infinite, on peut dire que leur essence demontre leur existence. (ibid., p. 124) Dire qu'il n'y a pas de nombre entier naturel plus grand que tous les autres revient a dire que cela n'a pas de sens de parler d'un nombre de ce genre. La theorie des ensembles eherehe a transformer une determination grammaticale en une enonciation descriptive portant sur une extension infinie. La raison ultime de I'hostilite de Wittgenstein a la theorie des ensembles est, de toute evidence, que le mode de pensee ensembliste contribue plus que n'importe quel autre a masquer le caractere de propositions grammaticales que possedent les propositions mathematiques et a rapproeher beaucoup plus qu'il ne le faudrait le statut des mathematiques de celui d'une science naturelle. L' origine et le fondement des possibilites et des impossibilites que les mathematiciens decouvrent sont supposees, en effet, resider dans des collections infinies donnees, qui ne peuvent etre explorees que par des moyens indirects. On ne sait pas actuellement s'il est possible ou non de trouver quatre entiers positifs x, y, z et n tels que n > 2 et x n + yn = zn. Mais cela signifie que I' on ne sait pas s'il y a ou non quatre nombres de ce genre dans l'ensemble des entiers positifs. Ils y sont ou ils n'y sont pas. Mais, pour l'instant, seul un etre omniscient comme Dieu, qui connait les ensembles infinis dans leur totalite, sait laquelle de ces deux choses est vraie. Dans le Tractatus Wittgenstein remarquait deja que: LE REEL ET SON OMBRE 71 La theorie des classes est en mathematiques tout a fait superflue. Cela se rattache au fait que l'universalite dont nous avons besoin en mathematiques n'est pas l'universalite accidentelle. (TLP, 6.031) Dans les Philosophische Bemerkungen, Wittgenstein souligne que la generalite qui interesse les mathematiques ne peut pas etre la generalite amorphe, ee1le d'une proposition eomme "Toutes les pommes sont mures", mais uniquement la generalite dominee, strueturee et finalement eonstituee par la demonstration. Si les nombres natureIs eonstituaient une totalite donnee en extension, il serait eoneevable qu'ils possedent tous une eertaine propriete aeeidenteIle, c'est a dire d'une maniere teIle que 1'0n pourrait de ｦ｡ｾｯｮ＠ seulement eonstater qu'ils la possedent en les examinant tous l'un apres l'autre, sans avoir les moyens de demontrer que cette propriete qu'ils possedent tous est une propriete qu'ils doivent necessairement posseder. (Gödel s'oppose directement a Wittgenstein lorsqu'il soutient que nous pourrions tres bien nous trouver dans une situation dans laquelle il est possible de eonjecturer a la fois que I' on sera en mesure de verifier, pour n'importe quel nombre donne, qu'il possede une eertaine propriete, et qu'il n'existe cependant pas de demonstration universelle de ce fait.) Wittgenstein ecrit: I1 n'y a pas tous les nombres, justement parce qu'il y en a une quantite infinie. Et parce qu'il ne s'agit pas ici du 'tous' amorphe, comme dans Ia phrase 'Toutes les pommes sont mÜfes', Oll 1'ensemble est donne par une description externe, mais de la totalite de structures qui doivent precisement etre donnees en tant que telles. Cela ne concerne pour ainsi dire pas la logique de savoir combien il y ade pommes, lorsqu' on parle de toutes les pommes. En revanche, il en va autrement dans le cas des nombres, de ceux-ci elle est responsable individuellement. (p. 148) Le mode de pensee extensionnel de la theorie des ensembles encourage l'idee desastreuse que, meme lorsqu'on a affaire a une collection infinie, i1 y a eneore deux facons d' etablir une proposition universelle: I'une qui consisterait apasser successivement en revue tous les elements et que des etres finis comme nous le sommes ne peuvent utiliser, et l'autre qui consiste aaboutir au resultat par un moyen indirect ou en prenant un raecourei dont on ne peut pas etre certain apriori qu'il existe, ce raceourci etant eonstitue en l'oecurrenee par la demonstration. Wittgenstein soutient que, s'agissant de l'infini, une telle distinction n'ajustement aueun sens. De la theorie des ensembles, il dit qu'elle raisonne eomme si tous les elements d'une collection infinie etaient en quelque sorte contenus dans une eaisse que l'on peut transporter avec soi. Si l'on emporte la eaisse, on est sur d'emporter tous les elements. Mais si l'on veut s'assurer qu'ils possedent tous une eertaine propriete, il faut ouvrir la eaisse et trouver un moyen quelconque de les en sortir tous: 72 JACQUES BOUVERESSE La description le saisirait [l'infini] ä. peu pres comme on porte emballees dans une caisse une quantite de choses que l'on ne peut pas tenir dans les mains. Elles nous sont alors invisibles, et pourtant nous savons que nous les portons (pour ainsi dire de ｦ｡ｾｯｮ＠ indirecte). La theorie des agregats vend en quelque sorte chat en poche. L'infini n'a qu'ä. s'arranger comme il veut dans cette caisse. (PB, p. 206) II L'aspect le plus deconcertant de la conception que Wittgenstein developpe au debut des annees trente apropos de la possibilite et de l'impossibilite (logiques) est certainement son idee que toutes les impossibilites correspondent ades interdits grammaticaux et que toutes les choses que la grammaire exclut sont exclues de la meme fayon et dans le meme sens. Dans son compte rendu des leyons que Wittgenstein a donnees au cours des annees 1930-1933, Moore avoue qu'il avait ete passablement surpris par l'usage que Wittgenstein faisait de l'expression "regle de grammaire", et qu'il avait tente en vain de lui faire admettre qu'en tout cas i1 n'utilisait pas cette expression dans son sens ordinaire. L'idee de Moore etait que l'on devrait distinguer au moins deux sens auxquels on peut dire qu'une regle de grammaire a ete violee: (1) le sens usuel, qui fait reference a une regle grammaticale reconnue et explicite comme celle qui interdit de dire, par exemple, "Trois hommes va en bateau"; (2) le sens philosophique wittgensteinien, en vertu duquel on peut qualifier de faute contre la grammaire l'affirmation que deux couleurs differentes figurent au meme endroit en meme temps dans un champ visuel. Moore ne consent pas a mettre sur le meme plan une simple incorrection grammaticale, au sens usuel, comme celle de la phrase "Trois hommes va en bateau" et une chose que la grammaire exclut comme inimaginable, inconcevable, impensable ou impossible (au sens logique), par exemple la co-occurrence de deux couleurs differentes au meme endroit du champ visuel. La reponse de Wittgenstein est la suivante: L'expression correcte est 'Cela n'a pas de sens de dire ... '; mais d'ordinaire nous utilisonsun mode d' expression defectueux qui consiste aparler d 'une regle de grammaire. Ainsi, cela n' a pas de sens de dire 'Cette table est aussi identique que I'autre' . Voir ce qui se passe si I' on utilise Ie meme echiquier et Ies memes pieces que nous Ie faisons pour Ies echecs, mais en faisant des mouvements qui ne sont pas prevus par Ies regles. Nous avons Ie sentiment que Ie premier mesusage auquel Moore fait reference est anodin, mais Ie second, vicieux. Mais en fait Ies deux sortes de regles sont des regles dans Ie meme sens. Ce qui se passe est simplement que certaines d'entre elles ont fait I'objet d'une discussion philosophique et d' autres non. Si nous discutons une regle, nous devons I' enoncer. Ces difficultes proviennent d'une fausse analogie. Ainsi Ie puzzle selon Iequel nous ne 'pouvons LE REEL ET SON OMBRE 73 pas' mesurer le temps est dü aI'analogie du temps physique. Nous sommes enclins adire que nous ne pouvons pas imaginer ou penser quelque chose, et nous sous-entendons que nous pourrions I'exprimer correctement si nous en avions I' experience. Dire que quelque chose est 'Iogiquement impossible' ressemble aune proposition. Ainsi, si nous disons que nous ne pouvons pas penser rouge et bleu ensemble dans le meme espace visuel, nous avons I'impression d' essayer de le faire, comme si nousparlions du monde physique; nous nous dupons d'une certaine ｦ｡ｾｯｮ＠ nous-memes et croyons que celapeut etre fait. Les regles grammaticales sont toutes de la meme espece, mais ce n' est pas la meme erreur si un homme viole l'une d'entre elles que s'il en viole une autre. S'il utilise 'va' au lieu de 'vont', cela ne provoque pas de confusion; mais dans I'autre exemple, I'analogie avec l'espace physique (cf deux personnes sur lameme chaise) entraine une confusion. Lorsque nous disons que nous ne pouvons pas penser adeux couleurs ala meme place, nous commettons I' erreur de croire que c'est une proposition, bien que cela n' en soit pas une; et nous n 'essaierions jamais de dire cela si nous n 'cHions pas induits en erreur par une analogie. Il est trompeur d'utiliser les mots 'ne peut pas' parce que cela suggere une mauvaise analogie. Nous devrions dire: 'Cela n'a pas de sens de dire .. .'. La regle concemant rouge et bleu «2) ci-dessus) est une regle concemant I'usage du mot 'et'; etnous ne dirions que 'va' «1) ci-dessus) n'a pas de sensque si quelqu'un disaitque cela pose un probleme philosophique. (WL1930-32, pp. 97-98) La raison pour laquelle la premiere impossibilite ne pose pas de probleme philosophique, alors que la deuxieme en pose un, est que, dans le premier cas, nous avons affaire, semble-t-il, a une convention purement grammaticale, dont la violation n'empeche pas la phrase de rester comprehensible et de continuer a representer un fait possible, alors que, dans le deuxieme cas, la regle grammaticale, si l'on consent a l'appeler ainsi, donne l'impression de refleter une impossibilite objective et d'enoncer un fait important qui resulte de la nature meme de Ia couleur. Une incorrection grammaticale de I' espece ordinaire est une chose que l'auditeur peut corriger implicitement et qui n' affecte generalement pas ou pas beaucoup la comprehensibilite de la phrase; mais Ia violation d'une regle grammaticale au deuxieme sens, ne laisse apparemment subsister aucune possibilite pour la phrase de signifier encore quelque chose. L'idee que Wittgenstein critique est celle qui consiste a croire que, dans le cas d'une phrase grammaticalement bien formee et representant neanmoins une impossibilite logique ou UD non-sens comme "A cet endroit du champ visuel figurent en ce moment deux couleurs differentes", nous essayons bel et bien de faire quelque chose que nous ne parvenons pas a faire, a savoir combiner entre elles des significations de mots qui, apremiere vue, se pretent tout a fait a ce genre d'operation et qui pourtant se revelent etre telles que le sens propositionnel obtenu n'en est pas uno Un sens qui avait l'air possible se trouve etre, en depit des apparences, un sens impossible, un sens qui n'a pas 74 JACQUES BOUVERESSE de sens. Wittgenstein soutient qu'un non-sens n'a rien avoir avec un sens qui n'aurait pas de sens. Lorsqu'on dit que "Cela n'a pas de sens de dire que p" on veut dans tous les cas simplement dire que la phrase 'p' est, pour une raison ou pour une autre, retiree de la circulation, et non qu'elle a en quelque sorte un sens qui n'en est pas un. Wittgenstein souligne que: Lorsqu'on dit qu'une phrase est denuee de sens, ce n'est pas pour ainsi dire son sens qui est denue de sens. En realite, une combinaison de mots est exc\ue de la langue, retiree de la circulation. (PI I, § 500) Les mots qui indiquent un etat de choses impensable sont elimines de notre langage, comme par exemple n'importe quel bruit arbitraire, et la raison de leur elimination expresse ne peut consister que dans la tentation ou nous sommes de les confondre avec une phrase de notre langage. (PG, p. 130) La raison pour laquelle un non-sens syntaxique caracterise comme "ah est a" semble beaucoup plus radical et ne nous pose nonnalement aucun probleme philosophique serieux, alors qu'un non-sens semantique comme "Cesar est un nombre premier" peut nous en poser un, est donc uniquement qu'une analogie trompeuse, qui est tentante dans le deuxieme cas, ne peut pas nous tenter et nous tromper dans le premier. La position de Wittgenstein sur ce point est evidemment liee de fa90n tout a fait directe a la critique qu'il fonnule au debut des annees trente contre l'idee de ce qu'il appelle le corps de signification (Bedeutungskörper) qu'un mot a derriere lui et dont il n' est que la surface apparente. 11 ecrit: Considerez, l'analogie suivante: entre un cube ou une pyramide avec une surface peinte, derriere laquelle il y a un corps invisible, et un mot et la signification derriere lui. Toute position dans laquelle cette surface pourrait etre placee dependra de la position du corps solide qui est derriere elle. Nous sommes tentes de croire que, si nous savons qu'un cube est derriere la surface peinte, nous pouvons connaitre les regles qui gouvernent l'arrangement de cette surface avec d'autres surfaces. Mais ce n'est pas vrai. On ne peut pas deduire la geometrie du cube de la contemplation d'un cube. Les regles ne suivent pas d'un acte de comprehension. (WLl932-35, p. 50) L'analogie du corps de signification est discutee egalement dans la Grammaire Philosophique (p. 54) et evoquee en passant dans les Recherehes Philosophiques: On aimerait parler par exemple de la fonction du mot dans cette phrase. Comme si la phrase etait un mecanisme dans lequelle mot a une fonction determinee. Mais en quoi consiste cette fonction? Comment se manifeste-t-elle? Car il n'y a en verite rien de cache, nous voyons bel et bien la phrase dans sa totalite! La fonction doit se montrer dans le cours du calcul (Corps de signification.) (PI I, § 559) LE REEL ET SON OMBRE 75 Ce que Wittgenstein conteste est justement l'idee que les possibilites et les impossibilites grammaticales qui s' expriment dans les regles que nous adoptons pour l'usage d'un mot soient le reflet de possibilites et d'impossibilites cacbees qui concement le corps de signification que le mot a derriere lui. L'imagerie du corps de signification suggere que si nous connaissons les possibilites de combinaison des solides dont les mots constituent les surfaces peintes apparentes, nous connaissons ipso facto les possibilites d'agencement de ces surfaces, c'est a dire les combinaisons de mots qui sont possibles et utilisables et celles qui ne le sont pas. Lorsque deux ou plusieurs expressions ne peuvent etre combinees en un tout signifiant, c'est parce que cette combinaison est exclue par la geometrie de leurs corps de signification respectifs. Une combinaison de mots est denuee de sens lorsque les sens qu'il y a derriere les mots sont de nature telle qu'ils ne se pretent pas a un arrangement spatial susceptible de constituer a nouveau un sens. Le principe de compositionnalite du sens dit que le sens d'une phrase est une fonction compositionnelle du sens des expressions qui la composent ou encore que le sens d'une partie de la phrase est une partie du sens de la phrase. Comme le constatait deja Frege, si la denotation de la phrase est egalement une fonction de la denotation de ses parties, ce n'est cependant pas une fonction compositionnelle, puisque la denotation d'une partie n'est generalement pas une partie reconnaissable de la denotation du tout. Mais aussi longtemps que l'on en reste au niveau du sens, cette fa90n de voir les choses semble parfaitement en ordre, et elle est tres tentante. Lorsqu'on se demande pourquoi la phrase 'Socrate est identique' n'a pas de sens, on a l'impression d'essayer de combiner entre eux deux sens partiels, ou, dans le langage de Wittgenstein, deux corps de signification et de s'apercevoir que c'est impossible. On pense que, si une phrase de ce genre n'a pas d'usage, c'est parce qu'elle ne peut pas en avoir; et si elle ne peut pas en avoir, c'est parce qu'elle represente une tentative pour signifier quelque chose, que la nature des significations partielles utilisees empeche d'aboutir. Wittgenstein voudrait, semble-t-il, que nous nous en tenions autant que possible a l'idee que la phrase en question n'a simplementpas d'usage, qu'elle a ete, avec une quantite d'autres, et pour une multitude de raisons qui peuvent differer considerablement d'un cas a l'autre, mise en quelque sorte sur une voie de garage, sans essayer de justifier les exclusions de ce genre en termes de signification et de possibilites ou d'impossibilites de signifier. Le rejet de l'idee du corps de signification est dirige justement contre l'illusion que les regles dont nous constatons la presence dans le langage decoulent de 76 JACQUES BOUVERESSE significations preexistantes et ont ete imposees par elles. Wittgenstein dit dans la Grammaire Philosophique qu'''on ne peut pas penetrer derriere les regles, parce qu'il n'y a pas de derriere." (p. 224) Les possibilites et les impossibilites de signifier auxquelles nous sommes tentes de faire appel pour expliquer et justifier les regles sont en realite le resultat immediat de I' existence des regles elles-memes et I' expression directe du fait que le jeu qui se joue est defini par les regles en question. La description la plus sobre et la plus appropriee de la situation serait d'ailleurs, selon Wittgenstein, ce1le qui consiste adire non pas "On ne peut pas ...", ce qui, en raison de la tendance acroire que toute possibilite doit reposer en fin de compte sur une factualite correlative, peut aisement nous induire en erreur, mais plutöt simplement "Dans ce jeu, i1 n'y a pas.. .".12 Dans la mesure Oll Wittgenstein estime que toutes les prohibitions de la grammaire sont finalement du meme type et qu'il n'y a pas lieu de considerer certaines d'entre elles comme refletant des impossibilites plus fondamentales et plus radicales que d'autres, i1 choisit de negliger ouverternent, du point de vue philosophique, des distinctions qui ont generalement ete considerees comme essentielles. Husserl, par exemple, soutient qu'il ne faut pas confondre la veritable absence de signification, ce1le des expressions du type "vert est ou", qui, aproprement parler, ne sont pas des expressions, avec ce qu'il appelle "l'impossibiliie apriori d'un sens remplissant"13, qui caracterise des expressions contradictoires comme cercle carre. Une expression a une signification dans le deuxieme sens "quand un remplissement possible, en d'autres termes la possibilite d'une illustration intuitive, formant une unite, correspond a son intention." (ibid.) Elle n'en a pas lorsqu'au contraire "nous apprehendons l'impossibilite ideale de ce remplissement de la signification en vertu du vecu de I' incompatibilite des significations partielles dans I'unite de remplissement intentionnee. "(ibid.) Husserl conteste la tendance que nous avons a qualifier de depourvues de signification au premier sens, des expressions contradictoires ou comportant des incompatibilites evidentes. 11 remarque que, si nous procedons de cette fa90n, nous devons egalement considerer comme depourvues de signification toutes les expressions dont les mathematiciens etablissent par des demonstrations que1quefois tres compliquees qu' "elles sont apriori sans objet" (ibid., p. 64) et etre prets a nier, par exemple, qu'un concept comme celui de decaedre regulier soit simplement un concept. Ce qui se passe en pareil cas, n'est pas que l'expression apparait au terme de la demonstration comme etant en realite denuee de sens, mais plutöt que son sens apparatt comme ayant ete construit d'une maniere telle qu'il est apriori exclu qu'un objet puisse lui correspondre. LE REEL ET SON OMBRE 77 En d'autres termes, il faut faire une distinction entre ce qui n'a pas de sens, le non-sens proprement dit (Das Unsinnige) et l'absurdite ou le contresens (Das Widersinnige), qui n'est pas une absence de sens: La combinaison un carre rond donne lieu vraiment a une signification unitaire, qui a son mode d'etre dans le monde des significations ideales, mais c'est une evidence apodictique qu'a la signification existante ne peut correspondre aucun objet existant. Si par contre nous disons un rond ou; un homme et est, etc., il n 'existe meme pas la de significations qui correspondent aces assemblages comme \eur sens exprime. Les mots qu'on a rassembles eveillent sans doute en nous la representation indirecte d'une certaine signification unitaire exprimee par eux, mais nous avons en meme temps l'evidence apodictique qu'une telle signification ne peut exister, que des parties de signification de cette nature et combinees de cette maniere sont incompatibles dans une signification unitaire. (Recherche IV, p. 121) Dans le premier cas, c'est simplement l'objet vise qui est impossible; dans le deuxieme, c'est la signification elle-meme. Dans le premier, "la signification elle-meme existe" (ibid., p.122), sans que l'objet puisse exister; dans le deuxieme, c'est la signification elle-meme qui ne peut exister. Ce que Husserl appelle la "morphologie pure des significations" a pour but d'exc1ure le non-sens, et non le contre-sens, c'est a dire de separer les significations qui ne peuvent exister de ce1les qui existent, de fa«on completement independante de la question de savoir si ces dernieres sont "objectives" ou "sans objet" (cf. p.129). La "grammaire pure logique", comme I'appelle encore Husserl, traite des lois du sens, au sens strict, ou des formes de signification possibles en general; et les lois qu' elle formule, qui correspondent a ce qui, dans la grammaire, est "purement" grammatical, c' est-adire apriorique, sont universelles. Elles peuvent l' etre dans la me sure Oll les entites sur lesquelles elles portent, les significations, existent ou n' existent pas, sont possibles ou ne le sont pas d'une fa«on qui est independante de la question de savoir si elles sont ou non exprimees dans une langue quelconque et de la fa«on particuliere dont elles le sont, si elles le sont. Par apriori, Husserl entend ici evidemment quelque chose de beaucoup plus fort qu'une simple universalite a Ia Chomsky. Ce qu'il designe par la n'est pas simplement "ce qui est universellement humain au sens empirique" (pp. 132-133), mais un apriori pur, c'est a dire purifie de tout element anthropocentrique et de toute relativisation implicite de ce genre. Dans la sixieme Recherche, Husserl distingue entre des significations possibles (reelles) et des significations impossibles (imaginaires). Mais il entend cette fois par possibilite (realite) d'une signification le fait qu'il lui corresponde un sens remplissant ou "une intuition complete dont la matiere est identique a la sienne" (p. 129): 78 JACQUES BOUVERESSE Quand, par exemple, nous realisons, sur la base de l'intuition, la signification de l'expression suiface blanche, nous avons I'experience vecuede la realite de ce concept, le phenomene intuitif represente veritablement quelque chose de blanc et une surface, et cela precisement en tant que surface blanche; et c'est ainsi que I'intuition remplissante ne represente pas seulement, en general, une surface blanche, mais fait de celle-ci, au moyen de son contenu, une donnee intuitive, et cela d'une maniere aussi complete que l'exige l'intention de signification. (p. 131) Quand a I'impossibilite, elle ne doit pas seulement etre definie comme negation de la possibilite, mais "reatisee au moyen d'un fait phenomenologique propre." (ibid.) Ce fait phenomenologique propre doit etre eherehe dans la sphere du conflit ou de I'inconciliabilite entre contenus. Le correlat de la conciliabilite des contenus est la "possibilite" des significations complexes. Par exemple, la conciliabilite des especes rougeur et rondeur fait que I' expression "rouge et rond" exprime une signification possible, meme si les deux especes ne se trouvent jamais reunies empiriquement. En revanche, I'incompatibilite du rouge et du vert rend impossible ou imaginaire la signification de l'expression "rouge et vert" (au meme moment et sur toute sa surface). Husserl prend soin de preciser que: ... Quand on parZe de conciliabilite en ce qui concerne les significations (Ies 'concepts'), il ne s'agit pas 11\ seulement, en general, de la possibilite qu'elles se concilient en un tout, serait-ce meme en un tout de significations - ce qui serait bien plutöt la conciliabilite grammaticaZe pure Zogique au sens de la IVeme Recherche - mais ( ...) de la conciliabilite de la signification en une significationpossible, c'est Adire en une signification qui est conciliable avec une intuition correspondante en l'unite d'une connaissance objectivement adequate. (p. 133) Apropos de I' exemple sur lequel Wittgenstein est revenu sans cesse, l'incompatibilite du vert et du rouge, Husserl note que le conflit produit une "separation" entre rouge et vert, mais contribue en meme temps a produire une nouvelle unite concemant les elements: conflit, rouge, vert. En revanche, i1 y adesunion entre les elements: conflit, rouge, rondo La synthese de la connaissance etait conscience d'une certaine concordance (Übereinstimmung). Or, A la concordance correspond, A titre de possibilite correlative, la 'non-concordance', le 'conflit' (Widerstreit). L'intuition ne 'concorde' pas avec I'intention de signification, elle 'est en conflit' avec elle. Un conflit separe, mais le vecu du conflit met en relation et pose une unite, il est une forme de lasymhese. Si la synthese precedente etait du genre de 1'identification, celle-ci est du genre de la distinction (nous ne disposons malheureusement pas ici d'un autre nom positif). (ibid., p. 58) Si I'on considere les choses a la ｦ｡ｾｯｮ＠ de Husserl, on peut dire que, si une expression comme "un objet rouge et vert" n' a pas de sens, c' est parce qu' elle tente d'exprimer une signification qui est en reatite impossible ou imaginaire. LE REEL ET SON OMBRE 79 En d'autres termes, l'absence de sens tire son origine d'une impossibilite; et l'impossibilite repose en derniere analyse sur un certain type de factualite ou de verite concernant en l'occurrence les significations elles-memes. Cellesci, dans I' exemple considere, bien que conciliables en un tout au sens de la grammaire pure logique, se revelent inconciliables avec une intuition correspondante dans une unite de type cognitif. Comme on I' a vu, toute la conception de Wittgenstein cherche au contraire a discrediter l'idee que les possibilites et les impossibilites de surface (le fait que certaines combinaisons de mots sont autorisees et d'autres non) refletent des possibilites et des impossibilites reelles concernant des choses qui se passent a l'arriere-plan du langage. Ce qui rend tres tentante une vision des choses comme celle de Husserl est justement le fait que, pour quelqu'un qui connait le langage, les mots n' ont pas seulement une surface, mais egalement une profondeur, dans laquelle se passent apparemment les choses essentielles. Wittgenstein est revenu suivante frequemment sur cette idee Oll cette impression, qu'il decrit de la ｦ｡ｾｯｮ＠ dans le § 594 des Recherehes Philosophiques: Mais les mots, prononces d'une fa<;:on qui a un sens (sinnvoll) n'ont tout de meme pas seulement une surface, mais egalement une dimension en profondeur: 'Quelque chose d'autre a bien Heu, justement, lorsqu 'ils sont prononces avec un sens, que ce qui a Heu lorsqu'ils sont seulement prononces.' - La maniere dontj'exprime cela n'est pas la chose importante. Que je dise qu'ils ont dans le premier cas une profondeur, ou que quelque chose se passe a ce moment-Ja en moi, dans mon interiorite, ou qu'ils ont une atmosphere - cela revient toujours aumeme. 'Si maintenant nous sommes tous d'accord la-dessus, est-ce que cela ne va pas etre vrai?' (Je ne peux pas accepter le temoignage d' 'autrui', parce que ce n'est pas un temoignage. Cela me dit seulement ce qu'il est enclin a dire.) 11 y a bien des fa90ns d'exprimer le fait que les mots que nous pronon90ns en voulant dire quelque chose par eux contiennent quelque chose de plus que ce que I' on voit et entend, quelque chose qui est justement I' essentiel. Mais Wittgenstein semble estimer que tout ce que I' on peut etre tente de dire sur ce point correspond davantage a une inclination pour une certaine forme de pensee ou un certain mode d'expression qu'a la solution d'un probleme. Et comme il le dit: "En philosophie, on doit distinguer entre des propositions qui expriment notre inclination de pensee et celles qui resolvent le probleme". 14 11 n'est pas contestable que la prononciation et l'audition de mots d'une langue qui nous est familiere sont accompagnees de sensations, d'impressions et d'experiences caracteristiques qui manquent totalement lorsque nous proｮｯｾｳ＠ ou enten dons des mots dans une langue que nous ne comprenons pas: Celui qui lit une phrase dans une langue courante pour lui ressent les mots des differentes categories verbales de fa<;:on tout a fait differente. (L'image du corps de signification). Nous 80 JACQUES BOUVERESSE oublions totalement que les images phoniques ou graphiques 'non' et 'table' et 'vert' sont des choses homogenes et ne voyons clairement I'uniformite des mots que dans une langue qui nous est etrangere. (Cf William James sur les sensations qui correspondent ades mots comme 'non', 'mais', etc.). (PG, p. 58) La realite des experiences mentales qui sont ici evoquees n' est evidemment pas en question. Ce qui l' est, aux yeux de Wittgenstein, c' est notre tendance a expliquer leur occurrence par I'action des corps de signification tres differents que les mots des differentes categories ont derriere eux, alors que celui qui ignore le langage n'est sensible qu'a l'uniformite de surface. Illui manque justement la dimension essentielle, a savoir la profondeur. Wittgenstein manifeste la plus grande mefiance a l'egard de I'idee que la signification d'une phrase douee de sens est quelque chose comme un tout compose des significations de ses parties: La signification n' est pas I' experience vecue que I' on a lors de I' audition ou de la prononciation du mot, et le sens de la phrase n'est pas le complexe de ces vecus. - (Comment le sens de la phrase 'Je ne I'ai toujours pas vu' se compose-t-i1 a partir des significations de ses mots?) La phrase est composee des mots, et cela suffit. (PI, pp. 500-501) Une fois rendue suspecte cette idee que la signification de la phrase est le complexe des vecus de signification associes a ses mots, il est evidemment difficile de continuer a dire que l' experience du non-sens est quelque chose comme l' experience de 1'impossibilite de combiner des significations partielles en un tout coherent: 'Le concept s. n'estpas un s.' Est-ce un non-sens? Eh bien, je ne sais pas ce que quelqu'un qui dit cela veut dire par Ja: c'est-a-dire de quelle fa90n iI a l'intention d'utiliser cette phrase. Je peux m'imaginer mainte utilisation facile a concevoir pour elle, - 'Mais tu ne peux justement pas I'utiliser ou meme simplement la penser d'une maniere teile que par les mots 'Le concept S.' et par le deuxieme 'S.' on veuille dire la meme chose que tu veux dire habituellementparces mots'. C'est la que reside I' erreur. On pense ici comme si on avait devant les yeux cette comparaison: les mots dans la phrase vont ensemble, c'est a dire on peut ecrire la suite de mots denuee de sens; mais la signification de chaque mot est un corps invisible, et ces corps de signification ne vontpas ensemble. ('Le vouloir dire donne a la phrase une dimension supplementaire' .). D'ou l'idee que I'on ne peut pas penser la phrase, car dans la pensee je devrais maintenant composer les significations des mots en un sens, et cela ne va pas. (Puzzle). Mais la contradiction n'est-elle pas defendue par la loi de contradiction? - 'non (p & non-p)' ne defend en tout cas rien. C'est une tautologie. Mais si nous interdisons une contradiction, alors nous excluons les formes contradictoires de notre langage. Nous eliminons ces formes. 15 Universite de Paris I Pantheon-Sorbonne LE REEL ET SON OMBRE 81 NOTES (*) Page numbers of the quotations of the German or English texts here translated into French refer to the Basil Blackwell edition ofWittgenstein's works. 1 The Blue and Brown Books (BLB and BRB), ed. and with aPreface by R. Rhees, B. Blackwell, Oxford 1958. 2 Cf. Wittgenstein 's Lectures. Cambridge, 1930-1932 (WL1930-32). From the Notes of J. King & D. Lee, ed. by D. Lee, B. Blackwell, Oxford 1980, p. 30. 3 Cf. Tractatus Logico-Philosophicus (TLP), ed. by D. F. Pears & B. F. McGuinness, Routledge & Kegan Paul, London 1961. 4 See Wittgenstein's Lectures. Cambridge, 1932-1935 (WLI932-35). From the Notes of A. Ambrose & M. Macdonald, ed. by A. Ambrose, B. Blackwell, Oxford 1979, p. 69. 5 Cf. Philosophische Untersuchungen. Philosophical 1nvestigations (PI), ed. by G. E. M. Anscombe & R. Rhees. Trans. by G. E. M. Anscombe, B. Blackwell, Oxford 1953. 6 Cf. Philosophische Grammatik (PG), hrsg. von R. Rhees, B. Blackwell, Oxford 1969. 7 Cf. Wittgenstein 's Lectures on the Foundations olMathematics. Cambridge, 1939. (LFM). Based on Notes by R. G. Bosanquet, N. Malcolm, R. Rhees & Y. Smyhies, ed. by C. Diamond, The Harvester Press, Hassocks, Sussex 1976. 8 Bemerkungen über die Grundlagen der Mathematik. Remarks on the Foundations 01 Mathematics, ed. by G. H. von Wright, R. Rhees & G. E. M.Anscombe. Trans. by G. E. M.Anscombe, B. Blackwell, Oxford 1964, p. 164. 9 Cf. Gottlob G. Frege, Kleine Schriften, hrsg. von I. Angelelli, G. Olms, Hildesheim 1967, p. 253. 10 (PB). Aus dem Nachlaß hrsg. von R. Rhees, B. Blackwell, Oxford 1964. 11 Cf. Georg Cantor, 'Mitteilungen zur Lehre vom Transfiniten', in Gesammelte Abhandlungen, mathematischen und philosophischen 1nhalts, hrsg. von E. Zermelo, G. Olms, Hildesheim 1966, p.390. 12 Cf. Zettel, ed. by G. E. M. Anscombe & G. H. von Wright. Trans. by G. E. M. Anscombe, B. Blackwell, Oxford 1967, § 134. 13 Cf. Edmund Husserl, Recherches Logiques, Tome I, traduit par Hubert EHe, Tome II et III, traduits par Hubert Elie avec la collaboration de Lothar Kelkel et Rene Scherer, P.U.F., Paris 1959-1962. 14 L. Wittgenstein, Letzte Schriften über die Philosophie der Psychologie. Last Writings on the Philosophy olPsychology, Vol. I, ed. by G. H. von Wright & H. Nyman. Trans. by C. G. Luckhardt & M. A. E. Aue, B. Blackwell, Oxford 1982, § 109. 15 Bemerkungen über die Philosophie der Psychologie. Remarks on the Philosophy 01 Psychology, Vol. I, ed. by G. E. M. Anscombe & G. H. von Wright. Trans. by G. E. M. Anscombe, B. Blackwell, Oxford 1980, §§ 42-44. CARLO CELLUCCI WITTGENSTEIN ON THE MEANING OF LOGICAL SYMBOLS 1. The nature of meaning was a major concern of Wittgenstein throughout his work. In this note 1 shall focus on a special aspect of the matter, the question of the meaning of the universal quantifier and more generally the question of the meaning of logical symbols. 2. Wittgenstein maintains that the meaning of logical symbols is given by their inference rules: We can conceive the rules of inference as giving the signs their meaning because they are rules for the use ofthese signs ([W4] VII 30.) It seems indisputable that the inference rules meant here by Wittgenstein are formal rules, i.e. mechanical rules: Every rule you give for the use of a symbol could be given in the fonn of a mechanism [... ]. Any rule can be imagined to be a description of a mechanism ([WS] XXX p. 282.) Of course, there are several kinds of inference rules. Which rules give the meaning of logical symbols? Wittgenstein provides an answer to this question in the case of the universal quantifier: One leams the meaning of'all' by leaming thatfa follows from (x).ft ([W4] 110.) The meaning of (x).ft is made clear by our insisting onfa's following from it ([W4] I 11.) The interesting case of generality is this: we often have a means of ascertaining the general proposition before considering particular cases; and we then use the general method to judge the particular case ([W4] V 50.) What Wittgenstein is claiming here is that the meaning of the universal quantifier is given by the rule of universal elimination: (\fE) \fxcp(x) <p(t) Such a view was conceived very early by Wittgenstein, as it appears from a letter to Russell of 1.7.1912: Will you think that I have gone mad if I make the following suggestion?: The sign (x) . q»c is not a complete symbol but has meaning only in an inference of the kind: from I- ｱﾻ｣ｾｸ＠ 'fIX . tp(a) follows lfIa. Or more generally: from I- (x) . tpx. soCa) follows tp(a). I am - of course -most uncertain about the matter but something ofthe sort might really be true ([Wl] p. 12) 83 R. Egidi (ed.), Wittgenstein: Mind and Language, 83-9l. © 1995 Kluwer Academic Publishers. 84 CARLO CELLUCCI This should be seen against the background of a 'consequentialist' theory of meaning according to which the meaning of a proposition is to be given in terms of its consequences: We must be able to understand a proposition without knowing if it is true or false. What we know when we understand a proposition is this: We know what is the case ifthe proposition is true, and what is the case if it is false. But we do not know (necessarily) whether it is true or false ([W2] p. 98). 3. Wittgenstein's view was a very isolated one. At about the time when he elaborated it, an alternative view was put forward by the German logician Gerhard Gentzen. According to hirn, the meaning of logical symbols is given by the introduction rules rather than by the elimination ruIes: The introductions represent, as it were, the 'definitions' of the symbols concerned, and the eliminations are no more, in the final analysis, than the consequences of these definitions. This fact may be expressed as folIows: In eliminating a symbol, we may use the formula with whose terminal symbol we are dealing only 'in the sense afforded it by the introduction of that symbol' ([G] p. 80.) Thus, in particular, the meaning of the universal quantitier is given by the rule of universal introduction: (V'I) cp(a) V'x cp (x) Gentzen's view has become popular through the writings of Dag Prawitz and Michael Dummett who have been building upon it (see [Su] for references). On the other hand, Wittgenstein's view has long been overlooked in the literature. As far as I know, it was first considered by Prawitz [P], without explicitly crediting it to Wittgenstein: Since the introductions and eliminations are inverses ofeach other, Gentzen's idea to justify the eliminations by the meaning given to the constants by the introductions may be reversed. Instead of interpreting the constants [... ] as asserting the existence of certain constructions that build up formulas with these constants, we may interpret them as stating the performability of certain operations ([P] p. 289.) Later on the possibility of such a view was also considered by Dummett: Would it not be equally feasible to adopt the reverse procedure and take the meaning of a sentence as given in terms ofthe consequence ofuttering it? [... ] It is a matter oftaking the meanings of the logical constants as given by the elimination rules governing them and formulating a procedure for justifying arbitrary forms of argument, including introduction rules, by reference to them ([Du] p. 17.) THE MEANING OF LOGICAL SYMBOLS 85 4. Prawitz, however, considers the view that the meaning of logical symbols is given by the elimination rules as generally untenable: in his opinion such a view can be maintained only for a restricted c1ass of formulas. He draws this conc1usion by pointing out the difficulties involved in the notion of validity based on the elimination rufes. Such a notion is introduced by Prawitz as follows. An atomic system S is a set of inference rules whose premises and conc1usi on are atomic. If S is an atomic system, then M(S) is the natural deduction system whose inference rules are the rules of minimallogic M plus the rules of S. A derivation !!fl in M(S) is closed if it has no open assumptions and no open parameters; !!flis open if it is not closed. Let !!flbe a closed derivation of a sentence qJ in M(S). The notion: !!flis valid w.r.t. S, is defined inductive1y as follows: (i) if qJ is atomic, then !!flis valid w.r.t. S if !!flis a derivation in S; (ii) if qJ == qJ/,qJ2' then !!flis valid w.r.t. S if the derivation !!fl (AE) qJI AqJ2 qJi is valid w.r.t. S for i = 1,2; (iii) if qJ == qJI-+qJ2' then !!flis valid w.r.t. S if, for each extension S' ofS and for each c10sed derivation ｾ＠ of qJI that is valid w.r.t. S', the derivation !!fl ｾ＠ (-+E) qJI-+qJ2 qJI qJ2 is valid w.r.t. S'; (iv) if qJ == \fxI/f(x), then ｾｩｳ＠ valid w.r.t. S if, for each name t, the derivation ｾ＠ (\fE) \fXI/f(x) I/f(t) is valid w.r.t. S. An open derivation ｾｩｮ＠ M(S) is valid w.r.t. !!flin M(S) if, for each extension S' of Sand for each c10sed derivation !!fl' obtained from !!fl by replacing (1) each open assumption '" by a c10sed derivation of", that is valid w.r.t. S' and (2) each open parameter by a name, !!fl' is valid w.r.t. S. 86 CARLO CELLUCCI This validity notion seems to apply only to derivations of formulas containing no disjunction or existential quantification: Disjunctions and existential quantifications seem impossible to handle in this way since the induction over the complexity of the end formulas used in the definition of validity breaks down in these cases ([P] p. 290.) Indeed the clause for disjunction would be of the following form: (v) if CfJ = CfJ1vCfJ2 , then §?is valid w.r.t. S if, for each open derivation and that is valid w.r.t. S, the derivation [CfJ1] [CfJ2 ] §? ｾ＠ ｾ＠ (vE) _CfJ1_V_CfJ_2 _ _"' ___ "'_ is valid w.r.t. S. '" Since '" is a formula whose complexity can be higher than that of CfJ1VCfJ2, here the induction on the complexity of CfJ breaks down. Similarly in the case of existential quantification. 5. Prawitz's claim, however, seems to depend on the assumption that the mIes of disjunction elimination (vE) and existential elimination (3E) can be formulated only in the form ofGentzen [G]. We know now that we can introduce systems of natural deduction - the so-called sequent natural deduction systems (see [Bo], [Cl], [C2], [C3]) - which are free of most limitations of Gentzen's systems. While the original formulation of such systems still contained an existential instantiation mle of a kind similar to Gentzen's, this is by no means necessary, as shown by the systems of [Cl], [C2]. In the latter systems the mIes of disjunction elimination and existential elimination take the form: (vE) (vE) I1,CfJ 1vCfJ2 11, CfJl'CfJ2 11,3xCfJ(x) 11, CfJ ( e) 87 THE MEANING OF LOGICAL SYMBOLS The notion of validity based on the elimination mIes can be formulated in terms of these mIes in such a way that induction no longer breaks down. An alternative approach is provided by the so-called multiple conclusion natural deduction systems (see [K], [KK], [Si], [SS]), which however are of a different character. 6. The view that the meaning of logical symbols is given by the elimination mIes seems to provide an approach to the theory of meaning of logical symbols that is at least as viable as the more popular Gentzen-Prawitz approach. I am not claiming that it is an adequate approach, but only that it is at least as adequate as the Gentzen-Prawitz approach. Indeed, in certain respects it may even be more adequate. For example, consider the rule of induction, (IND) _ep_(O_)__ｖ｟ｸＨ･ｰＩｾＧ＠ Vxep(x) Let PRA be primitive recursive arithmetic, formulated as an atomic system. Then (IND) is valid w.r.t. PRA in the following sense: for each c10sed derithat is valid w.r.t. PRA and for vation ｾ＠ of ep(O) and ｾ＠ of Vx( ･ｰＨｸＩｾＧ＠ each numeral n, the derivation ｾ＠ (IND) ｾ＠ Vx( ･ｰＨｸＩｾＧ＠ ep(O) Vx ep (x) epen) can be transformed into a derivation §:rthat is valid w.r.t. PRA. Just let §:rbe the following derivation: ｾ＠ (VE) ep(O) Vx( ･ｰＨｸＩｾＧ＠ ｾ＠ ･ｰＨｏＩｾｬ＠ ) ＨｾｅＩ＠ (VE) Vx( ･ｰＨｸＩｾＧ＠ ･ｰＨＱｾ＠ epen-I) ･ｰＨｮＭｬＩｾ＠ ＨｾｅＩ＠ epen) ｾ＠ 88 CARLO CELLUCCI Therefore (IND) is valid w.r.t. PRA in a sense that is perfectly in line with Wittgenstein's view about (IND): 'The proposition (n)!n follows from the induction' only means that every proposition ofthe formj('n) follows from the induction ([W3] VI 31.) Such a view of induction is al ready expressed by Wittgenstein in his conversations with Waismann (cf. [Wal] IV) and is similar to the one reported by Waismann (cf. [Wa2], end of chapter 9). On the other hand (IND) is not valid w.r.t. PRA in the sense of GentzenPrawitz, i.e. in the sense of the notion of validity based on the introduction rules. For, generally two closed derivations, ｾｬ＠ of ep(O) and ｾＲ＠ of \tx( ･ｰＨｘＩｾＧＬ＠ that are valid w.r.t. PRA are not sufficient to obtain an open derivation of ep(a) that is valid w.r.t. PRA, for a parameter a. However such a derivation of ep(a) is required in order to obtain a closed derivation of \txep(x) that is valid w.r.t. PRA in the sense of Gentzen-Prawitz. Therefore the Gentzen-Prawitz view ofthe meaning ofthe universal quantifier is essentially more restrictive than Wittgenstein's view. Dummett and Prawitz do not seem to question Wittgenstein's principle that meaning is fully manifested in use. On the other hand, admittedly the Gentzen-Prawitz meaning of the implication and the universal quantifier, which is constructive in a strong sense, is very se1dom used in mathematics. It must be admitted that such a strong meaning of::> and V is seldom used [ ... ] this notion of validity of a formula in a system does not have natural mathematical applications. For instance, the induction axioms are not generally valid ([P] p. 288.) On that account one might conclude that, if meaning is to be fully manifested in use, then Wittgenstein's view ofthe meaning ofthe universal quantifier is more adequate than the Gentzen-Prawitz view. 7. Prawitz' s notion of validity based on the elimination rules is not the only possible implementation of the view that the meaning of logical symbols is given by the elimination rules. An alternative implementation is provided by tableaux. It was already suggested by Beth that the [tableau] rules [... ] exactly reflect the meaning ofthe words 'all', 'some' , 'not', 'and', 'if . .. then', 'or', 'if and only if', and ofthe symbols by which they have been replaced ([Be] p.2\.) This view has been recently taken over by Mondadori (see [MI], [M3]) who maintains, again without special reference to Wittgenstein, that tableaux THE MEANING OF LOGICAL SYMBOLS 89 are a way of implementing the view that the meaning of logical symbols is given by the elimination rules: Although tableau proofs really are (quasi-mechanical) tests for validity [... ], their structure somehow reflects the meaning ofthe logical operators involved in them [... ]. In fact, tableau rules are just one way of making precise what can be called the consequentialist theory of meaning according to which the meaning of a logical operator • is connected with the consequences of propositions of the form ·(PI, ... , Pn), in which • is the main logical constant. This conception of meaning implies that elimination rules - as opposed to introduction rules - are to be taken as primitive meaning rules ([M3] p. 4) Mondadori (see [MI], [M2], [M3], [Da]) also proposes an alternative implementation of this view, in terms of a calculus KE consisting of elimination rules plus (a version ot) the principle of excluded middle: (EM) A <p -,<p He considers (EM) to be a structural rule reflecting the assumption of a classical notion of truth. This proposal, however, does not sound very convincing. The main reason for considering (EM) as a structural rule seems to be that it is strictly related to Gentzen's cut rule: (Cut) Indeed, reading (Cut) upside-down, what the rule says is that, in all circumstances and for all formulas cp, either cp is true or cp is false (and hence -,cp is true). There is, however, a basic difference between the role of (Cut) in Gentzen's calculus ofsequents LK and the role of(EM) in Mondadori's elimination calculus KE. By Gentzen's Hauptsatz, (Cut) is eliminable in LK, or equivalently (EM) is eHminable in the system oftableaux (see [Sm] pp. 110-115 or [BM] pp. 31-33, 79-83). On the other hand (EM) is not eliminable in KE because eliminating (EM) from KE would yie1d an incomplete system (see [MI]). This means that, while (Cut) does not contribute to the meaning of logical symbols in LK, (EM) contributes to such a meaning in KE in an essential way. Thus, while (Cut) plays a role of a proper structural rule in LK, (EM) does not play such a role in KE: it is rather an additionallogical rule, distinct from the elimination rules. 90 CARLO CELLUCCI The situation is somewhat similar to that of Gentzen's natural deduction system NK where, as stressed by Gentzen (see [G] p. 82), the law of the excluded middle qJv-,qJ plays a role of a logical mle distinct from both the introduction mIes and the elimination mIes. It is just because of such a special role that Gentzen's view that the meaning oflogical symbols is given by the introduction mIes does not extend to classical logic. Since (EM) plays a role of a logical mle distinct from the elimination mIes, thus contributing to the meaning of logical symbols in an essential way, KE does not seem to provide a viable approach to a consequentialist theory of meaning. Universita di Roma 'La Sapienza' REFERENCES [Be] [BM] [Bo] [C 1] [C2] [C3] [Da] [Du] [G] [K] [KK] [M 1] [M2] Beth E.W. (1969): 'Semantic entailment and formal derivability', in The philosophy 0/ mathematics, ed. by J. Hintikka, Oxford University Press, Oxford, pp. 9--41. Bell J. and Machover M. (1977): A course in mathematicallogic, North-Holland,Amsterdam. Boricic B.R. (1985): 'On sequence-concIusion natural deduction systems', Journal 0/ Philosophical Logic 14, pp. 359-377. Cellucci C. (1988): 'Efficient natural deduction', in Temi e prospettive della logica e della filosofia della scienza contemporanee, vol. I, ed. by C. Cellucci and G. Sambin, Clueb, Bologna, pp. 29-57. Cellucci C. (1991): 'Sequent natural deduction and intuitionistic logic', in Nuovi problem i della logica e dellafilosofia della scienza, vol. 11, ed. by G. Corsi and G. Sambin, Clueb, Bologna, pp. 259--266. Cellucci C. (1992): 'Existential instantiation and normalization in sequent natural deduction systems', Annals 0/ Pure and Applied Logic 58, pp.III-148. D' Agostino M. (1990): Investigations into the complexity 0/some propositional calculi, Oxford University Computing Laboratory, Technical Monograph PRG-88, Oxford. Dummett M. (1980): 'Comments on Prof. Prawitz's paper', in Logic and philosophy, ed. by G.H. von Wright, Nijhoff, Dordrecht, pp. 11-18. Gentzen G. (1969): 'Investigations into logical deduction', in The collected papers 0/ Gerhard Gentzen, ed. by M.E. Szabo, North-Holland, Amsterdam, pp. 68--131. Kneale W. (1956): 'The province oflogic', in Contemporary British Philosophy, 3rd series, ed. by H.D. Lewis, Allen & Unwin, London, pp. 237-261. Kneale W. and Kneale M. (1962): The development o/logic, Oxford University Press, Oxford. Mondadori M. (1988): 'Classical analytical deduction', Annali dell 'Universita di Ferrara, n,s, section I1I, Discussion Paper No. 1. Mondadori M. (1989): 'Classical analytical deduction, Part 11', Annali dell'Universita di Ferrara, n,s, seetion III, Discussion Paper No. 5. THE MEANING OF LOGICAL SYMBOLS [M3] [P] [Si] [Sm] [Su] [SS] [Wal] [Wa2] [Wl] [W2] [W3] [W4] [W5] 91 Mondadori M. (1989): 'An improvement of Jeffrey's deductive trees', Annali deli 'Universita di Ferrara, n,s, section IH, Discussion Paper No. 7. Prawitz D. (1971): 'Ideas and results in proof theory', in Proceedings of the Second Scandinavian Logic Symposium, ed. by J.E. Fenstad, North-Holland, Amsterdam, pp. 235-307. Sikic Z. (1990): 'Premiss tree proofs and logic of contradiction' , Zeitschrtji fiir mathematische Logik und Grundlagen der Mathematik 36, pp. 273-280. Smullyan R.M. (1968): First-order logic, Springer-Verlag, Berlin. Sundholm G. (1986): 'Prooftheory and meaning', in Handbookofphilosophicallogic, vo!. III, ed. by D. Gabbay and F. Guenthner, Reidel, Dordrecht, pp. 471-506. Shoesmith D.J. and Smiley T.J. (1980): Multiple-conclusion logic, Cambridge University Press, Cambridge. Waismann F. (1967): Wittgenstein und der Wiener Kreis; Gespräche, aufgezeichnet von F. Waismann, ed. by B.F. McGuinness, Biackwell, Oxford. Waismann F. (1959): Introduction to mathematieal thinking, Harper, New York. Wittgenstein L. (1974): Letters to Russell, Keynes and Moore, ed. by G.H. von Wright, Blackwell, Oxford. Wittgenstein L. (1979): 'Notes on logic', in L. Wittgenstein, Notebooks 1914-1916, 2nd edition, ed. by G.H. von Wright and G.E.M. Anscombe, Biackwell, Oxford, pp. 93-107. Wittgenstein L. (1974): Philosophical Grammar, Eng!. trans!. by AJ.P. Kenny, Biackwell, Oxford. Wittgenstein L. (1978): Remarks on the Foundations of Mathematics, Eng!. trans!. by G.E.M.Anscombe, Blackwell, Oxford. Wittgenstein L. (1976): Leetures on the Foundations of Mathematies, Cambridge 1939, ed. by C. Diamond, The Harvester Press, Hassocks. AVRUM STROLL ON FOLLOWING A RULE The question of what it is to follow a rule, much debated in the current literature on Wittgenstein, is more complicated than most of its discussants seem to have realized. In general, their writings exhibit a tendency to overlook or blur important distinctions, lapses that make it difficult or even impossible to determine exactly what is at issue. Baker and Hacker are certainly right when they say that unless it is understood what rules are "confusions about rules ramify into muddies about following rules".1 In addition, Baker and Hacker are to be commended for their own discussion of rules, a discussion that is extensive and illuminating. Among the examples of rules they give are rules of games, legal rules, the trafik code, and rules of etiquette; and they point out that the range of sentence types that may be used to state rules is very wide. It inc1udes deontic sentences, imperatives, and ordinary dec1arative sentences inter alia, and they go on to emphasize that such sentence-types can be used for altogether different purposes as weH (p. 39). They also bring out certain essential features of rules, for instance, i) that any rule can be expressed, ii) that it must be possible to follow (or violate) a rule, iii) that rules are creatures of the will, iv) that they are standards of correctness and guides to action, and v) that they must be more or less transparent to participants in a rule-governed practice (pp. 62-63). Yet as broad-ranging as is their coverage ofthe topic there are things they tend to underemphasize and more importantly, there are things they overlook. My intention in this paper is to carry the subject forward and in ways that are philosophically significant. Their significance can best be appreciated by comparing what follows with Kripke's approach to the issue of rule following. In my view Kripke's interpretation of Wittgenstein, as deep and interesting as it is, is marred by focusing on too narrow a set of examples, in particular those involving the expansion of mathematical formulas. A considerable portion of this paper will thus be devoted to Kripke's treatment of this problem. There is one important issue, and the subt1e distinctions connected with it, that I will not pursue here. This concems the relationship between ru1e f0110wing and the notion of a private 1anguage, inc1uding the question ofwhether rule following is simply a matter of regularity (so that in principle someone 93 R. Egidi (ed.), Wittgenstein: Mind and Language, 93-105. © 1995 Kluwer Academic Publishers. 94 AVRUM STROLL can be said to be following a rule ifhis or her behavior exhibits such regularity) or whether rule following presupposes regularity and also a human community. The question is directly relevant to such conundrums as what could be meant by a private language and whether someone could have or use such a language. In their discussions of rule following Baker and Hacker in effect contend that regularity is a sufficient condition for rule following. Their interpretation thus comes dangerously elose to asserting what Wittgenstein is generally thought to be denying; and even ifthis is not precisely so, the interpretation seems to me wrong on its own merits. On this point, Kripke seems to me to be right as does Norman Malcolm. Here is what Malcolm says in a recent paper, 'Wittgenstein on Language and Rules': To speak a language is to participate in a way of living in which many people are engaged. The language I speak gets its meaning from the common ways of acting and responding of many people. I take part in a language in the sense in which I take part in agame - which is surely one reason why Wittgenstein compares languages to games. Another reason for this comparison is that in both languages and games there are rules. To follow the rules for the use of an expression is nothing other than to use the expression as it is ordinarily used which is to say, as it is used by those many people who take part in the activities in which the expression is embedded. Thus the meaning ofthe expression is independent of me, or of any particular person; and this is why I can use the expression correctly or incorrectly. It has a meaning independent of my use of it. And this is why there is no sense in the supposition that a forever-solitary person could know a language any more than he could buy and seil (Inquiry, 32, 1989, p. 22). However, as I have indicated, this issue is too complicated to be explored here; so I shall set it aside. But there are at least three other, overlooked, strands in this complex web that must be discriminated if certain liabilities are to be avoided and ifthe issue, or at least certain permutations on it, can be made susceptible to resolution and it is these upon which I shall concentrate. First, there is the question of how to describe what it is that is being followed. To my knowledge no commentator has distinguished from one another such things as edicts, decrees, rescripts, orders, ordinances, statutes, maxims, resolutions, regulations, judicial decisions, canons, precepts, formulas, regimens, rituals, recipes, habits, customs, standards, criteria, principIes, and certain sorts of practices. Yet some or all of these can, in certain contexts, be referred to as "rules", whereas in other contexts some of them may be distinguished from or contrasted with rules. Most of them are the sorts ofthings that can be said to be followed, but some ofthem are not, such as criteria. We do not follow criteria but apply them. In addition, it is important to distinguish between things that express rules - such as road signs - ON FOLLOWING A RULE 95 and the rules that are expressed by these things, a point that Baker and Hacker do not emphasize. And there are more complex cases. Take "formula" for instance. A formula is defined, in one of senses of the term, as a set of words that indicate a procedure which is to be followed in order to bring about some particular result. There are both mathematical and non-mathematical formulas. One who follows the injunctions laid down in a mathematical formula - for example, by extending aseries in a prescribed way - may be said to be following the rule for the expansion of that series. Such a rule does not permit improvisation or flexibility. Among the non-mathematical formulas are recipes; these consist of sets of words which give directions for preparing foods, medicines etc., typically by describing the proportions which go into such preparations. But cooks would not generally describe recipes as rules. Nor are they formulae that have an obligatory force. A mathematical formula must be be followed exactly and no doubt that is true of certain recipes; but it is assuredly not true of others. Recipes are often construed by chefs or cooks as guidelines or suggestions that allow for considerable variation and creativity, and not as injunctions to be followed slavishly. A cook who is following a recipe is not following an injunction nor is a cook who changes or modifies a recipe violating an injunction. It is thus possible to assert that a chef who deviates from a recipe is nonetheless following it and also without inconsistency to assert that a student who deviates from a mathematical formula is not following it. Clearly then what counts as a rule will depend on the kind of thing that is being adhered to or deviated from. Furthermore, we can distinguish various degrees of"strength" in the set of terms 1 have listed above. Consider, for instance, the differences between regulations and precepts, both of which can be described as types of rules. Regulations tend to be rules of maximum strength and obligatory force. They are generally issued by some person or by some group that has the power or authority to do so, such as a govemor or a city council. Accordingly, they often have the force of law behind them and their violations frequently carry sanctions. Because regulations tend to be detailed and specific in what they mandate, such sanctions also tend to be specific and detailed. In many of these respects we can contrast regulations with precepts. Both, of course, being rules are things that can be followed or not followed. But precepts are, as it were, "softer" in character than regulations. They tend to be personal rather than institutional, are generally not "issued" by authorities, but are adopted voluntarily, and usuaUy after some reflection. They are on the whole advisory rather than obligatory and are not typically susceptible 96 AVRUM STROLL to penalties, let alone to specific penalties. One who fails to act according to his or her percepts may suffer, not from the imposition of an extern al sanction but rather from guilt or shame. Both regulations and precepts are principles that may directly affect the ways that ordinary people live, as distinct say from mathematical formulas which do not, but their impacts on people's lives differ. Regulations tend to be devised and to be laid down by others and are often regarded by those they affect as irksome or even as unfair. But precepts are hardly ever regarded as unfair, even if they are demanding or even if one finds hirns elf unable to live up to them. In such cases, a person will not infrequently blame hirnself and not his precepts, since these are taken to express ideals towards which he ought to aspire. Both regulations and precepts are subject to change and modification, but precepts are less subject to outright replacement and are more susceptible to deve1opment, broadening and growth than regulations. They are normally general in scope rather than detailed and often have a pronounced moral character instead ofbeing neutral, as when it is said that a person's life is being guided by certain precepts. In many circumstances ofuse, it is patent that we could not, while preserving meaning, replace the term "precept" by the term "regulation". What will count as mle following will therefore vary depending upon whether one is referring to regulations or to precepts in using the term "rule" and similar remarks to apply to most of the other items I have listed above. Furthermore, it is important to stress that such mIes differ in their functions and therefore are subject to differing kinds of assessments, such as what (if anything) counts as a violation or an infringement of a mle, or what sorts of inaptnesses or infelicities in its formulation or application it is heir to. For instance, in a game such as contract bridge, there are what might be called "constative" mIes. These are formative mIes that define the game of contract bridge. One who does not play in total compliance to these mIes is not playing contract bridge. They include the provision that the deck used in the game must consist of exactly 52 cards, divided equally between four suits and the rule that each player must be dealt exactly 13 cards if the course of play that follows is to be bona fide. What is called a "misdeal" would be a violation of the latter mle and the hand would be annulled. But there are also mies of strategy in bridge and these differ from constative mIes. One who fails to take a finesse is still playing bridge ifhe is in compliance with the set of constative mIes, but he may not be optimizing his chances of winning by such a lapse. But the hand is not nullified for that reason. Each ofthese types ofmles is thus subject to different infelicity conditions, mIes of ON FOLLOWING A RULE 97 strategy being connected with winning, whereas constative rules are not; and it may be that ambiguities can arise about rule following depending on which sorts of felicity conditions are observed or not observed. Rules also differ with respect to how, and whether, questions concerning them are decidable, and by whom. In contract bridge, questions about whether a certain constative rule has been followed can in principle always be decided, for example by counting the cards, but whether one should have taken a finesse may be the subject of disagreement even among experts and even after post-game analysis. In legal contests, for instance, even where there is no doubt about what a particular rule means there can be disagreement among lawyers about whether the rule has been violated in a particular case; and even different levels of appellate courts may disagree. The issue of what it is to follow a "rule" may thus vary depending on such factors as whether a constative or strategie rule is being referred to and upon various contextual factors. In thelnvestigations 2 Wittgenstein is comparatively selective in the number (five in all, according to my count) of the items he identifies as rules. He speaks of rules themselves, of course, and also of formulae, orders, criteria, and practiees. Most ofthose who write about Wittgenstein or about the problem of rule following are much more myopic. They tend to focus, as Kripke does, on one or two examples, generally mathematieal expressions, where the issue posed is how to extend a particular formula; but it may be quite misleading to generalize or to draw presumably important philosophieal conclusions - as say Kripke does - from such special or restricted cases. I shall say more about this later. Second, one must distinguish the diverse roles that the concepts of understanding, meaning, and explanation play when the issue of rule following is broached. And these in turn must be distinguished from the concept of applying a rule, and that from operating in accordance with a rule. It is often said that understanding what a term means is a sufficient condition for one's correct1y applying it. And then by analogy it is claimed that if one can understand what a rule means it follows that one can correct1y apply it. But one of Wittgenstein's main points is to reject the supposed tie between the concepts of meaning, understanding, and that of correct application. Consider this passage. Wittgenstein writes: I say "There is achair". What in go up to it, meaning to fetch it, and it suddenly disappears from sight? - "So it wasn't achair, hut some kind ofillusion." - But in a few moments we see it again and are ahle to touch it and so on. - "So the chair was there after all and its disappearance was some kind of illusion". - But suppose that after a time it disappears again - or seems to disappear. What are we to say now? Have you rules ready for such cases - rules saying whether one may 98 AVRUM STROLL use the word "chair" to incIude this kind of thing? But do we miss them when we use the word "chair"; and are we to say that we do not really attach any meaning to this word, because we are not equipped with rules for every possible application of it? (PI I, § 80) In the entry that follows, Wittgenstein extends the point to operating in accordance with a rule. All this, however, can only appear in the right light when one has attained greater cIarity about the concepts ofunderstanding, meaning, and thinking. For it will then also become clear what can lead us (and did lead me) to think that ifanyone utters a sentence andmeans orunderstands it he is operating a ca\culus according to definite rules. (PI I, § 81) Wittgenstein's analysis thus indicates that one can drive a wedge between the concepts ofunderstanding (or meaning) and that ofrule following andlor of operating in accordance with rules. Wittgenstein thus allows for the possibility that one might understand a rule without knowing how to apply it or without acting in accordance with it. The latter point is especially obvious since nonconformists are persons who understand rules they refuse to obey. Third, we must distinguish three different perspectives on rule following from one another: a first person from a third person perspective (a contrast that is also frequently expressed in a plural form, in which "we" is contrasted with "they"), and a neutral point of view that is different from either. We find Wittgenstein's discussion in the Investigations moving in exceedingly complex ways between these three different positions. He will frequently begin a query from a first person aspect and then without emphasis shift to one, or not infrequently, to both of the other aspects even within the same sentence. Here are some examples, beginning with a first person case. As the quotation will illustrate, the issue is posed in terms of how can I know that I am developing the series correctly. " ... WhenI say I understand the rule ofa series, I am surelynot saying so because Ihavefound out that up to now I have applied the algebraic formula in such-and-such a way! In my own case at all events I surely know that I mean such-and-such aseries; it doesn't matter how far I have actually developed it." (PI, § 147) But then we have a shift to a third person perspective, as in the following. Wittgenstein asks: How can he know how he is to continue a pattern by hirnself - whatever instruction you give hirn? - (PI I, § 211) And here is a mixed example: So how am I to determine the rule according to which he is playing? (PI I, § 82) ON FOLLOWING A RULE 99 Even more common are instances in which the point of view is neutral or in which the issue is raised from other than a point of view. In the quotation I am about to read no mention of a point of view occurs. To carry it out correctly! How is it decided what is the right step to take at any particular stage? (PI I, § 186) Or again, What is the criterion for the way the formula is meant? (PI I, § 190) In contrast, Saul Kripke's emphasis is invariably on the first person. His characteristic question is: "How do I know that I am following a rule?" Kripke imagines a speaker saying to hirnself, "I follow directions that determine how I should go on, but what are these directions since I never explicidy gave them to myself?" Note the emphasis upon the word "I" in the following quotation. The basic point is this: Ordinarily, I suppose that in computing '68 + 57' as I do, I do not simply make an unjustified leap in the dark. I follow directions I previously gave myselfthat uniquely determine that in this new instance I should say , 125' . What are these directions? By hypothesis, I never explicitly told myselfthat I should say '125' in this very instance. Nor can I say that I should simply 'do the same thing I always did', ifthis means 'compute according to the rule exhibited by my previous examples'. That rule could just as weIl have been the rule for quaddition (the quus function) as for addition. The idea that in fact quaddition is what I meant, that in a sudden frenzy I have changed my previous usage, dramatizes the problem. 3 I believe that Kripke's concentration upon the first person point of view and his apparent unawareness ofthe frequent shift in perspective that we find in the text of the Investigations is evidence of a failure to appreciate the complex problem that Wittgenstein is considering. We can bring this out, beginning with a simple case where the issue of rule following is raised from a first person perspective. Let us see how the matter of whether a rule has been followed can be decided in that case and then compare it with what Kripke says. We can then consider more difficult cases where there are shifts in perspective to see how they might be resolved. Our purpose in examining this spectrum of cases is to emphasize the complexities that are involved both in stating the putative problem of rule following and also of deciding how, if at alt, it might be resolved. Our first example, then, will be a simple case, where the perspective is first person. We shall see that the question is both easily stated and easily answered. Let us assurne that what is in question is whether the constative rule of contract bridge that each person must be dealt exacdy thirteen cards 100 AVRUM STROLL has been violated. In an important respect this is like the question of whether a mathematical formula has been followed correctly, for the question is dear and there is a determinate, correct solution to it. If I am the dealer how do I find out that I have violated the rule? This can oecur in a number of ways: I may note that I have a card left over when I finish dealing; or another player may tell me that he has counted and found either more or fewer than thirteen cards in his hand. If in doubt ab out the accuracy of what he has said, I can also count his cards. There seems to be no special problem about resolving this first person case. But Kripke sets up his problem in such a way that it is not dear what the problem is supposed to be or how it would be solved. Suppose I am the person adding two numbers according to the usual rule for the addition of the positive integers. As Kripke describes the situation the problem is that I do not know whether in calculating the sum '68 + 57' I am using '+' (i.e., 'plus') in the same way as I did in the past. Have I forgotten what '+' means? That would be a natural way of interpreting the situation. If that were so, there are various steps to find out whether I am now using '+' in the way I formerly did. It is, of course, frequently possible to find out that one has forgotten something, but once that has been discovered, one ean then go on. Yet Kripke explicitly denies that the problem arises because I have forgotten what '+' means. But then we have some puzzles. How is it that I know what '68' and '57' mean but not what '+' means? What has happened to prevent me from knowing what '+' means? Why should I suppose that I am following the rule for quaddition instead of the rule for addition? Kripke does not ans wer any of these questions. But he should beeause they are crucial. For example, if I, the rule follower, do not know whether I am following the rule for addition or the rule for quaddition then why should I suppose that I have violated the rule for addition? And how can I go about arranging for my future actions to conform to it? We can contrast Kripke's formulation of Wittgenstein's supposed problem with that we have just presented about whether a constative rule of bridge has been followed. In the latter ca se we not only know what it would like be to follow or to violate the rule, but also what procedure or procedures can be invoked in order to deeide the matter of whieh of these we have done. But now let us expand the range of examples and consider a situation where there is both a first person and a third person point of view at issue. We can ask with respect to the same constative rule ofbridge: "How do I know that he (where he is now the dealer) has violated the rule?" Onee again, the answer is ON FOLLOWING A RULE 101 readily at hand. I, a non-dealer, count my cards and if I do not have exactly thirteen then I know that the dealer has violated the rule; or if I find I have thirteen, then I can ask the other players if they do so as weIl. If some of them do not, then we know that amisdeal has taken place; both the question and the mode of resolution are c1ear. And finally, suppose that the question is posed from "the view from nowhere", to use Tom Nagel's felicitous phrase, namely in these words: "How can it be decided that the rule has been violated?" This question does not invoke first, second or third person perspectives. And yet it is easily settled. For the cards can be counted and in this way the question can be settled. If there is some unease feIt by my putting the point in the passive voice, we can simply say that anyone can count the cards and determine whether a violation has occurred. All ofthis is perfectly straightforward. It is instructive to contrast this simple case, where a constative rule has been under consideration, with a case, that oftrespass, where the issues are of a higher order of difficulty; but in so doing, we should recognize that there are, say, legal cases such as that involving the Miranda Rule or state regulations concerning aborti on that are even more complex. The first difficulty one faces with respect to trespass arises with respect to the question of what the concept itself means. There are both legal and nonlegal uses for this notion and all of them raise conceptual problems. In some broad sense, the concept oftrespass involves an unwarranted passage by someone or some group into a domain whose proprietor is another person or group. But how shall we describe a case of trespass - was it a simple crossing of boundaries, an intrusion, an infringement, an invasion? And what is meant by "domain", and "proprietor" in this definition? The notion of a domain, for instance, is normally interpreted broadly. It might inc1ude certain activities, such as those of a busy executive whose time is his domain and which can be trespassed upon by another. The notion of "being a proprietor" can be expanded to inc1ude such an executive; so one can be a proprietor of something susceptible to trespass other than land or real property. Shall we say that cases of trespass are always cases of intrusions? And what does "unwarranted" mean? Let's briefly explore some ofthe complexities these questions raise. The notion of an intrusion covers a broad spectrum of activities, ranging from simply stepping into or walking across (say) the property of another to an actual invasion of that property. We can distinguish within this range such differing cases as encroaching, entrenching, infringing upon, and invading the property of another. 102 AVRUM STROLL Each of these concepts will have different felicity conditions. One who, say, while hunting, and without realizing that he has entered restricted boundaries, walks across a field owned by a farmer is not to be described as invading that piece of property though, depending on the circumstances, he can be said to be intruding upon it. And depending on the circumstances, as we shall see, such intrusion may be characterized as a case of trespass. In contrast, invasion presupposes a specific intent to enter the domain of another whereas this is not necessarily true of intrusion. Nor is our hunter encroaching upon the owner's property, since "encroach" conveys the sense of doing something gradually or perhaps stealthily - the term would more suitably describe the actions of the white settlers in America who gradually displaced the native Indians, for instance. The hunter is simply passing across that piece of land without authorization. But even unauthorized entry into a piece of property is not to be identified with trespass. Trespass requires that the owner would specifically forbid certain c1asses of persons from entering his property without his permission. But in this case, if the owner does not object to strangers hunting upon his property, then we do not have a case of trespass. So in addition to unauthorized entry, the owner's attitudes about such entry will make a difference with respect to how the crossing of that piece of land is to be characterized; and in particular, to the determination of whether a rule has been violated or not. Furthermore, what does "unwarranted" mean? Here again we have a vast array of possible answers, ranging from deliberate violations of explicit regulations (such as those committed by conscientious objectors who breach a promulgated govemment regulation by entering a fenced area at a military establishment) to simply entering some facility without being invited by the host and where no laws are involved at all. The concept of "unauthorized entry", for instance, is one type of unwarranted entry yet cases of unauthorized entry are susceptible to differing assessments ranging from those not having the express approval of say, the owner of a piece of property, to those which violate specific ordinances. A tradesman intent upon soliciting business may enter a property without authorization, annoying the owner by so doing, and in that sense his entry is unwarranted; but is it trespass? Simply on the basis of what has been said so far, it is doubtful that we would describe it as such. Chutzpah it certainly is, but trespass - well it is hard to say without a more detailed scenario. And we should distinguish that case from the case of a poacher, who with a wholly different intent and purpose, may enter an area without authorization. Assuming the satisfaction of certain other conditions, ON FOLLOWING A RULE 103 it raises no doubts: the poacher is violating a specific ordinance, and is guilty of actions that are punishable by law, whereas those of the aggressive salesman are not. One could explore these notions at greater length if time permitted, but since it does not, let us see how even these distinctions affect questions about rule following when they have perspectival implications ofthe sort mentioned above. Let's make the case as simple as possible. Suppose I am a hunter intent upon shooting some birds and while following a flock I cross a farmer's fields. How do I know that I have violated the rule of trespass in this case? The answer is much more complicated than it would be in the case ofthe constative rule of bridge where I am the dealer. It will depend on all sorts of circumstances. Has the farmer posted signs stating that anyone entering his property is guilty of trespass, or signs stating that shooting is not allowed? Suppose he hasn't - then how do I know I have entered forbidden terrain? If there are no signs I may not know that I am in unauthorized territory. Weil perhaps there are no signs, but it may be that the field has been fenced with wire. Then if I have cut the wire, or crossed over or under it in order to enter the field, I would at least know that I have entered a territory which is enclosed. But the fact that the field is enclosed does not entail that one who enters it is committing trespass; that, as I noted earlier, requires the satisfaction of other conditions as weil. But suppose the field has not been fenced. Then how do I know that I have trespassed? One possible answer, given the present scenario, is that I may not know that I have, and indeed I may never know that I have. So the situation is different from the case where I have misdealt some cards or from the case of the correct expansion of a mathematical formula. The question of whether I have or have not followed the rule in these latter cases does not typically depend on the sorts of contextual factors I have been describing. Such queries differ from those about trespass in that they are in principle completely resolvable. And as we have seen in the case of constative rules, violations can be easily decided no matter what the point of view or perspective that is assumed. But this is not true of trespass. Here the point of view may make a substantive difference. Let us vary the hunting scenario to see why. There I am, within a fenced enclosure, gun in hand aiming at some birds. The farmer, going about his daily activities, suddenly sees me and orders me off his property even using the word "trespass". Have I trespassed? Of course I have. Do I know I have without a doubt. The question has been decided for me by the farmer. 104 AVRUM STROLL Now let us alter the point of view. Suppose the question is: How do I know that another has trespassed? Suppose the "I" in this case is the farmer. Then I know that the hunter has entered the enc10sed field without my authorization. I mayaiso have posted signs that the hunter has not seen. There is no doubt in my mind that the hunter has entered the field without permission; the prohibition against trespass has been violated. So once again the farmer has resolved the issue - but this time it is because I am the farmer. But these two cases are not generalizable, and if we change the scenario again, even slightly, we can see why. We shall ask the same question: "How do I know that another has trespassed?" But now suppose that the person who asks is the hunter. Further, suppose a friend and I are hunting together and that we have become separated. Suddenly I see my friend in an adjacent field. How do I know that he has trespassed? WeIl, I may not: I may not have seen any signs forbidding shooting, I may not from my position have seen any fences, and I may not have seen my friend crawl under the wire. Now as I look at hirn, I just see hirn walking in a nearby field. 1 don't believe he has trespassed and 1 certainly do not know that he has. My situation is thus different from that of the farmer. He knows things that 1 do not know - and the fact that he does makes a difference with respect to my answering the question: "How do I know that he has violated a rule?" Clearly then the circumstances make an enormous difference with respect to a correct answer. My point in this essay has been that the question of rule following that Wittgenstein poses in the Investigations is more complex than most commentators have realized. I am suggesting here that before we can state and then hope to answer this question the three factors 1 have described in this paper must be taken account of. First, we must broaden our range of examples, and in particular not allow mathematical formulae to become paradigms of what rules are; second, we must distinguish what is involved in understanding a rule from knowing how to apply it, and finaIly, we must distinguish from one another various perspectival questions about rule following. Ifwe follow these prescriptions - they are surely not rules - then perhaps some "muddies" about rule following can be avoided. University 0/ Calijornia, San Diego ON FOLLOWING A RULE 105 NOTES I G. P. Baker & P. M. S. Hacker, An Analytical Commentary on the 'Philosophicallnvestigations " vol. 2, Wittgenstein: Rules, Grammar, and Necessity, B. Blackwell, Oxford 1985, p. xi. 2 Philosophical Investigations. Philosophische Untersuchungen (PI), ed. by G. E. M.Anscombe & R. Rhees. Trans. by G. E. M. Anscombe, B. Blackwell, Oxford 1953. 3 Quoted in The Philosophy and Language, ed. by A. P. Martinieh, Oxford University Press, New York, p.478. DIEGO MARCONI FODOR AND WITTGENSTEIN ON PRIVATE LANGUAGE In his book The Language ofThought (= LoT),l Fodor makes, among many claims, the following two: (a) Wittgenstein's private language argument (= PLA) "isn't really any good" (p.70); (b) Wittgenstein's argument is anyway not really against Fodor's own theory of the language of thought (pp. 68, 70). Concerning the first claim, I shall try to show that Fodor's refutation of Wittgenstein's argument fails. On the second, Fodor is right but only trivially: Wittgenstein's is not an argument against the possibility of a "language ofthought" in Fodor's sense because the language ofthought is not a private language in Wittgenstein' s sense. Indeed, it is not a language in Wittgenstein' s sense. I do not mean that it is not a language for it is a private language, and a private language can't be a language for Wittgenstein. I think it can be shown not to be a language (for Wittgenstein, and perhaps for most of us) without appealing to its privacy. On the other hand, I shall say nothing against (or for) Fodor's main views in the book: I will not be denying (or asserting) that we have something like an innate system of internal representations which plays a crucial role in the learning of our first naturallanguage. I believe that such views are independent of(a) and (b). 1. According to Fodor, the reason why the PLA isn't any good is that it is a verificationist argument. Suppose we grant to Wittgenstein that "unless there are public procedures for telUng whether a term is coherently applied, there will be no way of knowing whether it is coherently applied" (p. 70).2 To go from this to the further conclusion that, therefore, the term cannot be coherently applied is to assume verificationism - i.e. to assume that it makes no sense to assert that a term is (or isn't) coherently applied unless one has some way of telling. But verificationism is surely wrong, so Wittgenstein does not prove that terms of a private language cannot be coherently applied - even if we grant that in the case of a private language we could not know that they are so applied. But this is not Wittgenstein' s argument. The trouble with a private language 107 R. Egidi (ed.), Wittgenstein: Mind and Language, 107-115. © 1995 Kluwer Academic Publishers. 108 DIEGO MARCONI is not that we have no way of assuring ourselves that a term belonging to it is being used correctly (or coherently); the trouble is that we can't be mistaken about whether it is being used correctly, for "whatever is going to seem right to me is right".3 The private language is not agame at which we can't win, it is agame at which we can't lose. 4 IfWittgenstein were a verificationist in the PLA, his argument ought to be the following: (1) It makes sense for me to say that 'S' is being used correctly (or coherently) (by me) just in case I know how to ascertain that 'S' is being so used; (2) But I don't know (and can't possibly know) how to ascertain whether 'S' is used correctly (coherently); (3) Therefore, it doesn't make sense for me to assert that 'S' is being used correctly (or incorrectly). But, independently of whether or not Wittgenstein believed in verificationism at the time ofthe Investigations (step 1), he wouldn't have taken step 2: for we know aIl too weIl how to ascertain that 'S' is being used correctly. Simply, ifit seems to me that it is so used, then it is, otherwise it isn't. In the case of correctness in a private language, verification, far from being precluded, is immediate. So, Wittgenstein's is not a verificationist (or, for that matter, falsificationist) argument: the point is not that 'S is being used correctly by me' cannot be verified (or falsified). It can. The point is rather that the private language thought-experiment has been so set up that it simply does not make sense to speak of my being right in the use of 'S' "whether I know it or not": for "being right in the use of 'S'" has been defined as 'conforming to my private convention conceming the use of "S''', and, by hypothesis, I am the only one who can "check" whether the convention has been respected or not. There is no being right per se, whether I know it or not. 2. Fodor is sensitive to this aspect of the argument. In fact, he appears to agree with Wittgenstein on two things: (a) if the notion of a language of thought is to make sense, some characterization must be given of being right in the LoT ("using it coherently", p. 71); (b) such a characterization must not be in terms of a private convention (pp. 73, 78). It must be said that Fodor does not set down the latter requirement in so many words: but he appears to accept it implicitly. In so doing, it seems to me, Fodor essentially accepts the gist of the PLA. What he rejects is (1) that the notion of a non-public language is essentially connected with the idea of a private convention as to its use, and (2) Wittgenstein's alleged argument from the impossibility ofknow- FODOR AND WITTGENSTEIN ON PRIVATE LANGUAGE 109 ing whether a non-public language is being used correctly to the impossibility of its being used correctly (or of asserting that it is being used correctly). Conceming (2), 1 have tried to show that there is no such argument in Wittgenstein. Concerning (1), the burden of proof is entirely upon Fodor, in the sense that it is up to hirn to show that the notion of a non-public language can be dissociated from that of a private convention: as he hirnself accepts, this must be done by satisfying (a) and (b), Le. by giving some account of being right in the use of a non-publie language (such as the LoT) which is not based on a private eonvention. Fodor adds the further requirement that such an aceount must show that the sense in whieh a term of a PL is employed eoherently is "reasonably analogous" (p. 71) to the sense in whieh a term of a publie language is so employed (otherwise, why speaking of a language at all?). Aeeording to Fodor, it is generally true, for both publie and non-publie languages, that terms are used eoherently if there is "a stable relation" between the way they are used and the way the speaker believes the world to be (p. 71).5 A speaker S uses a predieate 'P' eoherently iff S applies 'P' to an objeet x just in case he believes that x is P. The stable relation that Fodor has in mind involves S's beliefs, on the one hand, and his assent-behavior on the other. Thus for instance, Fodor says, Smith uses 'Jones is siek' to represent the state of affairs in which Jones is siek iff Smith assents to assertions made by employing the form ofwords 'Jones is siek' just in case Smith believes that Jones is siek (p. 72).6 Now, says Fodor, in the ease of a public language, the relation between a eertain form of words and a eertain belief is mediated by the eonventions ofthe language: S applies 'P' to xjust in ease he believes that xis P beeause the eonventions ofthe language have it that such are the cases in whieh 'P' (rather than 'Q') is to be applied. This is to say that (in a publie language) the only reason we may have to eall P's 'P' is a linguistie eonvention. All this leads Fodor to the following general statement (relative to publie languages): (C) S uses 'a is F' to represent a's being F just in ease it is eonventional that (S believes that a is F iff Sassents to 'a is F'). But this formulation eannot be right. For suppose S misuses 'a is F' (he uses it to represent a's being G, say): then aecording to eondition (C) it should follow that It is not eonventional that (S believes that a is F iff Sassents to 'a is F'). 110 DIEGO MARCONI But it clearly doesn't: whether or not anything holds as a linguistic convention does not depend on the correctness of any single person's linguistic behavior. What does follow is the falsity of the embedded biconditional: if S misuses 'a is F' , he will not assent to 'a is F' just in case he believes that a is F; he will assent to it, say, just in case he believes that a is G. It thus seems that Fodor's condition is ill-phrased. What he wants is something like (C') (Given the conventions ofthe language) S uses 'a is F' to represent a's being F iff (S assents to 'a is F' just in case he believes that a is F). or perhaps the more traditional "S uses 'a is F' (as an expression of L) ...". Now, of course, what Fodor is aiming at is the analogue of (C) for a nonpublic language such as the LoT. He believes he can get it simply by replacing linguistic convention with "nomological necessity", and assent with standing in a particular computational relation to a formula (ofthe inner language). Disregarding for an instant the fact that (C) is clearly wrong, this yields (D) S uses 'a is F' [a formula in the LoT] to represent a's being F iff it is nomologically necessary that (S R-s to 'a is F' just in case S believes that ais FV The similarity of(C) and (D) allows Fodor to claim that (1) the notion of ''using a term coherently" can be made clear even in the case of a non-public language (such as the LoT) without appealing to private conventions at all, but only to nomological connections between propositional attitudes, internal representations and computational relations; (2) the notion is analogous to the corresponding notion as defined for public languages. We can have no better than an analogy, for public languages are conventional whereas the language of thought is not. If one is struck by the disanalogy, Fodor admits, he will be inclined to deny that the LoT is a language. Even such a denial, however, will not affect the main point: that first-Ianguage learning involves the use of an unlearned internal representational system. Which last point, as I said, I am not going to challenge. Instead, let us consider condition (D), which is supposed to define coherent use of a language such as the LoT. First of all, it is not at all clear whether (D) can be transformed so as to make it analogous to (C') (the right condition for public languages) rather than to (C) (the wrong condition). For (D) as it stands does convey Fodor's idea ofhow the expressions ofLoT have meaning: 'a is F' has a certain meaning (it means that a is F) iff it embodies the content of S's beliefthat a is F: i.e. iffit is a fact ofnature ("nomologically necessary") FODOR AND WITTGENSTEIN ON PRIVATE LANGUAGE 111 that S stands in relation R to 'a is F' whenever he believes that a is F. It does not seem that the same point would be made by the obvious analogue of (C'), namely (0') (Given the facts of nature) S uses 'a is F' to represent a's being F iff (S R-s to 'a is F' just in case S believes that a is F). Indeed, one receives the impression that the mi staken condition (C) has been constructed after the pattern of (0), which is what Fodor wants for a language such as the LoT. (0') just misses the point: what is a fact of nature is exacdy the connection between S's R-ing to 'a is F' and his believing that ais F. So, it seems that at least part ofthe analogy between the LoT and plain, public languages is lost. This, however, isn't so bad in itself. The core ofthe analogy consists in this, that coherence of use in a public language can be expressed in terms of the coherence betweert a speaker's propositional attitudes and his assent-behavior; and similarIy, coherence ofuse in the LoT can be defined as the coherence between the subject's propositional attitudes and his computational relations with respect to formulas ofthe "inner code". This is preserved even between (C') and (0). Trouble comes from another source. Suppose again that 'a is F' is misused. Given (0), this entails that It is not nomologicaIly necessary that (S R-s to 'a is F' just in case S believes that a is F). I.e., there are exceptions: it may happen that S believes that ais F but does not R to 'a is F', or that he does stand in such a relation to 'a is F' and yet does not believe that a is F. WeIl, so it may. But, would we describe such an event as a case of misuse ofthe LoT by S? Would we even describe it as a case of incoherent use of 'a is F' on S' s part? No; we would rather say that, contrary to hypothesis, 'a is F' - a certain formula ofthe inner code - does not mean that a is F. Notice that this is consistent with Fodor's overt formulation of (0): for (0) says that "'a is F' is used to represent a's being F iff...", and in such cases we would say that 'a is F' is not used to represent a's being F. But this is just equivocating: for what we were defining was proper, or coherent, or correct use of 'a is F'. And in such cases, "S is not using 'a is F' to represent a's being F" cannot be taken to mean "S is not using 'a is F' coherently (properIy, correct1y)". This is the important disanalogy between (C') and (0): for in the case of (C'), the above paraphrase is entirely in order. 3. Can there be such a thing as improper use ofthe LoT? Can one go wrong in the LoT? If one cannot, it may be doubted that the LoT is a language in 112 DIEGO MARCONI Wittgenstein's sense; for Wittgenstein insists that a language can be misused (by an individual speaker), that its expressions can be understood only to a limited extent, that they can be misunderstood. I don't know whether this is to be regarded as "an essential feature of language" but it does playa crucial role in the phenomena of the use of language - of the languages we know, of course. Perhaps some light upon this question can be cast by Fodor's repeated (pp. 66,67) analogy between the LoT and the computers' machine-Ianguage. Can a machine-Ianguage be misused or misunderstood? Certainly, the programmer can misuse a machine-Ianguage: he can make errors of syntax, or he can write a program that does not have the intended effect because the programmer has misunderstood the notation. On the other hand, what would it be for the computer itselfto misuse the machine-Ianguage? Not operating in accordance with the intended effects of the commands (intended by the engineers who built the computer and designed its machine-Ianguage)? But in such a case we wouldn't say that the computer is misusing the machinelanguage: we would say that something went wrong with the design of the whole apparatus, computer-cum-machine-Ianguage. Notice that, from the programmer's standpoint, the machine language is not "private" in any plausible sense. Indeed, it doesn 't even look private. Many people understand the same machine language and can use it to communieate among themselves. I doubt that even from the computer's standpoint it makes sense to say that a machine-Ianguage is private, for many computers ''understand'' the same machine-Ianguage. This is all right with Fodor, for he claims that even the LoT is in no way "epistemically private", i.e. inaccessible in principle. Ultimately, the LoT is just neurological machinery, and "neurological events are public" (p. 70). But perhaps we have a disanalogy here: for, independently ofthe "ultimately it's all neurons" business, I don't know whether it makes sense to speak of the same language of thought being "understood" by many different minds. Anyway, to conc1ude this part of the discussion: from the programmer's standpoint, one can go wrong in a machine-Ianguage, but then from that viewpoint a machine language is not private, or "inner", or non-public. The computer, on the other hand, cannot be said to misuse the machine language; so that one wonders whether the machine-Ianguage is a language from the computer's viewpoint. For instance, does it make sense to say as Fodor says (p. 66) that computers have a machine-language "in which they talk to themselves"? Or: what do we mean exactly by saying that the computer "understands" the machinelanguage, as it cannot misunderstand it? Here, Fodor does have an answer, for he distinguishes between, not two senses of 'understand' , but two ways of FODOR AND WITTGENSTEIN ON PRIVATE LANGUAGE 113 understanding (p. 66). In the first (and usual) way, a device understands a predicate 'P' if it has and employs a representation of its extension, which is itself given in a language the device understands. In the other "way", the device understands 'P' in the sense that it conforms to the conditions that such a representation would specify; i.e. it behaves as we would behave, understanding 'P'. This is plausible: but, again, it hides the disanalogy between the two cases. In the first "way", a predicate 'P' can be misunderstood for it is possible to have the wrong representation of its extension, or a representation that is too partial for competent use, etc. Not so in the second "way": the device cannot lai! to conform to the appropriate specifications. Perhaps it would be more appropriate to say that the computer obeys the machine language. But this is not an important point. We may be willing to stretch the concept of understanding so as to cover the computer' s handling of the machine language. And one may even be willing (though not so uncontroversially) to stretch the concept of language so as to include the machine language as seen from the computer's viewpoint (as opposed to the programmer's viewpoint, from which a machine-Ianguage is unproblematically a language, and a public language at that). Ifthe analogy with the LoT holds (but we saw there are elements of disanalogy) this would entail that even the LoT may be regarded as a language, in spite of the fact that we can 't go wrong in it. What matters, however, is that it cannot be so regarded for the purposes of Wittgenstein's argument against private language: for the argument crucially revolves around the fact that a language must be liable to misuse (as public languages are). So, Fodor is right in holding that the PLA does not apply to his LoT: not, however, because the LoT is not private in Wittgenstein's sense, as Fodor claims (p. 70), but rather, as I hope to have shown, because it is not a language in Wittgenstein' s sense. 4. Finally, has Fodor succeeded in giving an account of "using a non-public language coherently" that would meet the two requirements of (a) not appealing to private conventions and (b) making being right in the LoT reasonably analogous to being right in a normallanguage? The first requirement may be said to have been met, if we take Fodor's use of 'coherent' (as opposed to 'correct') seriously. If the notion of a machine operating coherently makes sense - i.e. if we disregard Kripkensteinian problems with the notion of coherence in the use of a language - then Fodor has made sense of the notion of "using a non-public language coherently", without appealing to private conventions (or so it seems). On the other hand, it is doubtful that the other 114 DIEGO MARCONI requirement has been met. Being right in a non-public language turns out to be quite different from being right in a regular, public language: for being right in a non-public language admits of no alternative. One can of course go wrong about a non-public language such as the LoT, in the sense of making false hypotheses ab out its workings: for instance, by entertaining the hypothesis that S R-s to P - a formula in the LoT - iff S believes that q, whereas this is not the case. But one cannot go wrong in the LoT, except in the deviant sense of suffering brain injuries or other such inconveniences. Thus it seems that Wittgenstein was right in strict1y connecting "being right or wrong" with the publicity of standards. Universita di Torino NOTES Crowell, New York 1975. Here one would be tempted to object that Fodor should be speaking of'correct' application rather than of 'coherent' application. But Fodor's choice ofwords is not without consequence (as we shalJ repeatedly see): for we speak of correct use or application of a word with respect to some standard or convention, whereas Fodor wants to claim that no convention is involved in the appropriate functioning of the language of thought - especially not a private convention, for Fodor agrees with Wittgenstein - implicitly - that there is no such thing as a private convention. I shall try to show in each case what difference is made by Fodor's choiee ofwords. 3 Philosophische Untersuchungen. Philosophical Investigations (PI), ed. by G. E. M. Anscombe & R. Rhees. Trans. by G. E. M. Anscombe, B. BlackwelJ, Oxford 1953, § 258. 4 Of course, for Wittgenstein it ensues that there is really neither winning nor losing: there is no game, really - there is only the imitation of agame, or a ritual which can be mi staken for agame. S The reference to beliefs (rather than to matters offact, or shared beliefs) is puzzling, and may be one case in which speaking of"coherent" rather than "correct" use oflanguage does make a difference. For suppose I systematiealJy apply the predieate 'cat' to dogs: I would not be said to be using the word correctly, irrespective ofwhether I do so because I systematically mi stake dogs for cats or because I (falsely) believe that 'cat' refers to dogs (Le. that everybody else uses or should use the word as I do). On the other hand, perhaps I could be said to be using the word coherently. 6 Notice thatto use 'Jones is siek' to represent the state ofaffairs in which Jones is sick is notjust using it coherently; it is using it correctiy. Perhaps we should be more careful here, and make a distinction between using 'Jones is sick' with the intention ojrepresenting the state of affairs that Jones is sick, and using it to represent the aforementioned state of affairs. Fodor's definition applies to the former situation, not to the latter. 7 According to Fodor (pp. 76-77), the biconditional "S R-s to 'a is F' just in case S believes that a is F" is not to be taken as a (mentalistic) definition of belief, but rather as the prospective finding of a cognitivistic theory ofbelief and other attitudes: we hope we'lJ be able to discover that, as a matter of fact, whenever S believes that p he is in a certain computational state with respect to an internal formula q, the translation ofp in the internal code. CalJ such a relation 'R'. I 2 FODOR AND WITTGENSTEIN ON PRIVATE LANGUAGE 115 Clearly, that S believes that a is F iff S R-s to 'a is F' (a formula in the internal code) is not a matter of convention: it is rather, according to Fodor, a matter of "nomological necessity", a regularity of nature. MASSIMO DELL'UTRI MENTALESIANS AND WITTGENSTEIN'S PRIVATE LANGUAGE In his careful analysis of Fodor's interpretation of Wittgenstein's 'Private Language Argument' (henceforth 'PLA'), Diego Marconi claims that Fodor does not entirely succeed in finding a way out of the PL marsh. What I would like to do is to present some further considerations ab out Wittgenstein's PLA and Fodor's Language of Thought - often referred to as 'mentalese' - in an attempt to shed some further light on the issue. We can begin by drawing an analogy between a person who develops a PL, and the situation of a 'brain in a val' (henceforth 'BIV'), as described by Hilary Putnam. I Taking into account that both situations are deemed inconceivable by their respective authors, let us first look at the man with a language exactly identical to that described in § 243 of Wittgenstein' s Philosophicallnvestigations, a man who has his own terms and sentences to report his intimate feelings. Let us call him 'the PL-man', and ask whether he manages to refer to the private items ofhis interior monologue. Wittgenstein certainly does not deny the possibility ofhaving an inner life, i.e. inner sensations, processes, motivations and the like, but he clearly points out that starting from a mentalistic position - the trouble we are finally confronted with is that we lack any means for talking about what is internal and private. Because ofthat, it amounts to our lacking the very subject about which we are trying to talk, i.e. the private items, as § 304 reveals. But why can't we talk about our private items? Wittgenstein's answer can be expressed by another question, namely, "How could the words in a PL acquire meaning and refer to something?" He discusses the problem of the reference of sensation terms, and finds out that neither criteria nor rules of correctness for private expressions (§§ 258, 259, 261-265) or identity of sensations or mental representations in general (§§ 253, 376 ff) are available in "An 'inner process' stands in need of outward this internal world, ｳｴ｡ｩｾｧ＠ criteria" (§ 580). Rules and criteria are obtainable only on the basis of a consolidated practice or, in other words, on the basis of a social context. Wittgenstein's analysis therefore reveals the essence oflanguage as a social means of communication, and stresses the importance of establishing the social context of use. It is in this context that terms acquire a meaning and their 117 R. Egidi (ed.). Wittgenstein: Mind and Language. 117-123. © 1995 Kluwer Academic Publishers. 118 MASSIMO DELL'UTRI use a reference. 2 But what our PL-man is bereft of - by definition - is just such a context, and hence cannot use his private words referentially. In this century a more peculiar attention to the social dimension of human activity - whatever its object - has been the tenet of a number of philosophers. This dimension, in particular, is also of capital importance for Putnam's causal theory of meaning,3 a theory which - for what concerns us here plays a central role in his argument against the BIV, which we will now look at in order to see how far the analogy with the PL-man can be pushed. The fantasy ofthe BIV is the contemporary correlate ofDescartes' hypothesis of the evil genius. 4 This time however, instead of the evil genius at play, we have a mad scientist who causes a person to hallucinate the whole world. After extracting the brain from the skull of a miserable subject and placing it in a vat filled with nutrient fluids, the scientist connects the brain's nerveendings with a highly sophisticated computer programmed to let the brain experience any situation he deemes upon it. Everything appears perfectly normal to the brain: it lives its usual life and experiences its normal bodily sensations, has a mouth for speaking, eyes. for seeing and ears for hearing and, in short, seems to be the person he always was, perhaps even laughing as it listens to the amusing but horrible story of a deluded guy whose brain was kept alive in a vat! Usually, such a story is employed in epistemology in order to raise a skeptical doubt as to the knowledge of the external world: "How do you know that you are not aBIV?" is the challenge posed by the skeptic. However, Putnam uses the BIV story for a different but connected aim: that of investigating the notion of reference. "Have your words a reference to the actual world?", or "Is the language you are using actually connected with the real world or is it a mere abstract and empty syntactic play?" is the question he engages to answer. 5 His peculiar treatment of the problem leads hirn to conclude that the words the BIV utter and employ in complex sentences do not have any relationship with the real objects existing outside the apparatus they live in. This conclusion partly depends on the causal theory of reference we mentioned above. Very briefly, according to this theory the use of a word by speaker S refers to an object x if it is causally tied to x, either directly or indirectly, and S's intentions are such that his use of that word is to refer to the object which counts as the referent of the word in question within the collective body of speakers. Hence, one of the central features of this theory is deeply Wittgensteinian. The reference of a term must be public, and whatever individual description given of the referent of a term used ｩｾ＠ no good, unless it is the description obtaining throughout the society. MENTALESIANS AND WITTGENSTEIN'S PRIVATE LANGUAGE 119 But Putnam's argument against the BIV hypothesis contains something more in respect to Wittgenstein's PLA. What he tries to do is to derive the knowability of reference from its public and social character. He tries to show that, given Wittgenstein's argument, i.e. given that what counts as the reference of terms is what is socially acknowledged as such, this very reference has to be knowable at least in principle. Indeed, if reference were not knowable, it wouldn't even be public. But ifreference is at least in principle knowable - Putnam argues - it follows that not every causal connection is good for reference. Only those causal connections which are part of a social practice of knowledge will do. Therefore, the relationship BIV could have with the world outside the vat plays no role for reference, because their knowledge is limited to what appears in the image furnished by the computer, and thus all their references are within the image. From this, then, Putnam concludes to the self-refuting character of the BIV hypothesis. This conclusion reveals that the stress posed on the social dimension is a common feature in the arguments of both Wittgenstein and Putnam. Indeed, Putnam's philosophy has many Wittgensteinian traits, and this conclusion does not faH to show traces ofWittgenstein's criticism ofmentalism, a criticism avowedly endorsed by Putnam. Explaining why mental representations do not constitute understanding, Putnam reformulates Wittgenstein' s exampIes in §§ 159-160 ofthe Philosophicallnvestigations. He invites us to imagine a person who does not know Japanese, leams a Japanese text by heart, and under the influence of a hypnotist says that text in his mind with the impression of understanding it. Here it is clear that, even if this whole arrangement is so perfect as to cheat a Japanese telepath, that person does not understand what he is saying, because he is not able to give appropriate answers to any questions and use the words in the right contexts. 6 Therefore, mental representations do not bring about understanding by themselves - and moreover, as Putnam puts it: Mental representations no more have a necessaty connection with what they represent than physical representations do. The contrary supposition is a survival of magical thinking. 1 The step from the lack of understanding to the lack of referential power, as the quotation reveals, is at hand. But there are obvious differences in the way Wittgenstein and Putnam arrive at realizing the importance ofthe social dimension. Putnam's argument is based upon his proper explanation of the mechanism of reference, while Wittgenstein's is actually based upon the notion of rule, and seems to have but a derivative impingement on the notion of reference. As Marconi says, 120 MASSIMO DELL'UTRI The trouble with a PL is not that we have no way of assuring ourselves that a term belonging to it is being used correctly (or coherently); the trouble is that we can't be mistaken about whether it is being used correctly, for 'whatever is going to seem right to me is right'. (pp. 107--8) Yet Wittgenstein's argument also has his implications as to the notion of reference - as we have been saying - and it is possible to notice this by reformulating Marconi's wording. In the reformulation we can see Wittgenstein claiming that we can't be mistaken about whether a term in a PL has a reference, because whatever will seem to have a reference to me has a reference. Moreover, according to § 304 which we cited above, the entire question of reference is thereby suddenly tumed into non-sense for a PL theorist. 8 Bearing this in mind, what could we say ab out Fodor's mentalese - the inner code ofthe brain - upon whieh Marconi's paper dwelt at length? Is it possible to involve mentalese in the same analogy we held between the PL and the language spoken by the BIV? Well, imagine a community of people which possesses a language of thought but no spoken language, perhaps because they are bereft of a phonatory apparatus. Imagine, in advance, that they have a special telepathie faculty whieh allows them to talk to each other. There is nothing strange in this for, according to Fodor, mentalese is what makes possible the leaming of a historical language and must hence be in itself complete and fully structured. But even if these Mentalesians - as we may call them - manage to attach a definite reference to every term of their language of thought, this would not happen by virtue of mentalese itse1f. Here the situation is similar to that of the people playing the language-game of the beetle in a box (§ 293). It doesn't really matter what is in the box; what matters, actually, is that all the speakers use the word 'beetle' in the same way. "The thing in the box - as Wittgenstein says - has no place in the language-game at all." Equivalently, the possible reference which Mentalesians' terms could acquire would not depend on what is in their head, but on the agreement on which their practice converges. Therefore, contrary to Marconi himself - who displays an accurate examination of some of Fodor's contentions - 1 think that the argument Wittgenstein develops against the PL hits the language of thought as weIl. It has to be notieed that Fodor tries to safeguard himselffrom the start, for he claims that he is offering a particular reconstruction of Wittgenstein's PLA, and admits that other reconstructions could well prove Fodor's own analysis wrong. 9 So, it is Fodor himself - in a sense - that wams us against his own argument. Moreover, he claims that Wittgenstein's purpose is that of attacking the idea of a sense-datum language. 1O But the sense-datum language - as we already remarked - is taken by Wittgenstein only as a language which MENTALESIANS AND WITTGENSTEIN'S PRIVATE LANGUAGE 121 - were mentalism right - would be an example of PL, where the purpose is that of showing the absence of inter-subjective criteria. Fodor gives a two-fold reply to Wittgenstein: first, he notices that "there is no reason why a mentalist needs to assurne that mental operations exhibit epistemic privacy"; and, second, he points out that the PLA isn't really any good because it follows "on the verificationist princip1e that an assertion can't be sensible unless there is some way of telling whether it is true, but surely there is nothing to be said for that principle"Y The second point is rightly criticized by Marconi, but the first seems to miss the issue: the epistemic privacy is not something that a mentalist can assurne or not, but is rather something to which a mentalist is compelled. And here resides the core of Wittgenstein's argument, the point in which his real polemical target reveals itself as mentalism. If this is right, the whole project of finding a characterization of a "Language of Thought" leaves much to be desired just from the beginning - at least from Wittgenstein' s point of view (and - for that matter - from Putnam' s, since his criticism of mentalist theories of reference - which he labels "magical theories of reference" - is one of the core premises of the argument against the BIV hypothesis). Indeed, what Fodor tries to do, as revealed in his defence of Augustine,12 is to revive the mentalist tradition, a tradition whose roots go back to Aristotle with many formulations in the history of philosophy. Putting these formulations aside, it could be said that the leading idea pervading mentalism is that of the existence of ready-made representations in the mind which give meaning to spoken words and refer to things in the world. The relation between representations and words is conventional, for it depends on the particular historical language which tries to convey those representations, whereas the relation between representations and things in the world is intrinsic, built in the very representations themselves. Moreover, the set of mental representations is structured to form a sort of language: a 'natural' language which can be seen as private, because in itself - without the intervention of a spoken language - cannot serve for the purpose of communication. The utility of a spoken language, in this respect, resides just in its being a conventional means by which the speaker translates his mental code in order to exchange information with a listener who, in turn, translates the spoken language into his mental one. I mentioned before that Fodor's project is that of giving the mentalist tradition a modem style. 'Mentalese' is just the traditional inner code someway connected with the world and prior to the acquisition of any spoken language. The difference is that mentalese is explicitly located in the brain, and makes 122 MASSIMO DELL'UTRI the 1earning of a spoken language possible, whereas this latter feature is absent from the medieval lingua mentis. Marconi plausibly shows that one cannot misuse mentalese, or, as he puts it, being right in a non-publie language turns out to be quite different from being right in a regular, publie language: for being right in a non-publie language admits of no alternative ... One eannot go wrong in the LoT (p. 114). But Marconi uses this fact in order to conclude that mentalese is not a language in Wittgenstein's sense and, therefore, that the PLA does not touch it. What Wittgenstein would have said about mentalese, on the contrary, is not that it is not a language, but that it is not a public language. Impossibility of "misusing", impossibility of "using a term according to a (public) rule", impossibility of "using a term (in-)correct1y", are all one and the same thing from Wittgenstein's perspective. One could also grant, for the sake of discussion, that mentalese is not a language, but neither is PL. The PLA applies to mentalese because it hits its epistemological and semantical presuppositions, i.e. mentalism. Therefore, the very reason why - according to Marconi - the PLA does not apply to mentale se (i.e. the impossibility of misusing it), is precisely the reason showing that the PLA applies to it. Far from constituting an autonomous and complete representational system which precedes and makes possible the learning of any historical language, far from possessing a meaning and a reference per se, mentalese - if there really is such a thing - inherits meaning and reference from spoken language, and therefore all the problems connected with them. Mentalese does not solve any problems: it only duplicates them. Terza Universita di Roma NOTES See H. Putnam, Reasan, Trnth and History, Cambridge University Press, Cambridge 1981, eh. I. "Our words refer to things by the way they enter in diseourse; by their eonnexions with what people are saying and doing, for instance, and by the way they affeet what is said and done". (R. Rhees, 'Can There Be a Private Language?', in Wittgenstein. The Philosophical Investigations, ed. by G. Pitcher, Maemillan, London 1968, p. 267). 3 See H. Putnam, 'The Meaning of "Meaning', in Mind, Language and Reality, Philosophical Papers, Vol. 2, Cambridge University Press, Cambridge 1975, eh. 12, and Representation and Reality, The MIT Press, Cambridge 1988, eh. 2. 4 See R. Descartes, Meditations on First Philosophy, Liberal Arts Press, New York 1960. 5 See my 'Choosing Coneeptions ofRealism: The Case ofthe Brains in a Vat', Mind, 99,1990, pp. 79-90. I 2 MENTALESIANS AND WITTGENSTEIWS PRIVATE LANGUAGE 123 As Rhees (Op. cit., p. 281) puts it: "Speaking the language ... is not just uttering the words; any more than understanding the language is just 'recognizing' the words. It is carrying on a conversation, for instance; or it may be writing reports, or listening to a play in a theatre. It is being someone to whom the rest of us can speak and get an answer; to whom we can tell something and with whom we can make ajoke and whom we can deceive". 7 H. Putnam, Reason, Truth and History, eit., p. 3. R Tbe reformulation leaves it open whether terms of a PL do have a reference or not. Yet, sinceaccording to Wittgenstein - one could be said to use a term correctly, Le. understand it, only if one uses it following certain social standards (which state what is to count as the correct use ofwords and sentences), and since understanding a term involves grasping what its soeially determined meaning and reference are, it follows that those standards state also what is to count as the reference ofterms. Hence, failing to accord to the standards amounts to failing to have a (socia!) reference. Be this as it may, § 304 cou!d be interpreted as to show that having a reference about which nothing could be said - as in the case of a PL - would serve just as weIl as a lack of reference. 1fthis is true, by the way, § 304 takes on a verificationist flavour. 9 See J. A. Fodor, The Language ojThought, Crowell, New York 1975, p. 68. 10 Ibid., p. 69. 11 Ibid., p. 70. 12 Ibid., p. 64. 6 PART 111 INTENTIONALITY AND THE GRAMMAR OF MENTAL LIFE ALDO GIORGIO GARGANI INTENTIONALITY IN WITTGENSTEIN'S WORKS The problem ofthe logical form and ofits relationship with that ofthe understanding of propositions was worked out by Russell in his Theory 0/ Knowledge, where he developed the thesis that the knowledge of or acquaintance with a pure logical form, that is with an abstract complex conceived of as a single object would be the requisite for the understanding of a sentence. In his work Russell said: I think it may be shown that acquaintance with logicaJ form is involved before explicit thought about logic begins, in fact as soon as we can understand a sentence. 1 In the Tractatus and in the preliminary works Wittgenstein rejected Russell's thesis and maintained that logic cannot depend on the experiencing of a fact, whether it be logical or empiricaI.2 Criticizing Russell' s theory asserting the existence of an abstract formal structure the logical experience of which would precede the understanding of any proposition, Wittgenstein assigned the conditions of the meaning to the elements inherent in the very constituents ofthe propositions. 3 With regard to this, Wittgenstein remarked that ''the precedent to which we are constantly inclined to appeal must reside in the symbol itself'.4 Formerly in a entry of bis Tagebücher in 1914 Wittgenstein wrote that a proposition achieves its meaning "offits own bat" (au/eigene Faust).s This immanent conception of the proposition, according to which it reaches its sense on the basis of its own resources, constitutes a thematic premise of Wittgenstein' s later work, even where the main theses of the Tractatus have been dropped. But there was another premise connected with this conception which was bound to give rise to one ofthe main difficulties ofthe Tractatus, and it is the claim that every proposition is an independent picture of astate of affairs and it is not possible to infer one elementary proposition from another one. Using the metaphor of the proposition as a yardstick for reality, Wittgenstein wrote that only the end-points ofthe graduating lines ofthe scale are laid against the object to be measured. 6 This claim was to be dropped as soon as Wittgenstein realised that it prevented the proposition from being an univocal representation of reality. Logical products like "p . q" - where p and q are sentences determining the same co-ordinate of reality in two different ways - had to be excluded. 127 R. Egidi (ed.), Wittgenstein: Mind and Language, 127-136. © 1995 Kluwer Academic Publishers. 128 ALDO GIORGIO GARGANI Consequently in the course of the Twenties Wittgenstein became aware that logical constants alone cannot account for all inferences we make in language and that they have to be placed within a language now conceived of as a system, as a more comprehensive syntax of internal relations which direct our use of language, permitting some applications of it and excluding others. 7 Using the earlier metaphor again, Wittgenstein writes now that language is no longer a yardstick of which only the end-points of the graduating lines are applied to reality, but one where "all the graduating lines" are applied,8 with a view to warranting the univocality ofpropositions. Wittgenstein's well-known slogan, that the meaning of a word is its use, is strictly connected with his conception of language as a system of internal relations and rules directing the proposition taken as something with which we should operate. 9 Precisely, the meaning of an expression is achieved by knowing the procedure, the method to reach it. lO In this new conception a fundamental role is played by intentional acts (meaning, intending, expecting, wishing and the like) so that Wittgenstein goes so far as to write that "if you exclude the element of intention from language, its whole function then collapses."11 In his middle period, between the late Twenties and the early Thirties, Wittgenstein determined in the intention the rule governing the use of language, since it would be the intention that connects use with a system of symbols (for instance playing and score).12 Tackling the problem ofthe relation of intentional acts and their fulfilments, Wittgenstein was aware that it cannot be shaped on the terms of a rough comparison between the proposition expressing an intention, a wish, an expectation and a mere fact considered from the outside, as Russell did in The Analysis of Mind. 13 Russell analyzed intentional acts in terms of external relations involving three factors, the thought or proposition (expressing a wish, an expectation), the fact that verifies it and a third event which would be a sort of satisfaction, "the fulfilment of the expectation." Criticizing this doctrine, Wittgenstein points out that if the fulfilment ofthe expectation were to consist in an external relation, that is in a third event not describable in advance, then any event or fact might be considered as the fulfilment of the expectation. In his Philosophische Bemerkungen Wittgenstein rejects Russell's doctrine because it does not provide any place for the inner comparison holding between an intention and its object. 14 Russelllimits hirnself to setting up an external and rough comparison, relating a proposition to an event which is not describable in advance. For Wittgenstein, this comparison should be established on the basis of an internal relation of our language according to which expecting that p will be the case should be the same as expecting that this expectation will be fulfilled. ls INTENTIONALITY IN WITTGENSTEIN'S WORKS 129 For Wittgenstein, the fulfilment instead of consisting in a third unpredictable event, must be represented in advance as a costituent inherent in the very proposition expressing the expectation. 16 The expectation of the perception of a sense-datum, for instance, involves a sort of being prepared for the occurence of it. "If I expect to see red" - writes Wittgenstein - "then I prepare myself for red" .17 But, according to Wittgenstein, the connection between expectation and fulfilment doesn't fit a three-term relation (as Russell c1aimed), but a two-term relation. In the period corresponding to his verificationist phase, with special reference to the Philosophische Bemerkungen l8 and to his conception of language as a calculus and system of internal relations, Wittgenstein rules out the notion of a judging or perceiving subject who would confront his expectation with what happens,19 experiencing that third event, which Russell referred to, which would consist in a sort of satisfaction. For Wittgenstein, the relation of the expectation to its fulfilment no longer depends upon an act of interpretation of a judging or perceiving subject. With regard to this it is worth remarking that Wittgenstein prefers the expression "the use of signs" to "the interpretation ofsigns."20 The relation between an intentional act and its fulfilment depends on the rules of a language taken as a caIcuIus which directs our employment of expressions in their operative application to reality. With a shift which is relevant from the point of view of a new philosophical anthropology, Wittgenstein makes the connection between expectation and its fulfilment no longer depend on the so called act of interpreting or comparing peculiar to the traditional judging or perceiving subject or self, who would negotiate that connection, but directly on the system of the grammatical relations which regulate our operating with propositions; therefore the expectation must be situated in the same space as what is intended, expected and the like. 21 The fulfilment of an intentional act does not consist in a third event but directly in the immediate production of the event which taking place provides an answer or fulfilment. Expecting is connected with looking for: looking for something presupposes that I know what I am looking for, whithout what I am looking for having to exist ... How do I know that I have found that which I was looking for? ... I cannot confront the previous expectation with what happens. The event that replaces the expectation is a reply to it. 22 In a way which is analogous to the case of the relation of proposition and verification, in the connection between an intentional act and its fulfilment, there isn't any go-between, any Mittelglied, taken as the factor which would have the task of negotiating that fulfilment or verification respectively.23 In accordance with that, Wittgenstein in the Philosophische Bemerkungen puts 130 ALDO GIORGIO GARGANI forward a distinction between, on the one hand, the expectation of a proposition p considered as corresponding to the hollow shape of a body and, on the other hand, the occurence ofp assumed as corresponding to the solid shape of a body.24 This hollow shape translates metaphorically Wittgenstein's notion of expectation as the being grammatically prepared for something. 25 In these terms, the proposition of an expectation contains in advance the representation of what constitutes its fulfilment. But then Wittgenstein seems not to be completely satisfied with his solution. In the course of his lectures in Cambridge in the early Thirties, and then in the 'Brown Book', Wittgenstein questions his former thesis according to which the proposition expressing an expectation is applied to the point of reality which is present but at the same time it "points out beyond it ... in the direction of the expectation".26 Wittgenstein seems to be self-critical when he remarks during a lecture in Cambridge that "it seems to us as ifthe proposition was not simply itself, but pointed beyond itself and contained a kind of shadow of its fulfilment, which is neither the proposition nor the fulfilment, but something between them (e.g. my expectation that Mr. Smith will come into the room seems to foreshadow his so doing)".27 Thus Wittgenstein is precisely criticizing his recent view according to which the proposition of an expectation or order would outline its own fulfilment. "We look for this shadow - said Wittgensteinparticularly in an expectation".28 In this discussion a particular role is played by the concept of similarity. In fact while analysing the embarassing consequences involved in the use of the notions of "similar", "similarity", Wittgenstein has been working out a new theoretical arrangement of his doctrine on intentional acts. "Similarity" is a treacherous notion which drives us into a blind-alley. It is strictly interlocked with the concept of proposition as an expression foreshadowing its fulfi1ment. If, in accordance with the Philosophische Bemerkungen, the expectation was "the hollow shape" of a body and the occurrence of its fulfilment was "the solid shape" of it, then there seems to be a relation of similarity between the two shapes. But it is just this relation shaped on such terms which gives rise to some difficulties. If I expect to meet Mr. Smith, then in my expectation there is something which would be similar to Mr. Smith; retuming to his former example of the expectation of seeing red, which implied the being prepared for red, Wittgenstein remarks: IfI expect a red patch, my expectation cannot contain the red patch but only something similar to it ... But to expect something 'similar' is already to expect something different. The 'similar' element, the shadow, in the expectation, is different from the fulfi1ment: you are no closer to the fulfi1ment".29 INTENTIONALITY IN WITTGENSTEIN'S WORKS 131 Then, Wittgenstein underlines the paradoxical consequence that if the expectation were a shadow ofthe expected fact, then we couldn't expect for instance Mr. Smith to enter the room if he does noeo Furthermore, referring to the link between his former view on intentional acts and verification, Wittgenstein remarks that the symbol doesn't point to something outside itself, nor does it anticipate something else "in a shadowy way"; he finally remarks that a symbol doesn't imply a cognition of its being true or false, whereas it is supposed that the thought (belief, wish) foreshadows its verification; and this shadow we call the meaning ofthe proposition, and it seems to mediate between the proposition or symbol and the fact. But this assumption of an intermediate link does not help US".31 This shadow, which is conceived of as similar to its fulfilment, seems to correspond to what Wittgenstein in the Philosophische Bemerkungen had referred to as the hollow shape of the expectation contrasted with the solid shape constituted by the occurrence of wh at is expected. Wittgenstein' s analysis of the concept of similarity was bound to bring about important consequences on the development of his latest philosophy. In fact it is connected with Wittgenstein' s rejection of all those philosophical doctrines which are grounded on duplication, on the pattern of "a double", as Wittgenstein says.J2 The thesis underlying these duplications claims that an expression is in itself just a scribble and it achieves its meaning through its fitting a mould already prepared for it in the mind. Here Wittgenstein is attacking his former view on intentional acts, when he rejects the conception according to which "the words falls into a mould of my mind long prepared for it" ,33 and when he questions the model ofthe word's fitting a mould which involves a comparison ofthe hollow shape and the solid shape before they fitted together; these are in fact the very expressions used by hirn formerly. Wittgenstein rules out the resorting to any duplication: in this sense recognizing a scribble as a proposition or a drawing as a face does not involve two separate steps, that is the recognizing ofthem as scribble or drawing and the recognizing ofthem as meaningful sentence or a face respectively. We get hold ofthe sentence or face directly in the scribble or drawing. 34 That reminds us of Brentano's remark in the Psychologie vom empirischen Standpunktthat "to taste a piece ofwhite sugar does not mean to taste a piece of sugar as white" (ein weißes Stück Zucker schmecken, heißt nicht, ein Stück Zucker als weißes schmecken.)35 What Wittgenstein is here questioning is that sort of fictitious comparison which is peculiar to philosophical attitudes. But the philosophical employment of fictitious comparisons36 through duplications, prototype or double, is based on 132 ALDO GIORGIO GARGANI that notion of similarity which is mistakenly supposed to justify the use of language. Wittgenstein rejects both the false comparisons and the concept of similarity presupposed in the former by resorting to a strongly constructive conception of language. The conclusion to be drawn from his analysis is that it isn 't the similarity between experiences which is responsible for our use of words but, on the contrary, it is our use of words which establishes what is to be called similar or identical. "As similarity does not explain the shadow, it does not explain the relation to fact either".37 It is the circumstance that the same word is used to establish that an intentional act and its fulfilment have in common. The fact that I say that I expect Mr. Smith and then I say that Mr. Smith has arrived, justifies my saying that my expectation has been fulfilled. 38 "What expression and fulfilment have in common is shown by the use of the same expression to describe both what we expect and its fulfilment (the word 'red' occurs in both)."39 Differently from what was claimed in the Philosophische Bemerkungen, now the expectation is not a hollow shape foreshadowing in advance its fulfilment; and its being fulfilled by the occurence of a certain event is testified by the use of the same expression for both. In the "Brown Book" Wittgenstein writes that ''to say that we use the word 'blue' to mean 'what all these shades of colour have in common' by itself says nothing more than that we use the word 'blue' in all these cases."40 A constructive interpretation oflanguage practices is responsible for Wittgenstein's demolition of the traditional conception which made the use of such notions as those of similarity and identity dependent on mental or psychological representations. On the contrary, it is our constructive procedures in actual using language 'which define the identity or similarity of objects, or, for example, what following a mathematical rule consists inY In language, said Wittgenstein, "we don't look for anything. We construct something." A constructive procedure accounts for the bridging of sign and application. In fact, no explanation ever saves the jump between them. 42 Thus intentional acts are not directed at the shadowy being of external facts and events: The symbol is in same sense self-contained; you grasp it as a whole. It does not point out something outside itself, it does not anticipate something else in a shadowy way... Thought is therefore autonomous, complete in itself; and anything not given in my thought cannot be essential to it. Thought does not point outside itself. 43 The sense-datum, the immediately given which Wittgenstein referred to in the Philosophische Bemerkungen have now to be replaced by the linguistic expression of them: I am inclined to suggest to you to put the expression of your experience in place of the INTENTIONALITY IN WITTGENSTEIN'S WORKS 133 experience. 'But these two aren't the same'. This is certainly true, at least in the sense in which it is true to say that a railway train and a railway accident aren't the same thing. And yet there is ajustification for talking as though the expression "the gesture 'long,long ago'" and the expression ''the feeling 'long,long ago'" had the same meaning.44 In conclusion, in Wittgenstein's middle period, in accordance with his conception oflanguage as a ca1culus, the relation of expectation and fulfilment is identified with that of calculation and result. Wittgenstein said that "thought anticipates its fulfilment in the same sense that a calculation anticipates its fulfi1ment" .45 That accounts for Wittgenstein' s consequential claim that meaning is set up inside language and it is given by explanation. 46 This is true of Wittgenstein's point ofview to the extent that he assumes language practices to be directed and guided by rules along the rigid rails of grammar which would operate at a distance independently of our use.47 Later, when Wittgenstein realised that no rule by itself could determine any course of action, since any course of action might be considered as conforming to the rule,48 and worked out his doctrine of language-games, it is still exclusively within our grammar and ordinary practices that the relations of expectation and fulfilment, order and execution, wish and satisfaction, measurement and result are defined and ascertained. All these relations depend on weH defined criteria of what we ordinarily call "obeying a rule" and "going against a rule" in actual cases. 49 Consequently in the Philosophische Untersuchungen Wittgenstein ·comes to the conclusion that it is in language that expectation and fulfilment make contact (in der Sprache berühren sich Erwartung und Erfüllung.)SO Now, the inclination peculiar to the traditional philosophy, to resort to what Wittgenstein caHs "the double", that is duplications and processes parallel to the written signs or spoken sounds in order to explain and justify our use of words, has arisen from an ill-placed need for foundation which in fact drives us to an infinite regress. That inclination turns out to be based on the presupposition of conferring a transitive meaning on expressions actually having an intransitive use. In this sense, the philosophical metaphors concerning the hollow shape and the solid shape in the relation of expectation and fulfilment, or the image of the word's falling into our mind and fitting a mould into our mind long prepared for it,SI are bound to represent a class of expressions as transitive, which are in fact intransitive, by means of fictitious comparisons. While trying to set up these feigned comparisons, the philosopher does nothing but compare an expression with itself, bringing out a mere reflexive comparison, without contrasting it with another. The reason for the philosopher doing that depends on his reluctance to recognize that an expression is meaningful in itself and it does not contain its meaning outside itself. 134 ALDO GIORGIO GARGANI Wittgenstein conceives oflanguage now as a sort of grammatical space which can be divided but as such is not contiguous to anything else. 52 When we say that we are getting hold of the expression of this face or the meaning of a sentence, it seems as ifwe were getting hold ofsomething which is in the face or in the sentence respectively but at the same time different from them. While ruling out the primitive conception of any fictitious comparison,53 Wittgenstein restores the meaning and understanding of expressions in terms of an intransitive use. Instead of the model of a double or prototype of expressions as mould in our mind into which the pictures fall and fit it, Wittgenstein speaks of an intransitive use through which, as he writes, "we let the picture sink into our mind and make a mould there". 54 If one had asked what the colour pattern of the pansy said, the right ans wer would have seemed to be that it said itself. Hence we could have used an intransitive form of expression, say "Each ofthese colour patterns impresses one".55 And in letting the face impress itself on me and contemplating its 'particular impression', no two things of the multiplicity of a face are compared with each other; there is only one which is laden with emphasis. Absorbing its expression, I don't find a prototype ofthis expression in my mind; rather, I, as it were, cut a seal from the impression. 56 Thus, instead of being the requisite of the meaning of an impression, the alleged mould would be the consequence brought about by the impression. This originally intransitive form of our expressions turns out to be strict1y connected with Wittgenstein's frequent comparison ofthe proposition with a gesture or a musical phrase which are understood by themselves. 57 In conc1usion, in spite of all the changes it has undergone, Wittgenstein has remained coherent and faithful to his original, youthful conception which made hirn say that the proposition represents the state of affairs, as it were, off its own bat (so stellt der Satz den Sachverhalt gleichsam auf eigene Faust dar).58 But in the new version the circumstance that a proposition represents something offits own bat is connected with Wittgenstein's conception ofintention as a sort of way of using language, which in its turn doesn't find expression in language, since "the expression of the intention can't contain the intention".59 But it is connected as well with that constructive impulse which is asserted by Wittgenstein to bridge the gap holding between the sign and its application. An unavoidable gap which cannot be saved by any explanation, nor negotiated by any reason, since, as Wittgenstein said, "no reason compels us to learn language".6o Universita di Pisa INTENTIONALITY IN WITTGENSTEIN'S WORKS 135 NOTES B. RusselI, Theory of Knowledge, MS in the Russell Archives at MacMaster Univesity, p. 185. See R. W. Clark, The Life ofB. Russell, Weidenfels & Nicolson, London 1975, pp. 204ff; D. Pears, 'The Relation between Wittgenstein' s Picture Theory ofPropositions and Russell's Theories of Judgment', in Wittgenstein: Sources and Perspectives, ed. by G. Luckhardt, Harvester Press Hassocks, Sussex 1979, pp. 192-212. 2 L. Wittgenstein, Notebooks 1914-1916 (NB), ed. by G. H. von Wright & G. E. M. Anscombe. Trans. by G. E. M.Anscombe, B. Blackwell, Oxford 1961, p. 3; Tractatus Logico-Philosophicus (TLP), ed. by D. F. Pears & B. F. McGuinness, Routledge & Kegan Paul London 1961, prop. 5.4731, 5.552. 3 NB, p. 31. 4 TLP, prop. 5.525. 5 NB,p. 26. 6 TLP, prop. 2.1512, 2.15121, 4.211, 5.134, 5.135. 7 See Wittgenstein 's Lectures. Cambridge 1930-32 (WLl930-32). From the Notes of J. King & D. Lee, ed. by D. Lee, B. B1ackwell, Oxford 1980, p. 84. 8 Wittgenstein and the Vienna Circle (WWK), ed. by B. F. McGuinness. Trans. by B. F. McGuinness & 1. Schulte, B. Blackwell, Oxford 1979, pp. 63-64. 9 See WWK, pp. 167-168. 10 Ibid., pp. 86, 88, 167; L. Wittgenstein, Philosophische Bemerkungen. Philosophical Remarks (PR), ed. by R. Rhees, B. Blackwell, Oxford 1964. Trans. by R. Hargreaves & R. White, B. Blackwell, Oxford 1975, pp. 69-70. II PR,p. 63. 12 WLl930-32, p. 40. 13 B. RusselI, TheAnalysis ofMind, G. Allen & Unwin, London 1961, pp. 75-76 14 PR, pp. 64-65. 15 Ibid., pp. 63, 65, 66. 16 See Philosophische Grammatik. Philosophical Grammar (PG), ed. by Rhees, B. Blackwell, Oxford 1969. Trans. by A. J. P. Kenny, B. B1ackwell, Oxford 1974, p. 157. 17 PR, p. 71. 18 Ibid., pp. 77,200. 19 Ibid., p. 67. 20 Ibid., p. 70. 21 WWK, pp. 167-168. 22 PR,p. 67. 23 Ibid., p. 87. 1 Ibid.,p. 71. 1bid., pp. 70-71. 26 Ibid., p. 72. 24 25 27 WLl930-32, p. 30 Ibid. 29 Ibid. 30 Ibid., p. 44. 31 Ibid. 32 Ibid., pp. 29, 31, 33, 77, 82; L. Wittgenstein, The Blue and Brown Books (BBB), ed. and with a 28 136 ALDO GIORGIO GARGANI Preface by R. Rhees, B. Blackwell, Oxford 1958, pp. 163-164. BBB, p. 170. 34 Ibid., pp. 162·-171. 35 F. Brentano, Psychologie vom empirischen Standpunkt, F. Meiner Verlag, Hamburg 1973, Erster Band,p.131. 36 BBB, p. 163. 37 WLl930-32, p. 32. 38 Ibid., p. 33. 39 Ibid., p. 35. 40 BBB, p. 135. See L. Wittgenstein, Zettel, ed. by G. E. M. Anscombe & G. H. von Wright, B. Blackwell1967, p. 53. 41 Philosophische Untersuchungen. Philosophical Investigations (PI), ed. by G. E. M. Anscombe & R. Rhees. Trans. by G. E. M. Anscombe, B. Blackwell, Oxford 1953, I, §§ 208, II, p. 225; Bemerkungen über die Grundlagen der Mathematik. Remarks on the Foundations 01 Mathematics (RFM), ed. by G. H. von Wright, R. Rhees & G. E. M.Anscombe. Trans. by G. E. M.Anscombe, B. Blackwell, Oxford 1956, pp. 115,138-139,184. 42 WLl930-32, p. 106; see also pp. 4, 8,10; PG, pp. 184,244,256. 43 WLl930-32, pp. 43, 45-46; see also pp. 61-63, 87; PG, p. 88. 44 BBB, pp. 184-185. 45 WLl930-32, p. 62; see also WWK, pp. 178ff; PR, p. 178; PG, pp. 63,67,70,273,312. 46 WL1930-32, p. 62; PG, pp. 60, 66, 94-95. 47 PI I, § 219; RFM, pp. 34-35, 37-38, 167-168; G. Baker, 'Following Wittgenstein: Some Signposts for Philosophical Investigations §§ 143-242', in Wittgenstein: to Follow a Rule, ed. S. H. Holtzman & C. M. Leich, Routledge & Kegan Paul, London 1981, p. 64. 48 BBB, pp. 123ff; PI I, § 201, RFM, p. 184. 49 Wittgenstein Lectures. Cambridge 1932-35. From the Notes of A. Ambrose & M. Macdonald, ed. by A. Ambrose, B. Blackwell, Oxford 1979, p. 84; PG, pp. 184,244,272-274; PI I, §§ 199, 202; Bemerkungen über die Farben. Remarks on Colour, ed. by G. E. M. Anscombe. Trans. by L. L. McAlister & M. Schättle, B. Blackwell, Oxford 1977, p. 29; see G. Baker, op. cit., p. 63. 50 PI I, § 445; see also PG, p. 157. 51 PG, p. 170. 52 WLl930-32, p. 112; BBB, pp. 158ff, 178. 53 BBB, p. 165. 54 Ibid., p. 163; see also pp. 160, 169, 170, 174-178, 180. 55 Ibid., p. 178. 56 Ibid., p. 165. 57 Ibid., pp. 184-185, 177-178; L. Wittgenstein, Vermischte Bemerkungen. Culture and Value, ed. by G. H. von Wright in collaboration with H. Nyman. Trans. by P. Winch, B. Blackwell1980, p. 69; L. Wittgenstein, Lectures and Conversations on Aesthetics, Psychology and Religious Belief, ed. by C. Barrett, B. Blackwell, Oxford 1966, pp. 37-38. 58 NB, p. 26. 59 WLl930-32, p. 112. 60 Ibid., p. 67. 33 HERBERT HOCHBERG INTENTIONAL REFERENCE AS A LOGICAL RELATION: A VARIATION ON A THEME IN MOORE, RUSSELL, WITTGENSTEIN, AND BERGMANN In his 1910-11 lectures that became Some Main Problems of Philosophy, Moore made critical and explicit use of an intentional relation - referring which he took to hold between beliefs and facts: Obviously this expression 'referring to' stands for some relation which each true beliefhas to one fact and to one only; and which each false beliefhas to no fact at all; and the difficulty was to define this relation .... and we may know both that there is such a relation, and that this relation is essential to the definition oftruth. And what I want to point out is that we do in this sense know this relation .... Take any beliefyou like; it is, I think, quite plain that there is just one fact, and only one, ... which would have no being - would simply not be, if the belief were false. And as soon as we know what the belief is, we know just as well and as certainly what the fact is which in this sense corresponds with it. 1 The last sentence suggests that Moore held that the representational connection between "the belief' and the corresponding fact, which would "have no being" if the belief were false, is not a contingent relation that obtains or not as a matter of fact, but is a necessary connection. Moore's use of the term 'belief, rather than 'proposition', gives his discussion a misleading air, since Moore uses 'belief in two senses. In one sense, a belief is a particular mental state of a person, in the sense in which we speak of having a perception or a particular experience prior to another. Thus, it would make sense to speak of Othello having a belief at a particular time, just as it would make sense to speak of Othello perceiving something at a certain time. 2 In another sense, a belief is what a sentence may be said to express, as Frege speaks of a sentence expressing a thought or some speak of various sentences expressing the same proposition. While Moore explicitly denies that there are propositions, he implicitly thinks of beliefs, in this second sense, as types or characterizations ofbeliefs, in the first sense. In place of propositional entities, he thus implicitly appeals to universal attributes of particular mental states. Two people may be said to have the same belief at a certain time, if each ofthem has a particular mental state, where both mental states are instances of the same belief type, that is to say, the same universal attribute. 3 In his way, Moore replaced propositional entities with universal attributes, 137 R. Egidi (ed.), Wittgenstein: Mind and Language, 137-170. © 1995 Kluwer Academic Publishers. 138 HERBERT HOCHBERG which may suitably be called "propositional attributes". He also implicitly acknowledged non-existent facts. For his referring relation connects a propositional attribute (belief) to a fact, irrespective of whether the propositional attribute is true or false. 4 Moore's view, construing propositions as attributes, implicitly acknowledging non-existent facts, and explicitly adopting a basic intentional relation, is interesting in yet another way. He may be said to take the reference relation between a propositional attribute and its "intended" fact to be logical. We noted above that he takes the connection to be necessary when he holds that as soon as we know what the belief is we know what fact it refers to. But, he characterizes the formal nature of the connection a bit more explicitly when he holds: So that we might say: The fact to which a belief refers is always the fact which has the same name as that which we have to use in naming the belief.... We can, therefore, only name beHefs by means ofthese expressions beginning with 'that' - 'that Hons exist', 'that bears exist' and so on, or the equivalent verbal nouns. But, curiously enough, ifwe want to name the fact to which a belief refers - the fact which is, ifthe beliefbe true, and is not if it be false - we can only do it by means of exactly the same expressions. S Thus, when we have a true belief, and hence a fact, we take a sentence to represent both the belief and the fact, and, consequently, for any true sentence 'S', the beliefthat-S refers to the fact that-S. But, Moore's view is not confined to true beliefs, for as he, perhaps inadvertently, says in the initial quotation: "Take any beliefyou like; it is ... quite plain that there is just one fact, and only one, ... which would have no being - would simply not be, if the belief were false." Introducing signs formed by placing angle braces around sentences to represent propositional attributes, while taking such sentences to represent facts, Moore's analysis is then expressed by the pattern: (I) <....> refers to .... , where the result of replacing both occurrences of the dots in (I) by the same sentence is taken to be true in virtue of its form. If we limit the language schema to avoid multiple references to objects and properties and, for the time, limit the discussion to atomic sentences, we can say that any instance of the pattern (I) is a formal truth when and only when the same sentence replaces the two occurrences of the dots. We may then think of (I) as something like an axiom schema, a pattern for logicalor linguistic truths of the schema, and, hence, take the relation represented in (1) as a logical, linguistic, or formal relation. It may be called logical since the instances of (1) are taken to be necessary truths that are true as a matter of form. The sense in which (I), INTENTIONAL REFERENCE AS A LOGICAL RELATION 139 and refers to, may be said to be formal is clear, but the sense in which instances of (1) are necessary is not, for they involve a necessary connection between facts and propositional attributes. Since an atomic sentence occurs as a term of the relational predicate 'refers to', whether it is true or false, Moore is forced to recognize facts that have "no being" as terms of such a connection. Moore thus writes as if he specifies truth grounds by a pattern like: (1 *) Tr <aR'b> iff «aR'b> refers to aR'b & aR'b exists). (1) and (1 *) embody the four aspects of Moore's analysis; his recognition, first, of propositional attributes, second, of states of affairs which may obtain or not, third, of a basic reference relation holding between a propositional attribute and astate of affairs, and, finally, of the logicalor formal nature of such a referential connection. Russell sought to avoid reference not only to non-existent objects but to non-existent facts in his 1905 paper "On Denoting". In that paper he suggested using adefinite description of a fact and, in place of something like (1 *), employed a pattern along the lines of: (2) Tr <aR 'b> iff the complex aR'b exists. In (2), the sign <aR'b> can be taken to represent a proposition, a type of entity Russell still recognized in 1905. However, Russell's analysis is c1early problematic on other grounds, since he uses an atomic sentential form in the description in (2). That is, Russell does not avoid using a variant of an atomic sentence in specifying the truth condition for the atomic sentence in question, and, hence, he does not avoid questions about another problematic entity, a non-existent or possible but not actual fact (complex), as he does avoid a corresponding question about the problematic present King ofFrance. 6 In the immediately succeeding years, Russell rejected propositional entities and proposed his relational analysis of intentional contexts. Construing a context like 'Othello believes that a has F' in terms of a three-term belief relation holding of Othello, the property F, and the object a, and expressed by a sentence like 'B(O,F,a)" he spoke ofthere being a belief-complex and took Othello's beliefto be true ifthere was the further complex or fact, a's being F. He now took the correlation between the belief complex and the fact that was its truth ground to be established by the occurrence of all the constituents of the latter complex in the former. This manner of connecting complexes resembles Moore's in that the fact represented by 'Fa' is the truth ground for Othello's belief, where 'B(O,F,a)' represents Othello's having such a belief. The two sentences, 'Fa' and 'B(O,F,a)', display the connection 140 HERBERT HOCHBERG between such facts, just as a pattern like (1) displays the connection between a propositional attribute and a fact. We can make the connection even more transparent by "abstracting" from 'B(O,F,a)' to obtain the propositional function sign 'L(ß,F,a), with 'L' and 'ß' as variables ranging over attitudinal verbs and persons, respectively. In effect, in his 1913 manuscript Russen took such a function to replace the classical proposition expressed by 'Fa'. Given the obvious connection between the signs 'L(ß,F,a)' and 'Fa', we can say that Russell, like Moore, takes the connection between such a propositional function and the complex or fact that grounds the truth of the beliefto be logical. Since such a function was not a constituent of a belief fact, in versions of his relational analysis prior to 1913, Russell took hirnself to avoid acknowledging a propositional entity. It is easy to see why such a function is not a constituent of a belief fact. The function represented by 'L(ß,F,a), can take Band 0 for arguments, yielding the belief complex represented by 'B(O,F,a)'. On analysis, such a complex has the relation B, the property F, the objects a and 0 and the logical form L(ß,0,X) as constituents. The function L(ß,F,a) is not a constituent of the belief complex. Hence, Russell will speak of propositions being "incomplete symbols", as he will speak of class abstracts (and cl ass es) and definite descriptions (and the present King of France) in such terms. Russen not only sought to avoid propositional entities as wen as non-existent or merely possible facts. He also sought to avoid a basic relation, such as Moore's reference relation, that connects the complex that is said to be true (or faIse) to the complex or fact that is the ground of truth. To accomplish this he held the two complexes to be linked by their having common constituents, appropriately ordered. Such an appeal to order and to the constituents of facts raises several questions for Russell's analysis. In the 1913 manuscript Theory of Knowledge, Russell fundamentally altered his relational theory by returning to a crucial theme of 1905. Prior to 1913, in The Problems ofPhilosophy, for example, he had taken the order in a belief fact to determine the order in its ground of truth: The relation 'loving', as it occurs in the act ofbelieving, is one ofthe objects-it is a brick in the structure, not the cement. The cement is the relation 'believing'. When the beliefis true, there is another complex unity, in which the relation which was one ofthe objects ofthe beliefrelates the other objects. Thus, e.g., ifOthello believes truly that Desdemona loves Cassio, then there is a complex unity, 'Desdemona' s love for Cassio' , which is composed exclusively ofthe objects of the belief, in the same order as they had in the belief, with the relation which was one ofthe objects occurring now as the cement that binds together the other objects ofthe belief. On the other hand, when a belief is false, there is no such complex unity... 7 INTENTIONAL REFERENCE AS A LOGICAL RELATION 141 It is clear that Russell depends on the order ofthe constituents in the belief or judgment fact to provide the unity for the content of the belief, and hence to determine which fact is its truth maker. Recognizing that he was appealing to the order of the constituents in facts, and not just to the constituents, to correlate judgment facts with their truth grounds, he carne to believe that he must account for that order. This led to an important modification of his multiple relation theory in 1913. Convinced that he must give an account of relational order in facts, he took a non-symmetrical relation, RI, to determine (and be determined by) two relations, R* and R**, that held between particulars and complexes (facts), and which specified the position of a particular in the complex. The truth condition for 'aR1b' could then be expressed by: (3) Tr'aR1b' iff (1p)(aR*p & bR**p) exists, and Othello's beliefthat a stood in RI to b by: (3*) B(O, (1p)(aR*p & bR**p) exists). But, with (3 *) a familiar problem reappears. What does the belief relation hold between when the fact does not exist? While (3) offers a solution to one ofthe problems posed by false belief, the linking ofatomic sentences to what they are about, without acknowledging non-existent entities like possible facts, situations, etc., (3*) does not provide an analysis ofintentional contexts, like 'Othello believes that aR1b', that avoids such entities. In the Tractalus, Wittgenstein pronounced: At first sight it looks as if it were also possible for one proposition to oeeur in another in a different way. Partieularly with certain fonns of proposition in psyehology, sueh as 'A believes that pis the ease' and 'A has the thought p', ete. For ifthese are eonsidered superfieially, it looks as ifthe propositionp stood in some kind ofrelation to an objeetA. (And in modem theory ofknowledge (RusselI, Moore, ete.) these propositions have aetually been eonstrued in this way.) (5.541) It is c\ear, however, that 'Abelieves that p', 'Ahas the thought p', and 'Asays p' are ofthe fonn "'p" says p': and this does not involve aeorrelation ofafaet with an objeet, but rather the eorrelation offaets by means ofthe eorrelation oftheir objeets. (5.542) These remarks seem to be directed not so much at Russell's relational analysis or at any details of Moore's view, but at the taking of a person to be a constituent of a belief fact. Thus, Wittgenstein proceeds to remark: This shows too that there is no such thing as the soul- the subjeet, ete. - as it is coneeived in the superftcial psyehology ofthe present day. Indeed a composite soul would no longer be a soul. (5.5421) 142 HERBERT HOCHBERG In 5.542 Wittgenstein holds that there are occurrences, thought elements or "constituents", as he writes in his notebooks, that represent the objects a and band a relation between them that represents the relation RI.8 The occurrence of a thought-complex, an arrangement of thought elements in a relation, represents the situation that a stands in R 1 to b. (Whether a person is construed in some sort of Humean fashion, which the last sentence of 5.5421 suggests, is not relevant here.) Thus, Wittgenstein suggests something similar to Moore's pattern, as expressed in (1), in basing his discussion of intentional contexts upon his account of the representational role of atomic sentences and his correspondence account of truth, an account quite similiar to Moore's. Wittgenstein takes atomic sentences to represent situations, irrespective of whether such situations exist or obtain. Numerous passages in the Tractatus point to such a view, and it is such a view that Russell rejected in 1905 and 1913. There is another aspect to Wittgenstein's view. Like Moore and Russell, Wittgenstein takes the connection between a proposition and a situation to be logical. This is one theme involved in 5.542 and in his holding that the internal or essential properties of things determine their possibilities of combination. It is also involved in his notion of shared logical form, which involves more than the idea that what represents and what is represented must share the requisite multiplicity. In this connection it is relevant to note a comment of Wittgenstein's in his Philosophical Grammar that virtually reproduces Moore's way ofexpressing matters: Certainly, the proposition "p" determines that p must be the case in order to make it true; and that means: (the proposition p) = (the proposition that the fact p makes true).And the statement that the wish for it to be the case that p is satisfied by the event p, merely enunciates a rule fOT . 9 slgns ... Wittgenstein's comment about a "rule for signs" can be seen to take a pattern like (1) to be true by linguistic rule and not as a matter of fact. Russell was greatly influenced by 5.542, perhaps more so by this passage than by Wittgenstein's celebrated criticism of his relational theory that occurs in 5.5422, in the well-known letter to RusselI, and in the often quoted passage from the notebooks. lO For Russell continued to discuss the relational theory until he hirnself raised questions about it in 1918, and abandoned it in 1919 in "On Propositions" , (although there are indications in his papers that he already rejected it in 1918)." By contrast, Russell kept returning to Wittgenstein's remark in 5.542 from the time ofhis discussion of it in his introduction to the Tractalus through his lengthy treatment of it INTENTIONAL REFERENCE AS A LOGICAL RELATION 143 in An Inquiry into Meaning and Truth; and 5.542 was the basis for the view that Russell developed in an appendix to the 2nd edition of Principia. 12 In Principia, Russell took there to be amental occurrence or token, a*, that "meant" or represented the object a, another, F*, that represented the property F, and a complex mental occurrence that consisted of a* and F* in a relation or arrangement. This complex belonged to both the class or series that constituted a person, on his then current view of a self, and the class that could be said to be the belief that a is F. Russell's analysis was intended to be a variant of the view in the Tractatus, as that view was construed above. Like Wittgenstein, Russell believed that the complex occurrence (or sentence) is correlated to a fact in virtue of its constituents being correlated with constituents of the latter, which is the point of the phrase "the correlation of facts by means of the correlation oftheir objects." Yet, Wittgenstein, unlike Russell, takes a propositional complex to represent, irrespective of whether it is true. Thus, such a complex is correlated to a situation or possible state of affairs, though Wittgenstein construed such possibilities in terms of constituent entities having internal properties. In Principia, Russell avoided the problem of analyzing '''Fa' is true" and 'Othello believes that a is F' without introducing possible facts. Aside from the problems posed by explicitly taking atomic sentences to represent possibilities, it is not clear how the correlation of complexes is based on the correlation of constituents. Neither Russell nor Wittgenstein recognizes traditional propositional entities or Moore's propositional attributes. Yet, all belief complexes that have the same content - that are beliefs that a is F - have something in common. The problem is to say what that is without acknowledging propositional entities, whether in the form ofMoore' s attributes or Frege's thoughts. To avoid such entities Russell will eventually appeal to causalorigins and behavioral dispositions. Wittgenstein, in the Tractatus, appealed to internal properties of the sentential or thought elements, and the correlated essential properties of objects, to carry the burden. The possibilities of combination of the elements in the representing combination are taken to represent the possible states of affairs, which are determined by the natures or internal properties oftheir constituents. Wittgenstein undoubtedly thought that one avoided shadowy entities, such as possible facts, by taking the internal or essential properties of constituents to determine the possibilities of combination. Thus, the object a could be said to have as an essential or internal property, the property of possibly being F, rather than to be a constituent of a possible fact. But, wh ether or not such a move avoids possible facts and provides a viable way of construing possibilities, a question remains as to just how a particular propositional fact correlates with a possibility. 144 HERBERT HOCHBERG The correlation of the respective constituents will not suffice to correlate the complexes of which they are constituents, unless one takes it to be internal to being a token of an element-type, an a * for example, not only that it can combine with a token of another element type, F*, in a certain arrangement, but that the resulting arrangement represents a being F. Yet, to take matters in such a way is not to correlate complexes by correlating their constituents, it is to correlate complexes directly, though the correlation of complexes is dependent on the correlation of constituents. To be dependent on, however, is not to be implicit in or to be implied by. What Wittgenstein and Russell do is to interpret an arrangement of elements, token or type, an F* combined with an a*, as stating that a is F. And, in so doing, one substitutes such a complex token or type, interpreted as it is, for Moore's propositional attribute. An occurrence of a token of such a complex type, a specific occurrence of an arrangement ofan a* and an F*, represents a's being F precisely because an arrangement, whether the token or its type, is taken as a complex to represent a's being F. One can perhaps avoid Moore's propositional attributes by appealing to interpreted complex tokens or types, but one does not avoid linking the representing complex, as a whole, in addition to linking its parts. Hence, one does not avoid possible facts by declaring that the constituents provide the link between complexes. In aseries ofarticles beginning in the 1940s, Gustav Bergmann developed an analysis of intentionality similar to Moore's, while explicitly stressing the logical nature of an intentional relation and the need for propositional attributes. 13 Bergmann took mental states to exemplify such attributes, which are logically connected to possible facts. Expressions like '<Fa>' were construed as predicates satisfied by particulars, mental states, and his version of (1), (lb) '< ... > M .. .', was declared to be an analytic truth, with 'M' as a relational predicate representing the intentional relation means that. His claiming that such propositional attributes were simple properties, while being represented by complex signs, posed an obvious problem. Yet, Bergmann was forced to do this in order to support the claim that M-sentences, instances of (l b), were true or false in virtue of their form. Replacing such complex predicates by primitive predicate symbols, as he later did, rendered specious his claim that M-sentences were formal truths. Thus, like Moore, he was forced to accept a basic, necessary connection between propositional attributes and possible facts. But, more explicitly than Moore, Russell, or Wittgenstein, he still insisted on the connection being logical. 14 One motive for his taking it to be logical was derived from Wittgenstein: logical relations were merely pseudo-relations and not entities that need be acknowledged in INTENTIONAL REFERENCE AS A LOGICAL RELATION 145 an ontological assay of intentional contexts. The sign 'M' did not represent anything in Bergmann's world. The various views we have touched on suggest, but are not solutions to the problem of specifying truth conditions for atomic sentences and belief contexts involving them. A satisfactory solution will avoid a commitment to possible facts, in any form, while specifying what mental states with the same content have in common, without appealing to unanalyzable attributes of such states, to propositions, or to mysterious internal properties. To get at such a solution, let us assurne that atomic facts are the truth grounds for atomic sentences and that such facts consist, in the monadic case, of a property, a particular, and the logical form of monadic predication, a form Russell sometimes represented by the sign '0x'. A fact will be said to contain its constituents, and such a whole-part relation will be represented by the predicate 'C'. The truth condition for 'Fa' is then given by: (4) Tr<Fa> iff(3p)(pCa & pCF & pC0x). In (4), the angle braces are retained, but we can now simply read them as metalinguistic quotes. As the signs 'a' and 'F' are taken to represent the object a and the property F, respectively, this can be expressed by instances of (5) ''' .. .' represents ... ", where the dots may be replaced either by proper names or primitive predicates. When and only when tokens of the same type replace the two occurrences of the dots, we have a formally true pattern, a pattern true by the interpretation rules for the schema. Thus we have a formal expression ofthe referential role of names and basic predicates. Atomic sentences are another matter. Following Russell's early view, the specification of truth conditions, in patterns like (4), replaces such a role for atomic sentences. Thus, whereas basic predicates and names may be said to represent in a basic and fundamental way, atomic sentences represent only in the sense that their constituent terms refer to entities and in that truth conditions are specified for them by patterns like (4). No primitive representational relation, like Bergmann's M or Moore's referring, is involved in the sentential case. This eliminates Wittgenstein's situations, Moore's non-existent facts, Bergmann's potential facts, and Russell's possibilities, that may be positive or negative facts, of 1918-1919 (a view that is suggested in terms of"neutral facts" in an appendix to the 1913 manuscript).15 We now see in what clear and unproblematic sense representational connections are 10gically necessary: '''a' represents a" and "'F' represents F" are HERBERT HOCHBERG 146 instances of a pattern reflecting interpretation rules for the schema, and (4) is an instance of the definitional pattern for a truth predicate for atomic sentences. The use of(4) might suggest modifying Bergmann's analysis by using either - (i) <Fa> M (lp)(pCa & pCF & pC0x) - or - (ii) <Fa> M (1p)(pCa & pCF & pC0x) exists. But, (i) will be false if 'Fa' is false, while (ii), employing a sentence as a term for the predicate 'M', raises a familiar question about its truth condition: What does M relate when 'Fa' is false? Thus, such a modification would face the same problem we noted in connection with (3) and (3*) on Russell's 1913 analysis. However, while neither (i) nor (ii) will do as unproblematic replacements for Bergmann's "Fa' M Fa', Russell's analysis of descriptions would allow (iii) <Fa> M (1p)(pCa & pCF & pC0x) iff (3p)(pCa & pCF & pC0x) to be unproblematically taken as a formal truth, with the occurrence of the description having secondary scope. This recaptures an idea of Bergmann's regarding the logical nature of intentional reference, but not as he understood it, for in (iii) 'M' can be said to express a representational relation between an atomic sentence and an atomic fact only if there is such a fact. Bergmann, when he later acknowledged logical relations, was forced to accept possible facts as terms of his M relation-entities (iii) avoids. Our first problem is resolved. In keeping with our earlier reading of the Tractatus and Russell' s Principia adaptation of it, we can proceed to the second. Consider amental state ofbeliefthat is astate ofOthello's. We need not worry about the details of its being astate of belief, rather than of understanding or doubt, or of it being astate of Othello rather than of Iago. Our concern is with the content or intention of the state. We must state what it is for the intention of a thought being that a is F. In one sense of content or intention, we can do so quite simply. To say that amental state or thought has a content expressed by 'Fa' is to say that it consists of constituent elements, representing Fand a, in a relation such that the resulting complex or fact is true iffthe fact containing a, F, and 0x exists. As Bergmann explicitly, and Moore implicitly, took propositional attributes to be true or false, and Russell and Wittgenstein, in their ways, took propositional facts to be true or false, we here speak of complexes or facts having truth values. More important, however, is that instead of recognizing possibilities, whether implicitly, as internal or essential properties of objects, or explicitly, by taking the occurrence of a propositional complex to refer to a possible fact, the present analysis takes such a complex to represent something only in the sense that the complex is taken to be true under certain conditions. While we thus recognize the need to INTENTIONAL REFERENCE AS A LOGICAL RELATION 147 interpret a complex, in addition to interpreting its constituents, we avoid propositions and mysterious propositional attributes, as well as situations or possible facts. To have an interpreted atomic complex, a thought content, is to have its condition of truth specified in addition to having its constituents interpreted. However, to handle the familiar puzzles posed by quantifying and substituting into intentional contexts, we shall have to acknowledge further senses of "has the content expressed by 'Fa'''.16 In his Philosophical Grammar, Wittgenstein pointed to a crucial feature of the "interpretation" of thought contents: If a thought is observed there can be no further question of an understanding; for if the thought is seen it must be recognized as a thought with a certain content: it doesn't need to be interpreted! - That really is how it is; when we are thinking, there isn't any interpretation going on. 17 Wittgenstein's point is correct in that there is not an interpretation in one sense, namely, an accompanying thought doing the interpreting. But this is not to say that the complex content is not interpreted. In short, the occurrence ofthe complex as an act ofthought is an occurrence as an interpreted complex. In acts of thought one can refer to something and take a complex content to have a truth condition without also thinking that one is doing so. Relational contents are somewhat more complex, for we need to account for the order in relational facts that involve non-symmetrical relations. There are alternative ways of doing this. One way is to introduce two relations, I and S, that, like Russell's R* and R**, obtain between elements ofa fact or complex and the complex. An element stands in I to a relational complex when it is the first term ofthe relation, in S when it is the second. land S are, in asense, generalizations of Russell' s R * and R **, since they indicate the place of the term in the fact without specifying the constituent relation. To specify the truth condition for a dyadic relational sentence, where R l is a non-symmetrical relation, we can have: (6) Tr <aR lb> iff (3p)(pCa & pCb & pCRI & pCR(x,y) & alp& bSp), with 'R(x,y)' representing the form of dyadic relational exemplification. A corresponding complexity enters into the case of beliefs with relational contents. (6) is based on themes from Russell's 1913 manuscript, the latter parts ofwhich he unfortunately abandoned, supposedly in the face ofWittgenstein's criticism. Aside from Wittgenstein's objection to Russell's and Moore's earlier views regarding a self, his often cited criticism, stated in the Tractatus, the letter to Russell, and the Notebooks, focuses on the early form ofthe relational 148 HERBERT HOCHBERG theory that Russell preserved for some contexts in the 1913 manuscript. The point, simply put, is that patterns like 'B(Othello,0x,F,a), presuppose that the subject terms 'a' and 'F' are understood to be combinable into a meaningful sentence, and the mere addition of a sign for a form does not, as a further subject term, serve to combine 'a' and 'F'. This being so, Russell's analysis is forced to implicitly take 'Fa' to be a meaningful sentential expression. Hence, Russell either implicitly employs a propositional sign or cannot viably rule out nonsensical beliefs. He is thus forced to abandon his relational theory. Such a line of criticism is not relevant to the way the non-symmetrical relational contexts are treated on the modified theory, employing descriptions with variables ranging over complexes. But, as we noted in the discussion of (3) and (3*) above, there is another line of criticism that can be directed against the modified relational account of 1913. To consider the relevance of such problems to Wittgenstein's attacks on Russell, and Pears' recent reiteration ofthose attacks, we must look more closely at Russell's 1913 view. In the 1913 manuscript Russell took propositions to be "forms": ... the total complex which occurs when the subject has the relation U to the objects in question may be symbolized by U(S, x, R, y, y) If we now proceed to the "form" which results from varying U and S, Le. to "there is a U and an S such that U(S, x, R, y, y)" we arrive at something which is the same for all subjects and for all propositional relations which we should regard as concerned with the same proposition. Thus there is no formal obstacle to defining this as the proposition. 18 The form, which he "defines" as the proposition, is represented by an existentially quantified sentence (the quantifiers being expressed in English) with (bound) variables replacing the attitudinal verb and the term for the person having the attitude. Taking forms, and hence propositions, to be existentially general facts posed a problem Russell recognized: The above definition has some merits and some demerits. Its chief merit is that it provides propositions, both true and false, as fast as we can think ofthem ... Its chief demerit is that we cannot be sure that there are propositions in all cases in which logic would seem to need them. It is not necessary to our definition that there should actually be a subject which has one of the familiar mental relations to the objects, but it is necessary that there should be some term and some relation by which a complex results having the requisite form and containing the objects in question. It may be possible to prove that there always are such complexes, but I do not at present see how such a proof could be found. In its absence, we cannot know ofthe existence of propositions other than those that have been actually thought of.... we must admit, I think, that the objection in question is serious. I cannot, however, INTENTIONAL REFERENCE AS A LOGICAL RELATION 149 think of anything better calculated to fulfil the purposes for which we want propositions, and I therefore propose to continue to use the word in the above sense. 19 It is strange that Russell did not explicitly consider propositional functions here, since they would not only fulfill his purposes but, as we noted above, they would allow hirn to avoid taking propositions to be basic entities. This is perhaps due to Russell's being ambivalent about propositional entities in the 1913 manuscript. Early in the rnanuscript, assuming that the first six chapters belong to the manuscript, he appears to reject propositions as he did in The Problems of Philosophy and earlier papers. And, at the beginning of Chapter I of Part 11 of the manuscript, he continues in the same vein: And this in turn, since it is repugnant to admit the reality of false propositions, forces us to seek a theory which shall regard true and false propositions as alike unreal, i.e. as incomplete symbols. 2o This rejection was a dear motive for his relational analysis of judgment in writings prior to 1913. But, later in the manuscript, he is concerned to preserve propositional entities as "subsistents" and identifies them with existentially general forms or facts. Yet, there is something odd about his taking existential generalizations, containing constant terms, to signify forms, though he does not speak of logical forms in this connection. But, though he does not take such forms to be logical forms, his taking them to be existential facts or truths obviously sterns from his taking logical forms to be existentially general facts - something has some property, something stands in some relation to something, and so on. Just as Russell had earlier attempted to avoid Fregean propositional entities and Meinongian contents, he also seeks to avoid a basic intentional relation connecting a belief fact, Othello's beliefthat ais F, with the fact that grounds the truth of the belief, a's being F. The link between the two facts is now furnished by their having Fand a as common constituents and by the logical form, 0y, ofthe one fact, a's being F, being a constituent of the other, the belief fact. In focusing on order in relational facts, Russell took the difference between monadic facts and relational facts, in virtue of which order was involved in the second, to be based on their different logical forms and different types of constituents. 21 Monadic facts were ofthe form 0y, while dyadic facts were of the form R(y, z). Facts ofthe first form did not involve order because they did not contain constituents ofthe same logical kind. Order could be involved in a fact of the second form since such a fact contained two constituents of the same logical kind. If, in addition, it contained a non-symmetrical relation, the terms of the relation must be ordered. This led to his modifying the 150 HERBERT HOCHBERG relational theory by taking intentional (belief, judgment, etc.) facts to involve logical forms as constituents, and thus to be more perspicuously represented by sentences like 'B(O,0y,F,a)' and 'B(O,R(y,z), RI, a, b)'. The form 0y, as a constituent of an intentional fact with a monadic content, was then taken to unify the constituent object and property, a and F, without appealing to a propositional entity. But such a modification of the relational theory of judgment did not provide for an analysis of the order involved in the case of relational contents. This led to Russell's second modification ofhis theory in 1913, the introduction ofrelations like R'" and R""" and descriptions like '(1p) (aR"'p & bR"""p)'. However, this second, major modification of his theory is not involved in Wittgenstein's critique ofit. Wittgenstein's attack was directed at only the first change: the use of forms like 0y, R(y,z), etc. to purportedly provide the unity of content for a judgment fact. Wittgenstein's criticism can be understood as two-fold. First, by claiming that a viable theory must prevent the judgment ofnonsense, he is pointing out that in a pattern like 'B(O,0y,0,y), the signs that replace the variables '0' and 'y' must be such that, taken in the appropriate order, they form a meaningful sentence. Thus, 'F' and 'a' are permissable replacements since 'Fa' is a sentence. But, then, Russell's analysis is purportedly misleading, for it seems that it merely employs a problematic rewriting of 'B(O,Fa)', which is the pattern ofthe propositional analysis. Second, to transcribe 'B(O,0y,F, a)' as 'Othello believes that a is F' presupposes that we have interpreted the sentential pattern 'Fa', andnotjusttheterms 'F' and 'a', alongwith '0y'. Thus, assume we have interpreted 'a' and 'F' as standing for an object and a property respectively. We must still appeal to a further rule that interprets their juxtaposition as a sentence. Thus we interpret 'Fa' as ascribing the property that 'F' represents to the object that 'a' represents. This is simply a variant ofthe point that a sentence is not a class of its constituent terms. By adding the form ofmonadic predication, 0y, to the members ofthe class {F, a}, to form the class {F,a,0y}, we do not form a complex consisting ofF and a standing in the form 0y, just as adding the sign '0y' to the class of signs 'F' and 'a' does not form a sentence. To have an atomic sentence we must take the signs to form a complex linguistic pattern that requires an interpretation rule in addition to the rules for the constituents ofthe complex. Russell presupposes that the co ordination of the constituents serves to coordinate the complexes, but the coordination of complexes is not accomplished by the coordination of their constituents, though the former may be dependent on the latter. We can consider the matter simply in terms of 'Fa' and 'RI(a,b)'. Assume we have interpreted the names and predicates into objects, a monadic prop- INTENTIONAL REFERENCE AS A LOGICAL RELATION 151 erty, and a non-symmetrical relation. We then have, reflecting such interpretation ruIes: 'a' refers to a, 'F' refers to F, 'RI' refers to R I , and so on. Such statements can be taken to be linguistic or semantical rules of our schema. We may then add: A subject predicate sentence states that the object referred to by the subject term has the property referred to by the predicate term; a two-term relational sentence states that .... , etc. These embody the interpretation ofthe atomic sentences ofthe schema, as 'refers to' is used to interpret the subject and predicate signs of the schema. We then face two questions. Is stating that a relation atomic sentences stand in, as referring is a relation that names and predicates stand in? And, if so, what are the non-linguistic terms ofthe relation? The obvious candidate for the other term of such a relation is a fact, as a truth ground or truth maker for atomic sentences. But, since there can be no such ground when the sentence is false, to what is a false atomic sentence related - a possibility, a situation, a non-existent fact? The question, and its various answers, are familiar. Such a question about false atomic sentences can also be raised about purported propositional entities or mental contents that such sentences supposedly express. In early versions ofhis relational analysis of intentional contexts, Russen sought to avoid not only propositional entities and content properties but a basic intentional or reference relation for sentences and thoughts, along with non-existent or possible facts as purported terms of such a relation. While the 1913 analysis, at places, ambivalently acknowledges propositions as "subsistent" forms or existential facts, he still seeks to avoid a basic intentional relation, between beliefs (or propositions) and their truth grounds, as wen as the possible facts that such a relation seems to involve. But his analysis faces Wittgenstein's two-fold line of criticism. Consider the point that objects of a certain kind must be taken to be the terms of intentional relations like B. In one sense there is no difference between contexts like 'B(x,j, 0, y)' (with 'f as a variable for logical forms) and contexts like 'Fy'. In the latter case, as in the first case, only a certain kind of sign can replace a certain kind ofvariable. In the first case, however, there is an additional feature. The ascription of the predicate in the "0-place" to the subject in the "y-place" in accordance with the form in the 'rplace" is built into the reading of 'B'. Russen seems to have thought that this was unproblematic because of the logical differences between the form 0y, the property F, and the object a; differences that permitted only one type of combination. But the point Wittgenstein makes is that this presupposes the recognition of a possibility of combination that is not explicitly recognized on Russell' s analysis. This raises a question about the significance of implic- 152 HERBERT HOCHBERG itly recognizing such a form of combination. Thus, in the Tractatus he holds that objects and signs have essential or internal or logical properties that determine their possibilities of combination. Yet, such a move is compatible with Russell's analysis, since such logical differences are precisely what determine the role an entity can play in the fact expressed by 'B(0,0y,F,a)'. Moreover, in addition to such specific belieffacts Russell has other entities that can represent the complex composed of a, F and 0y: the propositional function L(x,0y,F,a) and the form or existential fact (3L)(3x)L(X,0y,F,a), identified with the proposition. On the earlier multiple relation theory Russell sought to avoid propositional entities and took belief complexes to be truths or falsehoods, and a belief fact, like the fact consisting of Othello, F, and a in the relation B, to represent the fact of a's being F. The analysis of 1913 focused on forms or existential facts like (3L)(3x)L(X,0y,F,a), as propositions, and hence as the subjects of ascriptions of truth and falsity. Thus, like earlier versions of the theory, some facts are true, others false, on the basis ofthe existence or non-existence of other facts. As was mentioned earlier, it is not clear why Russell did not identify propositions with propositional functions. One can deny that such functions are basic entities in that they are not ultimate constituents of the judgment facts that are their values: they disappear on the analysis of such facts. But, while one might make such a claim for propositions construed as propositional functions, one cannot make such a claim for Russell's existentially general forms or facts. In the case of a logical form, like monadic exemplification, wh ether construed as (30)(3y)0y or simply as 0y, we clearly have an entity that does not disappear on the analysis of the facts of which it is the form, such as the fact that ais F: Ifwe take sorne particular dual cornplex xRy, this has three constituents, x, R, and y. Ifwe now consider "sornething has the relation R to y", we get a fact which no longer contains x, and has not substituted any other entity for x, since "sornething" is nothing. Thus our new fact contains only Rand y. For sirnilar reasons, "sornething has the relation R to sornething" contains no constituent except R; and "sornething has sorne relation to sornething" contains no constituent at all. It is therefore suitable 10 serve as the "form" of dual cornplexes. 22 This passage reveals that Russell does not take a logical form to be a constituent of the fact of which it is the form. If he did so take it he would face a version of Bradley's regress, for the consituents, including the form, would apparently require a further form to "unite" them. But, though the form of "dual complexes" is not a constituent ofthe fact that-aR'b, it does not thereby disappear. This suggests a motive for Russell's taking logical forms to be INTENTIONAL REFERENCE AS A LOGICAL RELATION 153 existentially general facts. He seems to have two choices, given the symbolism of Principia. Either such forms are logical functions, or logical relations, represented by signs like 'R(y,z)' and '0y', or perhaps by 'R(Y, z)' and '09", or they are general facts without constituents. Russell might have thought that there is no way to decide in such a case, and, hence, there is an advantage to construing forms as general facts. For, if he did not so take them and yet required, as he does, such facts to ground the knowledge of logical truths, then he would be forced to hold that a form, like 0y, was a constituent of a fact, like (30)(3y)0y. He would then face the problem of specifying the other constituent or constituents of such a general fact. But, as we just noted and will discuss below, he thinks of the existential quantifier, taken by itself, as representing "nothing" . Whether it is construed as an existential fact or simply as a function, a form like 0y does not disappear from judgment facts, like B(O,0y,F,a), of which it is a constituent. On the other hand, neither the form which is the proposition, (3:E)(3x):E(x,0y,F,a), nor the propositional function, :E(x,0y,F,a), seems to be a constituent of the belief fact. Thus, that fact does not consist of Othello standing in B to that proposition, and, while Russell recognizes propositions, they do not playa traditional role. Whether construed as existentially general facts or functions, his propositions play the role offunctions mapping a subject and a propositional "verb" (attitude) onto an intentional (belief, judgment, etc.) fact. This is not to say that Russell does not recognize such an entity. The point is simply that, on Russell's multiple relation analysis of intentionality, if one expressed the fact that Othello believes that a is F by (0) 'B(O,(3:E)(3x):E(x,0y,F,a»', where B appears to be a two-term relation holding between Othello and the proposition (form), (0) should be elliptical for 'B(O,0y,F,a)', where B is a four-term relation holding among Othello, 0y, F, and a. But, since Russell construes the proposition as the existential form (3I:)(3x)I:(x,0y,F,a), which is taken to be an existential fact, it is not c1ear' how such a fact can be held to disappear as a constituent of the intentional fact. This poses a critical problem. On the one hand, it would seem that, as in the case ofthe propositional ftmctionI:(x,0y,F,a), 'B(O,(3:E)(3x):E(x,0y,F,a»' should be elliptical for 'B(O,0y,F,a)'. That is, if we take the proposition to be the propositional function, we can consider a sentence pattern like 'I:(x,0y,F,a)(O,B), , - where '0' and 'B' are subject terms for the predicate expression 'I:(x,0y,F,a)' - instead of 'B(O,I:(x,0y,F,a»' to express the fact that Othello believes that a is F. And, as in Russell's early form of lambda abstraction in Principia, such a sentence would convert to 'B(O,0y,F,a)'. On the other hand, 'B(O,(3I:)(3x):E(x,0y,F,a»' expresses the fact that a relation, 154 HERBERT HOCHBERG B, holds between two entities, a person and a fact, and hence should not be merely elliptical for 'B(O,0y,F,a)'. If it is, there is little point to Russell's distinguishing propositions, taken as existentially general facts, from propositions taken as propositional functions. This poses one problem for Russell's analysis. There is another, related problem. Since only 0y, F, and aare the constituents of the existential fact, (3L)(3x)L(X,0y,F,a), one may ask what distinguishes that fact from the atomic fact that a is F? On Russell's view there are two things on which the distinction can be based. First, as the quotation just above reveals, 0y is a constituent of the existential fact, but not of the fact that a is F. It is the form of the latter fact, but not a constituent of it. Russell, to avoid problems with a Bradleytype regress, does not take the form of a fact to be a constituent of it. This leads to the second difference. The form of the one fact is then different from the form of the other. This points to a peculiarity of Russell' s use ofform that we noted earlier. He speaks of a form, such as (3L)(3x)L(X,0y,F,a), where we not only have constituents but specific particulars and attributes or relations as constituents ofsuch forms. This is not only a problematic use of'form', but there is a c1ear sense in which such an existential fact, qua form, itself has a logical form, (3j)(30)(3y)(3L)(3x)L(Xf,0,y). This in itself is an oddity of the view, but the problem is compounded in that such a logical form is also the logical form ofthe intentional fact expressed by 'B(O,0y,F,a)', assuming that we do not express that fact by 'B(O,(3L)(3x)L(X,0y,F,a»)'. That could suggest taking (3V)(3w)V(w,(3j)(30)(3y)(3L)(3x)L(Xf,0,y» as the appropriate form, but that would be absurd. There is another "entity" that could also be involved in Russell's analysis. Consider the c1ass of all intentional facts containing F, a, and 0y - those obtained as values of the two-place propositional function L(x,0y,F,a), i.e. the c1ass that is the extension of that function. Such a c1ass mayaiso be taken as areplacement for the c1assical proposition and, hence, as the subject of ascriptions of truth and falsity. If a is F, we can consider such a c1ass to be true, since every fact in the c1ass is true, and take it, like its generating function, to represent the fact that a is F. Since c1asses are not entities, for Russell at this period, identifying propositions with such c1asses would be another way of avoiding propositions as entities. Putting aside, for the time being, problems posed by Russell' s acknowledging propositional entities, and no matter how a proposition is construed - as a c1ass of facts, as a function like L(x,0x,F,a) or as an existential fact like (3L)(3x)L(X,0y,F,a) - we may note that in order to take either a proposition or a specific belief fact to have a truth value, we must acknowledge an under- INTENTIONAL REFERENCE AS A LOGICAL RELATION 155 stood connection between any of these and the fact that grounds its truth. Even ifthat connection is expressed simply by 'B(0,0x,F,a) is true iff ais F', we coordinate one complex to another in addition to the coordination oftheir respective constituents. One ofthe misleading features of Russell's analysis that hides this fact is that the representing complex contains constituents that are elements of the represented complex. In the case of a sentence (or a proposition or mental content) the representing complex does not share constituents with the represented complex. If one does not see that a and F, as they occur in the representing complex, B(0,0x,F,a), represent themselves, as they occur in the represented complex, the fact that a is F, one can easily overlook the need to connect the two complexes. Wittgenstein's criticism points to the need to connect the representing complex to the represented complex. He makes an additional critical point in that the relation B not only orders the terms, 0, F, etc., but in so ordering them it takes F to function as a property, and not just as a term of a relation. Russell will explicitly recognize this in his 1918 logieal atomism lectures when he will note that such facts problematically have "two verbs". However, Wittgenstein's criticism fails to recognize that Russell does have a complex representing entity. In fact, given and the function ｾＨｸＬＰｹｆ｡Ｉ＠ as well as the the form ＨＳｾＩｸＬＰｹｆ｡＠ latter's "value range", there are diverse things that can be taken to represent the fact that a is F, in addition to the individual judgment facts. Each ofthese can be taken to play the role of a representing complex, though they need not be construed as complexes in the sense in which a fact or a Fregean thought is a complex. That neither such a propositional function nor its value range need be taken to be abasie entity is a feature of a Russellian type analysis that Wittgenstein and Pears ignore. There is a further ironie feature of Wittgenstein' s argument. In the Tractatus and the Notebooks Wittgenstein appealed to the internal or logical properties of the constituents of atomic sentences and of thoughts to connect atomic sentences and thoughts to their correlated situations or possibilities. Supposedly such logieal properties were themselves corre1ated with (or identified with) internal-Iogical properties of constituents of possibilities. The possibilities of combination of the elements in the representing combinations were taken to represent the possible states of affairs determined by the natures or essential properties oftheir constituents. Wittgenstein seems to have thought that by having the possibilities of combination be determined by such logical properties ofthe combining elements, he avoided recognizing possible facts as entities, for one supposedly did not have to take one combination to represent another over and above the representative roles played by 156 HERBERT HOCHBERG the constituents of the complexes. In asense, the object a could be said to have the essential property of possibly being F, instead of being an element of a possible state of affairs, a being F. But, it is clear that the correlation of constituents does not suffice to correlate the complexes of which they are the constituents, unless one takes it to be a logical property of a constituent of a thought or sentence that it not only can combine with another constituent of a certain kind but that the resulting combination represents a possible state of affairs. It is then gratuitous to claim that one has correlated complexes by correlating their constituents, though the correlation of the complexes is de-:pendent upon the correlation of constituents. Wittgenstein's alternative to Russell's analysis of intentional contexts thus faces the very problem Pears takes hirn to have raised against Russell's analysis. The need to correlate complexes points to a two-fold problem for both Wittgenstein and RusselI. Since the interpretation rules for terms, such as (rl) '''a' refers to a" and (r2) "'F' refers to F", must be supplemented by a rule, like Moore's (1), for atomic sentences, (r3) "'Fa' represents the fact that a is F", they cannot avoid either the connection expressed by 'represents' in (r3) or the problem. about what is represented when 'Fa' is false. Wittgenstein believed that (rl) and (r2), along with our understanding of what sorts of things a and F are and what sorts of signs 'a' and 'F' are, given us the interpretation of 'Fa' without (r3). But this just packs (r3) into the construal of 'a' and 'F' as subject and predicate terms. For, to take them as such terms is to understand that they can combine into the sentence 'Fa', which is taken to state that a is F. Nothing is gained by his talk of internal or essential properties of signs and objects or of "shared logieal form" or of the "pieturing" of facts by atomic sentences. Russell's view fares no better. He must take the fact that B(O,0y,F,a) to represent the fact that a is F. Tbe relation B, unlike normal relations, thus plays a dual role. It combines with Othello, a, F, and 0y into a belief fact, which has a specific content since it is taken to represent a second fact composed of some of the terms that it relates. This dual feature reveals a problem with Russell's analysis. Consider the predicate 'B(x,0y,F,a), , formed, by abstraction from the sentence expressing Othello's judgment. To ascribe such a predicate to Othello is, then, to assert two things: that Othello has a belief and that what is believed is that a is F. Suppose we express this by the conjunction 'B#(O) & <Fa>(O)', where 'B#' is a monadic predicate transcribing 'x is in or has a belief state' and <Fa> is a monadic predicate expressing what the content of the belief iso Thus, we separate the attitudinal property - believing, judging, doubting, understanding - from the content, as Russell does. But, like INTENTIONAL REFERENCE AS A LOGICAL RELATION 157 Bergmann, we now explicitly employ a predicate, <Fa>, representing a content property. What Russell does is take the occurrence of F, a and 0x in a belief fact to express the content. Thus, he does not avoid two of the things he seeks to avoid. First, he does not avoid a variant of (r3) above, since he takes the occurrence of Fand a in an intentional fact (or in the form obtained by abstraction from such a fact) to represent the fact that a is F. Thus, he will still have to face the problem of non-obtaining states of affairs being represented by beliefs. Second, he makes use of a representational relation between complexes, over and above the connections expressed in (rl) and (r2). This is easily seen in the case of either the function l:(x,0y,F,a) or Russell's form (3l:)(3x)l:(x,0y,F,a). Russell must correlate such a function or form to the fact that a is F. Actually, his analysis not only recognizes something that plays the role of a proposition, but several things for each classical proposition. For, aside from the class of judgment facts, containing F, a and 0x, and the relational propositional function, l:(x,0y,F,a), that we have introduced in addition to Russell's form, (3l:)(3x)l:(x,0y,F,a), there will be various monadic functions like B(x,0y,F,a) and forms like (3x)B(x,0y,F,a). Such entities, however, are not crucial to the present issue. What is important to note is that Wittgenstein's appeal to logical or internal properties of constituents adds nothing to Russell's analysis. In fact, it is quite natural to take Russell's view as embodying such logical properties. That is why only logically proper types of signs can be substituted for variables in contexts like 'B(O,0y,0,y)'. What Wittgenstein and, many years later, David Pears fail to see is that Wittgenstein' s view suffers from the same difficulty that Russell's does. 23 Wittgenstein, no more than Russell, recognized the need to explicitly invoke a pattern like (r3) in addition to patterns like (rl) and (r2) - the need to correlate complexes as weH as their constituents. This is clear from 5.542 in the Tractatus. Pears takes Wittgenstein to argue that the addition of a logical form like 0y to thejudgment complex does not make Russell's analysis more viable. 24 This is correct, for one must still treat the relation B as being such that it can only relate certain kinds ofterms, and adding a logical form only redundantly specifies that appropriate terms must be of a 0-type and a y-type, respectively. The issue is whether the need to impose the condition that a sentence ofthe form 'B(xf,0,y)' is weH-formed only ifthe terms replacing the variables are of the requisite logical kinds reveals that Russell's analysis implicitly appeals to complex linguistic expressions as subject terms, as in 'B(x,0y)'. Russell was obviously bothered by this question. But, it is interesting that, in spite of his dramatic reaction to Wittgenstein's criticism, he discussed the relational theory, in its original form, in the 1918lectures on logical atomism, 158 HERBERT HOCHBERG and he still rejected propositional entities in The Monist essays of 1914, which are supposedly the first six chapters of the 1913 manuscript. In 1918, as noted earlier, he expressed reservations about the theory being forced to employ "two verbs" in patterns like 'B(O,L,d,a), and, ultimately, abandoned the theory for that reason and for the further reason, expressed in 'On Propositions: What They Are and How They Mean' of 1919 and in a 1918 entry in his papers, that he no longer believed in a subject, whether it be a believing person or amental act ofbelief. One can take his concern over the use oftwo "verbs" to be a variant ofWittgenstein's criticism, as we have construed that criticism, that Russell assumes that some of the terms of an intentional fact are of appropriate logical kinds to combine into a fact grounding a belief s truth. Pears, following Wittgenstein, focuses on the relatively insignificant and obviously problematic introduction oflogical forms into the analysis ofintentional facts. Russell mistakenly thought that the addition of such forms somehow connected the appropriate constituents. 25 The addition of such forms is insignificant not only in that it adds little to the earlier form of the multiple relation theory, but in that Russell introduced such forms, in part, to make a negative point: they did not suffice to specify truth grounds in the case ofnonsymmetrical relational facts. Moreover, in 1913, Russell no longer thought that he must account for and not simply appeal to the ordering of the constituents in a fact. For, to take facts to involve an order of the constituents requires an account or analysis of the basis for differences of order. This led to his introduction of relations like R* and R* * and the analysis of order in facts that we discussed earlier. RusselI' s handling of relational contents reveals a radical difference in the analysis of belief contexts involving non-symmetrical relations and those involving monadic properties or symmetrical relations. In the latter cases, the pre-1913 pattern still applies, with the relatively insignificant addition of logical forms as terms. In the former case, we now have an explicit variant of a propositional analysis, since the second subject term for the predicate 'B' in 'B(O,(1p)(aR*p & bR**p) exists)' is a sentence. (We might also note that 'a' and 'h' occur in the embedded sentential forms 'aR*p' and 'bR**p'.) Russell' s employment of such a radically different pattern for non-symmetrical relations is an awkward feature of his analysis, though we can understand why he makes such a move. There is no obvious way to adapt his treatment of relations for the monadic case, and if he distinguishes the positions in a symmetrical relational fact, he would recognize two such facts in any given case of the relation obtaining. The latter consequence he explicitly sought to INTENTIONAL REFERENCE AS A LOGICAL RELATION 159 avoid, while he saw no need to revise his earlier pattern for monadic contents. Aside from any questions about the awkwardness of employing two totally different patterns of analysis for intentional contexts, what is interesting, for OUf discussion, about his handling of non-symmetrie al relation al contexts is that Wittgenstein's criticisms are irrelevant to it. They are irrelevant since Russell employs a sentential expression as a subject term along the lines of a propositional analysis. This raises a question regarding just how much of Russell's 1913 manuscript Wittgenstein read or discussed with Russell. 26 Russell' s revised analysis, being a variant of a propositional analysis, faces an obvious problem, but one that is quite different from that raised in Wittgenstein' sand Pears' criticisms. In fact, Pears, like Wittgenstein, ignores the obvious change in Russell's analysis. The problem Russell's new analysis faces is the traditional one we noted earlier: since B is now a two-term relation between a person and a complex he requires a second term for the relation when the belief is false and the complex does not exist. As we noted earlier, Russell's new analysis thus fails, but not for the reasons Pears or Wittgenstein give. In his discussion, Pears focuses on the role of acquaintance in Russell's analysis and the question ofjust what logical forms are. The two questions are connected. Russell is concerned to ans wer the somewhat Kantian question: How is the understanding of a sentence possible? In the case of 'Fa' he takes it to be sufficient that we have acquaintance with F, with a, and with the form 0y. Fand aare objects that are unproblematically terms of the relation of direct acquaintance. The form 0y poses a problem, since it appears to be necessarily connected with a complex of that form. Hence to be acquainted with it appears to require acquaintance with a complex. But, as Pears understands RusselI' s problem, to be acquainted with and hence understand a complex presupposes acquaintance with its constituents. Thus, Russell supposedly faces an insoluble problem, since acquaintance with the form presupposes acquaintance with a complex of which it is the form; but acquaintance with such a complex presupposes acquaintance with its component form. Russell faces no such problem. Even if he holds that acquaintance with a complex implies acquaintance with whatever is involved in its analysis, which Pears attributes to hirn on the basis of a misunderstanding, there is no problem so long as one understands 'presupposes' or 'implies' in a logical and not a temporal sense. There is no more a problem with talking of acquaintance with logical forms than there is of speaking of acquaintance with universal properties by one who holds that only instantiated properties exist and can be 160 HERBERT HOCHBERG objects of acquaintance only as instantiated. All Russell need do is hold that to be acquainted with a form is to be acquainted with a complex of which it is the form and vice versa. But the criticism itself arises, as I mentioned above, from a misunderstanding by Pears. Russell holds that to understand a proposition requires knowledge of the form and constituents. Assuming he no longer recognizes propositional entities (even though we noted an ambivalence about this in 1913), we may understand this claim to mean that an atomic sentence like 'Fa' is understood only if the object a, the property F, and the logical form 0y are known. Assurne also that the form 0y is known only ifan atomic fact ofthat form is known. To be acquainted with an atomic fact still does not require prior acquaintance with a form. All that is required, at most, is that if there is acquaintance with such a fact, and awareness of it as such, there is also acquaintance with or knowledge of a form. There is no more a problem here than there is in the case of being acquainted with several instantiations of a property. What Pears does is simply raise a variant of Plato's old objection against abstraction and, hence, ofthe argument for the innate knowledge of the forms. A logical form must supposedly be known in order to recognize it in a fact. But it cannot be grasped in experience, Le. by acquaintance, for in order to experience it one must recognize it as the form it is, and, hence, it already must be known. Pears' argument is no more viable than Plato's. What gives it some credence is Russell's mingling questions about logical forms of facts with questions about logical truths. Russell, concerned about grounding truths like '(30)(3y)0y', held that such a truth was not grounded by a complex. In part this was due to his concern about the constituents of such a fact. What he did, as we noted earlier, was identify a logical form like 0y with the truth ground for the existential statement '(30)(3y)0y', the form R(y,z) with the existential fact (but not complex) that was the truth ground for '(3R)(3y)(3z)R(y,z)', and so on. The logical forms of monadic predication, dyadic predication, etc. are then existentially general facts that are simples. His argument for this was concise. Since neither the form nor the existentially general fact had constituents, both were simples. Hence, the existential fact could not contain the form but could be identified with it. Pears sees this as raising an inescapable regress for Russell, as he would have to be acquainted with the forms to comprehend the complexes. Yet, he could only be acquainted with the forms if he was acquainted with the relevant complexes. But, while Russell identifies the predicative forms with such existential general facts, he sees problems with the claim for two reasons. First, he is inclined to hold that for each such general fact or form to exist, there must be a basic relation and atomic fact of the appropriate logical kind, INTENTIONAL REFERENCE AS A LOGICAL RELATION 161 and it cannot be a matter of logic that such a relation and fact should exist. That is, the form must be the form of some fact. But it also seems evident to hirn that a form could be grasped even if there should be no such fact. Thus, the form of an n-term relational complex is grasped, or is an object of acquaintance, whether or not there are n terms standing in an n-term relation or whether or not there is such a relation. Second, identifying a logical form with a general fact reraises the problem he had earlier noted when he took propositions to be forms: But in saying that there are relations ofwhich there are no instances, what we should naturally suppose that we mean would be "there are propositions ofthe form xRy for values of R for which such propositions are false whatever x and y may be". But with our definition of "proposition", we cannot have any reason to believe that there is a proposition which has never been understood, and we cannot know that a proposition ofthe form xRy has ever been understood, if the R is one of which there is no instance, and with which consequently no human being is acquainted. This instance suggests, what is also suggested by many other considerations, that our definition of a proposition is inadequate. 27 Russell does not resolve the problem, but he does proceed to speak of "the form 'x and y in a relation'" and of 'xRy' as the "general form" of a relational proposition, and not of something having some relation to something. But this may merely be a concise form of expression. 28 Yet it suggests that he was tempted to abandon the identification of logical forms with existentially general facts and to hold that the knowing of a logical form does not involve being acquainted with a fact of which it is the form. In so far as logical forms are universals, this means that for logical forms Russell did not hold that the existence of uni versals requires that they be exemplified. Thus, we can know each form in the series 0y, R(y, z) ... , without it being true that there is some exemplified relation of the appropriate kind. This is in keeping with his Platonism in The Problems of Philosophy, where he claimed acquaintance with unexemplified universals. It is Russell's identification oflogical forms with the existentially general facts earlier in the manuscript that provides the basis for Pears posing his dilemma for Russell's analysis. But the purported dilemma, which is not really a dilemma even given that identification, completely dissolves without it. And that identification may have been abandoned by Russell in the later parts of the manuscript, or, as we shall note later, it may even have been a reason why he abandoned the manuscript. Given the problems posed by identifying the form of monadic predication with the existentially general fact that something has some property, one may wonder why Russell construes the form as 162 HERBERT HOCHBERG a logically necessary fact. Since he takes such facts to ground logical truths, but cannot see how such general facts can have constituents, he has one motive for so construing forms. In this connection he speaks of the existential quantifier ("something") as being "nothing". Since the quantifier sign does not represent an entity there is no constituent of the existentially general fact to distinguish it from the form. This feature also provides a clear cut distinction between logical facts and non-Iogieal facts that fits with the notion of logic being a matter ofform: logical facts, being forms, are facts that have no constituents. Russell also mentions, as another motive, wanting to take a proposition as "something which can subsist on its own account, and not only as a fictitious constituent of certain mental complexes", Le. not as an "incomplete symbol". Both reasons fit with his taking such a form to be an object of acquaintance, rather than an object of a belief or judgment, since the latter can be true or false, but in acquaintance cases there is no correlate of falsity. There is also another, quite simple reason, which Russell hints at, but does not explicitly state, for his identifying the form with the fact. He takes the fact that a is F to be composed of a and F united by the form. We use 'Fa' to state that a is F, but we could use '(30)(3y)(0y & 0 =F & Y =a)'. Ifwe do so and withdraw the clause 'y =a', and hence the object a, we are left with '(30)(3y)(0y & 0 =F)" which we can use to say that something has the property F. It is not far fetched to think ofthis fact as being derived from the first by abstracting or removing the object a. We are then left with the form and the property F. If we then withdraw the clause '0 = F', we are left with the expression '(30)(3y)0y'. It is then easy to think that this existential generalization represents what remains after the other constituent ofthe fact, the property F, is removed. Hence, we can think that what remains, the form, is represented by an existentially general sentence. Thinking along such lines one might think that abstracting from 'Fa' we arrive, in successive stages, at '(3y)Fy' and '(30)(3y)0y'. Russell takes the form 0y, or (30)(3y)0y, to be an object of acquaintance as a consequence of his holding that understanding the proposition 'Fa' impli es being acquainted with the elements ofthe fact that is its ground oftruth. To understand a monadic atomie sentence, in short, requires that one be acquainted with a property, a subject, and the form of monadic exemplification. How one obtains acquaintance with such entities is not something Russell goes into. What Pears raises as a problem, and what he took Wittgenstein to raise, is a question about how we obtain the understanding of what it is for a subject to exemplify a property - what it is to be a monadic atomic fact - by being acquainted with the elements of such a fact, including its form. Pears INTENTIONAL REFERENCE AS A LOGICAL RELATION 163 argues that Russell introduced logical forms to resolve this problem but that such things do not suffice for the task. We have already noted in what sense this is correct, for it is another aspect of the criticism that the relational view is merely a verbal variant of a propositional analysis, since we must recognize that the items in a fact are of a certain kind. As put just above, the point is that a and F must be taken to be appropriate logical kinds to stand in the form 0y, and the introduction of 0y c1early does not resolve the matter. This reiterates the point that the terms in the places occupied by 'F' and 'a' in 'B(O,0y,F,a), must be combinable into a subject-predicate sentence, a complex sign that represents a possible situation. We noted earlier that Russell's theory of descriptions does not allow him to resolve one problem of false beliefposed by 'Othello believes that ahas RI to b', since with 'B(O,( 1p)(aR*p & bR**p) exists)' Othello cannot stand in the relation B to a non-existent complex. Thus, Russell will have to reintroduce such complexes or hold that Othello cannot believe what is false. It will not do to modifY (3*) and use (3**) 'B(O, (1p)(aR*p & bR**p»)' - to employ the description as a term, in place ofthe senten ce stating that what is described exists. For, then, on Russell's theory of descriptions, Othello can still only have such a belief if there is an appropriate fact, Le. if the belief is true. It is interesting that this is not Wittgenstein's criticism nor is it raised by Pears. One might speculate that the failure of his theory of descriptions to provide a key for the analysis of belief contexts, not Wittgenstein' s criticism, was the reason Russell abandoned the 1913 theory. For, the criticism did not apply to the contexts treated according to the real innovation of the 1913 theory, the non-symmetrical relational contexts. Pears suggests that Wittgenstein's criticism of Russell's identification of logical forms with existentially general facts may have played a role in his abandoning the manuscript. 29 Russell, hirnself, raised problems about that identification in the manuscript, and it may weil be that he became even more dissatisfied with it as the manuscript progressed. Pears argues that identifYing forms with existentially general propositions, and taking these to be simple, makes "their truth unintelligible".30 But his argument presupposes that a ground of truth for a logical truth must be a complex, as is a ground of truth for a non-Iogical truth. Russell is not compelled to agree. On his account there is nothing unintelligible ab out taking an existentially generalfact to be a simple and a ground of truth. The problem sterns, in part, from Pears' focusing on Russell's use ofthe term 'proposition'. This is somewhat misleading, since as Russell says, a logical form is a fact: "if a logical form, Le. a fact containing no constants ... ".31 The point is that if we state Russell's view more carefully than Pears does, or than Russell himself does in the passage 164 HERBERT HOCHBERG Pears quotes, the proposition would be represented by the expression 'l:(S,(30)(3y)0y)', read as 'There is al: and an S such thatl:(S,(30)(3y)0y)'. The proposition is a complex, and it is true, since the object, the form, (30)(3y)0y, which is its ground of truth, is an object of acquaintance and, hence, exists. This object is not, like the true proposition, complex, but simple. The true proposition is complex since it contains the form (30)(3y)0y as a constituent. The difference between a proposition, whose ground oftruth is a logieal form or logical fact, and a proposition like 'l:(S,(30)(3y)0y,F,a)' is two-fold. First, in the former case, only the form is a constituent of the proposition, and, second, its being a constituent of the proposition means that the proposition contains its own ground of truth. One may take that to "show" that it is a necessary truth, as the existentially general fact (the logical form) is a logical or necessary fact. Pears has not only not shown that Russell's view is unintelligible, he has not set it out correctly. Pears does not discuss an obvious problem connected with Russell's identification of a form with a general fact. It is a problem that Russell undoubtedly came to notice. He cannot identify a form like (0)(y)0y with a logieal fact, since the corresponding statement is a logieal falsehood. Yet, he has no way to analyze (0)(y)0y, and he cannot deny that it is simple, by his line of reasoning in the existentially quantified case. There is no evidence in the text that Russell considered this problem, but it seems too obvious for hirn to have overlooked. It is useful to recall that in his book Mental Acts, Peter Geach criticized Russell for ignoring the problem of relational order in dealing with intentional contexts. Geach was obviously and understandably unaware ofRussell's attempt to deal with such problems in detail in the 1913 manuscript. Yet, given Russell's early concem with the logic of relations, one would expect hirn to be aware of problems concerning relational predication and order: the problems that were taken up in detail in the sections of the 1913 manuscript that he abandoned. As in the case of relational order, it would be unreasonable to think, as Geach did, that Russell was oblivious to an obvious problem. In any case, if Russell had separated questions about logieal form from questions of logieal truth, he could have abandoned the view that logical forms were general facts and returned to construing them as uni versals or functions; a view more in line with his treatment of expressions like '0y', '0" y" " and '(3y)0" y' in Principia. Since Russell introduced a propositional complex, in the form of a complex mental occurrence, in his Principia adaptation of the view cryptically expressed in 5.542, he could have employed patterns like those introduced earlier in (4) and (6) to resolve both problems - the problem of specifying INTENTIONAL REFERENCE AS A LOGICAL RELATION 165 truth grounds for atomic sentences, without recognizing situations, as weIl as the problem posed by the analysis of atomic intentional contexts. For, such patterns provide a link between the complex sentential signs and their truth conditions that is applicable in both cases. Given such a link, along with the representational role of its constituent signs, an atomic sentence may be said to represent, even though there will not always be something it represents, a fact. Once complex contents are acknowledged, the same pattern can be used in explicitly intentional contexts. Though Russell took one of the two steps necessary in Principia, by following Wittgenstein and introducing complex mental contents, he did not then explicitly take the step of linking such states to their truth conditions via definite descriptions or existential claims. This is ironic since he had done so in 1905. What has been proposed here, as a solution to the problems posed by false belief, is thus derived from aspects of Russell's views at two distinct stages ofhis thought: the 1905 use of definite descriptions (or existential claims) to avoid non-existent facts and the 1927 Principia analysis, influenced by Wittgenstein; avoiding propositional entities. One must still account for intentional contexts involving complex contents: truth functionally complex and quantified contexts. To do so one must first consider what entities are recognized in the specification of truth conditions for statements expressing such contents. As I have considered this in some detail elsewhere,32 I shall simply state, without argument, what is involved. Since we acknowledge a set of atomic facts, we can specify truth grounds for atomic sentences and their negations in terms of facts with certain constituents belonging to that set. Thus, the truth ground for '-,Fa' is provided by: Tr ',Fa' iffnot-(3p)(pCa & pCF & pC0x). Conjunctions, disjunctions, and negations of molecular statements can then be treated in the familiar reductive way without recognizing conjunctive, disjunctive, etc. facts. A conjunction, for example, is true if and only if its conjuncts are true. If the conjuncts are atomic or negations of atomic statements, the specification ofthe conjunction's truth condition reduces to claims about the existence and non-existence of atomic facts. If they are not, then such statements themselves involve claims about the truth (or denial) of further statements, until the claims reduce to such claims about existence and nonexistence, which are claims about the members of the set of atomic facts. One can deal with quantified statements in a similar mann er, given that we acknowledge domains of facts, of particulars, and of properties. For primi- 166 HERBERT HOCHBERG tive predicates, the following patterns are available: Tr'(3x)Fx' iff (3x)(3p)(pCx & pCF & pC0x) Tr' (x )Fx' iff (x)(3p )(pCx & pCF & pC0x) with the obvious treatment ofthe negations. For quantification over complex contexts, the analysis can make use of the familiar construal of satisfaction, as in the case of molecular compounds. We can now turn to the consideration of complex contents for intentional contexts. Consider amental state whose content is expressed by 'Fa & aR1b'. How are we to characterize such astate so that we may say its content is expressed in such a way? Clearly we would expect to have, at least, what would characterize states whose contents were expressed by the individual atomic sentences. The problem is to avoid taking the occurrence of '&' to represent something, as the tokens of the names and predicates are taken to represent objects and properties. But we can avoid doing that. What we have is a complex content that is taken to be true under the conditions for the conjunction being true. This, after all, is what is involved in understanding conjunction. Thus, our earlier pattern can be applied in specifying more complex truth conditions: the occurring conjunctive content will be true if and only if the constituent atomic sentence tokens are true. Similar patterns are applicable in the case of other molecular and quantified statements and, in particular, the case of negation. One need not introduce, as Russell once did, basic attitudes (characteristics of mental states, in our terms) of dis belief or rejection to link negative contexts to truth conditions. This is not to say that one cannot also recognize such states of disbelief as further psychological kinds. There would simply be a difference, corresponding to any generic difference among mental states, between states of disbelief that a is Fand states ofbeliefthat a is not-F. The relation between them would be captured by the connection between the respective truth grounds for the statements expressing the contents. Thus, in one sense one may say that a person who is in a state of belief that a is F is in astate of disbelief that a is not-F. In another sense this is not so, just as one who is in a belief state expressed by 'a is F' is not, in one sense, in astate ofbelief expressed by 'a is F & ais F', but is, in another sense, in such astate. The present analysis can accomodate these diverse senses ofbelief (thought, judgment) without recognizing either possible facts or propositional entities. The use of biconditionals specifying truth grounds might lead areader to see a resemblance between the present analysis and attempts to offer a theory INTENTIONAL REFERENCE AS A LOGICAL RELATION 167 of meaning in terms of so-called truth conditions in the manner of Donald Davidson. Nothing could be more misleading. Davidson claims that 's means that p' can be rendered as 's is true if and only if p' .33 Put in terms of our discussion, we would then have '''Fa' means that Fa", (or Moore's (1) or Bergmann's (tb», rendered as "'Fa' is true iff Fa". Thus, Othello's understanding or knowing the meaning of 'Fa' would supposedty be analyzable in terms of his knowing or understanding that: 'Fa' is true iff Fa - where the right side of the biconditional, here 'Fa' itself, furnishes a statement of the truth condition for 'Fa'. We need only recall that the appeal to facts and situations by the logical atomists and other correspondence theorists involved the taking of truth conditions (grounds) to be facts to appreciate the utter triviality of such an approach. Davidson attempts to avoid the acknowledgment of a correlated fact as a truth ground by producing a token of a Tarskian T-sentence. But if one is to deal with the philosophical issues involved, one must probe into the role 'Fa' (or 'p') plays in Davidson's analysis and face the problem of connectitig such a sentence or sentence token to a ground of truth in the sense in which a fact is a condition or ground oftruth. All Davidson does is utter, as a truth condition for a sentence (sentence token), another sentence token - either ofthe sentence itself - 'snow is white' is true iff snow is white - or a transcription of it. On such a view one gives a truth condition by repeating or transcribing a sentence: one does not specif.y what it is to be a truth condition as traditional theories of truth do specify what truth conditions are - facts, coherence relations, and so forth. If one takes up the philosophical issues involved, then the problem about the meaning of 'Fa' becomes a problem about whatever one puts for 'p' in '''Fa' is true iffp". This is one reason Davidson's analysis is trivial. It is also trivial in that it plays on the obvious fact that when one asks for the meaning of a sentence in ordinary contexts, a satisfactory answer is provided by furnishing another sentence. If we recall that Russell and Wittgenstein sought to avoid meanings, in the form of propositions, by connecting atomic sentences (or acts of thought, belief, etc.) directly to facts (states of affairs) without employing intermediary propositional entities, we see why Davidson's analysis lacks originality as weil as substance. Under Tarski's influence, he alters the pattern of Russell and Wittgenstein so that in place of "Othello knows that 'Fa' represents Fa" explicating "Othello knows the meaning of 'Fa''', we have "Othello knows that: 'Fa' is true iff Fa". Davidson thus copies the basic pattern ofthe logical atomists who take facts as truth grounds for atomic sentences, while avoiding propositions as meanings. But, as he stops short by failing to finish the pattern and provide truth grounds for atomic sentences, 168 HERBERT HOCHBERG he prematurely withdraws from the philosophical controversies involved. What he does is adopt the form of the atomists reference theory of meaning, which avoids special entities as "meanings", without its substance: the recognition of facts as truth grounds. The present analysis not only takes facts to be truth grounds but acknowledges a basic intentional or representational (meaning) relation between terms of atomic sentences (or elements of mental states that are thoughts) and the constituents ofatomic facts. Yet, by not employing such a relation for atomic facts and sentences (or thoughts), it resolves the problems that Davidson dismisses. University 0/ Texas at Austin NOTES I G. E. Moore, Some Main Problems 0/ Philosophy (SMPP), G. Allen & Unwin, London 1953, pp. 267-68. 2 Thus Moore writes " ... in what would this belief ofhis consist? What is the correct analysis ofthe event that would be happening in his mind?" (SM PP, p. 258). 3 Moore does not explicitly appeal to such attributes, since, after rejecting propositional entities, he proposes "to give up the attempt to analyze beliefs" (SMPP, p. 266). But, he does implicitly appeal to such attributes in his discussion by speaking of"the beliefthat lions exist" and he recognizes that he must account for different particular beliefs having the same content as well for particular states ofbeliefwith different contents (SMPP, pp. 258-59). For a detailed discussion ofMoore's attempt to analyse belief see H. Hochberg, Thought. Fact. and Re/erence: The Origins and Ontology 0/Logical A tomism, Minnesota University Press, Minneapolis 1978, pp. 3-12; 53--86. 4 SMPP, p. 267. 5 SM PP, pp. 256,257. 6 Russell obscured matters by speaking of"such an entity as the relation R" existing or not existing in such a case. What he had in mind was c1early the obtaining or existing ofthe relation between a and b. His speaking of"the relation" perhaps prevented hirn from seeing the problem. On this matter, see my 'Descriptions, Situations, and RusselI' s ExtensionalAnalysis ofIntentionality' ,Philosophy and Phenomenological Research, 49, 4, 1989. 7 B. RusselI, The Problems 0/Philosophy, William & Norgate, London 1956, p. 128. 8 L. Wittgenstein, Noteboo!Gs 1914-1916 (NB), ed. by G. H. von Wright& G. E. M.Anscombe, B. Blackwell, Oxford 1961, pp. 82, 129. There is a familiar problem concerning the interpretation of the Tractatus and passages of the notebooks regarding Wittgenstein' s views about properties and relations, on the one hand, and predicates, on the other. For my purposes here, I will simply treat his view in the Tractatus as "realistic" in that he would take an ideallanguage to contain predicates representing properties and relations. This is not to say that the realistic interpretation is correct, as it is usually presented. Rather, as I see it, Wittgenstein thinks of predicates as both representing functions, in Frege's sense, that are "completed" by objects and as being functions that are completed by objects, the latter objects being names of the former. It is the ambiguous nature of Fregean functions (concepts), being incomplete entities that are contrasted with objects, that accounts for the INTENTIONAL REFERENCE AS A LOGICAL RELATION 169 realistic remarks as weIl as for the nominalistic tone of other remarks in the Tractatus. See 'Frege on Concepts as Functions: A Fundamental Ambiguity', in H. Hochberg, Logic, Ontology, and Language, Philosophia Verlag, Munich 1984. 9 L. Wittgenstein, Philosophical Grammar (PG), ed. by R. Rhees, Trans. by A. Kenny, University ofCalifomia Press, Berkeley-LosAngeles 1978, p. 161. \0 The letter is reproduced in NB, p. 121; the discussion occurs on p. 96. 11 B. A. W. Russen, The Collected Papers ofBertrand Russell, vol. 8, ed. by J. G. Slater, G. Allen & Unwin, London 1986, pp. 253-284. 12 A. N. Whitehead and B. A. W. Russel1, Principia Mathematica, vol. I, Cambridge University Press, Cambridge 1960, p. 662. Ironical1y, this analysis is essential1y repeated, in a confused version, by P. Geach as his "theory" in MentalActs (B. Blackwell, Oxford 1957). Geach, writing many years later, does not note or mention the connection, while offering criticisms ofRussell's 1912 version of the multiple relation theory that, in part, repeat Wittgenstein's criticisms and, in part, complain about the supposed inability of the multiple relation theory to handle relational contents. Geach can be excused for not being aware of Russell's 1913 manuscript dealing with the latter issue, though even the first edition of Principia has the technical devices for handling relational contents in terms ofthe early form of lambda abstraction Russell employs; the failure to mention the 2nd edition Principia discussion is another matter. 13 The earliest of Bergmann 's papers on the subject is 'A Positivistic Metaphysics of Consciousness' , Mind, 45,1945. But, it is not until 'Bodies, Minds and Acts', in The Metaphysics ofLogical Positivism, Longmans, Green & Co., New York 1954 and 'Intentionality', in Semantica. Archivio di Filosofia, 1955, reprinted in Meaning and Existence, University ofWisconsin Press, Madison 1959, that he explicitly develops the views we are considering. 14 To preserve his claim that M-connections were analytic, Bergmann proposed a revised analysis of the concept of analyticity, seeking to provide an ontological ground for analytic truths. Thus he not only followed Moore's lead in his analysis ofintentional contexts, but he returned to issues that Russell had tried to resolve in 1912-13, issues concerning the ontological ground oflogical truth. For a discussion ofthe development ofBergmann's analysis of intentionality and ofvarious problems with it, see my "Intentions, Facts, and Propositions" in Classics ofAnalytical Metaphysics, ed. by L. Blackman, Macmillan, NewYork 1984; 'Intentionality, Logical Structure, and Bergmann's Ontology',Nous, 15, 2,1981; 'Beliefand Intention', in Hochberg (1984); and 'Gustav Bergmann' in Handbook ofMetaphysics and Ontology, ed. by H. Burkhardt & B. Smith, Philosophia Verlag, Munich 1991. 15 B. A. W. Russell, Theory ofKnowledge, v. 7, The Collected Papers of Bertrand Russell, ed. E. Eames et al., G. Allen & Unwin, London 1984, p. 195. 16 For a detailed discussion of the various senses of 'same content' involved in such matters see 'Beliefand Intention', pp. 399-406 and 'Descriptions, Situations, and Russell's ExtensionalAnalysis ofIntentionality'. 11 PG, p. 144. 18 Russell (1984), p. 115. 19 lbid., pp. 115-116. 20 Ibid., p. \09. 21 Discussions ofRussell's various theories oftypes often ignore a basic sense of'type' in which Russell takes monadic, dyadic, triadic, etc. attributes and relations to be oflogically different types. On the importance ofthis sense of'type' and its relevance for Russell's paradox see my 'Russell's Paradox, Russellian Relations, and the Problems ofPredication and Impredicativity', in Rereading 170 HERBERT HOCHBERG Russell: Essays on Hertrand Russe/l's Metaphysics and Epistemology, ed. by C. Savage & C. Anderson, in Minnesota Studies in The Philosophy 0/ Science, Vol. 12, Minnesota University Press, Minneapolis 1989, pp. 63-87. 22 Russen (1984), p. 117. 23 D. Pears, 'Russell's 1913 Theory ofKnowledge Manuscript', in Rereading Russe/l: Essays on Bertrand Russell's Metaphysics and Epistemology, cit., pp. 169-182. 24 Pears (1989), p. 179. 25 Russen (1984), p. 116. 26 On a page (Russen 1984, p. 197) cited as Appendix B. 1 Folio 2, by the editors, Russell speaks of positive and negative facts and the neutral fact that is involved in judgrnent and which replaces the/orm ofhis old theory. The idea seems to be that the neutral fact (situation) can exist as a positive or negative fact, thus providing a truth ground for an atomic sentence or its negation. Propositions, or atomic sentences, would then represent neutral facts irrespective oftheir truth or falsity. Thissuggests Wittgenstein's influence on Russell, as culminating in the 10gical atomism lectures, 'On Propositions' of 1919, and the appendix to the 2nd edition ofPrincipia cited above. It should be noted that Wittgenstein' s 'Notes on Logic' are printed as an appendix to the NB and are dated September, 1913, when Wittgenstein dictated some of them to Russell and supplemented them with a typescript sent to Russell a "few days later". See R. Monk, Wittgenstein: The Duty 0/ Genius, Jonathan Cape, New York 1990, p. 93. 27 Russell (1984), p. 134. 28 Ibid., p. 135. In the 191810gical atomism lectures Russell takes expressions for logical forms to employ free variables, thus retuming to hisPrincipia notation for such forms. Following Russell's use at different places I use 'xRy' as weil as 'R(x,y)' for the dyadic relational form. 29 Pears (1989), pp. 176-178. Russel1 might have been influenced by Wittgenstein's views as recorded in the latter's notebooks. Though the relevant entries date from later than Russel1's manuscript, Wittgenstein and Russel1 had been discussing and communicating about such matters in the period prior to and during Russel1's work on the manuscript. Thus, in a letter of January, 1913 Wittgenstein writes: "I now think that qualities, relations (like love) etc. are al1 copulae! That means I for instance analyse a subject-predicate proposition, say, "Socrates is human" into "Socrates" and "something is human", (which I think is not complex)." (Wittgenstein (1969), appendix iii, pp. 120--121.) This could have suggested the idea that forms are existential1y general facts. In the NB, p. 17, Wittgenstein also writes that he "had thought that the possibility ofthe truth ofthe proposition 0a was tied up with the fact (3x)(30).0x", and on p. 22, "(Similarly (3x)0x would be the form of 0a, as I actually thought)." 30 Pears 1989, p. 178. 31 Russel11984, p. 131. In the fol1owing discussion I wil1 fol1ow Russel1's notation and use 'S' in place of'x' for the variable that would be replaced by a constant representing aperson. J2 See 'Truth Makers, Truth Predicates, and Truth TYPes', in Language. Truth and Ontology, ed. by K. Mulligan, Kluwer Academic Publishers, Dordrecht 1991, pp. 87-117. The crucial issues concern (i) the use of a negation in specif)ring that no fact of a certain kind is among the atomic facts; (ii) the use ofthe quantifiers in specif)ring truth conditions for generalized sentences; (iii) the appeal to sets ofatomic facts, particulars, and properties; (iv) possible regresses ofthe kind Frege and Bradley raised. 33 D. Davidson, 'Truth and Meaning', Synthese, 17, 1967, p. 309. ROSARIA EGIDI WITTGENSTEIN BETWEEN PHILOSOPHICAL GRAMMAR AND PSYCHOLOGY It is known that Wittgenstein's interest in philosophical problems ofpsychology goes back to Tractatus and his first writings. It certainly has its origins in the commitment which he shared with Frege, to challenge the long-established psychologistic solutions to the problems oflogic and ofthe theory ofknowledge. Nevertheless, the very "plan for the treatment of psychological concepts" assumed a specific form only later, when Wittgenstein lost his confidence in the anti-psychologistic strategy of Frege as weH as the belief which he had briefly held - in a phenomenological analysis of internal experience, which was in a way similar to the one already developed by Brentano and theAustrian-German schools ofphenomenology. This treatment, together with reflections on the foundations of mathematics, was intended to form the 11 part of Philosophicallnvestigations l and constitute therefore a coroHary or - one might say - the application of the new theory of meaning elaborated in the I part. As we know, such a project was never carried out; only two vast fragments exist: the writings on the foundations of mathematics 2 and the later materials on the philosophy of mind coHected in the two volumes of Remarks on the Philosophy 0/ Psychology3 and in the first volume of Last Writings on the Philosophy 0/ Psychology. 4 Even if these ultimate texts do not contribute much to Wittgenstein's general perspective, already weH known from the lnvestigations and ZettelS the range of the arguments as weH as the richness and sophistication of presented details, make their study an advantageous starting point for Wittgenstein's announced conception of perceptive, emotional and cognitive language ofman in the frame ofnon-Fregean theory of meaning. Perhaps Wittgenstein never intended to write a systematic book that would coHect his thoughts on mathematics and psychology. It is certain, however, that already in the beginning ofthe 1930's he had in his mind a clear picture of what such work should be like, how it should be structured and which subjects it should present. The manuscript on which he worked in these years, published more than thirty years later under the title of Philosophische 171 R. Egidi (ed.), Wittgenstein: Mind and Language. 171-184. © 1995 Kluwer Academic Publishers. 172 ROSARIA EGIDI Grammatik can be considered an example of this ideal book, a sort of sketch book avant la lettre of the great work in which the two parts would be united. In an appendix to this work Wittgenstein explicitly declares what the structure and the content of a book dedicated to the philosophical grammar should be: If we had grammar set out in the form of a book, it wouldn't be aseries of chapters side by side, it would have quite a different structure. And it is here, ifI am right, that we would have to see the distinction between phenomenological and not-phenomenological. There would be, say, a chapter about colours, setting out the rules for the use of colour-words; but there would be nothing comparable in what the grammar had to say about the words "not", "or", etc. (the "logical constants"). It would, for instance, be a consequence ofthe rules, that these latter words unlike the colour words were usable in every propositions; and the generality belonging to this "every" would not be the kind that is discovered by experience, but the generality of a supreme rule of the game admitting of no appeal. 6 Two principal theses of the late Wittgenstein are presented in this passage; they are mutually supporting and both deserve our attention. One concerns (i) the distinction between "phenomenological" and "non-phenomenological", the other intro duces (H) the kind of higher "generality" which belongs to the rules of philosophical grammar in contrast with empirical generality characteristic of physical laws. (i) The purpose of grammar is, then, to separate the physical, the empirical, the causally determined from the domain of the phenomenological and to attribute to non-phenomenological concepts occurring in mathematical and psychological propositions, a peculiar requisite of generality, not to be confused with the one that belongs to empirical generalisations, but one that is typical of grammar or of rules of a game. These propositions which obey the rules of grammar or of a game are in fact different from empirical generalisations; they do not have the explicative power of a deductive or inductive type and cannot be analysed by means oftruth-functionallogic. It is important, however, to bear in mind that before attributing to philosophical grammar the task of analysing the constructs resistant to the treatment with truth-functionallogic, Wittgenstein had thought that this could be achieved by an intermediate discipline (ein Mittelding) that would He in between logic and empirical science, by a discipline Wittgenstein called in the Philosophical Remar!,s and later in the Remarks on Colour "phenomenology". The focal position that the distinction between phenomenological and not-phenomenological or physical occupies in the first work 7 shows that Wittgenstein's idea of grammar is still bound to the context of ideas, where BETWEEN PHILOSOPHICAL GRAMMAR AND PSYCHOLOGY 173 the analysis of propositions concerning immediate experience was the attribute of phenomenology. Actually, it is the function of the latter to analyse the special class of phenomena of immediate experience and constituting the language of data which Wittgenstein indicated as "primary language" (das Primäre). It is distinct from the "secondary language" (das Sekundäre), which consists of concepts like those of physical sciences. 8 All propositions containing psychological verbs (intend, understand, believe, expect, desire, doubt, etc.) belong to the primary language. In the Tractatus they had been called SatzJormen der Psychologie, which include the propositions Russell called "propositional attitudes": "A believes/thinks/intends/hopes that p", as well as the hypotheses and modal sentences (the necessary and probabilistic propositions).9 They form the complex of non-extensional contexts which cannot be treated by means of truth-functionallogic. The idea that phenomenology is in power to analyse primary language is discussed in the Philosophical Remarks 1o , but the definition of phenomenology as a form of "conceptual analysis" (Begriffsanalyse) appears in the Remarks on Colour: Phenomenological analysis (as e.g. Goethe would have it) is analysis of concepts and can neither agree with nor contradict physics. 1l In the same work Wittgenstein claims that phenomenology is aMittelding between logic and physics and its field of application is not that of physical propositions: But what kind of a proposition is that, that blending in white removes the colouredness from the colour? As I mean it, it can't be a proposition of physics. Here is the temptation to believe in a phenomenology, something midway between science and logic, is very great. (p. 15) Even though Wittgenstein admitted the existence of"phenomenological problems", those late notes of the Remarks on Colour denied, all the same, the existence of a "phenomenology": There is indeed no such thing as phenomenology, but there are phenomenological problems. (pp. 9, 49) As we see, Wittgenstein qualifies as a "temptation" the belief that the phenomenology should be able to provide a language adequate to describe primary phenomena of immediate experience. Such a qualification seems still more meaningful if one realises that in the Remarks on Colour he is clearly alluding to a view that he hirnselfheld, when in the early phase ofhis thought he fell victim to the phenomenological temptation. 174 ROSARIA EGIDI The broader and better articulated context for the reconstruction of Wittgenstein' s conception of phenomenology can be found in the Philosoph ical Remarks. It is on the basis ofthe transitional works, however, that we can reconstruct the three arguments with which he renders plausible his idea of phenomenology.12 The first of these arguments consists of the distinction between "phenomenological language" (phänomenologische Sprache) and "physical language" (physikalische Sprache).13 The second argument refers to the attribution to propositions of physics the power to determine "laws" (Gesetze) while the phenomenological propositions should have the power to determine "possibilities" (Möglichkeiten) .14 The third argument, finally, which derives from the second, affirms that phenomenology concems the "sense" (Sinn) of the propositions in contrast with the physics which deals with "truth values" (Wahrheit and Falschheit).15 The three arguments interweave in the Philosophical Remarks and reappear in the materials concerning lectures and conversations held by Wittgenstein between 1929-1935 and collected by Moore, Waismann and others. 16 We know that most of the ideas that they were intended to convey were eventually repudiated, but this does not diminish the validity of Wittgenstein's attempt to offer an analysis ofthe language ofintemal experience free from any collision with psychologism as weH as logicism. This attempt gives rise to what is often indicated as the philosophy ofthe intermediate Wittgenstein. One of its characteristics is the idea of a specific analysis ofphenomenological sentences to be distinguished from that ofphysical sentences and whose reference seems to be expressed by "senses" rather than "truth values" or, better, by "sense-impressions" rather than by "physical objects". This distinction lays foundations to the conception of two languages discussed by Wittgenstein in thePhilosophical Remarks: the phenomenological and the physicallanguage and to the phenomenological theory of colours (die phänomenologische Farbenlehre) which he opposed in the same work to the physical one. 17 It is weH known that the attempt developed particularly in the 11 part of the Investigations, as weH as in the writings conceming philosophical psychology, differs considerably from the context of ideas of Philosophical Remarks in both the elaboration and the sophistication of arguments. One of the differences lies in the fact that the idea of phenomenology disappears from this text and with it the illusion, which Wittgenstein had held for a short time, that phenomenological analysis could playa role in the anti-psychologistic strategy, or, more exactly, in the analysis of sentences expressing sense-impressions or, according to the version of The Blue and Brown Books "personal experiences". 18 BETWEEN PHILOSOPHICAL GRAMMAR AND PSYCHOLOGY 175 The problem was to determine the sense of those propositions with nonFregean arguments and to preserve their objectivity, i.e. their not-subjective and public character with more adequate arguments than those offered in the Philosophical Remar/rs. Even ifit is an exaggeration to claim that the abandonement ofthe idea of phenomenology constitutes the only qualifying point of Wittgenstein's mature conception of senses and indicates something more than just a change of strategy, there is no denying the utility of its accurate examination in clarifying Wittgenstein's antipsychologistic project. Particular emphasis should be made on the reasons that made hirn abandon the first attempt of a phenomenological analysis and to undertake a new one, whose qualifying points would be a consequence of the rejection of the Fregean conception of Sinn and a treatment of intentionality and of psychological concepts in a grammatical context. While Frege and then Carnap respectively, speaking about "senses" and "intensions",19 were convinced to be moving within the domain of philosophical logic, Wittgenstein, talking about belief, intention, understanding, desire, doubt, hope and expectation thought to be dealing with the problems of the Erkenntnistheorie, which he identified with the philosophy of mind or of psychology. Wittgenstein's theses related to it can be reconstructed starting with the analysis of sense and intentionality sketched in the I part of the Philosophical Grammar 20 and with the theory of intentional or voluntary action, whose traces are found in the 11 part of Investigations and in the writings on philosophical psychology.21 They can be summarised with the following theses which still today influence the contemporary theories of meaning. The one is that (a) the meaning of intentional propositions does not derive from psychological or mental phenomena or from internal experiences, but has a "grammatical", normative nature; the other thesis is that (b) the explanation of actions, the "intentional explanation" does not have the causal character typical of explanations of facts. The thesis that the intentional sentences (or propositional attitudes) of the type "A believes/doubts/hopes/is certain that p" do not refer to internal experiences and to mental phenomena, is a doubtless testimony to the continuity of Wittgenstein's anti-psychologistic approach with Fregean aim of preserving the objectivity of senses, as weIl as to the affinity with Carnap' s theory of intensions as expressing their senses in one of the possible worlds. It shows, in other words, that the analysis of intentionality is inspired by the common anti-subjectivistic and anti-mentalistic trend. However, Wittgenstein did not justify the objective content of intentional sentences at the cost of admitting 176 ROSARIA EGIDI senses or intentions. Their "objectivity" is not achieved by reference to objects, of both internal and extern al nature but depends on wh ether those sentences obey the system of rules of which they are part. According to Wittgenstein, "an intention is embedded in the situation, in human customs and institutions"22 and the meaning of sentences expressing intentions, desires, doubts, certainties and hypotheses is based on the way in which we use them, on the manner in which "we are taught to use [them]".23 The leading idea of Wittgenstein's analysis is that beliefs, intentions, hypotheses, hopes, doubts and expectations are not part of a primary language, the propositions expressing them do not refer to presumed mental phenomena, both internal and private, but that they playa peculiar function of the language. Such a function takes place in linguistic games in which such sentences perform a normative, pragmatic role, different from the descriptive function characteristic of the propositions whose meaning is based on direct knowledge referred to facts of extra-linguistic nature. Their unity of sense is not a determined objective content, but a whole system of rules which they obey and, hence, not a complex of semantic but of pragmatic criteria of significance. Within such contexts the term "pragmatic" means "referring to norms of actions", "according to rules", therefore such a requisite becomes pertinent to the sentences whose meaning is not based on reference to extra-linguistic facts or to mental, internal events, but is determined in relation to the rules or norms that govern social games, customs and institutions in which communicative behaviours or intentional or voluntary actions are embedded. They do not presuppose a direct but a background knowledge, according to which they can assume asense, a meaningful relevance for uso The sense is so given by the complex of activities and behaviours presupposed by "the understanding of a proposition" and not by those involved in the procedures to ascertain "the truth or the falsity" of a proposition. The experience determines the truth or falsity of a proposition but not its sense. 24 Intentional constructs do not, therefore, behave like sentences whose truth and falsity can be conc1usively determined, but they reflect the institutionalised, inherited forms of behaviour that cannot be expressed within the stable and defined domain of factual knowledge: But I did not getmy picture ofthe world by satisfYing myselfofits correctness; nordo I have it because I am satisfied of its correctness. No: it is the inherited background against which I distinguish between true and false. 25 Exactness and completness do not constitute a requisite or an ideal for BETWEEN PHILOSOPHICAL GRAMMAR AND PSYCHOLOGY 177 intentional sentences: rather they express a fluid and imperfect blend of criteria, at the same time ethical, aesthetic, economic, social, inspired by consolidated habits of a society or of a group and by considerations of an essentially pragmatic nature: We judge an action according to its background within human life, and this background is not monochrome, but we might picture it as a very complicated filigree pattern, which, to be sure, we can't copy, but which we can recognize from the general impression it makes. (RPP H, § 624) This is the reason why it is necessary to refute the deep-rooted conviction that the analysis of psychical or mental processes which should constitute the reference of intentional propositions, can be realised by a phenomenological language in the same way as the analysis of facts performed by means of physical language. It is, in particular, the idea of a phenomenology as the analysis of internal experiences, referring to the domain of sui generis pro positions that must be given up. It is necessary to resist to the temptation to mythologize the sense of expression deprived of a direct denotation, to admit the existence of mental processes as their objects and thus to avoid not only the mythology oflogicism but also the subtler mythology of psychologism. In this way Wittgenstein's flight from the Fregean senses serves the aim to avoid the double threat: the ontologization of mathematical concepts and numbers and the parallel reification of psychological concepts and mental states. As far as the other thesis of Wittgenstein is concerned, the one about the not-causal character ofthe explanations ofpropositional attitudes and ofsentences expressing intentional actions, one can say that it is a kind of corollary ofthe preceding thesis, i.e. ofthe idea to deny the identification ofintentions, desires, beliefs, doubts, etc. with mental phenomena and to claim the grammatical and pragmatic nature ofthe class of propositions which express them. Such a thesis is an alternative to the arguments which distinguish the causal theory of intentionality maintained, above all, by neoempiristic epistemology. As it is known, according to such a theory, the explicative relation between the intentions of an agent and his actions could be substantially analysed as a causal link. In the Investigations Wittgenstein develops his thesis of the causal inefficiency of intentions in respect to the actions 26 and the concept of "intention" is contrasted with that of "prediction" (Voraussage, Prognose) as an inference of the future actions of a person on the basis of a presumed psychological or mental process. In empirical sciences prediction is in fact a proposition based on inductive procedures and formulated in terms of extensionallogical models. In intentional contexts, on the contrary, what determines the possibility 178 ROSARIA EGIDI of a prediction offuture actions - or of an explanation (''justification'') of our past actions - is separate from all experiences and depends on the system of our linguistic conventions, institutions and rules. Wittgenstein emphasizes in On certainty: The squirrel does not infer by induction that it is going to need stores next winter as weIl. And no more do we need a law of induction to justify our actions or our predictions. (§ 287)27 This is the reason why Wittgenstein in On Certainty calls them, in contrast to empirical propositions, "systematic propositions" or propositions which have "the character of a rule" .28 A systematic as well as an intentional construct can have the linguistic form of an empirical proposition, "but of course it is not one that is based on experience".29 Wittgenstein's insistence on the sui generis character of such a relation, seems to create the need to distinguish between the methods of explanation of intentional and empirical propositions. It is impossible to give a causal explanation of intentional propositions, since intentions, desires, hopes and beliefs, i.e. the "motives" of an action are of a different category than the so called "causes" of a physical event are. 30 In the former case the link between cause and effect is not of logicalor rational nature; vice versa the relation between intention and action - the intentional explanation (die intentionale Erklärung) - is systematically or logically linked to the intended action. 31 However, for Wittgenstein this does not mean to propose a new dualism between facts and actions, between physical and mental events, in a word, between causality and intentionality. The difference is not ontological but a grammatical one; it simply indicates that the sentences constituting contexts in which intentional, rather than descriptive functions obtain, are not true or false statements but normative sentences or value judgments. The old purpose attributed to the phenomenology to distinguish between the phenomenological and the non-phenomenological is now absorbed by the philosophical grammar, not in the sense that a peculiar domain of propositions is attributed to grammar but in the new sense of distinguishing or c1arifying different functions oflanguage, "stable" and "fluid" forms of our modes of expression. According to the line of thought that continues up to On Certainty the complex of not-empirical, intentional contexts does not constitute a peculiar and independent ontological sfere. To Wittgenstein it is always possible that expressions of rules or norms, whose meanings do not depend on experience but on the system in which they occur, in determined contexts and for determined communicative purposes, achieve astability that makes them similar BETWEEN PHILOSOPHICAL GRAMMAR AND PSYCHOLOGY 179 in use to empirical propositions. Sometimes. the latter ones are used for communication purposes and are given the status of rules or prescriptions for actions, that is of propositions of grammatical rather than empirical nature. The fact that "the same proposition may get treated at one time as something to test by experience, at another as a rule oftesting"32 is underlined in another famous passage of On Certainty: It might be imagined that some propositions, ofthe form of empirical propositions, were hardened and functioned as channels for such empirical propositions as were not hardened but fluid; and that this relation altered with time, in that fluid propositions hardened, and hard ones became fluid. (OC, § 96) Analogically, the distinction between intentional and causal constructs is viewed not in dualistic terms since their distinction is not ontological but functional. (ii) According to Wittgenstein the entire conceptual tool set to work by mental sciences of his time to describe and explain the "real machinery of mind" and so to give a meaning to intentional contexts ceases to belong to the domain of experimental methods and is attributed to the system of rules goveming the use of propositions containing mental concepts. The treatment of psychological concepts shifted to the level of grammar, whose rules have a higher kind of generality than that attributed to empirical methods. Any attempt to describe, understand and explain a psychologie al phenomenon, an individual behaviour, amental process by means of empirical methods fails because when we believe to describe or explain a determined behaviour or process, what we actually have described or explained is the general context against which it emerges in its interplay with a number of other behaviours and processes. In a passage of the Remarks on the Philosophy 0/Psychology and Zettel Wittgenstein says: How could human behaviour be described? Surely only by sketching the actions of a variety of humans, as they are all mixed up together. What determines our judgment, our concepts and reactions, is not what one man is doing now, an individual action, but the whole hurly-burly ofhuman actions, the background against which we see any action. (RPP 11, § 629; Z, § 567) As we know, the rules of grammar do not apply to single words and sentences but to every word or sentence where "every" is not empirically determined.Analogically, mathematical and psychological concepts do not function as descriptions of singular states and processes but can be applied in a multiplicity of cases. Our concepts and conceptual schemes do not exercise absolute power to represent a single case but a whole network of cases, do not 180 ROSARIA EGIDI describe the behaviour of an individual but serve to insert it in a behavioral scheme shared by many individuals. Wittgenstein says that by means of one of his fascinating similitudes: Seeing life as a weave, this (pretence, say) is not always complete and is varied in a multiplicity ofways. But we, in our conceptual world, keep on seeing the same, recurring with variations. That is how our concepts take it. For concepts are not for use on a single occasion. And one pattern in the weave is interwoven with many others. (Z, §§ 568-569)33 In this sense the normative (rule-guided and aim-directed) character of our conceptual world immunizes Wittgenstein's idea ofinterdependence and variety of our concepts from subjectivism and relativism. What is the aim, one might ask, of Wittgenstein's provocative thesis on the grammatical character of concepts and propositions of mathematics and psychology and what is the meaning of his idea of generality which appeals to the higher rules of the game? At the first sight, his aim seems polemical, as it results from the severe evaluation in the weIl known passage of the Investigations: The confusion and barrenness of psychology is not to be explained by calling it a "young science"; its state is not comparable with that of physics, for instance, in its beginnings. (Rather with that of certain branches of mathematics. Set theory). For in psychology there are experimental methods and conceptual cotifusion. (As in the other case conceptual confusion and methods of proof). The existence of the experimental method makes us think we have the means of solving the problems which trouble us; though problem and method pass one another by. (PI 11, p. 232) The fact that the idea of the grammatical nature of psychological concepts was presented without a precise argumentation, but emerges from dialectical confrontation with the science of his time is not of secondary importance in the frame of Wittgenstein's thought. Polemic is rather a constitutive character of the Wittgensteinian philosophizing. It is linked with the therapeutical and not doctrinal aim of philosophy and indirectly also of the philosophical grammar. Philosophy and grammar are not systems of doctrines created in splendid isolation but activities obeying a system ofrules and developing in a communicative context whenever there are problems to solve, mi stakes to correct and misunderstandings to clarify. When Wittgenstein says that psychological propositions - and we know that it applies to mathematical ones as weIl - have grammatical nature underdetermined by experience and not govemed by explanation models of causal kind, he clearly alludes to the peculiarity of a language, that of immediate experience, whose aims and intentions are profoundly diverse from BETWEEN PHILOSOPHICAL GRAMMAR AND PSYCHOLOGY 181 physical language and experimental sciences. But in what way could grammar appear adequate, in Wittgenstein's gaze, to exemplify these aims? His appeal to grammar is more a programmatic manifest than the mention of a determined discipline. Moreover, Wittgenstein is not the first to present the relationship between grammar and psychology in a polemic function. This relationship was the focus of Austrian-German schools oflinguistics and philosophy of language between the eighteen and nineteen hundreds and was recalled in the Sprachphilosophien inspired by Brentano to juxtapose intentional and pragmatic use of language to tradition al expressive, semantical functions. 34 In contrast to Fregean logic and Carnap's syntax, Wittgenstein's grammar is not really a formal discipline; its propositions are not logically determined but obey a received system of rules; it does not follow inflexible criteria of truth or falsity of a particular proposition but is ready to submit to the need of multiple use and adjust to a variety of situations. Tbe so called psychological or intentional propositions express basically the rules of behaviour, codices, habits and institutions which as such refer to general, not singular cases. So they have a higher character than the empirical and inductive generality, as they depend on the standards established as rules or norms by the entire community. As a result of rule-guided and intentional (aim-directed) character as weH as of peculiar generality of grammar the psychological concepts do not aspire to be correct and precise as it is required for physical concepts: The pedigree of psychological concepts: I strive not after exactness, but after a synoptic view. The treatment of all these phenomena of mentallife is not of importance to me because I am keen on completeness. Rather because each one casts light on the correct treatment of all. (Z, §§ 464-465) The fact that philosophieal psychology is part of grammar sheds light on the peculiarity of the notion of meaning in the application to psychological concepts and propositions and show that it is controversial to connect to them the double aspect of meaning Frege called Sinn and Bedeutung. The meaning of propositions in which psychological concepts occur not only cannot be determined by the means of truth functional logic but is also independent of any reference to mental objects, to internal psychical processes. Wittgenstein refuted the idea of objective contents of thinking as refercnce of psychological or intentional propositions, of Gedankeninhalte of the type Frege called "senses" (Sinne). For Wittgenstein Sinn is not in the mind, but it is an eminently intentional, rule-guided, aim-directed notion. Probably he 182 ROSARIA EGIDI saw in the Fregean objective notion of Sinn a hidden tendency to psychologism. For this reason he did not accept it and rejected, at the same time, the idea of a phenomenology as a temptation to acknowledge an extra-linguistic meaning to senses. From the Philosophical Grammar to the last writings on philosophical psychology Wittgenstein's aim is to reformulate the problem ofthe sense of a proposition and, in general, the problem of understanding and of intentionality within the context of linguistic uses and conventions. The Wittgensteinian appeal to grammar is first of all polemical to Frege's strategy to give to the senses an ontological status and to Russell's thesis that propositional attitudes have the structure and the meaning of "external relations".35 The task of grammar is not to express thoughts and to determine the meaning they carry. The not-extensional propositions do not express internal experiences, mental activities or processes. So the Bedeutung is lost. Grammar aims at the reconstruction of rules which allow for accomodating of our linguistic means to the aim of communication and understanding. And so the Fregean Sinn is lost as well. The meaning is not a ready made notion but it is the result of deliberate activity interconnected with aims, thoughts, feelings and desires which constitute our form oflife. The meaning is not "found" but "given", because it belongs to the intentional, pragmatic element without which language cannot work. This is the way, according to Wittgenstein, through which we can in some way maintain the objective character of our concepts, the inner rationality of our conceptual world. Wittgenstein espresses this in the following way: Our investigation does not try to find the real, exact meaning of words; though we do often give words exact meanings in the course of our investigation. (Z, § 467) Terza Universita di Roma NOTES I Philosophische Untersuchungen. Philosophicallnvestigations (PI), ed. by G. E. M. Anscombe & R. Rhees. Trans. by G. E. M.Anscombe, B. Biackwell, Oxford 1953. 2 Bemerkungen über die Grundlagen der Mathematik. Remarks on the Foundations of Mathematics, ed. by G. H. von Wright, R. Rhees & G. E. M. Anscombe. Trans. by G. E. M. Anscombe, B. Biackwell, Oxford 1956. 3 (RPP I), ed. by G. E. M. Anscombe & G. H. von Wright. Trans. by G. E. M. Anscombe, B. Blackwell, Oxford 1980; (RPPII),ed. by G. H. von Wright& H. Nyman. Trans. by C. G. Luckhardt & M. A. E. Aue, B. Biackwell, Oxford 1980. 4 (LWPP I), ed. by G. H. von Wright & H. Nyman. Trans. by C. G. Luckhardt & M. A. E. Aue, B. Blackwell, Oxford 1982. BETWEEN PHlLOSOPHlCAL GRAMMAR AND PSYCHOLOGY 183 (Z), ed. by G. E. M.Anscombe & G. H. von Wright. Trans. by G. E. M.Anscombe, B. Blackwell, Oxford 1967. 6 Philosophische Grammatik. Philosophica/ Grammar (PG), ed. by R. Rhees, B. Blackwell, Oxford 1969, p. 214. Trans. by A. J. P. Kenny, B. Blackwell, Oxford 1974. 7 Philosophische Bemerkungen. Philosophica/ Remarlrs (PR), ed. by R. Rhees, B. Blackwell, Oxford 1964, §§ 1-9,57-66,218-224. Trans. byR. Hargreaves & R. White, B. Blackwell, Oxford 1975. 8 PR, §§ 11,19,53,216. 9 Tractatus Logico-Philosophicus (TLP), ed. by D. F. Pears & B. F. McGuinness, Routledge & Kegan Paul, London 1961,5.541-5.542. 10 §§ I, 53, 57,60. In the Philosophische Bemerkungen Wittgenstein assimilates phenomenology sometimes with "grammar" (§ 4), sometimes with "theory ofknowledge" or Erkenntnistheorie (§§ 57,60). po 11 Bemerkungen über die Farben. Remarlrs on C%ur, ed. by G. E. M.Anscombe. Trans. by L. L. McAlister & M. Schättle, B. Blackwell, Oxford 1977, p. 16. 12 Wittgenstein's transitional works include a considerable quantity ofwritings composed betweeen 1929--1935. The first ofthem is the conference 'Some rernarks on logical form' ,Proceedings olthe Aristotelian Society, Suppl. Vol. 9, 1929, pp. 167-71; the manuscripts later published as PR and PG; the texts of his Cambridge lectures: Wittgenstein 's /ectures. Cambridge 1930-32. From the Notes of J. King & D. Lee, ed. by D. Lee, B. B1ackwell, Oxford 1980; Wittgenstein 's /ectures. Cambridge 1932-35. From the Notes of A. Ambrose & M. Macdonald, ed. by A. Ambrose, B. B1ackwell, Oxford 1979; The B/ue and the Brown Boolrs (BBB), ed. and with aPreface by R. Rhees, B. Blackwell, Oxford 1958. \3 PR, § 1; 51-53; 57-61; 68-69; 75. 14 PR, § 1. See also PG, §§ 45; 82-84 and PI I, § 90. 15 PR, 10-19; 28; 39-46 and PG, §§ 45; 69-84. 16 See in particular G. E. Moore,Wittgenstein 's Lectures in 1930-1933, in Philosophica/ Papers, G. Allen & Unwin, London 1959, pp. 252-324 and Wittgenstein und der Wiener Kreis (WWK). Aus dem Nachlaß hrsg. von B. F. McGuinness, B. Blackwell, Oxford 1967. 17 PR, §§ 218-24. See also WWK, pp. 63-67. 18 See in particularthe arguments presented in the 'Blue Book', pp. 44-74. 19 See R. Camap'sMeaning and Necessity, Harvard University Press, Chicago 1947, chap. 1-11. 20 See in particular §§ 1-44,85-108. 21 See PI I, §§ 611-629; 11, xi (see also parallel arguments in LWPP I, §§ 183,848,893,905--913); RPP I, §§ 705-715, 830-851; RPP 11, §§ 116,222,258,267- 275,624-631. 22 PI I, § 337. 23 lbid., § 190. 24 PR, § 23. 25 See Über Gewißheit. On Certainty (OC), ed. by G. E. M. Anscombe & G. H. von Wright. Trans. by D. Paul & G. E. M. Anscombe, B. Blackwell, Oxford 1969, § 94. 26 PI I, §§ 629-647. See also 188-208 27 See also PI I, §§ 624-28; 650; 11, p. 224. 28 OC, § 494. 29 lbid., § 385. 30 See PI 11, pp. 224-225 (LWPP I, § 907-9087; RPP I, §§ 631,633). 31 See in particular PI I, § 220, where Wittgenstein underlines "a difference between being causally determined and being logically determined". 5 184 32 ROSARIA EGIDI oe, § 98. See also PI 11, p. 174 and RPP 11, § 624. See for example Anton Martys theory of the intentional origin ofthe language (der absichtlichen Sprachbildung), first outlined in his 1875 Über den Ursprung der Sprache (Minerva GMBH, Frankfurt) und later developed in die Untersuchungen zur Grundlegung der allgemeinen Grammatik und Sprachphilosophie"M. Niemeyer, Halle 1908. Ｓｾ＠ See the alternative proposal of "internal relations" to Russell's conception in PR, §§ 20-30. 33 34 ROBERTO CASATI NOTES ON PHENOMENOLOGY AND VISUAL SPACE The Philosophical Remarks 1 belong to those intermediary works of a philosopher, which seem to be fatally destined to remain unknown due to their experimental character and a sketchiness which lacks a precise aim. On reading Wittgenstein, one often retains the impression of an immense effort oriented towards details and nuances, as well as of possible hidden consequences of an argument or a hypothesis. Nonetheless, critical literature directed towards the most important works, such as the Tractatus and the Investigations, has aptly shown the structural force ofthe former and the coherence and thematic unity of the latter. Even aseries of short amorphous notes such as the Bemerkungen über die Farben, has shown proof of an impressive originality; the edition of this work was without doubt an important cause for a renewed interest in the c1assical theme of colour in contemporary philosophy. None of this can be said of the Remarks and with good reason, or so I believe, for this is a work we may term experimental, in which it is difficult to find a thematic line pursued till the end. I do not believe I am proposing anything new by substituting the chronological classification ofWittgenstein's works with a thematic division. Roughly speaking, this is a division between two fundamental foci: language and psychology of mental states respectively. The Tractatus and a good part of the Investigations center around the first focus, whereas the Philosophical Remarks and the Remarks on Colour center on the second one. Having said that, I shall analyse in some detail one theme amongst those of the Philosophical Remarks which belong to the second focus, which has not been extensively treated, though its originality deserves some interest.2 1. My topic is the problem of the visual field, treated in §§ 205-217. This group of remarks consists of an argument designed to show that visual space is provided with absolute coordinates and a similar argument in favour ofthe demonstration of the existence of an absolute distance in the visual field. The second argument is directly linked to the discussion of the problem of quantities and measures, which is subject to a long series of remarks in the book and is mostly treated in theRemarks on the Foundations ofMathematics. Given the fact that there exists a literature specifically dedicated to the 185 R. Egidi (ed.), Wittgenstein: Mind and Language, 185-192. © 1995 Kluwer Academic Publishers. 186 ROBERTO CASATI Wittgensteinian treatment of measures, my central question is therefore the first one, i.e. whether the visual field can be considered an absolute space. This question is introduced during the development of another problem, namely the complexity or the simplicity of the patch in the visual field. Can we say that a patch is simpler than a larger one? (§ 205) Indeed, the problem is not so much to know whether one can have simplicity in the visual field, but rather whether one can have composition: whether the smaller of the two patches can be considered apart of the bigger one, in other words, whether there is composition in the visual field. Wittgenstein has an argument for giving a negative answer to this question, even though the answer which first comes to mind seems to be exact1y the opposite. His argument is quite direct: It seems as if it is impossible to see a uniformly coloured patch as composite, unless you imagine it as not uniformly coloured. The image of a dissecting line gives the patch more than one colour, since the dissecting line must have a different colour from the rest ofthe patch. (§ 205) At this point some of us will probably be tempted to object that Wittgenstein mixes considerations about what there is in the world with considerations about what we can represent of the world to ourse1ves. Even if we cannot represent to ourselves the parts of a patch, it would be presumptuous to conclude that the patch is not composed of parts. If it is not, we need a better argument. This objection allows us a better evaluation of what is at stake for Wittgenstein. What interests Wittgenstein is the patch as seen, the visual patch. He suggests that the logic of the operations which one can perform on this object differs radically from that which govems the operations which can be performed on material objects. Thus, according to Wittgenstein, this object - the patch - belongs to a completely different category. Wittgenstein opposes the case of the non-composition of a visual patch to the case of the composition of complex configurations, such as a chessboard. Contrary to the previous case, here we have to account for composition. The figural unity is not a primitive element, it is inseparable from the composition of the figure. An intellect which takes in the component parts and theirrelations, but not the whole, is a nonsense notion [ein Unding]. (§ 205) Once more I use a proposition which at first sight is quite surprising, in order to better isolate the central thesis. Every philosopher knows about the existence of the phenomenon of composed colours, a phenomenon which is striking when we look at pointillist paintings: that which from afar appears NOTES ON PHENOMENOLOGY AND VISUAL SPACE 187 like a uniform orange patch turns out to be a chaotic heap of small red and yellow points. It is therefore possible to perceive only the component parts of a composition. Hence the intellect which apprehends the whole without apprehending the parts is not an absurdity, it is actual. Nonetheless, we must maintain once more that the simplest and most coherent interpretation of this remark is the one which considers visual objects as the only referents of the Wittgensteinian discourse. A visual orange patch cannot be composed of small red and yellow visual patches, as the latter are not visible: they are not visual patches themselves. 2. These preliminary remarks introduce our subject: the visual field as absolute space. Let us see the three arguments Wittgenstein gives for the absoluteness ofposition in the visual space. Ifthe visual patch has no visual parts, a fortiori one cannot predicate a colour of a part of the patch. Nevertheless, we would say that the corner at the bottom left of the patch is the same colour as the corner at the top right. What is it that permits us to make the predication implicit in this identity statement? A simple solution would be to accept the existence of absolute places in visual space, that is to say, to accept that we can individuate places like the top right of the visual field without directly or indirectly referring to visual objects occupying that place. Previously, we were talking about the possibility of distinguishing one part of the patch in relation to other parts, this being a necessary condition for a predication concerning the patch itself. We were therefore obliged to use a relational style of individuation, without which no predication would have been possible. It is only now that we avoid relational individuation and the problems which are linked to it by maintaining that the individuation of a place in visual space is absolute, that is to say, that it does not depend on the individuation of other places. I shall call this argument (a) for the absolute properties of visual space. A second argument (b) is an argument from movement. Wittgenstein says: It's obviously possible to establish the identity of a position in the visual field, since we would otherwise be unable to distinguish whether a patch always stays in the same place or whether it changes its place. (§ 206) This argument is insufficient, unless we make a restrietion, for we can grasp the movement as long as it concerns a movement in relation to another object. The intervening restrietion thus concerns the presence of a single patch in the field. If one patch only3 changes in place, we can see it, even if there is nothing else to see. Therefore we retain the possibility of individuating 188 ROBERTO CASATI plaees independently of the presenee of objeets whieh either oeeupy them or oeeupy other places. We eould have a better grasp of the diffieulty if we were to imagine the visual field as equipped with an asymmetrie form. In this ease the determination of a plaee eould be obtained by naming eertain sides of the field with proper names, and by indieating towards whieh side the patch moves. Now, the visual field is de facta symmetrie in relation to two axes and this renders such a method inapplicable. Thus we must provisionally accept the possibility of absolute position in the field. (e) We cannot imagine a visual space in which we would only perceive certain spatial positions but no absolute position. (§ 206) Here a good example is that ofthe watch the dial ofwhich shows only dots without numbers. If it were impossible for us to distinguish a dot in relation to its successor, it would be equally impossible to notice the movement ofthe pointer. I think the three arguments are valid. But the fuller consequences drawn by Wittgenstein - which I believe to be false - are worthy of discussion. The central remark is the following: since visual space is an absolute space, [w]e can also say visual space is an oriented space, aspace in which there is an above and below and a right and left. And this above and below, right and left have nothing to do with gravity or right and left hands. It would, e.g., still retain its sense even ifwe spent our whole lives gazing at the stars through a telescope. (§ 206) In this dense passage there are three points to discuss. Wittgenstein suggests (1) that to say ofvisual space that it is absolute is the same as to say that it possesses a system of absolute coordinates. (2) That these coordinates are associated with the directions of the field of bodily orientation: left, right, top, and bottom. (3) This association is however said to be independent of bodily orientation directions. I believe that none of these three points is correct in itself, that each requires specification, as I shall now try to show. First, the present state of the discussion: owing to (a), (b) and (c) we have been forced to admit the existence of absolute position in visual space. Is it possible to say that this is equivalent to admitting the existence of a system of coordinates? My answer is no. There is no doubt an equivalence between the capacities or the behaviour of two beings, one ofwhich disposes ofa system ofreference in which pI aces are individuated separately, whilst the other disposes of a system in which there are basic individuals, such as the point on the two axes ofreferences, as weIl NOTES ON PHENOMENOLOGY AND VISUAL SPACE 189 as individuals which are individuated by means of a reference to basic individuals. It matters little whether 1 give you a map with numbers spread in all places (used as proper names ofthe places) or a map with an abscissa and an ordinate. Nevertheless an operational equivalence does not imply anything interesting on the ontological level. In particular, it does not imply anything conceming the existence of a system of individuation of pI aces by means of coordinates in the visual field. 1 must here emphasize that 1 am not talking about the factual presence of a visible abscissa and ordinate in the visual field. This is not a fact and, moreover, it concems a possibility which is envisaged and criticised by Wittgenstein himself at § 206: But even that isn't an accurate representation, since ifwe really saw such a set ofaxes ofcoordinates [... ], we would in fact be in a position, not only to establish the orientations of objects relative to these axes, but also the position ofthe cross itselfin the space, as though in relation to an unseen co-ordinate system contained in the essence ofthis space. What 1 am talking about is thus rather that orientation which is contained 'in an invisible way' in the very essence of space. And what 1 object to is that the existence of such an orientation should be implied by the existence of absolute pI aces in the visual field. My argument is very simple: As far as individuation is concemed, the existence of any system of coordinates is operatively equivalent to the existence of absolute places in vi'iual space. Thus even if it were possible to positively infer the existence of a system of coordinates by basing ourselves upon the existence of absolute places, it would not yet be possible to infer the existence of this system of coordinates to which Wittgenstein refers. But this goes for any system of coordinates. Therefore we cannot infer the existence of any particular system of coordinates. Curiously, Wittgenstein appears to grasp this difficulty when he says that "this above and below, right and left have nothing to do with gravity or right and left hands" (§ 206). This is quite puzzling, but either Wittgenstein wants to say that this system is arbitrary, and therefore there is no reason to use specific proper names for these directions - and names such that they resemble the names for left and right; or he intends to keep the analogy with the directions of the orientation field; and in both cases the previous argument applies. Indeed, Wittgenstein seems to choose the second interpretation, which is to consider visual coordinates as being somehow analogous to orientation coordinates. His example is: IfI hold a book the wrong way, I can't read the print at all, or only with difficulty. (206) 190 ROBERTO CASATI This takes us to the discussion of point (2): I should like to say that the visual field is oriented according to our Bewegungsgejühlsraum. If this is true, the association of such coordinates with the structure of the visual field (although extrinsic, as the previous argument shows) is strongly conditioned by the fact that the visual field belongs to a being which has a body and a capacity of movement similar to ours. Let us think about what happens when we look at a wall in front of us as we slightly balance our head from left to right. If there were a line on the wall and if this line could be considered a visual line, we would say that it has kept its orientation, even if the visual field has changed position in relation to the orientation of the line. By this I mean that the geometrical axes of the field turn with the balancing of our head, but the axes of orientation are not modified. Hence, factual orientation ofthe visual field is (at least in one case) linked to that of the orientation field. In other words, a change in position of the visual field does not always entail a change of orientation of the field in relation to the environment, but it does entails a change of the positions of the different parts of the field relative to the orientation of the field itself. Therefore, it entails the impossibility of linking the absolute position of the parts to orientation, because orientation remains, even if the absolute position changes. Nevertheless, we have the possibility of making sense of the idea that the visual field has coordinates which move in concert with the movements of the field. This possibility is rendered effective by associating the system of coordinates ofthe visual field with the system of head-centered coordinates. 4 If we accept a pictorial version of this association, we could consider "visual right" as synonymous with "the parts of the visual field closer to the right ear", "visual top" as synonymous with "the parts ofthe visual field closer to the forehead", etc. In my opinion, the visual field can be given this other oriented structure, which is mainly head-oriented, but this depends on the fact that this visual field belongs to an animal which has a head with such and such capacities of movement. The orientation does not result from the fact that there are absolute positions (as I tried to show against (1)): and therefore it has to base itself (so to speak) on another system of orientation. Hence orientation is not independent of this other system, contrary to the point Wittgenstein tries to make in stating (3). Consequently, the last paragraph of the remark 206 presents an unsound argument: NOTES ON PHENOMENOLOGY AND VISUAL SPACE 191 I can obviously see the sign V at one time as v, another as an A, as a 'greater than' or 'less than' sign, even if I were to see it through a telescope and cannot compare its position with the position of my body. Perhaps someone might reply that I feel the position of my body without seeing it. But position in feeling space s (as I'd like to put it for once) has nothing to do with position in visual space, the two are independent of each other, and unless there were absolute direction in visual space, you couldn't correlate direction in feeling space with it at all. The false premise of this argument is the last one, which pretends to establish that the fact that the directions in visual space are not absolute would hinder us in correlating them with the directions of the orientation space. We have indeed no need of an absolute direction in visual space: it is sufficient that there are absolute positions in order to construct an orientation of the field. 6 3. This conc1udes my examination of Wittgenstein's argument for the absoluteness of some properties of visual space. Let me briefly comment upon Wittgenstein's attitude towards this kind ofproblems. We know that one of the reasons which motivated Wittgenstein to go to Cambridge to continue his philosophieal work in 1929, was precisely his interest for the visual field. As he writes in a letter to Schlick: 7 Ich habe mich entschlossen ein Paar term hier in Cambridge zu bleiben und den Gesichtsraum und andere Dinge zu bearbeiten. Wittgenstein's interest concerns, I believe, the discussions on the phenomenallanguages envisaged in several texts by Russell and the logieal positivists. Wittgenstein's phenomenological approach thus belongs to areaction to this type of approach to the problem of perception and perceptual description; and not exclusively to a critical return to the metaphysical problems left open in the Tractatus. I think that the meticulousness of Wittgenstein's argument on this point indicates a profound effort of reflection whieh seems to me exceedingly difficult to square with the received historiographieal schema. In the 1930's Wittgenstein is not only the philosopher who extends the horizons of his philosophy of language by refusing the monolithic model of the Tractatus in favour ofthe multiplicity ofthe language games; he is moreover interested in phenomenological problems in themselves. CNRS Aix-en-Provence 192 ROBERTO CASATI NOTES I Philosophische Bemerkungen. Philosophical Remarks (PR), ed. R. Rhees, B. Blackwell, Oxford 1964. Trans. by R. Hargreaves & R. White, B. Blackwell, Oxford 1975. 2 Tbe two short commentaries dedicated to the Philosophische Bemerkungen do not treat the topic I am concerned with here. N. Malcolm gave a survey ofthe fundamental themes ofthe Bemerkungen in 'Wittgenstein 's Philosophische Bemerkungen' ,Philosophical Review, 76, 1976, pp. 220-229 by underlining above all the connections with other works by Wittgenstein. He does not mention Wittgenstein' sinterest in visual space. H. Spiegelberg also commented on the Bemerkungen: 'The Puzzle ofLudwig Wittgenstein's Phänomenologie', American Philosoph ical Quarlerly, 5,1968, pp. 224-256, but from a metaphilosophical point ofview. These two articles have been re-edited in Ludvig Wittgenstein: Critical Assessments, ed. by S. Shanker, Croom Helm, London 1986, Vol. I, pp. 213-221 and 222-240. 3 For this reason the following paragraph, concerning the two stars which pivot around each other is irrelevant, since it concerns relative movements. 4 A possibility envisaged by Husserl. See Ding und Raum. Vorlesungen 1907, ed. by U. Claesges, M. Nijhoff, The Hague, 1973, Abh.B. 'Stufen der Raumkonstitution', pp. 309-321. 5 Gefohlsraum. It is precisely that which in the context of the present essay I called orientation space. 6 Perhaps one ofthe reasons which led to Wittgenstein 's error was his use ofthe example conceming the letters ofthe alphabet. In my opinion it is difficultto do withouta component ofthe Gefohlsraum in the perception ofthe letters. 7 18.2.1929, quoted in the introduction by Brian McGuinness to vol. III of Schriften, Suhrkamp, Frankfurt 1967, p. 17. PART IV BELIEF, DOUBT, AND RELATIVISM NORMAN MALCOLM DISENTANGLING MOORE'S PARADOX In the two volurnes of Remarks on the Philosophy of Psychologyl and in the first volurne of Last Writings on the Philosophy of Psychology 2 Wittgenstein wrote approxirnately one hundred and thirty rernarks on the verb 'to believe' . This concentration on a single verb was chiefly provoked by what Wittgenstein called 'Moore's Paradox'. What is Moore's Paradox? G. E. Moore noticed that such a sentence as, 'I went to the pictures last Tuesday, but I don't believe that I did, '3 would be an absurd thing for anyone, yourself for instance, to say. Moore rernarked that this is a perfect1y absurd thing to say, although what is asserted is something which is perfectly possible logically: it is perfectly possible that you did go to the pictures and yet you do not believe that you did ...4 In a later essay Moore referred again to this 'absurdity'. Taking the exarnple of sorneone's saying 'I believe he has gone out, but he has not', Moore's view continued to be that this is an absurd thing to say, but that what is said is 'logically possible'. As he put it, 'This, though absurd, is not self-contradictory; for it rnay quite well be true. '5 The central idea in Moore' s view of the Paradox is that such a sentence as 'It is raining and I don't believe it is raining', is an intelligible sentence. A nurnber ofphilosophers have taken the same position. Like Moore, they have thought that the sentence itself is intelligible; it rnakes sense; we understand it; it expresses a logical possibility; it could even be true. What is 'absurd', or 'strange', or 'logically odd', is to say it. I think this position is plainly wrong. If I asked someone what the weather is outside, he rnight reply 'It's raining', or 'I believe it's raining', or 'I believe it's not raining'. I would understand these replies. But ifhe said, 'It's raining and I don't believe it is', then I would not understand the sentence. I would think that I had rnisheard hirn. If he insisted on repeating the same words, and if he seerned to be serious, I would be stunned. Certainly I could not take his sentence as areport about the weather; nor as information about anything. I rnight react to it as to a piece of deliberate nonsense, a grarnrnatical joke. Why should a philosopher think that this cornbination ofwords is an intelligible sentence? There are a nurnber of considerations that I shall take up one at a time. 195 R. Egidi (ed.), Wittgenstein: Mind and Language, 195-205. © 1995 Kluwer Academic Publishers. 196 NORMAN MALCOLM 'ft fsn't Self-Contradictory'. Moore was obviously influenced by the idea that a sentence ofthe form 'p, and 1 don't believe p', is not self-contradictory. It is true that it doesn't look self-contradictory. But perhaps this appearance is deceptive. Before we can be sure that such a sentence is not self-contradietory, we must reflect on the ordinary meaning ofthe expressions 'I believe p' and 'I don't believe p'. 1 will return to this point. The Past Tense. The sentences that figure in the Paradox are sentences in whieh 'believe' occurs in the first person present tense. Examples are, 'It's raining and 1 don't believe it is', and 'I believe it's raining and it isn't raining'. The correspondingpast tense does make sense: e.g., 'I believed it was raining and it wasn't raining'. The fact that the past tense makes sense, inclines one to think that the present tense must make sense. One could reply that there is no justification for this 'must'. But far more illuminating is Wittgenstein's observation about the way in whieh the present tense, 'I believe p', is actually used. What he says is: 'I believe p' means roughly the same as 'p'. (RPP I, § 472) He go es on to say: 'I believe it' s going to rain' means something like: 'It' s going to rain' , but 'I believed then it was going to rain' doesn't mean anything like 'It rained then'. (RPP I, § 473) Wittgenstein asks the following question: What does it mean to say that 'I believe p' says roughly the same as 'p'? We react in roughly the same way when someone says the first and when he says the second; ifI said the first, and someone didn't understand the words 'I believe', I would repeat the sentence in the second form, and so on. As I would also explain the words 'I wish you would go away' by the words 'Go away'.6 Wittgenstein adds the following important explanation of why 'I believe p' can have approximately the same meaning as 'p'. He points out a significant logieal difference between the first person and the third person of 'believe': That he believes such-and-such, we gather from observation of his person, but he does not make the statement 'I believe .. .' on the basis ofself-observation. And that is why 'I believe p' can be equivalent to the assertion of 'p'. (RPP I, § 504) I will return to this remark later. Right now 1 will say that Wittgenstein's statement that 'I believe p' means roughly the same as 'p', is an accurate observation of the use of 'I believe'. Most of the confusion about Moore's Paradox is due to the failure to notice this fact. On the surface it seems that DISENTANGLING MOORE'S PARADOX 197 'p' and 'I believe p' have entirely different 'content'. But if we go beneath the surface and remind ourse1ves ofthe way we actually use the expression 'I believe p' in daily speech, we see that what it commonly amounts to is a tentative or hesitant assertion of 'p' . Suppose 1 am indoors preparing to go out. 1 wonder whether 1 should take an umbrella; so 1 ask a person near the window to look out to see if it' s raining. It is dark outside and hard to see through the window-glass. The person says, 'I believe it's raining'. I would take this as a cautious assertion that it's raining, and would fetch my umbrella. My reaction would be roughly the same as it would be ifthe person had simply said, 'It's raining'. Wittgenstein later puts this point in a slightly different way: Consider: the words 'I believe it is raining' and 'It may be raining' (Es dürfte regnen) can say the same: inasmuch as in certain contexts it makes no difference which ofthe two sentences we use ... That is: in order to say that perhaps it is raining we do not need the concept 'believe', although we can use it for that purpose. 7 Suppose there were a characteristic tone of voice in which we said things of which we were not entirely certain. 'I believe .. .' is used to play the role of such a tone of voice. We have other linguistic devices that do the same. For example, 'It may be raining', or 'Perhaps it is raining'. Of course these sentences are also used in other ways. But they can be used to make the cautious assertion that it's raining. 'I believe it's raining' is typically used to do the same thing. As Wittgenstein says 'I believe' does not have to be brought in to do that job. But it can be, and often is, used for that purpose. A statement ofthe form 'I believe p', is not always a hesitant assertion of 'p'. There are contexts in which it is an emphatic assertion of 'p'. Suppose that in a certain town there have been rumours that the Mayor is engaged in fraudulent practices. A supporter of the Mayor, speaking before an audience, dec1ares 'No one here believes that the Mayor of our city is dishonest' . Someone shouts from the audience: 'I believe he is dishonest' . In that context this would be a forceful assertion that the Mayor is dishonest, not a hesitant one. The fact that 'I believe p' is commonly an assertion of 'p', destroys what might be called 'the argument from the past tense'. This argument assurnes that 'I believe p' and 'I believed p', differ only in tense. This is an incorrect assumption. 'I believe p' is normally used to assert 'p', sometimes hesitantly, sometimes emphatically. But the past tense, 'I believed p' is never an assertion of 'p'. Moore's Paradox exists only in the present tense. The Supposition. When Moore decIared that a sentence ofthe form 'p, and 198 NORMAN MALCOLM I don't believe p' or 'I believe p, and p is not true', expresses 'a logical possibility' and 'may quite weIl be true', could be have confused an assertion with a supposition? For the corresponding supposition does make sense. There is no absurdity involved in saying, 'Let us suppose that I believe it's raining and in fact it isn't raining'. What this supposition proposes is 'logically possible'. In this connection Wittgenstein makes the following remark: Moore's Paradox may be expressed like this: 'I believe p' says roughly the same as 'p'; but 'Suppose 1 believe p .. .' does not say roughly the same as 'Suppose p .. .'.8 In saying that Moore's Paradox "may be expressed" in this way, Wittgenstein does not, of course, mean that Moore expressed it in this way. What Wittgenstein might mean by this curious remark is that if it were not for this logical difference between 'I believe p' and 'Suppose I believe p', then the sentences that Moore spotted would not be nonsensical. Another thing that Wittgenstein's remark may suggest is that this difference between the assertion and the corresponding supposition could be called 'paradoxical', not because there is anything logically dubious about it, but because it is surprising: it goes against our expectations. For in other cases the relation between an assertion and the corresponding supposition is that what is asserted and what is supposed is the same. The assertion 'The ship sails at dawn', is the assertion ofwhat is supposed by 'Suppose the ship sails at dawn'. The assertion says neither more nor less than what is supposed in the supposition. But this is not so in the case of 'I believe'. This is why Wittgenstein says: So it looks as ifthe assertion '1 believe ... ' were not the assertion ofwhat the supposition '1 believe' supposes. (RPP 1, § 493) To suppose 'I believe it's raining' is not to suppose 'It's raining'. But to assert 'I believe it's raining' is to assert 'It's raining'. The assertion asserts something that is not supposed by the supposition. The assertion does not exactly match the supposition; it brings in another element. But would it be right to say that 'I believe it's raining' is not the assertion that corresponds to the supposition 'Suppose I believe it's raining'? Surely not. What other assertion would be the corresponding one? There is no answer. We must simply accept this peculiar feature of the logical grammar of 'I believe'. In this connection Wittgenstein makes a penetrating remark: With the words 'Suppose 1believe that' you already presuppose the whole grammar ofthe word 'believe'. We are not supposing something that, so to speak, is given to you unambiguously through a picture, so that you can tack on to this supposition something other than the ordinary DISENTANGLING MOORE'S PARADOX 199 assertion. You would not know at all what you were supposing here, ifyou were not already familiar with the use of 'believe' .9 Ifyou substituted some assertion other than 'I believe p', as the assertion that 'correspond' to 'Suppose 1 believe p', then you would not understand what you were supposing. We have seen already that the pas! tense of 'believe' presents the same 'paradox' as does the supposition. The present tense, 'I believe p', is an assertion of 'p'; the past tense, 'I believed p', is not an assertion of 'p'. So it looks as if 'I believed it was raining' is not the 'correct' past tense of 'I believe it is raining'. But is there a 'more correct' match? Certainly not. We can apply here the same point that Wittgenstein made about the supposition. The assertion 'I believed' presupposes the whole gramm ar of 'believe', including that of the first person present. Otherwise you would not know what you were asserting in saying 'I believed'. Is 'I Believe P' a Description 0/ aMental State? We come now to the main source ofperplexity about the absurd formulation, 'p, and I don't believe it', or 'I believe p, and pis not true'. On the one hand these conjunctions strike us as nonsensical. On the other hand, it seems that they should not be nonsensical: for the two parts of such a conjunction appear to be dealing with separate subject matters; they appear to be logically independent of one another. Their conjunction in a single sentence should not produce any logical conflict. As regards the conjunction 'I believe it's raining, and it isn't raining', one wants to say that the first part is about amental state, while the second part is about the weather. So why shouldn't 1 be able to put together, without absurdity or inconsistency, areport of amental state of mine, and a report of the weather? Let us agree that when 1 say 'It isn't raining', 1 am speaking about the weather. The problem lies in the assumption that when 1 say 'I believe it's raining', 1 am reporting or describing amental state of mine. We have already seen that 'I believe it's raining' is a cautious assertion of 'It's raining'. Now someone might agree with that, but hold that 'I believe it's raining' also reports or describes amental state or attitude ofthe speaker. Wittgenstein imagines someone putting this view as folIows: At bottom, with these words I describe my own mental state - but here this description is indirectly an assertion ofthe state of affairs that is believed. lO At first sight this appears to be a very plausible view. It amounts to saying that the statement 'I believe it's raining', does two jobs at once. It is directly 200 NORMAN MALCOLM areport or description of amental state, and indirectly an assertion about the weather. This view would provide an explanation of why the whole conjunction is absurd: for the 'indirect' assertion of 'It's raining', carried by the first member of the conjunction, 'I believe it's raining', contradicts the explicit assertion by the second member of the conjunction, 'It isn't raining'. The interpretation is attractive: for it attributes the absurdity to a logical contradiction; yet it also preserves the inclination to think that 'I believe p' describes or reports amental state. But a pressing question arises. Why should what is primarily, or 'directly', a description or report of amental state, be at the same time an 'indirect' assertion about the weather? What connection is there between a description ofa mental state and a description ofthe weather? Why should 'I believe it's raining', construed as a description of amental state, be an indirect assertion of 'It's raining' - rather than an indirect assertion of 'Today is Tuesday', or of 'Turkey wants to join the Common Market', or of anything whatever? 1 think it is this problem that leads Wittgenstein to suggest that the interpretation in question is treating the sentence 'I believe it's raining' as if it were a photograph. Commenting on the idea that 'I believe it's raining' is an 'indirect' assertion of 'It's raining', he says: As in certain circumstances I describe a photograph, in order to describe what it is a photograph of. 1I But if one does this one is assuming that the photograph is reliable. Wittgenstein turns this point into a profound criticism ofthe view at issue. He says: Hut ifthis analogy holds good, then 1should have to be able to say further, that this photograph (the impression on my mind) is trustworthy. So I should have to be able to say: 'I believe that it's raining, and my beliefis trustworthy, so I trust it'.As ifmy beliefwere a kind ofsense-impression. 12 We are sometimes aware of having illusory sense-impressions. It looks as if the railroad platform is moving, but we know it is only our train that is moving. So often we trust our sense-impressions, but sometimes not. Now 1 can trust or distrust another person's belief. But can 1 do this with my own belief? Can 1 firmly hold a belief, but at the same time distrust it? Wittgenstein deli vers this smashing remark: One can mistrust one's own senses, but not one's own belief. (LWPP I, § 419) I think that Wittgenstein has completely destroyed the conception that 'I believe p' is, first of an, adescription or report of amental state, and secondly, an assertion of 'p'. For why does this mental state refer to 'p' in particular? How DISENTANGLING MOORE'S PARADOX 201 does it pick out 'p' from all ofthe things that might be asserted? Presumably, beeause this mental state is or eontains a 'pieture' or a 'representation' of 'p'. If lasserted 'p' I would be assuming that my mental state (my belief) is a reliable pieture, a trustworthy representation. This in turn implies that I ean think of a present belief of mine as being reliable or unreliable - which is nonsensical. Wittgenstein has undermined the notion that the sentence 'I believe p' is 'primarily' or 'direetly' a deseription of amental state. The chief souree of our confusion about a sentence such as 'I believe it's raining and it isn't raining', is our inelination to think that the first part is an assertion about what is within me, while the second part is an assertion about what is outside me. It seems that there ean be no eonflict here; yet we feel that the whole sentence is absurd. Wittgenstein fixes his attention on this image of the contrast between something within me and something outside me. He says: 'It is raining and I believe it is raining'. Turning to the weather I say that it is raining: then, tuming to myself, I say that I believe it. - But what do I do when I turn to myself, what do I observe? Suppose I say: 'It's raining, and I believe it will soon stop' -do I turn to myself atthe second part ofthe statement? - Of course, ifI wantto find out whether he believes that, then I must turn to him, observe hirn. And ifI wanted to find out by observation what I believe, I should have to observe my actions,just as in the other case I have to observe his. Now why don 't lobserve them? Don 't they interest me? Apparently not. I hardly ever ask another person, who has been observing me, whether he has the impression that I believe such and such: that is, in order in this way to make inferences to my future actions... 13 From the actions and utteranees of other persons I continually make inferenees about their beliefs and likely future aetions. But I don 't do this in my own ease. Why not? Wittgenstein says: How is it that I cannot gather that I believe it's going to rain from my own statement 'It's going to rain'? Can I then draw no interesting conclusions from the fact that I said this? If someone else says it, I conclude perhaps that he will take an umbrella with hirn. Why not in myown case? Of course there is here the temptation to say: In my own case I don't need to draw this conclusion from my words, because I can draw it from my mental state, from my belief itself. (RPP I, § 704) The temptation to give this 'explanation' is nearly irresistible. One wants to say: 'I know what I believe; I don't have to infer it from my actions or my words'. This attempt at explanation ean be criticized in various ways. One thing Wittgenstein does here is to scrutinize the meaning ofthe words, 'I know what I believe'. He begins with a sampie statement ofthe 'explanation': 'Why should I draw conclusions about my behaviour from my own words, when in any case I know what I believe'? And what is the expression ofmy knowing what I believe? Isn't it 202 NORMAN MALCOLM expressed in this: that 1 do not infer my behaviour from my own words? That is the fact. 14 This brief criticism exposes the emptiness of the proposed explanation. And the criticism applies equally to the idea that 1 don't have to infer what I believe from my actions or words, since 1 know what 1 believe without inferring it from anything. The point can be put like this: the only meaning that the words 'I know what 1 believe' will bear in this context, is just that 1 do not infer what 1 believe from my words and actions. So the proffered explanation is circular. That a person can determine by observation what others believe, but rarely, if ever, by self-observation what he himselfbelieves, is something that applies not only to beliefbut to many psychological concepts. Wittgenstein's insight into this feature of the grammar of those concepts, is one of his great contributions to the understanding ofwhat it is to be a human being. 1 want to try now to give an account ofthe part that this difference between the first person and the third person of 'believe', plays in the disentangling of Moore's Paradox. 1 will return to aremark of Wittgenstein's that 1 quoted previously: That he believes such-and-such, we gather from observation of his person, but he does not make the statement 'I believe .. .' on the basis of self-observation. And that is why 'I believe p' can be equivalent to the assertion of 'p' Y (RPP I, § 504) This compressed reasoning can be expanded as follows: Since the statement 'I believe it's raining' is not based on self-observation, the only observation it is likely to be based on is an observation ofthe weather. So it can be roughly equivalent to the statement 'l1's raining', when this is based on an observation of the weather. Let us return to the problematic sentence, 'I believe it's raining, and isn't raining'. Moore thought that 'what is said' makes sense, expresses a logical possibility, could be true. The 'absurdity' lies only in saying it. It is as if one could think it, but not say it. Wittgenstein's reflection on the way in which the expression 'I believe p' is commonly used, leads to an entirely different view ofthe Paradox. 'I believe p' is a cautious or hesitant assertion of 'p'. The statement 'I believe it's raining, and isn't raining', would be both an assertion and a denial of 'It's raining'. It would have the normal incoherence of a self-contradictory sentence. Each part of the sentence cancels the other part. The outcome is nonsense. Moore was wrong in claiming that 'what is asserted is something which is perfect1y possible logically'. For nothing is being 'asserted' : in a clear sense, nothing is being said. DISENTANGLING MOORE'S PARADOX 203 Moore insisted that the sentenee is not self-eontradietory. But this was beeause Moore did not pereeive those surprising features ofthe logieal grammar of 'believe' whieh Wittgenstein brought to light. I want to eonsider further the problem of whether a belief is a 'mental state' . Some philosophers have assumed that this is so. David Armstrong, for exampIe, says that 'Belief is a dispositional state of mind which endures for a greater or lesser length of time ... ', and that 'there must be some differenee in A's state of mind if he believes p from his state of mind if he does not believe p' .16 I find this a puzzling assertion. Armstrong is speaking in the most general terms. He is not saying just that some beliefs are mental states (or 'states ofmind'), but that all ofthem are. And when Armstrong says that there must be some difference in a person's mental state ifhe believes p from his mental state if he doesn't, Armstrong is speaking with eomplete generality. Apparently he is holding that to believe p (regardless ofwhat pis) is itself amental state. And this would appear to imply the view that every statement of the form 'I believe p', describes or expresses amental state. I do agree that sometimes when people say things of that form they are deseribing or diselosing mental states. If a friend said to me, 'I believe I am going mad', that would tell me something about his mental state. Or if he said, 'I believe that all the people I see in the streets are spying on me'. There ean be a different kind of case in which a statement of the form 'I believe p' expresses, not a 'mental state', but an attitude toward aperson. Suppose a man is accused of intending to avoid payment of a debt. A friend of his says in his defence, 'I believe he will pay it'. I am supposing that he does not say this on the basis of any special information, but instead that he is declaring his faith in the integrity of his friend. As I am imagining the situation, this would not be a cautious or hesitant assertion, but a firm statement of trust. His statement, 'I believe he will pay the debt', does two jobs here: it asserts 'He will pay the debt', and it expresses an attitude towards the accused person. Whether a statement ofthe form 'I believe p' describes or expresses a mental state, or mental disposition, or an attitude, depends on the subjeet matter and the partieular cireumstanees. Sometimes sentenees of that form are used in that way. What is wrong is the assumption that this is always so: that the words 'I believe p', by their very meaning, describe or disclose or express, some mental state or disposition or attitude. This is an example of thinking without looking. Wittgenstein's treatment of Moore's Paradox should teach us better. 204 NORMAN MALCOLM When a farmer studies the sky and then says, 'I believe it will snow this afternoon' , his remark is tantamount to the cautious assertion, 'It will snow this afternoon'. In saying, 'I believe it will snow this afternoon', he need not be ascribing to hirnself, nor expressing, any mental state, disposition, or attitude. He need not even be saying that he is in a hesitant frame of mind about whether it will snow. As Wittgenstein points out, A hesitant assertion is not an assertion ofhesitancy. (LWPP I, § 522) Although I have seen emphasizing the fact that there is a common use of sentences of the form 'I believe such & such', in which the speaker is not talking about hirnself, is not describing some attitude of his, but is talking about the such & such - yet it is true that in certain situations someone who says 'I believe such & such' , is talking about hirnself. I do not wish to give the impression that Moore' s Paradox arises only when the speaker is talking about the such & such, and not when he is tal king about hirnself. Let me give an example. Suppose that a committee of a trade union has been debating the question ofwhether the members ofthe union should go on strike. The advantages and disadvantages of a strike have been thoroughly discussed. There seems to be no point in continuing the debate, even though there still is dis agreement. The chairman then says: 'It would be fruitless to hear again the reasons for or against a strike. What I want to hear from each ofyou is whether you personally believe there should be a strike'. He puts to each member the question, 'Do you believe there should be a strike?' When a member replies, 'I do believe there should be a strike', it seems right to say that here he is talking about, or for, hirnself - that the primary thing he is doing is expressing his own attitude on the issue. Yet it would be a logical absurdity for hirn to say, 'I do believe there should be a strike, but there shouldn't be one'. The important question, therefore, is not whether when one says 'I believe p' he is talking about hirnself, but rather how he is talking about hirnself, when that is what he is doing. Wittgenstein has reminded us that usually 1 say of someone, 'He believes p', on the basis of observing hirn, but that usually 1 do not say 'I believe p' on the basis of observing myself. 'He believes p, but not-p' is not an absurd form ofwords. Moore's Paradox exists only in thefirst person ofthe present tense. But it does not exist even there, when (as sometimes happens) the statement 'I believe p' does arise from self-observation and recollection. If 1 remember certain things I have said and done in the recent past, or certain ways I have reacted, and if solelyon the basis of these recollections 1 exc1aim, 'It appears that I believe p', or 'I seem to believe p', DISENTANGLING MOORE'S PARADOX 205 or 'So I believe p' - then there would be no absurdity in adding, 'but not-p'. It would not be absurd, for there I would be speaking of myself as ifI were another person. Thus, at the most fundamental level, Moore's Paradox arises from the fact that there is an essential difference between the normal use ofthe first person singular present tense of 'believe' , and the use of the third person form. The grammar ofthe word 'believe' reflects a striking feature ofhuman life, namely, the fact that there is an attentive observation which we direct on other people, and not on ourselves. King 's College London NOTES 1 Vol. I (RPP I), ed. by G. E. M. Anscombe & G. H. von Wright. Trans. by G. E. M.Anscombe, B. Blackwell, Oxford 1980; Vol. H (RPP H), ed. by G. H. von Wright & H. Nyman. Trans. by C. G. Luckhardt & M. A. E. Aue, B. Blackwell, Oxford 1980. 2 Vol. I (LWPP I),ed. by G. H. von Wright& H. Nyman. Trans. by C. G. Luckhardt& M.A. E.Aue, B. Blackwell, Oxford 1982. 3 See G. E. Moore, 'AReply to My Critics', in The PhilosophyofG. E. Moore,ed. byP.A. Schilpp, Tudor Publishing Co., NewYork 1952, p. 543. 4 Ibid. 5 Philosophical Papers, G. Allen & Unwin, London, 1959, p. 175. 6 RPP I, § 477 [The translation from German by Norman Malcolm] 7 RPP I, § 821 [The translation from German by Norman Malcolm] 8 RPP I, § 478 [The translation from German by Norman Malcolm] 9 RPP H, § 416 [The translation from German by Norman Malcolm] 10 RPP I, § 481 [The translation from German by Norman Ma1colm] 11 Ibid. 12 RPP I, § 482 [The translation from German by Norman Malcolm] 13 RPP I, § 715 [The translation from German by Norman Malcolm] 14 RPP I, § 744 [The translation from German by Norman Malcolm] 15 RPP I, § 504 [The translation from German by Norman Malcolm] 16 See D. M. Armstrong, A Materialist Theory of the Mind, Routledge & Kegan Paul, London 1968, p. 214. EVA PICARDI MALCOLM ON MOORE'S PARADOX As I find myself in agreement with most of what Professor Malcolm has said in his paper, my task as a discussant is comparatively easy. Accordingly, I shall confine myself to a few brief remarks and pause on a statement of Professor Malcolm's which in my opinion stands in need of qualification. Let me first mention what I particularly liked about Professor Malcolm's interpretation, namely the emphasis he has laid on the decisive role which the asymmetry between the first and the third person present tense plays in Wittgenstein's argument to the effect that the words "I believe" in their natural setting are not descriptive of the state of mind of the speaker. This however does not rule out that they can be used to say anything about the speaker: indeed, as Professor Malcolm points out in the last section ofhis talk, they are sometimes used for disclosing an attitude towards people or events. Disclosing an attitude, however, is very different from describing amental state: we can account for this use ofthe words "I believe" without endorsing the naive pieture of mental states as something which is there to be inspected, recorded or photographed by the mental eye ofthe speaker with his privileged access. When I say "I believe it's raining" 1 am talking about the weather and not about me. If one wants to say (and Moore surely wanted to claim at least this much) that "I believe" illuminates (beleuchtet) the state ofthe speaker, then we have to say the same thing of the assertion "p".1 This is why "I believe that p" as an isolated assertion can mean the same as "p". The office of these words is to hint (andeuten, instead of straightforwardly saying) that an assertion is being made. As Frege stressed, in language we do not have any special mark to signal assertoric force, and Frege's assertion sign is no better off in this respect. It is only seen but not heard and this is why you could not detect its presence in Professor Malcolm's quotation. Here I am just drawing your attention to the fact that the notions and the terminology which Wittgenstein uses for formulating Moore's paradox derive from Frege. In the Remarks on the Philosophy 0/ Psychology, but not in the Philosophicallnvestigations, both Frege's name and his assertion sign are explicitly mentioned. This sudden resurrection of Fregean notions should be found perplexing in the extreme and alert the reader to the very complex gambit Wittgenstein is trying to play here. For as we know from the Tractatus and the Philosophical Investigations, 207 R. Egidi (ed.), Wittgenstein: Mind and Language, 207-210. © 1995 Kluwer Academic Publishers. 208 EVA PICARDI Frege's idea of splitting an assertion into a thought (functioning as Annahme) and a force indicator has never satisfied Wittgenstein because of the accompanying naive picture of an inner mental theatre in which a thought is first grasped, then judged and finally asserted. Luckily, many of these Fregean connections have been pointed out by J. Schulte in his paper 'Es regnet, aber ich glaube es nicht'.2 But something still remains to be said. Wittgenstein's chief move is that first he uses Frege's embedding test to detect the source ofMoore's puzzle and then uses this result to controvert both Moore's and Frege's naive belief that words function in one way only. And Wittgenstein also argues for the converse ofhis claim in Philosophicallnvestigations, § 28 according to which it is not the case that every assertion contains an assumption. This rather mysterious statement is illuminated by some of his later remarks. Here he wants to show that not every assumption contains a potential assertion, an assertion which matches the content put forward for consideration. To appreciate the first part of Wittgenstein's claim it is sufficient to see that which is asserted in "I believe that the train is late" is not the same as that which is assumed when these words are used in the antecedent of a conditional. As isolated assertions "The train is late" and "I believe that the train is late" can say the same thing but in the sentences "IfI believe that the train is late, then we'll miss the plane" and "Ifthe train is late, we'll miss the plane" obviously do not. As regards the second claim - that not every assumption contains a potential assertion - Professor Malcolm says: The assertion asserts something that is not supposed by the supposition. The assertion does not exactly match tbe supposition; it brings in anotber element. But would it be rigbt to say that 'I believe it's raining' is not tbe assertion tbat corresponds to tbe supposition 'Suppose 1 believe it is raining'? Surely not. What other assertion would be tbe corresponding one? Tbere is no answer. We must simply accept tbis peculiar feature of tbe logical grammar of 'I believe'.3 I am not entirely happy with this remark. The answer "Surely not" sounds a bit too sweeping, for it suggests that we have al ready got hold of an explanation ofthe remarkable grammatical fact, whereas at this stage ofthe discussion that is not yet the case. We still need an argument which shows that we do not stand to our words in the same way in which we stand to other people's words. This argument is in fact that which Professor Malcom provides in the second part of his talk. My next point is a little bit more substantial and concerns the observation that the "assertion does not quite match the supposition; it brings in another element". Now it seems to me that this is not a correct description of the results achieved so far. For it seems to me that our understanding of these MALCOLM ON MOORE'S PARADOX 209 conditionals presupposes that we are conversant not only with the use of the verb "to believe" in the first person present tense but with a good deal more, indeed with the whole grammar of the word "believe", as Wittgenstein says. The argument given so far has established that "I believe p" can say the same as an assertion of "p"; moreover, in order to understand what it is to make an assertion we need the concept of belief. Thus for an understanding of the assumption embedded in such a conditional we need a fuller grasp of what is it to ascribe beliefs to our fellow creatures, a grasp which is not exhausted by the observation that to assert p and to assert "I believe that p" can amount to the same thing. We have not yet found the appropriate contexts whence our grasp of the words "I believe" as they occur in the assumption might have been derived. 1 think that we derive this understanding from our practice of ascribing beliefs to other people, an understanding which we then derivately apply to our own statements. Here we must not suppose that it is by introspection that we come to grasp the notion of belief needed to formulate the thought expressed by the assumption. We should ask what it means to ascribe beliefs to others and on which basis such ascriptions count as correct. And in this context we must advert to the outer criteria which every ascription of mental states stands in need of. 4 Thus the first lesson to be learned is that in grasping the words "I believe" occurring in the assumption we also bring to bear our understanding of "he believes ... " used as an ordinary assertion, that is, our understanding ofwhat it is to attribute beliefs to others on the basis of what they say. But now you may ask: What is the difference? Do not we proceed in the same way when ascribing beliefs to other people and when ascribing them to ourselves? The answer is: No, we do not. 1 do not normally infer my beliefs from my utterances, which is what we tend to do concerning the beliefs of others. We do not guess our future actions from listening to what we are saying, nor do we treat beliefs as sense impression. This answer in turn leads to the second important lesson implicit in Wittgenstein's argument, namely that we should look in vain for an assertion in the first person present tense which exactly corresponds to that which the assumption assumes. In this sense also the converse claim seems vindicated, viz. that not for every assumption there is a corresponding assertion. For if tbis were the case it should be possible to make an assertion and in the same breath to take a distance from it in the way we do with respect to the utterances of other people. This mistaken image is the source of Moore's paradox which 1 find successfully dispelled by Wittgenstein's analysis. Universita di Bologna 210 EVA PICARDI NOTES 1 Philosophische Untersuchungen. Philosophicallnvestigations (PI), ed. by G. E. M. Anscombe & R. Rhees. Trans. by G. E. M.Anscombe, B. Blackwell, Oxford 1953, H, p. 515. 2 Teoria , 2, 1985, pp. 187-204. 3See p. 198. 4 PI I, § 580. WOLFGANG eARL WITTGENSTEIN'S REFUTATION OF SCEPTICISM IN 'ON CERTAINTY' Since the days of antiquity the refutation of scepticism is a highly estimated occupation of philosophers, but it was not till Descartes that it was regarded as a condition of any justified claim to knowlege. Whoever did not attempt to refute the sceptic and succeed in doing so could not be certain to be in a position to reject his doubts and was, therefore, for Descartes a concealed sceptic. 1 The refutation of scepticism became the best way ("optima via") of reaching knowledge which meets philosophical standards2 and, in this way, scepticism itself changed radicaUy: it didn't stand any more for a way of life free of an dogmatic claims to knowledge, but it was a philosophical standpoint taken for tactical reasons. It was not the end, but a means ofphilosophical reasoning. The doubt of Descartes is a "staged", methodological doubt. There are other modifications of the sceptical doubt and the strategies to reject it connected with this change in the role of the sceptic. 3 I don't pay attention to them because one can make sense of the idea 'no knowledge without refutation ofscepticism' without Descartes' distinction between the mental and the physical and without claiming any epistemological priority ofthe mental. It is this idea which makes his 'First Meditation' so attractive and which seems reasonably plausible, as Stroud has shown rather recently.4 Since Descartes it seems to be a truism that the sceptic doubts the reality of an external world; for Moore and Wittgenstein this kind of doubt is the characteristic mark ofthe sceptic. But how is it to be understood? Wittgenstein observes: "'Doubting the existence of the external world' does not mean for example doubting the existence of a planet, which later observations proved to exist".5 Such a doubt does not concern the existence of a certain object or a certain kind of objects. If Descartes attempts to justify a doubt whether he is sitting by the fire or is touching a piece ofpaper, he doesn't want to doubt some fact or other. Rather, he is trying to doubt what is presupposed by every critical examination of our beliefs and thus cannot be itself under consideration; he is trying to doubt what "stands fast" in an our revising and examining of beliefs. The sceptical doubt doesn't concern any "white spots" of the increasingly comprehensive and detailed map of our knowledge; it aims at those beliefs the truth of which is obvious to everybody and presupposed by an critical examination of our beliefs. So, the doubt of the reality of an 211 Rosaria Egidi (ed.), Wittgenstein: Mind and Language, 211-222. © 1995 Kluwer Academic Publishers. 212 WOLFGANG eARL external world is not remarkable with regard to its subjeet; only by taking aceount ofthe seeptie's intentions ean one understand his doubt. He wants to keep a distanee between himself and what everybody who entertains belief takes for granted.6 But why does he seek this distanee? An answer may be found by eonsidering the reasons given by Descartes for doubting his beliefs. The so-called "argument from dreaming" is an attempt to justify a doubt eoneerning beliefs which are acquired under ideal conditions of observation. I will not discuss whether he really succeeds in doing so, but confine myselfto pointing out that this argument doesn't direcdy attack my beliefs about myself and my situation, but only indirectly by way of a critical considemtion ofmy cognitive capacities. The question 'How do I know that I am not dreaming?' amounts, from Descartes' point ofview, to the question, rather offensive for the reader, 'How do I know that I am not mad?', Le. to a question which is intended to undermine my confidence in beliefs acquired under ideal epistemic eonditions. The faith in the assumptions is shaken that these conditions are satisfied and that I am able to know anything at alt. This doubt doesn't concern my beliefs about what I see or come to know otherwise, but my beliefs about my epistemie capacities. What is at stake is my conviction that I am able to know anything. This is the subject of the doubt called by Descartes "metaphysical" or "hyperbolic doubt". You ean 't dissolve this doubt by referring to your established knowledge of what is going on in the world; you have to challenge this doubt by raising the fundamental question whether I am in a position to know anything at alP The sceptic makes an attempt to mise this question. Because he wants to mise it, he keeps his distance from the established standards of making and justifying claims of knowledge. But why does he want to raise this question? The 'first meditation' gives an answer to this question too. The well-known genius malignus is said to have arranged the world in such a way that whatever I take to be true is false. Nowadays this thought-experiment, lacking its theological content, is replaced by the assumption that I am a "brain in the vat" manipulated by a malicious seientist and his computer. Descartes eonsiders the possibility of man being so imperfect that he is always wrong. 8 This imperfeetion does not only concern the truth of my beliefs but also their material adequacy. Descartes eonsiders the possibility that his representation of the world doesn't fit it at all. However, to be able to consider such a possibility he has to assume that the world is totally independent from oUf representations of it and from the kind of representations we have. This assumption states a condition which seems to be essential and indispensable for our idea of what knowledge really iso Following Stroud, it shall here be called the REFUTATION OF SCEPTICISM IN 'ON CERTAINTY' 213 "very idea of objectivity".9 This seems to be a trivial claim: what is known is conceived as being independent of our knowledge. What we know is not produced by our knowledge and doesn't depend upon it.1t is possible that what is known is itself some kind of knowledge or an effect of it; but in this case too, knowledge doesn 't produce what is known. To understand the sceptical doubt we don't have to consider this complex case; we may confine ourselves to the simple case ofthe knowledge ofthe world, i.e. the knowledge that something which is what it is, independently of our beliefs and knowledge. It is the subject of our knowledge which is independent in this way. The fact that we know that there is no life on the moon, of course depends on our other beliefs and knowledge. It requires a system of beliefs and assumptions giving that opinion its proper place. But this doesn't apply to what we know. This independence inspires the sceptic to take seriously the possibility considered by Descartes, and to wonder how we do know that the world isn't totaUy different from the best of our knowledge. The trivial claim for the objectivity of the subject of our knowledge admits this possibility. Since this claim is essential to a realist view oftruth and knowledge, Stroud's and Nagel's thesis that scepticism and realism are two sides of the same coin, suggests itse1f. So far, 1 have tried to give a rough sketch of the motives and intentions which stimulate the "modern", "Cartesian" sceptic. 1 haven't given an argumentative exposition of scepticism, and 1 have refrained from a critical examination. My intention was to make clear the meaning of the sceptical question and thus to indicate a material condition of adequacy for any refutation of scepticism. This condition may be expressed as foUows: You can't chaUenge the sceptical doubt conceming what is obvious to everybody by just putting the obvious in its proper position, as Moore, for example, did. You have to account for the intention of the sceptic to caU in question the very possibility of knowledge; and you can't do that without considering the idea of objective knowledge underlying that question. Without such a discussion the refutation of scepticism can't succeed. Whether Wittgenstein, in On Certainty, managed to refute scepticism, must be discussed with reference to this idea. 1 will sum up his remarks with regard to their subjects and concentrate on Wittgenstein's considerations on the concept of doubt and his analysis of knowledge. "The child leams by believing the adult. Doubt comes after belief' (§ 160). The connection between believing and leaming makes it quite clear that Wittgenstein's remark does not concem only the genesis of doubt. The child leams "to react" (§ 538); it acquires a piece of"knowledge ofnatural history" (§ 534), consisting in certainties which are not only earlier than justified 214 WOLFGANG CARL knowledge and critical doubt but establish the very foundations of them. Doubt, seen from a genetic point of view as weIl as from a conceptual point of view, is an epiphenomenon which appears only after having "got hold" of a "world-picture" by one's cultural and intellectual socialization. This worldpicture is, as Wittgenstein emphasizes, not deliberately chosen (§ 167). A genetic dependence of doubt upon beliefwon't puzzle the sceptic very much. Descartes hirnself repeatedly underlined this dependency. Not wanting to confine hirnself to believe what he believed he claimed that one should, once in a life, try to doubt every belief hitherto held for more or less obscure reasons. This attempt - eversio opinionum - presupposes that one already has opinions, that one believes something or other. But Wittgenstein does not only want to state the genetic posteriority of doubt; he rather tries to draw the boundaries of doubt and to refute Descartes' project in this way. For, the "game of doubting itself presupposes certainty" (§ 115). Or, "doubt itself rests only on what is beyond doubt" (§ 519). Thus the thesis 'no doubt without belief must be understood as stating that we only can doubt something if we don't doubt something else but believe it. This thesis confines the sceptical doubt twofold: it is not possible only to doubt, and it is not possible to doubt everything. Let's consider first the assertion that doubt without belief is not possible. Wittgenstein argues as folIows: Admittedly, ifyou are obeying the order "Bring me a book", you may have to check whether the thing you see over there really is a book, but then you do at least know what people mean by "book"; and ifyou don't you can look it up, - but then you must know what some other word means.And the fact that a word means such-and-such, is used in such-and-such a way, is in turn an empirical fact, like the factthat whatyou see overthere is a book. (OC, § 519; see § 114-115; § 370) The belief which is in question here is a knowledge consisting in the understanding of a language; and the argument says that doubting presupposes the understanding of a language and thus the knowledge of the meaning of the word. lO The sceptic who, by relying on the method of doubt, poses the question of the possibility of knowledge thus has to presuppose some kind of knowledge. Doesn't this project refute itself? I don't think so. Wittgenstein claims that understanding a language is some kind of empirical knowledge, i.e. no knowledge of analytic truths. The sceptic will agree but call attention to the difference between "empirical facts" depending upon our linguistic usage, and empirical facts being part of a world which, independently of our talking and thinking, is what it is.u We saw that the sceptic poses the question of the possibility of knowing something about the world; and this question can't be answered by appeal to linguistic usage and linguistic intuitions. REFUTATION OF SCEPTICISM IN 'ON CERTAINTY' 215 In other words: the sceptic won't deny that the English word 'moon' is intended to signify a ce1estial body circ1ing 'round the earth; but it doesn't follow from this that the moon exists. The kind of knowledge which is mentioned by Wittgenstein and has to be considered as a condition of doubt, doesn't belong to that kind of knowledge the possibility of which is called into question; and the sceptical doubt isn't yet removed by the belief which Wittgenstein takes to be its condition. Wittgenstein's second objection to the sceptic says that doubt essentially belongs to a language-game and can only be understood within the context of a language-game (see § 24). Two anti-sceptical consequences are supposed to follow from that: scepticism doesn't correspond to any conceivable form of life; and every doubt is an "internai" doubt. Let's consider the first consequence. Doubting requires a certain "behaviour" (§ 354). The usage of linguistic expressions articulating a doubt or indicating that it has been removed has to be connected with an identifiable, specific ''practice of the language" (§ 524). Doubting something or being certain of something requires specific epistemic circumstances. If these circumstances are not given and if the doubter's behaviour can also be understood without these sceptical "modifiers", his way of talking is "merely a bit perverse, and does not affect the issue" (§ 524). Ifthe dec1aration of a doubt or the indication that it has been removed doesn't touch upon my acting and behaving, no form of life corresponds to the sceptical attitude and it is nothing but the "fifth whee1 of a coach": "it does not infect practice" (§ 524; see § 428). Such a doubt is "hollow" (§ 312), it is no doubt. Doubting is part of a language-game; hence there must be a form of life corresponding to it. This may be understood as an explication of a connection ofknowledge and action which Wittgenstein critically uses against Descartes' project ofuniversal doubt. "Would my knowledge still be worth anything ifit let me down as a c1ue in action? And can 't it let me down?" (§ 409). These questions are rhetoricaiones and, moreover, not well-expressed ones. For, the knowledge referred to is a kind of knowledge which Moore considers in "A Defense of Common Sense"; it consists in certainties which cannot be separated from my acting and which thus can't justify those activities. They constitute the "basis ofaction, and therefore, naturally, ofthought" (§ 411; § 414; see § 87). My actions, however, reveal - "with a certainty that knows no doubt" (§ 360) -, that I have those certainties. This acting doesn't refute the sceptic's doubt, it doesn't even let it arise: "My life shows that I know or am certain that there is achair over there, or a door, and so on. - I tell a friend e.g. "Take that chair over there", "Shut the door", etc. etc." (§ 7). As there is 216 WOLFGANG eARL no reason or justification for this life, there are none too for these certainties: they are "there" (§ 559). It isn't difficult to answer the sceptical question; being a senseless question it can't even arise. Wittgenstein regards the language-game as a symbiosis of certainties and actions; the actions of men, therefore, show "that they believe certain things definitely" (§ 284). The sceptical question concerns the possibility of knowledge and it is asked from astandpoint which doesn't lie - as Stroud, following Clarke, says - "within the practical contexts which give our words their social point" .12 Descartes always stressed the specifically theoretical nature of the sceptical question. 13 But does such a question make sense? Wittgenstein denies this: acting is not possible without knowledge and this knowledge must be usable as "a clue in action" (§ 409). It seems to me that Wittgenstein hirnself rather critically comments on this connection ofknowledge and action; he says: "So I am trying to say something that sounds like pragmatism. Here I am being thwarted by a kind of Weltanschauung" (§ 422). But irrespective ofthis the sceptic may reply that it is far from clear that the certainties necessary for action are more than fast-standing opinions and unshakeable beliefs. Why must it be a knowledge of that kind which the sceptic is looking for? Why isn't it enough here "to know 'for all practical purposes'''?14 However, the sceptic isn't interested in that kind of knowledge. In other words: Wittgenstein's antisceptical objections concerning the connections of actions and certainties is convincing only if the debate is transformed into a fundamental discussion of the concept of knowledge. 1'11 come back to this issue soon. Let's now consider the second objection to the sceptic which rests on the concept of a language-game. It starts with the claim that a doubt must be justified (see § 122; § 458). Reasons can justify knowledge in showing that one "is in a position" to know such a thing (§ 555); and respectively the reasons justifying a doubt must show that error is possible. A doubt makes sense ifthe correction ofan error is conceivable; thus doubt presupposes the possibility of error. Wittgenstein maintains against Descartes first that reasons are part of a system of beliefs in which "some things stand unshakeably fast and some are more or less liable to shift" (§ 144). Thus a reasonable doubt must regard many things as "absolutely solid" (§ 151). The thesis discussed above - 'no doubt without belief - had after all to put limits on what can be doubted. The claim 'no doubt without the possibility ofjustification' implies that this is what must stand fast relative to all questions and justifications "as it were like hinges on which those turn" (§ 341). Thus, every doubt is an "internaI" doubt, is relative to a frame of fast-standing beliefs which cannot REFUTATION OF SCEPTICISM IN 'ON CERTAINTY' 217 be doubted. Second, Wittgenstein says that not all of our beliefs can in the same way be exposed to the risk of error: "For it is not true that amistake gets more and more improvable as we pass from the planet to my own hand. No: at some point it has ceased to be conceivable" (§ 54). To assurne the possibility of error in this case is a symptom of "a mental disturbance" (§ 71); to have this doubt is to be a "half-wit" (§ 257). The sceptic will reply to this reasoning by asking a further question. What stands fast? What is the status ofthat which must stand fast in order to render justification and doubt possible? Is this, again, a matter of truth and knowledge concerning our linguistic usage and our habits ofthought? Wittgenstein asks: "Who decides what stands fast?" (§ 125). This question is obviously treated as senseless as the possibility of "learned judges being asked whether a doubt was reasonable or unreasonable" (§ 453). Rather, "regarding it as absolutely solid is part of our method of doubt and enquiry" (§ 151). But does this imply its truth? Let's once more go back to the "argument from dreaming". The conviction that I am standing here is doubted by asking whether I know, after all, that I am not dreaming. Saying that we only know something about the external world if we know that we are not dreaming implies, as Stroud has tried to show, that every belief acquired under normal conditions of sense-perception may be doubted. 15 I'm not going to discuss Stroud's argument except to point out a weakness of Wittgenstein's thesis. The assumption that every doubt is "internaI" doesn't meet the point of the sceptical strategy; for this strategy doesn 't proceed by balancing the pros and cons within a homogeneous c1ass of sentences which have the same errorproof conditions. The sceptic picks up one of these conditions which are supposed to be fulfilled if one knows the truth of any sentence of this c1ass. How can Wittgenstein criticize this challenge? How can he justify the thesis that it must "stand fast" that this condition is fulfilled? He says that assuming the contrary would be a symptom of a "mental disturbance" (§ 71) and would seem to hirn "madness - of course, ... also in agreement with other people" (§ 281). But are these little favourable estimations sufficient to reject the sceptical question whether I could know that I am not dreaming? Wittgenstein remarks on error: "In order to make amistake, a man must al ready judge in conformity with mankind" (§ 156). This mayaIso be said about doubt. But then the truth of what is taken "to stand fast" rests upon its being held true; and the sceptic has good reasons to reject this. So the thesis that every doubt is "internaI" seems to be confronted with the difficulty that it either leaves open the truth of what "stands fast" - thereby not at all answering the sceptical question - or that it justifies its truth by a collective holding to be true - thereby 218 WOLFGANG eARL giving an unsatisfactory answer to the question. There remains the possibility to deny the thesis by adopting the sceptical strategy of the "argument from dreaming". So did Descartes and Stroud. Let's now consider the "anti-sceptical potential" ofWittgenstein's conception of knowledge; let' s concentrate on his thesis that "the concept of knowing is coupled with that of the language-game" (§ 560). I can't explain this thesis here at length and with the necessary attention; so I shall confine myself to selecting two characteristics of the concept "language-game" which are of particular importance to the understanding ofthe thesis. According to it, knowledge must be regarded as holistic: it is not possible that one should only have one beliefbut there always is a manifold ofinter-connected beliefs forming a "system" (§ 102). This system is the "element" in which they "have their life" (§ 105), - i.e. it determines the roles of the single beliefs in the totality of knowledge. Second, our knowledge is connected with a form of life which lies "beyondjustified or unjustified" (§ 318-319) and thus "has to be accepted", "is given", "like our life" (§ 559).16 In the following, these two theses about knowledge will be explained in more detail. Since our knowledge forms a system, not all of our beliefs have the same "status" (§ 167): not all ofthem are "equally subject to testing" (§ 162). Some ofthem are the "foundation ofmy research" (§ 167) and serve as the "background against which I distinguish between true and false" (§ 94); they have a "peculiar logical role" (§ 136) and can't be identified by semantic criteria (see § 308) or by rules in general. Their role consists in determining a "worldpicture" which serves as a condition and a framework of my knowledge. Wittgenstein compares such sentences with the rules of a game (see § 95). His analysis of how we know or are familiar with such sentences is of crucial importance to his criticism of scepticism. Here he is using the term 'certainty', having in mind a peculiar, non-epistemic attitude or mode of presentation. With reference to Kant' s terminology one could speak oftranscendental conditions ofknowledge; for Wittgenstein, however, as opposed to Kant, there is no knowledge of such conditions which is open to justification and thus there is no transcendental philosophy. The certainties which reveal that I have a world-picture, and what kind of world-picture I have, are certainties which I have "in agreement with other people" (§ 281) and can only have insofar as I "belong to a community which is bound together by science and education" (§ 298). The certainty I have is a certainty we have: it is a certainty I can't choose but which we receive through our cultural tradition and our socialization in which we "grow up": those certainties are the historically conditioned truisms-of a community which in REFUTATION OF SCEPTICISM IN 'ON CERTAINTY' 219 our case includes some knowledge of the various sciences. The critique of scepticism based upon Wittgenstein's concept of knowledge says that those certainties constitute a form of life and can't be conceived as a kind ofknowledge. In the words ofMarie McGinn, they constitute "a non-epistemic attitude to judgments" Y The non-epistemic status of the certainties follows from the fact that the world-picture which 1 share with others is the "inherited background against which 1 distinguish between true and false." (§ 94) This background is, therefore, neither true nor false (see § 205) and lies beyond what is accessible by a knowledge which rests on reasons (§ 243). So, "Moore's mistake" in his critique of sceptical doubt consists in 'countering the assertion that one cannot know that, by saying "I do know it'" (§ 521). The sceptic, as Wittgenstein says, is quite right in pointing out that 1 can't know whether that is my hand; but he gives the wrong interpretation of it. He regards it as a deficiency of justification whereas it is important to conceive those certainties as something beyond knowledge which rests on reasons. Our having such certainties doesn 't mean believing certain propositions which "strike us immediate1y as true, i.e. it is not a kind of seeing on OUf part"; rather it reveals itself in OUf "acting, which lies at the bottom of the language-game" (§ 204). With regard to the language-game of knowledge this aeting eonsists in our method of caleulating (§ 212) or in OUf practice of judging: "From a child up 1 learnt to judge like this. This is judging. This is how I learned to judge; this 1 got to know as judgment" (§ 128-129). The quest for reasons for my judgments doesn't arise any more, because our practice of judging is exactly what we understand as judging. Our concept of judging eonsists in OUf making such judgments; we would not have this concept ofjudgment ifwe didn'tjudge this way. And we can't ehoose this concept or any language-game; we have no alternative to it. Wittgenstein's remark "The truth of my statements is the test of my understanding of these statements" (§ 80) must be formulated precisely: My taking something to be true reveals my having understood or not having understood what I take to be true. The sceptic, in this ease refraining from judgment, doesn 't understand the language he uses; and Wittgenstein won't try to refute hirn but will treat hirn like a ehild having to learn "to reaet in sueh-and-sueh a way" (§ 538) -, like a child having to learn the use of the expressions "before" it is able to pose the problem ofthe justifieation of its use (see § 548). As doubt is later than belief, knowledge only starts on a later moment (see § 538) - after we have already got certainties. How will the seeptic react to this pedagogical therapy? Supposing that teachers are open at all to arguments (see § 314-315), he will try to evade it with 220 WOLFGANG eARL the foHowing argument. Marie McGinn said about Wittgenstein's critique of scepticism "that it has provided a model for understanding certainty as a nonepistemic attitude to judgments that have a non-empirical role" .18 This "model" implies that our knowledge fits into a world-picture we have since we have certainties which we aren't able to justify or give reasons for. The "groundlessness of our believing" (§ 166) should not be regarded as a deficiency of knowledge l9 , but as being prior to all our knowledge and at the same time as being its "foundation" (§ 167; op. cit., p. 135). How is this possible? If certainties reaHy shall serve as the foundation of our knowledge they can't be conceived as quite independent from it. Wittgenstein himself concedes the possibility that we compare "our system of knowledge" with that of other people and that theirs "is evidently the poorer one by far" (§ 286). It is far from clear what the criteria are according to which we can compare and evaluate different systems of knowledge; but Wittgenstein himself mentions the fact that our system is acquired "by means of observation and instruction" (§ 279), and that we "belong to a community which is bound together by science and education" (§ 298). It is evident that the richness or poverty of a world-picture is not, or is not only a matter of faith. The superiority of our knowledge has many sources; and it's not necessary to be "converted" or "persuaded" (§ 612, 262) to realize this. Therefore, it must be possible not only to accept the world-picture which our knowledge fits on faith but also to justify it with regard to its adequacy and explanatory power. What serves as the foundations ofknowledge can't be totaHy independent from what is rendered possible by these foundations. What only can be accepted by faith, cannot serve as the foundations of our knowledge. Thus, insofar as certainties are considered as the foundations of our knowledge, they can't lie beyond what is rendered possible by them. Therefore, we can't consider certainty, as Marie McGinn proposed, as a "non-epistemic attitude to judgements" and justify in this way the exclusion of the possibility of doubt. However, if one sticks to the idea that certainties are prior to all knowledge and, following Marie McGinn, interprets Wittgenstein as advocating a "conception of certainty that is connected with, and prior to, the idea ofknowing", one gets into other kinds of trouble. 20 Concerning those who don't ask the physicist but the oracle whether water boils at 100°C, he remarks: "Where two principles really do meet which cannot be reconciled with one another, then each man declares the other a fool and heretic" (§ 611). "Conversion" seems to be left open: it is possible to be "persuaded" to accept certain beliefs (§ 262, 612), to adopt another view of the world (see § 92). But how shall we understand the possibility of a plurality ofworld-pictures having the REFUTATION OF SCEPTICISM IN 'ON CERTAINTY' 221 same status and leaving only the possibility of conversion open? I'm not going to treat the objection raised by Davidson and others against the assumption of such a possibility: how are we, within and with the help of our world-picture, able to recognize that other people have another world-picture? For the sceptic, the certainty of the world-pictures and the possibility of a plurality of certainties corresponding to the possible plurality of worldpictures is more interesting. Wittgenstein's distinction between knowledge which is open to justification and may be doubted, and certainty where this is not the case, may be regarded as an epistemic version of Carnap's semantic distinction between internal and external existential statements. 21 As the existence of a piece of paper on my desk depends for Carnap on the acceptance of a thing language,z2 for Wittgenstein the certainty that the earth existed before I was born depends on the belief in a world-picture in which our knowledge fits. Other world-pictures correspond to other certainties. This, however, means that the epistemic status of statements varies or can vary from one world-picture to another. In other words: The historical background and the c:ultural socialization of the knowing subject determine what is knowledge, open to doubt and justification, and what are certainties which have to be accepted on faith. The idea that our knowledge gives a more or less adequate and unfeigned representation ofthe world loses its sense; and this too paralleis Carnap's view that being true or wrong depends on a chosen language-system. Following Carnap, we have to give up the idea of a world which is what it is, no matter how and whether we talk about it; following Wittgenstein, we have to renounce the idea that our knowledge of the world is determined by something which exists independently of the social and historical standpoint of the knowing subject. As Carnap misses what is "factual" ab out the world, Wittgenstein totally loses sight ofthe objectivity-claim ofknowledge. The sceptic will neither accept Carnap's ontological idealism nor Wittgenstein's epistemological relativism as a refutation; on the contrary, he may justify the sceptical question by appealing to the implications of these attempts to reject the question as senseless. Let's conclude! I have been concerned with Wittgenstein's discussion of scepticism in On Certainty, raising two issues: his critique of doubt and his conception of certainty which he developed in connection with his analysis of knowledge. Whether his remarks are to be considered as a refutation of scepticism depends on how one understands the sceptic. I have described hirn as raising the fundamental question of the possibility of knowledge, assuming that what is known is independent of our knowledge and the conditions under which, for us, knowledge is possible. Wittgenstein's attempt to draw the 222 WOLFGANG CARL boundaries of doubt either doesn't meet this question or doesn't meet the objectivity-claim ofwhat knowledge is ab out. His conception of certainty as a "foundation for my research" (§ 167) doesn't offer good reasons for rejecting the possibility of doubt; ifwe.take certainty, however, as a "background" or a framework of our knowledge (§ 94), the problems of an epistemological relativism arise and he will not meet the objectivity-condition of knowledge without having discredited it. Contrary to the claim that "doubt gradually loses its sense" (§ 56), the sceptic's doubt still hasn't lost its sense. Universität Göttingen NOTES Meditationes, Objeet. Sept. (AT VII.550). Principia Philosophiae I. 8; Meditationes, Synopsis (AT VII.l2). 3 See M. Bumyeat, 'Idealism and Greek Philosophy', Philosophical Review, 91, 1982, pp. 3 ff. 4 B. Stroud, The Significance 01 Philosophical Scepticism, Clarendon Press, Oxford 1984. 5 Über Gewißheit. On Certainty(OC), ed. by G. E. M.Anseombe & G. H. von Wright. Trans. by D. Paul & G. E. M. Anseombe, B. Blaekwell, Oxford 1969 § 20. 6 See M. MeGinn, Sense and Certainty: A Dissolution olScepticism, B. Blaekwell, Oxford 1988, p.3. 7 B. Williams, Descartes: The Project 01 Pure Enquiry, Humanities Press, Atlantie Highlands, N.Y. 1978, pp. 61ff. 8 Meditationes (AA VII. 21). 9 See Stroud, op.eit., pp. 81-82. 10 See A. Kenny, Descartes. A Study ofhis Philosophy, Random House, New York 1968, p. 21. II See Deseartes, Principia Philosophiae, I. 10. 12 Op. eit., p. 8 ｾＮ＠ 13 Principia Philosophiae I. 3; Meditationes, Respon. Quintae (AT VII. 350/1). 14 See Stroud, op. eit., p. 68. 15 Op. eit., pp. 19-23. 16 See also Philosophische Untersuchungen. Philosophical Investigations, ed. by G. E. M. Anseombe & R. Rhees. Trans. by G. E. M.Anseombe, Oxford, B. Blaekwelll973 3, II Part, p. 226. 17 Op. eit., p. 138. I 2 181bid. See C. McGinn, op. eit., p. 146. Op. eit., p. 135. 21 See Empiricism, Semantics, and Ontology, in R. Camap, Meaning and Necessity: A Study in Semantics and Modal Logic, Chieago University Press, Chieago 1964, pp. 205 ff. 22 Op. eit., pp. 206-207; see Stroud, op. eit., pp. 192 ff. 19 20 RUDOLF HALLER WAS WITTGENSTEIN A RELATIVIST? Let me first confess a certain uncertainty about the title of my paper: it is a simple question, and a simple question may be bewitching, in that one might seek an answer as simple as the question appears to be. So it is necessary first to get clear ab out the question - and then to try and find an answer, which I hope will also help us to understand the position of Wittgenstein better. Relativism, as we know, has many faces. To start with I shall mention only some of them. Re1ativism is also not highly esteemed among philosophers, since it is very often presented only in a negative form. Mostly it has been presented in a rather distorted form, for example, by Richard Rorty, who states that 're1ativism' is the view that "every belief on a certain topic, or perhaps about any topic, is as good as any other." And he adds, "no one holds this view".l Clearly, if you present a view held by 'many relativists' in such a way that its characterization does not fit any actually existing relativist you are indeed entitled to say that no one holds this view. But this could not be said of the view propounded by Protagoras. Perhaps it is this view which many take for granted, and which we find expressed in the famous words of the old Sophist. According to this view, what is true is not what is true per se, but what seems to be true to anyone to whom it seems to be true. This outstanding example of the attempt to invest the individual with the power to decide what is true or false has been attacked - as have been almost all the other forms of relativism - by way of arguments which attempt to exhibit the essential inconsistency and even absurdity of such a view. The main argument against Protagoras' thesis was that if no belief or no judgment is objectively true (or true per se), since truth (and falsehood) are only functions of what is believed or judged to be so, then the thesis itself is either self-refuting or absurd. If the thesis itself is taken as non-relativistic, that is, as absolute, then it is self-referentially inconsistent, because there would be an exception or a counterexample to it before the thesis itself can be considered. If, however, the thesis is taken to be relativistic, then it does not falsify the claim of the objectivity of truth per se, but presupposes it. However, relativism has - as I said - many faces. Indeed, it seems that we are nowadays confronted with several new forms of relativism which are 223 R. Egidi (ed.), Wittgenstein: Mind and Language, 223-231. © 1995 Kluwer Academic Publishers. 224 RUDOLF HALLER perhaps less objectionable than those held hitherto. Perhaps best known in present-day philosophy are two other kinds of relativism: we might call them ontological and paradigm-relativism. 2 The first is concerned with the question how we individuate the constituents of our ontological framework if we are hindered in principle by "the indeterminacy of translation of identity". The way out of the prison of one language is the translation into another one. If we stick to the question what objects (of a theory) - absolutely speaking are, we are lost, because there is no fact of the matter and no method which enables us to decide in favour of one interpretation of the reference of a term or utterance or against another. There may exist different, mutually incompatible translations, better systems of translations which accord with the behavioural data. This point ofview is akin to the (successful) attempt to convince the scientific community that scientific theories are very often not only mutually inconsistent, but not even commensurable. And the main reason is that the conceptual means, the conceptual apparatus, is essentially different, and not translatable from one into the other. If we project paradigms to the systems of languages then the words of one language may not even be translatable into another; and if neither facts or minds can decide between them, they remain incommensurable. The upshot of the incommensurability thesis is - among other things - that he or she who "moves" within one conceptual system will not be able to understand what is meant in the other. In order to find a way to bridge the gulf between two incommensurable systems, be they theories or languages, one would require the complete switch of the systems one has learned by entering, that is, learning the other from scratch. To the extent that truth is taken as a topic in tliese views, it certainly remains a predicate relative to the system, and has no meaning outside a theory or paradigm. But this does not mean that the distinction and application of truth and falsehood turns out to be entirely subjective, like personal taste. Learning one from the other ab ovo remains the only way to bridge the gulf. Even Paul Feyerabend,3 the jolly anarchist and relativist, does not accept the idea that the simple individualistic belief "existent or true for me" could be substituted for objective existence or truth per se. This is and remains a non sequitur. But when I mentioned other faces of relativism, I tacitly referred to some well-known and widely accepted views, respected and used even by those who strongly oppose relativism in any of its subjective versions. Actually the rise of conventionalism in mathematics and physics at the beginning of this century was a first big step into the direction in which I am alluding. That the deepest principles of science - axioms - are neither synthetic WAS WITTGENSTEIN A RELATIVIST? 225 apriori, as Kant thought, nor empirically true, but are the result of human decision, and therefore subject to change, was a step towards relativism. The idea that the different geometrical systems which may be used to describe physical space do not contradict one another, but are to be viewed as more or less convenient means of our mathematical or physical language, the idea that free choice does not mean arbitrariness or anarchy, marked a turn in our understanding of the conventional character of the language of theories. Moreover, conventionalism had a much wider impact than we usually assurne. I shall not embark on the journey needed to show or remind you of the places where one can hear the echo of its main theses: in Neurath, Hahn and Frank; in Schlick, Reichenbach and Carnap, not to mention Einstein, it is easy to identify this train of thought. All of them were fascinated by the new idea that decisions and conventions can do the work which, according to the traditional view, only apriori intuition, synthetic judgments apriori or, at best, tautologies could do. An important consequence ofthe same line ofthinking was Wittgenstein's understanding of what grammar and grammatical roles are. Like the conventionalists, Wittgenstein stated that, for instance, an "axiom about number is a postulate or role according to which we proceed". 4 Here he takes roles on a par with postulates. Rules of grammar, like postulates, do not describe reality: they neither conform to nor contradict facts. 5 It is not my intention here to impose on anyone the belief that Wittgenstein should have been a conventionalist, in spite of the fact that I cannot see how one can put some of the main ideas of his philosophy of mathematics on the opposite (non conventionalist) side. I only want to stress the fact that from the time of his return to philosophy in 1929, Wittgenstein defended the possibility of turning any empirical proposition into a role, and insisted that it is our decision whether we do the one or the other. It is "the roles in the language-game, the kinds of use" to which necessary statements and non-necessary statements are respectively put. 6 When Wittgenstein points to decisions in explaining the acceptance of a proof as a proof, he merely extends the idea that any move in a language-game might be taken as the subject of adecision. Since postulates and roles do not come into existence out of the blue, it is not at all out of place to ask where they do come from, and to whom they are addressed. A quick answer would be that it cannot be individuals, since individuals proper, that is, individuals taken in isolation, could not be the subjects of roles of meaning invented by themselves, since - as Wittgenstein has said and Kripke has shown - for the individual proper, any actual decision to mean or to use a word in a certain way cannot determine its meaning (in the 226 RUDOLF HALLER future) except in so far there are outer criteria or methods applied by other members of the language community to ensure the correctness of the use. Kripke's so-called "sceptical solution" points to one side ofthe answer to the question of where rules and postulates - to mention only two of the main pillars of our language-games - get their regulative power from - the answer is: the community. The main feature ofthe connection between the individual and the Ianguage community to which the individual is linked is an agreement in regard to the meaning of words and signs. It is not the purpose of this talk to enter into a discussion of the so-called sceptical paradox and its solution. But there is one particular idea to be found in Kripke's approach which is ofthe utmost importance also for our question here, namely, the idea that the individual taken as a private human being in its privacy could not follow a rule on the basis of his or her own beliefs, since he or she is in principle unable to distinguish by any facts between meaning something and meaning nothing at all. Wittgenstein says that "following a rule is analogous to obeying an order" (PI I, § 206); but from the examples given, it is clear that the analogy works only when there exists a certain regularity in the reactions of those following a rule (or obeying an order). Without regularity we could not 'see' the pattern of behaviour which connects the rule-following actions with the rule or the obeying with an order. Regularity is thus presupposed when we attribute to individuals that they are following rules as weIl as when we say of them that they are giving themselves orders, ask themselves questions or are answering them. But regularity is not enough. Wittgenstein' sanswer to t1ie question where the rules themselves come from not only points to the practice \>Ut states that it is the practice together with the drill and training belonging to it which establish the link between meaning and use. A practice consists of established rules for doing something in a certain way by doing it in a certain way, and of provisions for the continuation ofthis way of doing that thing by teaching others how to do it. At a deeper level the practice comprises within itself, therefore, the harmony between words and deeds. Thus, when Wittgenstein says "The practice gives words their meaning",7 he points to the difference words make at various situations in life. If saying something does not have a relation to our actions which are - as Wittgenstein says - "interwoven" with the language, then the words used do not mean what they should mean according to the rules for their use. Thus the meaning ofthe term 'practice' can be restricted neither to a technique of bringing something about according to a model, nor to actual behaviour, without taking into ac count the historical point of view. Because to state that some kind of behaviour is customary, or WAS WITTGENSTEIN A RELATIVIST? 227 rather, is a custom in a society, necessarily brings a historical dimension into the analysis. Calling a form of behaviour a custom presupposes that in a society or community it has been, from some time onwards, usual to act according to this form or model of behaviour, either always or on certain appropriate occasions. Thus, for example, if the custom concems only some sub-group of a society it usually remains unnoticed or even unknown by the majority of members of that society. But if the custom is one like the use of the common language, shared by all the members who leamt to speak the language, then the custom is known to all ofthem, even though the knowledge is in most cases only implicit or practical. To say that there is agreement conceming the application of the language-rules, however, does not mean that such an agreement has been achieved or has to be achieved. Agreement, as Wittgenstein uses the notion in the relevant text, is not something the language-user has to attain, rather, it has to exist in order that one may playa language game. Thus - except in the period ofleaming a language-agreement is a condition for the application of the rules. This is important and often misunderstood. 'Agreement' (Übereinstimmung) and 'rule', Wittgenstein says, are re1ated to one another. And the main point, as I read the passage, is: that it is same means which is concemed by communicating in the institution named "the language-game": namely that our agreement and the rules we are obeying are working when we act. The "queer" statement of Wittgenstein he himself called it such since it may sound so - is then that in order for human discourse to function properly there should exist agreement not only in definitions but also in judgments. 8 This is puzzling, because it is not true that people do agree with one another in their judgments, and since judgments are either true or false like the sentences expressing them, it is not true that they agree about the same when one person states the opposite of another. Thus, one has to find out what is really at stake, when it is stressed that agreement in judgments should be part of the communicational undertaking. The methods we apply to measure something may certainly be subject to descriptions which may be shared by those who enter a discourse. So, what has to be described may be judged in the same way. But not necessarily so. Rather, two people communicating about astate of affairs may (and very often will) judge it differently and still there will not be a dispute over the question whether both have obeyed a rule or not. Therefore, agreement in judgment can be asked for if agreement in concepts is at stake, because the grammar does not deal with.ready-made judgments, but with the composition of them in the language, and with their application. So, the definition of words is one thing, and if there is not agreement about it, we can not meaningfully proceed in 228 RUDOLF HALLER communication without making c1ear in what sense a word or term is used. And if we accept this, we may think that the agreement about the rules of grammar together with the definitions suffices to decide what is true and what is false. But Wittgenstein strongly denies that any agreement does decide what is true or false (PI I, § 241). It is on the basis of agreement in definition, judgments and rules, that people say things which are true or false, but they do not agree in what to say. 'Agreement' is used in a much wider sense, which leads to the very problem I have used as a title of this consideration. "Agreement", Wittgenstein teIls us, does not refer to opinions but to a form of life. Thus, truth pertains only to sayings, agreement in this context to much more, namely, the sharing oflanguage and the sharing of a form oflife. Now, I cannot go into an extended and deeper analysis of this concept. But this much seems c1ear to me, that it does not refer to "the presumption of a common nature", as C. McGinn has suggested. 9 'Lebensformen' or 'forms of life' is a plural, therefore there are more than one. On the other hand, there is only one common way of human acting; therefore the expression: 'form of life' should not be taken as meaning 'the common way of human acting'. Neither should one fall into what Hintikka has dubbed the "fallacy of verbal language-games".lo Discourse and speaking a language is only "Part of an activity or a form of life" (PI I, § 23), neither itself the form nor identical with the language-game. According to Wittgenstein's descriptions, forms oflife are at the very bottom of any analysis of language, and that without an understanding of the ways people act and live, language cannot be understood, or better, fully understood. When Wittgenstein uses the example of an explorer coming to aland in which the language spoken is totally unknown to hirn, Wittgenstein states that "he [the explorer] could leam to understand it only through the connection with the rest ofthe life ofthe inhabitants".u Thus, even ifyou have only mastered the language but have not reached the connection with the actions and the life of the people, you may still remain unable to understand the people in a country with totally different traditions. 12 Here we touch on the extreme possibility which Wittgenstein has put into the simile: "If a lion could talk, we could not understand hirn". 13 In his last writings, namely those which are published under the title On Certainty and Remarks on Colour, Wittgenstein seems to be replying to the question, "On what basis or ground do we decide between true and false propositions?" But I do not see here a big change of attitude, amounting to a new phase of Wittgenstein's philosophy. It is true that some ofthe already familiar notions, like the central concepts of a "picture" or of a "form of life", have WAS WITTGENSTEIN A RELATIVIST? 229 been given a wider application. For example, when we are told that our pieture of the world "is the inherited background against whieh I distinguish between true and false" .14 So the following question suggests itself: What if the background - e.g. your picture of the world - is different? This I think is one door where relativism concerning truth and falsity may enter. If distinguishing between true and false presupposes the acceptance or assumption of certain beliefs, then the acceptance or assumption of beliefs, which - as the case may be - are contrary to those, cannot serve as background for the same distinctions between true and false. Since in learning a language we are not only taught the use of words, rules of their use and examples, but since we are also taught judgments, other judgments may make up another system ofbeliefs when other judgments are taught. Since Wittgenstein says that neither experience nor success is the ground for the game of judging when we cannot rely on the fonner experiences and not wait for success we remain without any ground for the decision between conflicting judgments based on different world pictures. Of course we may try to invent some instruments for comparing them, like simplicity, but there is no guarantee that the invention will work and, if it works, that it will be accepted as a criterion for deciding between conflicting judgments. If we want to convince someone who does not share our picture of the world, we may try to instruct hirn, teach hirn all that we believe, but since no ground for adecision is there, our last resort will be persuasion. And, whether we like it or not, this is the end of the line of arguments. If we do not believe in an eternal discourse, which will make clear which part of our prejudiee has been destroyed by the grace of nature, we shall have to rest our system of beliefs on apresupposition, whieh nonnally will not be called onto question, but which is of the utmost importance. It is the fact that all human beings speak a language, even if the language spoken may not be understood by others: "We talk and act. That is already presupposed in every thing that I am saying" .15 There is after all a fact which is common to all human beings and which in our language may be expressed in the statement "We talk and act". To fill the gap between the unbridgeable systems ofbeliefs and their different backgrounds or fonns of life, there is the common way of human acting, whieh comprises assent to and dissent from the action and talk of others. Even when we imagine a Wittgensteinian language-game ofthe most primitive fonn, like the one with the four words for building-stones ('block', 'pillar', 'slab' , 'beam'), we have to presuppose that it is used among human beings and therefore that they may fail. So dissent is needed and assent is tacitly built into the actions. How else could they have learned the language? 230 RUDOLF HALLER Now, if the main point of relativism is that judgments presuppose some test procedure or criterion for deciding whether they are true or false, and that since there are different procedures, relying on different basic beliefs or systems of beliefs, no one of which may be called the fundamental one, the acceptance of one cannot be grounded on any other belief than those, which are presupposed by the acceptants: There are different language-games, different forms of life. So it seems we are bound to the relativity of our belief-systems, which rest on or may be embedded in different forms oflife. If someone does hold such a view, a view we mayaIso ascribe to Wittgenstein's conceptions ofit, it would be hard to deny that he or she is a relativist. Stephen Hilmy in his rich book on the later Wittgenstein,16 a book which uses so much material from the manuscripts and typescripts, has named a chapter 'Wittgensteinian Relativism and the Dynamic View of Language'. He describes the fundamental shift in Wittgenstein's philosophy of language as a shift away from an absolutistic and static to a relativistic and dynamic conception oflanguage. More important is even that Hilmy reminds the reader that Wittgenstein hirnself on different occasions uses the label "relativity theory" in regard to his own conception of language: "A step is necessary which is analogous to that of relativity theory" (MS 108, p. 270). So, when Wittgenstein does associate his own findings to that of a relativity theory, surely he has in mind the Einstein paradigm which changed the meaning of the key concepts and therefore of the systems by reshaping the question. The question now concerned also the relativity of the observer to the system itself. Abandoning the old idea of absolute concepts, like space, time and motion, was only a first important step. At least as important is the change in the general outlook covering concept-formation and theory-building. And, indeed, if the reference to language-games - the only way we act and use language - has a deeper role than just serving as a substitute for calculus, the role will be reflected in its use. Here I think the analogy with relativity theory comes in: If it is true what Wittgenstein says that everything is to be done in language, we do not have aresort to our inner life, be it a lingua mentalis, a private language, or the inner world of impressive emotions. The 'Nothing is hidden', indeed, is not a battle-cry but the signpost which points to the new relativistic point of view. Whether from this a theory emerges is another question or a question for another occasion. Universität Graz WAS WITTGENSTEIN A RELATIVIST? 231 NOTES 1 R. Rorty, •Pragmatism, Re1ati vism and Irrationalism' , in Proceedings andAddresses oftheAmerican Philosophical Association, 53, 1980, p. 727. 2 See W. V. O. Quine, Ontological Relativityand Other Essays, Co1umbia University Press, New York 1969, 19712; N. R. Hanson, Patterns of Discovery: An Inquiry into the Conceptual Foundations ofScience, Cambridge University Press, Cambridge 1958; T. S. Kuhn, The Structure of Scientific Revolutions, Chicago University Press, Chicago 1962; S. Toulmin, Human Understanding, I: The Evolution of Collective Understanding, Princeton University Press, Princeton 1971; P. K. Feyerabend,Against Method, New Left Books, London 1975; N. Goodman, Ways ofWorldmaking, Hackett Pub!. Co., Indianapolis-Cambridge 1978; H. Putnam, Reason, Truth and History, Cambridge University Press, Cambridge 1981; M. Hollis & S. Lukes (eds.), Rationality and Relativism, B. Blackwell, Oxford 1982; J. Meiland & M. Krausz (eds.),Relativism: Cognitive and Moral, Notre Dame University Press, Notre Dame 1982; P. K. Feyerabend, Farewell to Reason, Verso, London 1987; R. Egidi (ed.), La svolta relativistica nell'epistemologia contemporanea (with bibliography), F. Angeli, Milano 1988; W. Sauer, Aspekte und Probleme des BegrijJsrahmenRelativismus, Habilschrift, Graz (unpublished) 1989. J See Farewell to Reason, eit. 4 Wittgenstein 's Lectures. Cambridge, 1932-1935, From the Notes ofA. Ambrose & M. Macdonald, ed. by A. Ambrose, B. Blackwell, Oxford 1979, p. 224. 5 Ibid., p. 65 6 C. Wright, Wittgenstein on the Foundations ofMathematics, Duckworth, London 1980, p. 365. 7 L. Wittgenstein, Bemerkungen über die Farben. Remarks on Colour, ed. by G. E. M.Anscombe. Trans by L. L. McAlister & M. Schättle, B. Blackwell, Oxford 1977, III, § 317, p. 59. (I diverge from the English translation, which trans1ates the singular "Die Praxis gibt den Worten ihren Sinn" with "Practices give words theirmeaning".) 8 Philosophische Untersuchungen. PhilosophicalInvestigations (PI), ed. by G. E. M. Anscombe & R. Rhees. Trans. by G. E. M.Anscombe, B. Blackwell, Oxford 1953, I, § 42. 9 C. McGinn, Wittgenstein on Meaning: An Interpretation and Evaluation, B. Blackwell, Oxford 1984, p. 55. 10 M. B. Hintikka & J. Hintikka,lnvestigating Wittgenstein, B. Blackwell, Oxford 1986, p. 218. 11 L. Wittgenstein, MS 165, p. 97f, cit. by S. St. Hilmy, The Later Wittgenstein, B. Blackwell, Oxford 1987, p. 291f. 12 PI, 11, xi, p. 223. 13 Ibid. 14 Über Gewißheit. On Certainty, ed. by G. E. M.Anscombe & G. H. von Wright. Trans. by D. Paul & G. E. M. Anscombe, B. BlackweJl, Oxford 1969, § 94. 15 Wittgenstein 's Lectures on the Foundations of Mathematics, ed. by C. Diamond, Harvester Press, Hassocks-ComeJl University Press, Ithaca, N.Y. 1976, VI, § 17, p. 321. 16 Op. eit. PART V TOPICS IN THE PHILOSOPHY OF PSYCHOLOGY S. STEPHEN HILMY WITTGENSTEIN ON LANGUAGE, MIND AND MYTHOLOGY The notions oflanguage, mind and mythology are intertwined in certain passages in Wittgenstein's later writings. Even when these intertwined themes are not expressly identified, their intertwinement often seems to serve as a general thematic backdrop for many ofWittgenstein's later philosophical deliberations. However, there is a danger of misinterpreting the intertwinement of these themes in Wittgenstein's later philosophy as an indication that he shared the outlook of others who have tended to intertwine these themes in their writings. This danger is especially acute, because the intertwinement of these themes in the writings of others has tended to be intimately linked to an outlook which runs counter to some of the central critical thrusts of Wittgenstein' s later philosophy. This paper offers some reflections on the nature of the intertwinement of the themes of language, mind and mythology in Wittgenstein' s later writings, and examines the import of that intertwinement for an appreciation of his later philosophy, and in particular his fairly recently published volumes of remarks in the philosophy of psychology. Among those in whose writings one finds an interweaving ofthe themes of language, mind and mythology is Wittgenstein's contemporary Bertrand Russell. So Russell can conveniently serve both as aspringboard and as a foil. In the context of a discussion of the concept of 'meaning' , Russell once wrote that he believed "there is one method of acquiring knowledge, the method of science; and that all specially 'philosophical' methods serve only the purpose of concealing ignorance." His goal, he indicated, was to "endeavor to construct a theory of meaning after the model of scientific theories" . 1 Likewise, he regarded "thought as merely one among natural processes, and hope[d] that it may be explained some day in terms ofphysics".2 His ambition, as he explained in the context of a discussion of his view of 'belief, was to "reject everything mythological" and instead present a "theory" of belief which is empirically verifiable. 3 This is certainly an alluring outlook. It was sufficiently alluring and adaptable that it remained relatively undisturbed as Russell periodically shifted in and out of materialist, phenomenalist and behaviourist phases during his protean career. Arecent form of this outlook, although rejecting the specific theories of the mind that Russell propounded, has gone so far as to suggest 235 R. Egidi (ed.), Wittgenstein: Mind and Language, 235-248. © 1995 Kluwer Academic Publishers. 236 S. STEPHEN HILMY that the "primitive psychological taxonomy in ordinary language" ought to be eliminated and in its stead we ought to replace our ordinary psychological concepts with a "neuro-psychological taxonomy" which would "allow us to better describe, conceive and introspectively apprehend one's inner life".4 The psychological concepts of ordinary language themse1ves constitute, as it were, bad scientific theories, myths if you will, about the nature of our inner life, and we ought to replace this nexus of pseudo-scientific theories about 'mental states and processes' with modern neuro-psychological theories and vocabulary about the states of our brain and nervous system. For reasons which 1 will explain presently, 1 will call this overall outlook, shared by Russell and many others, the 'taxonomie view' of psychological concepts. One cannot deny the hold that this outlook has over uso It would seem that surely science will accomplish with respect to man's psychologieal constitution what it has accomplished with respect to the forces of the physical world around us - as it were, debunking mythologies ab out the 'mind' just as it has debunked mythologies about the forces of nature. The question arises as to whether Wittgenstein subscribed to a form of the taxonomic view of psychological concepts, and what implications that would have for an appreciation of the status of his last writings and remarks on the philosophy of psychology. If he accepted the taxonomie view, it would seem that the topics which are the subject matter ofhis last writings in the philosophy of psychology really ought to be the province of empirical psychology, not philosophy. And yet, if he rejected the taxonomie view of psychological concepts, that would seem to entail that his remarks in the philosophy of psychology serve, as Russell might put it, "only the purpose of concealing ignorance", because Wittgenstein, it would seem, employs a philosophieal method whieh is futily in competition with the theories and discoveries of the empirical sciences. It is important for appreciating Wittgenstein's last writings in the philosophy of psychology that one both recognize that he rejected what 1 call the 'taxonomie view' of psychologieal concepts, and, even more importantly, understand why he rejected that view. An understanding of why he rejected the taxonomic view would pave the way for an appreciation of his last writings in the philosophy of psychology as categorically different from but yet at least potentially of constructive value for empirical psychology. There are, though, some striking, but misleading, paralleIs between what 1 have called the taxonomic view of psychological concepts and general remarks Wittgenstein hirnself has made about the nature of the philosophical problems that concerned hirn. For instance, he often wrote of there being a WITTGENSTEIN ON LANGUAGE, MIND AND MYTHOLOGY 237 'mythology' in language, and of language presenting us with false pictures which hold us captive. In his chapter on "Philosophy" in the 'Big Typescript' he suggests that "In our language lies a whole mythology".5 In that same typeseript one also finds aversion of the similar familiar remark, whieh has since found its way into Zettel, namely, the suggestion that "In philosophy one is in constant danger of producing a mythology of symbolism, or of psychology [mental processes]. Instead of simply saying what anyone knows and must admit".6 More often than not, sueh remarks are made in the context of his attaeks on mentalism, and it is very tempting to read sueh remarks as involving a form of the taxonomie view. It is quite possible that very early in his career Wittgenstein may have entertained an outlook not very far removed from a form of the taxonomie view, one of the symptoms of which is the attitude that ordinary language is defective and in need of reform. In one of his later notebooks he seems to admit that as he first expressed hirnself, "it seemed as though ordinary forms of expression wouldn 't really accomplish what we wanted of them. As if they were an (old) form of historical rubbish ...that would be replaced by something gi.HlCJ, modern".7 However, it is clear from his last writings in the philosophy of psyehology, as it is clear in his later writings in general, that the goal of refining or reforming language was abandoned by hirn. This is obvious from Philosophical Investigations, § 98, where he expressly indicates that every sentence in our language is 'in order as it is'. That is to say, we are not striving after an ideal, as if our ordinary vague sentences had not yet got a quite unexceptionable sense, and a perfect language awaited construction by us. s This theme is eehoed in the notebooks from which Remarks on the Philosophy oj Psychology was drawn, where in the context of a discussion of the concept of 'seeing' Wittgenstein advises that we "must just take the concept of 'seeing' as one finds it; not want to refine it. .. Because it is not our problem to change it, to introduee (as seienee does) [a sense] adapted for some purpose or other, but rather to understand it; in order not to produce a false conception of it".9 He goes on to explain that we need to "accept the familiar language-game, and to note ja/se explanations of the matter as false. The primitive language-game we originally learned needs no justification, and false attempts at justification, which foree themselves on us, need to be rejected". \0 Those to whom Wittgenstein wishes to direet his advice clearly are not empirical psychologists per se, since there is implicit recognition here of the legitimaey of the psyehologist's task in its own right, as distinguished from the concerns of philosophy. Rather, Wittgenstein is directing his advice 238 S. STEPHEN HILMY to phi!osophers, by suggesting that the problems that ought to be the concern of philosophers, as he envisions their proper task, are confusions about our ordinary concepts, not confusions of language itself, which he claims is "in order as it is". At the root ofthis, lies both the basis for Wittgenstein's repudiation ofthe 'taxonomie view' ofpsychological concepts, and the possibility of constructively construing the potential relationship between his last writings and the tasks of empirical psychology. It was not Wittgenstein's view that ordinary language is in some sense an inherently inadequate and flawed repository of false theories, as it were, a repository of myths or false pictures about psychological processes. Quite to the contrary, the mythology and the 'false pietures' of which he speaks are misconceptions or confusions conjured through our misunderstandings of the actual function of our expressions. Thus the mythology would consist essentially of our confusions about language, not confusions 0/ language - and this would be the case even if our confusions about language were spawned by misleading but inessential analogies within language. Misleading paralIeis in language may give rise to our confusions, but it is the latter confusions which are the problem, not language itself. This is why Wittgenstein can consistentiy claim that philosophy is a struggle against the bewitchment of our intelligence by means of language, 11 and yet without contradiction also claim that language is in order as it iso Both of these claims are made in the Philosophicallnvestigations (§ 109 and § 98 respectively), and indeed they were originally drafted on the same page in his notebooks. 12 The myths and false pictures are thus constructs of our misunderstanding of our language, not defects in language per se. As Wittgenstein quite explicitly put it in one of his manuscripts: "philosophy struggles against myths [Mythen], pictorial tendencies in our understanding" [my emphasis]Y This is the real import ofhis frequent references to the "false pictures" and "myths" conjured by language - it is an indictment of our understanding, not an indictment of language itself. However, simply recognizing certain myths as by-products of our misunderstandings or confusions ab out our language is not enough to avoid the risk of falling prey to what I have called the taxonomie view. Even if one acknowledges these myths as misunderstandings of our language, one might nevertheless make the further mistake of simplistically dismissing those myths as the false theories of mentalism, and arguing that our psychological concepts do not function as reports of mythical mental events at all, but rather merely as reports of our behaviour or our tendencies to behave in certain ways. WITTGENSTEIN ON LANGUAGE, MIND AND MYTHOLOGY 239 It is tempting to interpret Wittgenstein as having held such a view, particularly if one is inclined, as many have been,14 to interpret hirn as advocating a form ofbehaviourism - in which case he would be advocating a behaviouristic understanding of our psychological concepts in place of a mentalistic one. The exegetical ease with which this interpretive move can be made is perhaps exactly why so many have tended to interpret Wittgenstein as a behaviourist, and why, in spite of his apparent protestations to the contrary in the Investigations, it is so easy for it to seem that he is advocating a behaviourist position. I am aUuding here to the Investigations, § 307, where Wittgenstein has an imaginary interlocutor ask hirn: "Are you not really a behaviourist in disguise? Aren't you at bottom really saying that everything except human behaviour is a fiction?" And Wittgenstein replies: "If I do speak of a fiction, then it is of a grammatical fiction." The import of Investigations § 307 is more explicit in the broader context of Wittgenstein's initial draft of the remark: But aren't you really only a behaviourist in disguise? For you say that nothing stands behind the expression [iußerung] of experience. Aren't you at bottom really saying that everything except behaviour is a fiction? So do 1 therefore believe that we don't really feel pain, but rather only make faces?! But the fiction is the object behind the expression [iußerung]. It is a fiction that our words in order to be meaningful must refer to asomething that 1 nevertheless can exhibit to myself even if not to someone else. (Grammatical fiction.) The statement "Nothing stands behind the expression [iußerung] of experience" is a grammatical one - it doesn 't therefore claim that we experience nothing.... My critique consists in that I show bring out that the whole (common, primitive) view of the ftlllctioning function of words (in the language game I/in the use of language/I) is too narrow. IS By calling a statement "grammatical", Wittgenstein means that it is a statement about the use or function of a linguistic term. He is not denying the existence of mental entities per se. Thus when he makes the grammatical claim that 'nothing stands behind a given expression of experience', he is not denying that there is anything going on inside us as we utter the expression, rather he is making a comment on the function or use of the 'expression of experience', name1y, that it does notfunction to report a hidden psychic thing or object behind it. 1 have argued elsewhere,16 that an interpreter, who concludes from Wittgenstein's repudiation of a mentalistic explanation of the 'meaning' of our linguistic expressions that Wittgenstein was therefore a behaviourist and was claiming that all 'psychological expressions' report (or ascribe) behaviour, may have in effect fallen under the spell of an illusion or general fiction about the nature of language and 'meaning' that Wittgenstein was bent on dispelling. 240 S. STEPHEN HILMY Ifbehaviourism were Wittgenstein's position, his view would be quite eonsistent with the taxonomie view of psyehologieal eoneepts, beeause he would simply be replaeing the mentalistie taxonomy whieh language allegedly eonjures, by a behaviouristic taxonomy. However, that was not his position at all, at least not in his later philosophy. Indeed at the very outset of the manuseripts from which Remarlcs on the Philosophy 0/ Psyehology was compiled Wittgenstein at one point expressly indicated: "I am attacking the pneumatic [mentalistic] view... The antithesis ofthe pneumatic view is the behaviouristic one, and they are both bad [sehleeht]"Y It may seem odd to group together such diverse standpoints as mentalism and behaviourism, but this is possible because both mentalism and behaviourism, and for that matter also materialism and phenomenalism, are all instances of a taxonomic approach to psychological concepts, and it is the taxonomic view itself which Wittgenstein would consider fundamentally flawed. The affinity between mentalism and behaviourism is clearly illustrated, for instance, in the philosophy of Russell. During the course of his career, Russell seemed to shift with relative ease from a Humean, phenomenalist account of 'meaning' to a behaviouristic account. The ease of that transition was indicative of the affinity lying in not only the causal but also the taxonomie nature of both the mentalistic and behavioural explanatory paradigms (to both of which explanatory paradigms Wittgenstein was critically reacting). Russell's mentalistic explanation of meaning was encapsulated fairly early in his career in the associationist principle he called the "law of mnemic causation": If a complex stimulus A has caused a complex reaction B in an organism, the occurrence of apart of A on a future occasion tends to cause the whole reaction B.18 Unpacking this explanatory principle as a mentalistie explanation of 'meaning' is quite straightforward. 19 This is accomplished simply by interpreting the 'complex reaction B' in mentalistic terms. If one is repeatedly subjected to a complex stimulus A which is composed of a verbal stimulus (a word) and an 'object' or thing along with the word, and one's eomplex mental reaction (B) to the complex stimulus is to have the roughly simultaneous sensations of a specific 'word' and a specific 'objeet', then, according to the above explanatory law, if at a later time one were to be subjected to only part of complex stimulus A (e.g. only the word), one's reaction should be the whole of complex reaction B, namely, both asensation of the word and a copy or mental image of the sensation of the object associated with the word. Thus the 'meaning' of a word would be the effect the word has on our mind, namely WITTGENSTEIN ON LANGUAGE, MIND AND MYTHOLOGY 241 the image or 'mental object' that it conjures, and the latter would be a causal by-product of our history of conditioning. Such an explanation gives not only a causal, but also a taxonomie aeeount of the notion of meaning. But the very same sort of eausal and taxonomie explanatory paradigm ean be adapted to a purely behavioural account of meaning. In fact, Russell did precisely that when later in his eareer he shifted from mentalism to behaviourism: People used to speak of'association ofideas', but now-a-days association is rather between bodily movements. The essential phenomenon is what Dr. Watson calls a 'leamed reaction'. Two stimuli A and B occur together, and Beauses a bodily movement C. Later on, A may eause C, though it previously had no tendeney to do so. All words are 'learned reactions' in this sense. There is no need to postulate a 'mental' intermediary between the stimulus and the reaetion.20 Russell simply replaeed 'mental reaetion' in his earlier aecount of 'meaning' by 'behavioural reaetion'. One' s taxonomy of 'things' has simply shifted from a mentalistie taxonomy to a behaviouristic one. The causal eharaeter of the explanation has of course remained constant, but so has the essential taxonomie nature ofthe explanation of 'meaning' also remained unehanged. Nor would the taxonomie nature of the explanatory paradigm be essentially altered by ineluding under the rubrie 'behavioural reaction' the neurophysiologieal effeets of verbal stimuli rather than solely the gross behavioural effects (the 'aetions' elicited). The taxonomie view is eharaeterized not so mueh by a theory about the nature of our 'inner life', but rather by a fundamental supposition about the use of our vocabulary of so-called 'psyehologieal eoneepts'. The fundamental supposition is that our ordinary psyehologieal voeabulary funetions as a taxonomy of things or processes. The various taxonomie standpoints differ only in what they claim the taxonomy is or should be, not in whether the primary funetion of our psychological eoncepts is to present us with a taxonomy of things or processes in the first place, whether mental, physieal or phenomenal. Viewed in this light, mentalism, phenomenalism, behaviourism and materialism are all birds ofa feather (ifyou will pardon the taxonomie idiom), in that they all eonstitute taxonomie standpoints vis-a-vis an interpretation of the funetion of our psychological voeabulary. It is for this reason that it was possible for Russell's overall taxonomie outlook to remain essentially undisturbed when he shifted between materialism, phenomenalism and behaviourism. And likewise the basie taxonomie outlook remains similarly undisturbed whether one simply advocates that the old so-called mentalistie taxonomy ean be trans la ted into a new materialistic, or phenomenalistic or behaviouristic 242 S. STEPHEN HILMY taxonomy, or whether one instead advocates the more extreme view that we ought eompletely to eliminate the old allegedly mentalistic taxonomy and start afresh with a new neuro-psychological taxonomie voeabulary. If mentalism stems from a misunderstanding of our vocabulary as taxonomie, then by rejeeting our psychological voeabulary beeause it allegedly presents us with a false mentalistie mythology, and advoeating replaeing that voeabulary with a neuro-psyehologieal voeabulary, one eommits two errors (aeeording to Wittgenstein): 1) the error ofmiseonstruing our psyehologieal voeabulary as funetioning as a taxonomy of mental events or proeesses, and 2) the error of perpetuating, by means of a materialistie substitute, the underlying taxonomie myth or eonfusion that gave rise to mentalism in the first plaee. A materialistie or behaviouristie taxonomie outlook and voeabulary would not be a eure for the so-ealled mentalistie mythology of language, but rather, sinee mentalism is simply an instanee of the taxonomie outlook, behaviourism and materialism would simply be perpetuations of the taxonomie myth which underlies mentalism - namely, the simplistie taxonomie view that the primary funetion of our psyehologieal voeabulary is to name or report things or proeesses, whether of a behavioural, mental, phenomenal or material nature. Although Wittgenstein did not use the expression "taxonomie myth", it ean safely be said that one of the eentral eritieal thrusts of his later philosophy was the attempt to debunk sueh a myth about the funetion of our psyehological voeabulary. Soon after his return to philosophy in the late 1920s, he wrote the following illuminating remark which bears on the nature of the mistake underlying the taxonomie myth: Ifit is charged that language can express everything by means ofsubstantives, adjectives and verbs, then we must say that it is then in any case necessary to differentiate between wholly different kinds of substantives (etc.), for different grammatical rules apply for them ... Thereby it is evident that their substantive character is only a superficiality [Außerlichkeit] and that we are actually dealing with wholly diverse families ofwords. The word-family is (first/now) detennined by alt grammatical rules which apply to a word. And thus viewed our language has a countless number of different kinds ofwords.2' The last two deeades ofWittgenstein's philosophical aetivity ean without distortion be eharaeterized broadly as an exploration of the rieh diversity and multiplicity of grammars that eonstitute our language, but whieh diversity and multiplieity tend to be coneealed by the misleading simplieity and superfieiality of the surfaee grammar of our language. And it is a failure to grasp the grammatical diversity and riehness of our language that is at the root of the 'mythology' oflanguage against which Wittgenstein was struggling. Indeed WITTGENSTEIN ON LANGUAGE, MIND AND MYTHOLOGY 243 in his chapter on "Philosophy" in the 'Big Typescript', almost immediately after identifying the mythology of language as a fundamental problem, Wittgenstein went on to explain the way in which language gives rise to such mythology: "The primitive forms of our language - substantive, adjective and verb - show a simple picture to which it tries to make everything conform".22 The mi staken, simplistic picture to which the primitive forms of our language give rise is perhaps most clearly characterised in the 'BIue Book', where Wittgenstein indicates that: The mi stake we are liable to make could be expressed thus: We are looking for the use of a sign, but we look for it as though it were an object co-existing with the sign. (One of the reasons for this mistake is again that we are looking for a 'thing corresponding to a substantive. ')23 Furthermore, the way in which this fundamental mistake manifests itself more specifically in the domain ofthe philosophy ofpsychology is, as Wittgenstein explained a few years later in one of his notebooks: The philosophical confusion conceming the idea of thinking, together with the problems of psychology, is presented to us by the pieture of a hidden (invisible) mechanism ... It seems: 'believing' describes something which occurs with the sentence - as 'digesting,' something which occurs with the meal. One could then understand the belief if one knew what really went on at the time. One would then have analysed the 'process ofbelief.' ... It is absolutely not unexplored processes ofbeliefthat interest us, but rather the use of our well-known processes ofbelief, e.g. ofthe expression ofthe sentence 'I believe .. .'24 The assumption that our psychological concepts function as names of hidden processes or events, rests on the simplistic assumption that the meaning of a word is an object, thing or process to whieh it refers. From Wittgenstein's perspective, the issue ofwhether the thing named is amental process, a physical process or a behavioural process is really a derivative controversy, and one which is played out by philosophers from within the taxonomie speIl that the surface grammar of language has allegedly thrown over uso To use Wittgenstein's own illustration, the surface grammar would suggest that believing is something which occurs with the sentence, just as digesting is something which occurs with the meal; and similarly for a host of other psychological concepts. The taxonomic view seems to fall prey to just this sort of surface grammatical analogy. Let me illustrate this in more detail with respect to the concept of 'expecting'. A mentalistic taxonomie view might explain the meaning of expressions of the form 'I expect. . .' as that of reporting the occurrence of a 244 S. STEPHEN HILMY hidden mental process accompanying such an utterance (for example, perhaps the occurrence ofhidden phenomenal pictures or images in one's 'mind'). On the other hand, a behaviouristic taxonomie view might explain an expression ofthe form 'I expect. . .' as areport ab out how one is behaving or how one is disposed to behave (perhaps one's tendency to pace, fidget, make tea, carry an umbrella, and so on) when one utters such an expression. (One could if one wanted also throw in here neuro-physiological behaviour.) Wittgenstein denies such explanations, but in doing so he is neither denying 'the mental' (for example that we may have mental images that accompany speech), nor denying that we may pace, fidget and make tea (or be disposed to doing so) when saying we 'expect' something to happen. Rather Wittgenstein is denying that, in the context of the conventional language games in which the expression 'I expect. . .' is embedded, the ftmction (the use, the logic, the grammar) of the expression is to report one's own mental or behavioural processes or states. As for what specific sort of function such an expression does serve, Wittgenstein hints: IfI say "I expect ... " -am 1 remarking that the situation, my actions, thoughts etc. are those of expectancy of this event; or are the words "I expect ... " part of the process of expecting? In certain circumstances these words will mean (will be replaceable by) "I believe suchand-such will occur." Sometimes also: "Be prepared for this to happen ...."25 I have indicated elsewhere,26 that Wittgenstein can be taken in this remark to be suggesting that the expression 'I expect. . .' is itselfthe expecting, and we might ascertain this by noticing that the utterance 'I expect. . .' in certain circumstances does the same sort of job as (means, is replaceable by) the utterance 'I believe such-and-such will occur' or 'Be prepared for this to happen.' That an 'expression of expectation' might in certain circumstances involve the same type oflinguistic move as an 'expression ofbelief' does not of course help us very much, if one is inclined to view an 'expression of belief' as areport about one's own mental events or behaviour. But to recognize that in certain circumstances the expression 'Be prepared for this to happen' plays the same roIe, serves the same duty, as 'I expect.. .' is more directly heIpfuI in that it is an indication that the linguistic job performed by 'I expect. . .' is not that of reporting about the occurrence of mental or behavioural processes but rather a matter of offering some assurance of the likelihood of a predicted state of affairs (and perhaps a matter of offering a waming to that effect). To offer assurance or to forewarn that something will happen obviousIy has a quite different function from that of reporting what one WITTGENSTEIN ON LANGUAGE, MIND AND MYTHOLOGY 245 is pieturing in one's mind's eye, one's nervous fidgeting, one's blood pressure, or one's disposition to do any of these. Thus the claim is that in some eireumstanees, although the expression 'I expect. .. ' may seem to be areport about a psyehologieal or behavioural aceompaniment to one's utteranee, it in faet offers assurance or forewarns - and in this sense Wittgenstein suggests that the expression of expeetation is the expeetation, for, if the language game aetually being played were that of offering assuranee or forewarning, elearly we would call an 'expression of assuranee' an 'assuranee' and an 'expression of warning' a 'warning'. Just what the meaning (the logie or grammar) of any given linguistie expression is will depend on the partieular language game or system of eommunieation in whieh it is embedded. It was Wittgenstein's view that this is the ease for any linguistie expression, even those that might be called 'psyehological expressions ' (e.g. of the form 'I expect.. .', 'I believe. . .', 'I hope. . .', and so on). Upon investigating the logie of psyehologieal expressions in the eontext of the multiplicity of language games which are constitutive of their sense, he was led to the conclusion that it is an erroneous and simplistie view of language to assurne that the function of such expressions is that of naming or reporting psyehologieal or behavioural meehanisms, proeesses or events. The taxonomie myth is preeisely the simplistie view of the grammar or funetion of our psyehologieal eoneepts as solely a matter of naming or reporting the oecurrenee of proeesses and events. Mentalism, behaviourism, materialism and phenomenalism bieker about the nature of what is reported, but they all share this grammatieal presumption. The train of thought that 'if the meaning of a linguistie expression is not an 'inner' mental thing associated with the expression, it must be an inner physieal thing or perhaps an assoeiated 'outer' behavioural thing', is a form of reasoning that remains under and is held eaptive by the grammatical illusion, the taxonomie myth that all words serve (alllanguage funetions) to name or report something (whether objeet, proeess or state, whether mental or behavioural, inner or outer). An essential part of Wittgenstein's eritieism of, for example, both mentalism and behaviourism is that the grammar or 'meaning' (the function) of our so-ealled 'psyehologieal expressions' in the language games in whieh they are embedded may not be that of reporting (or naming) an assoeiated something at all (whether a hidden 'inner' something, or an 'outer' behavioural one). Wittgenstein's remarks in the philosophy of psyehology during the last yearS of his life are a sustained effort to debunk this taxonomie myth by probing the rieh and diverse grammars of our psyehologieal eoneepts. As he 246 S. STEPHEN HILMY explained in a revealing note that did not find its way into the published versions of his last writings: In order to be able to get a surview ofthese concepts, you must compare them with something other than what their surface grammar [Obe1j1ächengrammatik] suggests to you.... The concepts are disguised. The verbs and substantives ofpsychology (seeing, hearing, fearing, willing, hoping, believing; dread, will, fear, belief, wish) are so very misleading. One must again and again return to the concepts to describe their conduct and its occasions, in order not to be misled about the nature of the language-games [Sprachspiele] by the surface of grammar. 27 The terrain of language whieh he explored so extensively during his last years is eertainly very eomplex, and his published remarks in the philosophy of psyehology are far from finished produets, but this may not neeessarily detraet from what he perhaps would have wished to be one ofthe effeets ofhis last writings in the philosophy of psyehology, namely, that the more closely seholars serutinize his exploration of the grammatical terrain of our psyehologieal eoneepts, the more they would eome to appreeiate the riehness and diversity of that terrain. And with that appreciation, perhaps would eome the abandonment of the simplistie though bewitehing taxonomie eategories and mind-sets that have traditionally vied for dominanee on the philosophieal scene. I indieated at the outset that there is the potential for a positive relationship between Wittgenstein's last writings and the tasks of empirieal psyehology. Seienee's task is an empirie al one, yet its questions, partieularly in the area of empirie al psyehology, are to a great extent spawned by and eouehed in terms of those very psyehologieal eoneepts whose eonfusing grammar was the subject of Wittgenstein's investigations. Consequently, seienee itself is to some extent vulnerable to the eoneeptual eonfusions whieh our psyehologieal eoneepts engender, and thus a clear surview of the logieal eontours of our psyehologieal eoneepts may be useful to the empirieal psyehologist as weIl as to the philosopher. To this extent, Wittgenstein's efforts to clarify the grammar ofthose eoneepts in order to debunk the grammatieal myths they engender ereate at least the possibility for eonstruetively eonstruing the potential relationship between his last writings in the philosophy of psyehology and the tasks of empirie al psyehology - in spite of the methodologieal ehasm whieh seems to separate his grammatieal investigations from the task of empirieally investigating man' s psyeho-physieal nature. Austin, Texas WITTGENSTEIN ON LANGUAGE, MIND AND MYTHOLOGY 247 NOTES 1 B. Russe1\, Symposium on 'The Meaning ofMeaning', Mind, 29 (Oct. 1920), p. 401. Ibid., p. 400. B. Russell, 'On Propositions: What They Are and How They Mean', Proceedings ojtheAristotelian Society, 1919, Suppl. Vol. H, pp. 27-28. 4 P. M. Churchland, Matter and Consciousness: A Contemporary Introduction to the Philosophy ojMind, MIT Press, Cambridge Mass. 1988, pp. 179-180. 5 TS 213, p. 434. All references to Wittgenstein's papers are to the Cornell microfilm version, 'The Wittgenstein Papers' , Corne1\ University Microfilms, Ithaca, N. Y. The system of reference is that of G. H. v. Wright, 'The Wittgenstein Papers', in Wittgenstein, Oxford, B. Blackwell 1982. Within quoted material, double obliques mark offterms or phrases that Wittgenstein entered as possible alternate wording. Broken or wavy underlines (reproduced here simply as broken underlines) were used by Wittgenstein to indicate that he was not sure ab out the appropriateness of a word or phrase. 6 TS 213, p. 145; see Zettel (Z), ed. by G. E. M.Anscombe & G. H. von Wright. Trans. by G. E. M. Anscombe, B. Blackwell, Oxford 1967, § 211. 7 MS 160, pp. 43-44; see my discussion in S. Hilmy, The Later Wittgenstein, B. Blackwell, Oxford 1987, nn. 174-5 and pp. 55 tT. 8 L. Wittgenstein, Philosophische Untersuchungen. Philosophicaiinvestigations (PI), ed. by G. E. M. Anscombe & R. Rhees. Trans. by G. E. M. Anscombe, B. Blackwell, Oxford 1953. 9 MS 137, p. 29; see L. Wittgenstein, Vermischte Bemerkungen. Culture and Value (CV), ed. by G. H. von Wright in collaboration with H. Nyman. Trans. by P. Winch, University ofChicago Press, Chicago 1980, p. 44. 10 MS 137, p. 30; published in Bemerkungen iiber die Philosophie der Psychologie. Remarks on the Philosophy ojPsychology, Vol. II (RPP II), ed. by G. H. von Wright & H. Nyman. Trans. C. G. Luckhardt & M. A. E. Aue, B. Blackwell, Oxford 1980, § 453. II PI I, § 109. 12 MS 152, p. 94. 13 MS 158, p. 54. 14 See, for example, G. Bergmann, 'The Glory and the Misery ofLudwig Wittgenstein' , in Essays on Wittgenstein's Tractatus, ed. by L Copi & R. Beard, Hafner Press, NewYork 1973, pp. 354-355; C. Mundle,A Critique ojLinguistic Philosophy, Glover & Blair Limited, London 1979, p. 199; G. Bowie, 'Wittgenstein's "Behaviorism" and Psycho-physical Laws', in Wittgenstein and his Impact on Contemporary Thought: Proceedings 01 the Second International Wittgenstein Symposium, ed. by E. Leinfellner et al., Hölder-Pichler-Tempsky, Vienna 1978, p. 445. 15 MS 124, pp. 5--6. See MS 161, pp. 79-80. 16 S. S. Hilmy, 'Wittgenstein and Behaviourism', Grazer Philosophische Studien: Wittgenstein in Focus -Im Brennpunkt: Wittgenstein, ed. by B. McGuinness & R. Haller, Rodopi, AmsterdamAtlanta,Ga, 33-34, 1989, pp. 335-352. 17 MS 130, p. 3. 18 B. RusselI, The Analysis ojMind, G. Allen & Unwin, London 1921, p. 86. 19 See my discussion in S. S. Hilmy, The Later Wittgenstein, cit., p. 116. 20 B. RusselI, 'The Meaning ofMeaning', The Dial (August 1926), p. 116. See Carnap's adoption of a form of this causal explanatory principle, R. Carnap, 'Psychology in Physical Language', in Logical Positivism, ed. by A. J. Ayer, The Free Press, London 1963, pp. 196-197. 21 MS 107, p. 210. 2 3 248 S. STEPHEN HILMY TS 213, p. 434. L. Wittgenstein, The Blue and Brown Booles, ed. and with a Prefaee by R. Rhees, B. Blaekwell, Oxford 1958; Harper& Row, NewYork 19652 , p. 5. 24 MS 117, pp. 127-129. For a more extensive diseussion ofthis passage in the broadereontext of Wittgenstein's later view ofthe nature oflanguage, see S. S. Hilmy, The Later Wittgenstein, cit., pp. 163 ff. 25 Z, § 65. 26 S. S. Hilmy, The Later Wittgenstein, eit., pp. 146 ff. 27 MS 134, pp. 126-129. 22 23 JOACHIM SCHULTE EMOTION: REMARKS ON WITTGENSTEIN AND WILLIAM JAMES In a remarkable passage ofhis Principles 0/Psychology William James writes: My theory [ ... ] is that bodily changesfollow directly the perception ofthe excitingfact, and that our feeling ofthe same changes as they occur IS the emotion. Common-sense says, we lose our fortune, are sorry and weep; we meet a bear, are frightened and run; we are insulted by a riyal, are angry and strike. The hypothesis here to be defended says that this order of sequence is incorrect, that the one mental state is not immediately induced by the other, that the bodily manifestations must first be interposed between, and that the more rational statement is that we feel sorry because we cry, angry because we strike, afraid because we tremble, and not that we cry, strike, or tremble because we are sorry, angry, or fearful, as the case may be.! Then James comes to what he calls 'the vital point' ofhis whole theory and mentions what has since come to be called a thought-experiment: /f we fancy some strong emotion, and then try to abstract from our consciousness ofit all the feelings of its bodily symptoms, we find we have nothing lejt behind, no 'mind-stuff out of which the emotion can be constituted and that a cold and neutral state of intellectual perception is all that remains. It is true that, although most people when asked say that their introspection verifies this statement, some pers ist in saying theirs does not. Many cannot be made to understand the question. When you beg them to imagine away every feeling oflaughter and oftendency to laugh from their consciousness ofthe ludicrousness of an object, and then to tell you what the feeling of its ludicrousness would be like, whether it be anything more than the perception that the object belongs to the c1ass 'funny' , they persist in replying that the thing proposed is a physical impossibility, and that they always must laugh if they see a funny object. Of course the task proposed is not the practical one of seeing a ludicrous object and annihilating one's tencency to laugh. It is the purely speculative one of subtracting certain elements offeeling from an emotional state supposed to exist in its fulness, and saying what the residual elements are. [... ] What kind of an emotion of fear would be left ifthe feeling neither of quickened heart-beats nor of shallow breathing, neither oftrembling lips nor ofweakened limbs, neither of goose-flesh nor ofvisceral stirrings, were present, it is quite impossible for me to think. Can one fancy the state ofrage and picture no ebullition in the chest, no flushing ofthe face, no dilatation ofthe nostrils, no clenching ofthe teeth, no impulse to vigorous action, but in their stead !imp muscles, calm breathing, and a pI acid face?The present writer, for one, certainly cannot. [... ] Every passion in turn teUs the same story. A purely disembodied human emotion is a nonentity. I do not say that it is a contradiction in the nature of things, or that pure spirits are necessarily condernned to cold intellectuallives; but I say that for us, emotion dissociated from all bodily feeling is inconceivable.2 Wittgenstein mentions this theory and this thought-experiment from William James more than once in his later writings on the philosophy of psychology. Of particular interest, however, is an earlier passage from the 'Brown Book' , 249 R. Egidi (ed.), Wittgenstein: Mind and Language, 249-262. © All Rights Reserved. 1995 Kluwer Academic Publishers. 250 JOACHIM SCHULTE which was dictated in the academic year 1934-35. There, in the context of discussing the question of the correctness of our translations of foreign, or radically foreign, words, Wittgenstein writes: You will find that the justifications for calling something an expression of doubt, conviction, etc., largely, though of course not wholly, consist in descriptions of gestures, the play offacial expressions, and even the tone ofvoice. Remember at this point that the personal experiences of an emotion must in part be strictly localized experiences; for ifI frown in anger I feel the muscular tension ofthe frown in my forehead, and ifI weep, the sensations around my eyes are obviously part, and an important part, ofwhat I fee!. This is, I think, what William James meant when he said that a man doesn 't cry because he is sad but that he is sad because he cries. The reason why this point is often not understood, is that we think ofthe utterance of an emotion as though it wefe some artificial device to let others know that we have it. Now there is no sharp line between such 'artificial devices' and what one might call the natural expressions of emotion. Cf. in this respect: a) weeping, b) raising one's voice when one is angry, c) writing an angry letter, d) ringing the bell for a servant you wish to scold. 3 This passage contains several themes which keep cropping up in Wittgenstein's laterphilosophy ofpsychology and, interestingly enough, even the combination of themes is typical of his later thought. The first point concerns the fact that we justify our ascriptions of psychological states to people by referring to facial expressions, gestures, and the tone of voice. This is important, not merely because these expressions often are our only grounds for attributing certain states to others, but also because often they are more reliable than what people say. After all, many of these expressions are what we call involuntary ones: they are difficult or impossible to suppress or imitate, and frequently our attempts at suppressing or imitating them are not successful, so that others can perfect1y easily and with confidence tell how we feel. The second point is a very involved one. Here Wittgenstein talks about our experiences of emotions and the questions arising in this context are largely coloured by problems about observing one's own state of mind. But here Wittgenstein seems to intend something narrower than what is suggested by the word 'observe' . When he speaks of experiencing a certain emotion he seems to mean that the occurrence of some emotions tends to go with certain typical feelings, and that this way of experiencing OUf emotions is possible only if these feelings concern a well-circumscribed part of our bodies. Wittgenstein mentions the frown which is feIt in my forehead and the weeping feIt in the area around my eyes. But this does not seem to apply to all kinds of emotions. Grief, for example, is not feIt in a particular spot of one's body. Certain more specific manifestations, on the other hand, are feIt in this way. Thus it may be the case that, being very sad, I weep; and this of course is closely related to my sadness and at the same time something that I do feel. EMOTION: REMARKS ON WITTGENSTEIN AND JAMES 251 Tears, however, can also be shed from joy, so the mere fact of crying is not very indicative ofthe emotion in question. But in the case ofjoy, at any rate, there normally are other typical feelings in certain parts of our face which tend to correspond to natural and thus reliable outer expressions of this emotion. The third point mentioned by Wittgenstein again regards the question of expressing an emotion. He says that the relation between feeling or experiencing an emotion and a certain place at which it, or part of it, is feit is often overlooked because 'we think ofthe utterance ofan emotion as though it were some artificial device to let others know that we have it' . Now this connection may not appear altogether obvious. It consists, I suppose, in this: that the typicallocalized feeling going with a certain emotion is at the same time and through the good offices of our gestures, weepings, groans, and stammerings related to that which other people perceive as expressions or utterances of that emotion. And Wittgenstein wants to stress that these expressions or utterances tend to be natural, not artificial, ones. What appears as a way of communicating a certain emotion normally is not primarily intended as such. When I sit at the breakfast-table and notice that my father's nose is twitching, his fingers jerking, his brows knit, and his face purple, I literally see that he is angry. These unmistakable signs ofhis anger are not artificial devices he uses to tell me that he is angry; they are his natural, even instinctive, expressions of anger which however do tell me more clearly than words could that he is angry. Of course, he may, on another occasion, try to use these signs or symptoms as devices for giving me the impression that he is angry. This may work, especially ifhe manages to 'feei' part ofwhat he is attempting to convey. But it is nevertheless a very different situation. However, the roles of these expressions of our emotions are not always clearly separable; they tend to shade into each other. This is why Wittgenstein asks us to consider four different cases where our expressions are more or less natural or, as the case may be, more or less artificial. The first case is that of weeping which, I take it, is a paradigmatic natural expressIon. The second case is that of raising one' s voice when one is angry. Normally this would also be a fairly natural expression which, however, can relatively easily be used, that is, artificially employed, to give the impression that one is angry. The third case is that of writing an angry letter. In this case it is the medium· of writing which by itself takes away a lot of the naturalness of the expression. Writing may come fairly natural to some people, but even they have to look for the mot jus te and to adjust their grammar, and these activities tend to take one's mi nd offthe emotion that originally inspired the letter. The fourth case, 252 JOACHIM SCHULTE in which you are ringing the bell for a servant you wish to scold, is a curious one. Here the means used, viz. to press a buzzer or to pull astring, is decidedly an artificial one but it may still be a natural way of expressing your anger. The servant, however, will not normally be able to tell from the way the bell sounds whether you are angry or not. We can, as we have seen, use what normally are natural expressions of emotions as more or less artificial means of giving the impression that one feels a certain emotion. But it is also possible to rely on our capacity to feign such expressions in order to cause oneself really to feel that emotion. This possibility is c1early suggested by James's words, and Wittgenstein mentions it in his later remarks on the philosophy of psychology. He writes: And how does it come about that - as James says - I have a feeling of joy if I merely make a joyful face; a feeling of sadness, if I make a sad one? That, therefore, I can produce these feelings by imitating their expression? Does that show that muscular sensations are sadness, or part of sadness?4 Wittgenstein's last question concems the problem of whether there is a conceptual or merely an empirical connection between statements about emotions and statements about the expressions of emotions. If there is such a conceptual connection, then it will be admissible to say that having certain feelings in your facial muscles is part of sadness. If, on the other hand, there is no such connection, that is, if statements about the expressions of emotions only say something about empirical relations between those expressions and then it will be incorrect to claim that those muscular the emotions ｴｨ･ｾｳｬｶＬ＠ sensations are part of sadness. Thus Wittgenstein writes: Suppose someone were to say: 'Raise your arm, and you will feel that you are raising your arm.' Is that an empirical proposition? And is it one if it is said: 'Make a sad face, and you will feel sad'? Or was that meant to say: 'Feel that you are making a sorrowful face, and you will feel sorrow'? and is that a pleonasm? (RPP I, § 452) Now what about the first question: Is 'Raise your arm, and you will feel that you are raising your arm' an empirical statement? One thing we must remember is that this statement can be false. This is the case, for instance, when the person we are talking to has taken a drug which renders his limbs numb, so that he won't feel anything. Another possibility is that he is bombarded with so many exceptionally strong stimuli that the feeling of raising his arm won't register at all. The possible falsity of the statement seems to speak in favour of thinking that it is an empirical one. On the other hand, it will be false only under very special circumstances whereas in normal situa- EMOTION: REMARKS ON WITTGENSTEIN AND JAMES 253 tions it will be true. And we have the feeling that normally somehow it cannot help being true, that feeling that one raises one's arm is simply part ofraising one's arm. Yet it surely is not a purely conceptual truth, that is, it is not what we may wish to call an analytic statement. Is the same true of the statement 'Make a sad face, and you will feel sad'? Now of course, when we are really sad we do not normally bother to find out about what it feels like to be sad. But in many cases of, as it were, minor sadness we do observe ourselves, trace our feelings and may even compare them with our feelings on other occasions. In such situations you may realize that you are making a sad face and literally feel that you are sad. But if you are not sad but attempt to make a sad face and succeed in making a face which looks like the face of a sad person, does that really mean that you feel sad? Wittgenstein reformulates the question in order to make it sound even more pleonastic and says, 'Feel that you are making a sad face, and you will feel sad'. But even this formulation, I think, is not fully pleonastic; it is not, if true, a mere conceptual truth. After all, it seems quite possible to make a sad face, so to speak savouring the various aspects of the sensation, and still be quite aware that you are not really sad. In that case the statement in question would be false and hence not a conceptually true one. This impression is reinforced by Wittgenstein's next remark: Does one say: 'Now I feel much better: the feeling in my facial muscles and round about the corners of my mouth is good'? And why does that sound laughable, except, say, when one had feit pain in these parts before? (RPP I, § 454) Of course one does not normally say things like that. And it is also clear that such a statement sounds silly because when tal king about myself feeling weIl or better I do not intend to talk about what it looks like to feel well or better. That would be treating myself as if I were a different person, which of course is a possible attitude to take but not in talking to others. Wittgenstein does not really come to a conclusion at this point, but he does throw doubt on the claim that certain sensations or feelings are to count as constituents of our emotions. He writes: Now granted - although it is extremely doubtful - that the muscular feeling of a smile is a constituent part of feeling glad; - where are the other components? Weil, in the breast and belly etc.! - But do you really feel them, or do you merely conclude that they must be there? Are you really conscious ofthese localized feelings? -And if not - why are they supposed to be there at all? Why are you supposed to mean them, when you say you feel happy? Something that could only be established through an act of looking - that's at any rate not what you meant. For 'sorrow', 'joy' etc. just are not used like that. (RPP I, §§ 456-457) 254 JOACHIM SCHULTE Our words for emotions like sorrow, joy, etc. are used for states which we need not find out about. And whatever is part of these emotions is something we are immediately aware of, not through making an introspective effort or an inference. In considering the problems conceming our concepts of the various emotions as welt as their relations to sensations, muscular feelings, etc. special attention must be paid to the possibilities of arousing emotions artificially and of pretending to be in a certain emotional state. Looking at these nonnatural situations, as we may call them, might help us to see what is really essential to such states. An example which, I think, springs to mi nd in this connection is that of the actor who tries to play his dramatic part as convincingly as possible. William James quotes a number of actors who have given accounts oftheir experiences. Thus one actress says: I often turn pale in scenesofterroror great excitement. I have been told this many times, and I can feel myself getting very cold and shivering and pale in thrilling situations. (James, op. cit., p. 464) And another actress gives the following answer: Playing with the brain is far less fatiguing than playing with the heart. An adventuress taxes the physique far less than a sympathetic heroine. Muscular exertion has comparatively Iittle to do with it. The example of the actor is used by Wittgenstein too in his attempt to get clearer about what an emotion consists of and whether it is legitimate to describe it in terms of various components. He writes in his Remarks on the Philosophy of Psychology: But wouldn 't I say that the actor does experience something like reallonging? For isn 't there something in what James says: that the emotion consists in the bodily feelings, and hence can be at least partially reproduced by voluntary movements? Is it so disagreeable, so sad, to draw down the corners of one's mouth, and so pleasant to pull them up? What is it that is so frightful about fear? The trembling, the quick breathing, the feeling in the facial muscles? - When you say: 'This fear, this uncertainty, is frightful!' - might you go on 'If only I didn't have this feeling in my stomach!'? (I, §§ 727-728) I take it that it is unproblematic to say that an actor who is acting his emotional part convincingly really experiences something like the relevant emotion. I suppose it is very much a question of degree. On the other hand, there is something puzzling in saying that a man who has no reason to yearn for anything feels something like yeaming; or that a person who has absolutely no reason to be angry feels something like anger. But still, this kind of EMOTION: REMARKS ON W1TTGENSTE1N AND JAMES 255 experience does exist, and actors are not the only people who know it; we all know it from occasions like reading thrilling books, watching romantic films, or hearing ghastly stories which arouse emotional feelings in us, although in our real lives there may be nothing to cause them. And in such situations of what one may call vicarious emotional feelings we can easily observe the palpitations of our hearts, the trembling of our hands, the biting of our lips, etc. But the relations between these occurrences and our emotions is still far from eIear. We have seen that it is difficult or impossible to imagine certain emotions without their typical physical accompaniments. Thus it seems for example quite meaningless to imagine fear without fear-feelings in your heart and your stornach. When one is in the grip of a strong emotion one does say things like, 'I'm so afraid; I cannot stand this feeling in my stornach any more',or, 'Ifthis state ofuncertainty will continue much longer, I shall die of heart-failure', etc. And yet, when Wittgenstein asks whether it is this kind of feeling which is so frightful when one fears something, we tend to answer, 'Of course not. The frightful thing is that which arouses the feeling'. A similar point is made by Wittgenstein: The expression 'This anxiety is frightful!' is like a groan, a cry. Asked 'Why do you cry out?' , however - we wouldn 't point to the stomach or the chest etc. as in the case of pain; rather, perhaps, at what gives us our fear. When anxiety is frightful, and when in anxiety I am conscious of my breathing and of a tension in the muscles of my face - does that mean that I find these feelings frightful? Might they not even signify an alleviation? (RPP I, §§ 729-730) There can be little doubt that it is not the feelings accompanying one's state ofanxiety which are frightful. Or, to change our example, think ofthe case of exhilaration. It's not the tension caused by the upward moving corners of your mouth which you find exhilarating, nor the extraordinary lightness you feel in your chest. Rather, it may be the good weather, or the good news, or the fact that you slept marvellously which fills you with high spirits. In such cases we are dealing with the causes of our emotion, and they are easy to indicate. Other states have no straightforward causes in this sense. If I am worried about the future in general, it would be foolish to suggest that the uncertain state of the future is the cause of my worrying. It may be possible that the future is in a certain sense the object of my emotion in a similar way in which the cook who has spoiled my dinner is the object of my wrath. But that does not get us very far because there obviously are many emotional states of excitement, nervousness, languor, etc. which have neither a eIear cause nor anything one might wish to call an object. But in these cases too we should not find it natural to say that the feelings of nervousness are that 256 JOACHIM SCHULTE which we find enervating, or that the palpitating heart is that which we find exciting. The distinction between emotions with a clear cause or object on the one hand and emotions without any identifiable cause or object on the other is, I think, what Wittgenstein has in mind when he asks us to compare fear and anxiety with care. The original German word translated as care is Sorge. And Wittgenstein continues by asking hirnself what sort of description is given by words like 'Ewiges Düstre steigt herunter', translated as 'Perpetual cloud descends'. These words are spoken by the figure of Sorge in the last act of Goethe's Faust. Sorge, or care, represents an emotion which is typically causeless and objectless. It is astate you can be in in spite of material and physical well-being; it is directed at nothing. It does not go with very strong typical accompanying muscular or other physical feelings. And still you can describe it. That is the point of Wittgenstein's quotation from Goethe. He wants to bring out that we can describe an emotion or, if you like, the content of an emotion by using words in a certain way, even though you cannot refer to causes or objects or accompaniments of that emotion. Now it would of course be interesting to use this idea in looking at the other emotions, that is, those which do have causes, objects, and accompanying feelings. And this is what I propose to do after I have gone a bit more thoroughly into some further questions arising from the theory suggested by William James. In his most extended discussion of James' theory Wittgenstein writes: James says it is impossible to imagine an emotion or a mood without the corresponding bodily sensations (ofwhich it is composed). Ifyou imagine the latter absent then you can see that you are thereby abolishing the very existence of the emotion. This might happen in the following way: I imagine myself sorrowing, and now in the imagination I try to picture and to feel myself rejoicing at the same time. To do that I might take a deep breath and imitate a beaming face.And now indeed I have trouble forming an image of sorrow; for forming an image of it would mean play-acting it. But it does not follow from this that our bodily feeling at that point is sorrow, or even something like it. - To be sure, a person who is sorrowful cannot laugh and rejoice convincingly, and ifhe could, what we call the expression of sorrow would not really be that, and rejoicing would not be the expression of a different emotion. - If the death of a friend and the recovery of a friend equally caused us to rejoice or - judging by our behaviour - both caused us sorrow, then these forms ofbehaviour would not be what we call the expressions of joy or sorrow. Is it clear apriori that whoever imitates joy will feel it? Couldn't the mere attempt to laugh while one was feeling grief bring about an enormous sharpening of the grief?s The passage in James to which Wittgenstein here alludes is evidently the quoted one in which James claims that it is impossible to abstract from our EMOTION: REMARKS ON WITTGENSTEIN AND JAMES 257 consciousness of an imagined emotion the feelings of its bodily symptoms. But while James argues his case by simply pointing out that if we subtract those physical feelings from our image of a given emotion nothing will remain of the original emotion, Wittgenstein introduces a new element into the discussion. He describes a case in which a person tries to imagine being in a certain emotional state and in that state trying to imagine to be in a different emotional state. His example is that of a man who imagines being sad and then tries to form amental picture of hirnself being in a good mood. It does not work. And that it does not work is at least partly due to our being unable to combine the imagined physical symptoms and feelings going with one emotion with the symptoms and feelings of the other. But as the whole experiment is not areal one but entirely taking place in the imagination, it is not owing to any really feIt physical sensations that the combination ofthose two states appears impossible. It looks more like a formal or Gestalt incompatibility and reminds one of our incapacity to see a picture or hear a tune in two different ways at the same time. This part ofWittgenstein's remark concerns imagination. But then he goes on to real cases and observes that someone who is in astate of sadness cannot . convincingly laugh, even if what he is confronted with is exceedingly funny. If he were capable to laugh, we should think that something was wrong with hirn, that his mind had suffered from the terrible event that brought on his sadness. This, however, has nothing to do with the physical feelings which James calls symptoms of our emotions. It is connected with what we regard as our natural ways of expressing our emotions. And some of these expressions exhibit a pattern in time which is incompatible with interruptions of a certain type. üf course, a great many emotions are quite compatible with other ones. Thus there is nothing strange in laughing about a joke when one is in a generally hopeful mood. But there are other emotions, notably sadness, melancholy, or grief, which form a pattern that is not compatible with interruptions by outbursts of good humour. A person who is terribly sad can nonetheless try to show joy or pleasure, for instance because he does not want to disturb others by showing his true feelings. In that case he may, as Wittgenstein points out, feel his sadness even more acutely than before. And that again speaks against the existence of a special logical relation between statements about emotions and statements about the feelings or sensations normally accompanying these emotions. For in cases like that of the sad man trying to laugh about a joke the attempt at imitating the typical behaviour of a person in high spirits can, even though the muscular sensations normally accompanying such behaviour are feIt, have 258 JOACHIM SCHULTE the effect that he senses his grief even more sharply than before. Thus it is, to use Wittgenstein's words, by no means 'clear apriori that whoever imitates joy will feel it' . However, all these considerations should not make us forget that the point from which we started-viz. the closeness ofthe connection between emotion and certain bodily symptoms - is incontrovertible. It is, although we cannot really find anything in the nature of an apriori relation, confirmed by the incompatibility of what Wittgenstein in the Philosophical Investigations calls the patterns of our emotions with certain other expressions of emotions, sensations, or feelings. This state of things is emphasized by Wittgenstein when he writes the following remark: Yet still I mustn 't forget that joy goes along with physical well-being, and sadness, or at least depression, often with being physically out of sorts. - If I go for a walk and take pleasure in everything, then it is surely true that this would not happen ifI were feeling unwell. But ifI now express my joy, saying, e. g., 'How marvellous all of this is!' - did I mean to say that all ofthese things were producing pleasant physical feelings in me? In the very case where I' d express my joy like this: 'The trees and the sky and the birds make me feel good all over' - still what' s in question here is not causation, nor empirical concomitance, etc. etc. (RPP 11, § 322) The first part of this remark is aimed at making clear that in talking about, or verbally expressing, a certain emotion we do not mean to speak of the physical feelings accompanying the emotion in question, although these feelings are in a certain sense inseparable from the emotion. And it is a sort of corollary ofthis that an expression of a given emotion mentioning those physical feelings is not meant to be a statement about 'causation' or 'empirical concomitance', as Wittgenstein says. This is naturally not intended as a denial of the existence of causal relations between that which makes me feel a certain emotion and the bodily symptoms accompanying that emotion. It should bring out that in expressing our emotions by way of mentioning what may possibly be their causes we still do not make a causal claim; at most we can be said to be trying to make our expressive behaviour specific. The role of causes is made clearer by the following passage: [... ] Possibly one could be sad because one is crying, but of course one is not sad that one is crying. It would after all be possible that people made to cry by application of onions would become sad; that they would either become generally depressed, or would start thinking about certain events, and then grieve over them. But then thesensations of crying would not thereby have tumed into apart ofthe 'feeling' of grief. (RPP 11, § 323) The causal story told by Wittgenstein is this: People are made to cry by some artificial means; the crying .causes them to become depressed or to think EMOTION: REMARKS ON WITTGENSTEIN AND JAMES 259 of sad events and thus come to feel sad. The sensation of crying, which presumably is indistinguishable from that feIt in normal situations of sadness, is here c1early separated from the emotion of feeling sad. And it becomes increasingly c1earer that what James calls the bodily symptoms of an emotion are in a certain sense not part ofthe emotion itself. They are closely related to the emotion but not an essential ingredient of it. William James was led to his position, according to which the physical feelings accompanying an emotion count as its essential component, by starting, so to speak, from the inside. His most convincing argument is his thought-experiment, in which we try in our imagination to subtract the physical feelings from the emotion and do not succeed. But that still does not show that these feelings are particularly characteristic of our emotions. That which is really characteristic of them may become clearer by changing our perspective and looking at the emotions not so much from the inside but from the outside, that is, from the point of view of a man who tries to understand how another person feels by observing his behaviour, his expressions, his gestures, his utterances. Some things are obvious. We judge that somebody is in a certain emotional state because his behaviour, or part ofhis behaviour, is a natural expression of a certain emotion. These natural expressions are often direct1y connected with the bodily symptoms William James talks about: the crying, the trembling, the raised voice, and the facial expression. But oddly enough, these phenomena now seem to be of much greater re1evance when looked at from the outside. After all, they often are our only criteria for telling what state another person is in. They are the basis on which we judge whether someone is lying or telling the truth ab out hirnself. It is as seen from the outside that these natural expressions can be useful for understanding and predicting other people's behaviour. Now it cannot be exc1uded that I can look at my own behaviour as it were from the outside and on the basis of that predict my future states of mind. But this would still be an exception. And equally rare or even rarer would be the type of case in which through noticing certain physical feelings I get c1earer ab out my emotional state. But the real point I want to make is this: Natural expressions like groaning, crying, trembling and so on are still only very rough indications of what another person may feel. Taken by itself, weeping does not tell you whether the person concemed is crying from joy or grief, jealousy or relief. And the same goes for the other natural expressions, like groaning, trembling, blushing, etc. In order to understand what they express we need to know more, either about the history of the person in question or about his present state. 260 JQACHIM SCHULTE Knowing more about his history, telling a story on the basis of that and the natural expression observed would amount to telling a plausible causal story. But if this type of knowledge is unavailable or insufficient we shall have to know more about the other person' s present state. One way of finding out about that is by looking at the finer shades of his behaviour, at additional gestures, and especially at the peculiar tone in which he utters his expressions. This point is of the greatest importance, for only now are we approaching that which is really specific and characteristic of certain emotions. It is the nuances, the shades and colourings of my expressions which make it possible for others to tell what I feel. An inkling of this is given by the following passage from the' Philosophical Investigations 6 : I say 'I am afraid'; someone else asks me: 'What was that? A cry of fear; or do you want to tell me how you feel; or is it arefleetion on your present state?' - Could I always give a c\ear answer? Could I never give hirn one? We ean imagine a11 sorts ofthings here, for example: 'No, not I am afraid!' 'I am afraid. I am sorry to have to eonfess it.' 'I am still a bit afraid, though I won'teonfess itto myself.' 'I tonnent myself with a11 sorts of fears. ' 'Now,justwhen I should be fearless, I amafraid!' To eaeh ofthese sentences a special tone ofvoiee is appropriate, and a different eontext. (PI n, ix, pp. 187-188) It is the specific tone of our expressions which allows other people to understand how we really feel. And it is this tone which is so difficult to imitate when you try to feign or to play-act a certain emotion. It is the existence of a certain tone of expression which enables us to make relatively fine distinctions between types of emotions and feelings, and it is the specificity of the tone which ac counts for cultural variations in the expression of emotion. Many feelings can be distinguished from others only because there is a tone going with them. As an example Wittgenstein mentions the feeling of conviction and writes: One speaks of a feeling of eonvietion beeause there is a tone of eonvietion. For the eharaeteristie mark of all 'feelings' is that there is expression ofthem, Le. faeial expression, gestures, offeeling. (RPP II, § 320) I think, however, that there are at least two further outstanding features of our expressions of emotions which ren der them specific and help us to understand them. The first of these features is connected with what Wittgenstein likes to call the pattern of a certain emotion, for instance when he writes in the Philosophical Investigations: EMOTION: REMARKS ON WITTGENSTEIN AND JAMES 261 'Grief describes a pattern which recurs, with different variations, in the weave of our Iife. If a man 's bodily expression of sorrow and of joy alternated, say with the ticking of a clock, here we should not have the characteristic formation ofthe pattern of sorrow or of the pattern ofjoy. (PI 11, i, p.l74) The expressions of our emotions follow a certain pattern in the sense of having a kind of form or rhythm which can sometimes be imitated in another medium. An example would be music where we sometimes speak of sad or joyful, melancholic or boisterous tunes, or movements, or even entire pieces. And this possibility is surely due to the fact that something of the rhythm or pattern - and certainly also something of the tone - of these emotions or feelings can be expressed by musical means. The third and last feature of the expression of emotions I want to mention is related to what was said about the quotation from Goethe's Faust concerning Sorge, or care. Here the point is that there are words or phrases by means of which one can contrive to describe exactly the characteristics of a certain emotion, although these expressions are descriptive neither of a cause nor of an object of these emotions. Goethe's poetic evocation of Sorge is a particularly accomplished description; but there are many more conventional, even hackneyed, phrases which serve to bring out the specific characteristics of our emotions. Thus we often use colour words to describe what we feel, for instance blue, grey, black, and yellow. But there are many more expressions - and sometimes quotations from poetry or song - which serve the same purpose. These three features - tone, pattern, descriptive phrase - are not merely means of telling how another person feels. They have a lot to do with the perception of our own feelings. The rhythm or pattern, the expressive tone, and the visual or tactile impression corresponding to certain descriptive phrases are feIt by the person who is in a certain emotional state. I think one could repeat William James' thought-experiment and ask whether it would be ｰｯｳｾ＠ sible to subtract the typical greyness, the subdued tone and the slow rhythm from our emotion of melancholy and still retain an idea of what the feeling was like. And the answer, I suspect, would again have to be that an emotion without those ingredients would be a 'nonentity'. 7 Bologna, Italy 262 JOACHIM SCHULTE NOTES W. James, The Principles ojPsychology, Volume Two, H. Holt, NewYork 1890, p. 499. Ibid., pp. 451 f. 3 L. Wittgenstein, The Blue and Brown Books, ed. and with aPreface by R. Rhees, B. Blackwell, Oxford 1958,p.103. 4 L. Wittgenstein, Bemerkungen über die Philosophie der Psychologie. Remarks on the Philosophy ojPsychology, Volume I (RPP I), ed. by G. E. M.Anscombe & G. H. von Wright. Trans. by G. E. M. Anscombe, B. Blackwell, Oxford 1980, § 451. 5 L. Wittgenstein, Bemerkungen über die Philosophie der Psychologie. Remarks on the Philosophy oj Psychology, Volume 11 (RPP II), ed. by G. H. von Wright & H. Nyman. Trans. by C. G. Luckhardt & M. A. E. Aue, B. Blackwell, Oxford 1980, § 321. 6 Philosophische Untersuchungen. Philosophicallnvestigations (PI), ed. by G. E. M. Anscombe & R. Rhees. Trans. by G. E. M. Anscombe, B. Blackwell, Oxford 1953. 7 This paper is an earlier version of Chapter 8 of J. Schulte, Experience and Expression: Wittgenstein's Philosophy oj Psychology. Claredon Press, Oxford 1993. 1 2 GUIDO FRONGIA WITTGENSTEIN AND MEMORY 1. One of the most significant explicit references to the philosophical problem of memory in Philosophical Investigations l is contained in § 265 which occupies a very central position in the discussion ofprivate language. Because of its position, that section has been quoted and paraphrased innumerable times in the context of the wide debate provoked by Wittgenstein's thesis concerning the latter issue, and in this way it has become quite widely known. On the other hand, it has been far less considered in itself as containing certain notions which, while they are connected with other more general aspects of Wittgenstein's theory, put forward a viewpoint on memory which has its own intrinsic interest. The reason for this probably lies in the fact that in Philosophical Investigations the other references to the specific philosophical problem connected with memory are somewhat rare and unsystematic. The recent edition of the previously unpublished works of his late years provides new material of considerable interest to this question for the first time, and this leads us to attempt a summary of what Wittgenstein thought on this specific issue. I shall limit myself here to considering certain aspects of the problem of memory which may have broader implications, useful for understanding his general theory on language and the bases of knowledge. Although I shall begin with abrief comment from PI, § 265, I shall seek to leave out as far as pO!:lsible questions direcdy connected with private language. There has been so much written and said on this subject that it would prove not merely impossible but also in a way inappropriate to reintroduce this question direcdy here; if for no other reason than the complexity of the twists and turns of debate which it has recently produced. 2. The most important theoretical indication contained in PI, § 265 is that the deep grammar which regulates the uses oflanguage connected with memory is not in fact homogeneous, as it might seem to be. A close analysis shows rather that there exist substantial differences within it, which make it possible to define two quite different notions of "memory", each characterised by its own rules which cannot be brought into line with those of the other. In order to demonstrate this theory, Wittgenstein attacks what is probably 263 R. Egidi (ed.), Wittgenstein: Mind and Language. 263-277. © 1995 Kluwer Academic Publishers. 264 GUIDO FRONGIA the most commonly accepted presupposition (though this acceptance may weIl be unconscious) ofthe opposite position, Le. the idea that it is possible in each case, and following the same procedure, "to appeal from one memory to another", or even to "justify" the assumption that both refer to a "same" event or object. Naturally, the critical notion is not necessarily restricted to two memories only, but can be extended to a greater number of them, and may present complex ramifications. In any case, I shall limit myself here to the simplest case, and for reasons of c1arity I will introduce an elementary formalisation. Let us call MI and M2 two successive memories which, we shall presume, come from the same event or object. The idea which Wittgenstein intends to criticise is the one which maintains that in the case of uncertainty about the reliability of the more recent record (M2), it is possible to have recourse to a memory which has arisen previously (Mt), and thus to proceed with a sort of comparison, and from that to a choice between the two options. This operation should make it possible in some way to confirm or deny M2 in authoritative manner by use of MI. The example which Wittgenstein puts in the mouth of his ideal objector is that of someone who has to take a train and being in some doubt as to its departure-time recalls to mind an image of a railway timetable which he presumably consulted some time in the past. The second memory (M2) - the one over which there is some doubt - would thus be confronted with a memorisation (MI) which has been previously acquired and which deals with the same source of information, the physical railway timetable. Ifthe two should coincide, then MI would provide a confirmation of the correctness of M2. Wittgenstein's counter-argument consists in observing that the example of his ideal opponent is certainly plausible enough in itself, but it is not pertinent to the purpose of demonstrating that one can speak of "justification" of the memory M2 through memory MI. If there is any "justification" here, it depends uniquely on the fact that the example envisages the possibility "of appealing to something independent" of both the memories (the physical railway timetable), which makes it possible to go on to a conclusive test of the correctness and thus eliminate all further doubts. In general, then, Wittgenstein seems to be stating, recalling to mind amental image (Mt) with the aim of "justifying" a memory of the same event (M2) has meaning if, and only if, MI can be considered "a memory which is actually correet." But this obviously does not always happen: indeed, it probably does not happen in the majority of cases. WITTGENSTEIN AND MEMORY 265 3. As we ean see, in this diseussion Wittgenstein is assuming a partieularly strong aeeeptanee ofthe term "justifieation" (Rechtfertigung). Tbis provides the outline of a eonfirmation needing the support of a proof whieh, at least in the speeifie eontext in whieh it is produeed, ean be held to be eertain and eonelusive. The paradigm of eertainty taken here as the measure of proof is that offered by the pereeption of a physieal objeet, observed in normal eonditions. Thus any further objections whieh may arise are not eonsidered on this oeeasion; Wittgenstein diseusses these elsewhere, at length, in his later writings. These may be put forward by a radieal seeptie, who maintains that, in the example cited above, the further verifieation by MI, whieh refers back to an independent text (the page ofthe railway timetable), would not be definitive either. The reasons for this would probably be those whieh are generally adopted in this type of diseussion: the person who proeeeds with the verifieation may still be in doubt as to whether he is not dreaming, or having haHueinations, ete. But ifwe exc1ude this type ofreservation, Wittgenstein's eonelusions seem to be fairly c1ear and eonvineing. Onee the type of eonfirmation whieh ean be provided by an "independent" physieal objeet (i.e. an objeet whieh has an existenee not dependent on the memory whieh one possesses of it) has been assumed as a paradigm of "justifieation", then no memory ean itself justify another. Naturally, if we eonsider the great variety of situations to whieh the funetion of memory must be adapted, we ean easily imagine an infinity of intermediate grades between the definitive "justifieation" and the insoluble doubt. Wittgenstein, as is weH known, does not deal thoroughly and in unified fashion with this variety of eases, though here and there he provides interesting footnotes whieh may eoneem (sometimes indireetly) different aspeets of their grammar. It would eertainly be very useful to reeonstruet these rather fragmentary and teehnieal referenees of his analysis in a systematie way. Simply in order to provide examples of these grades of eertainty whieh ean be diseemed in the use of memory in various effective eontexts, we may refer here to eertain eonsequenees whieh ean be extrapolated from the ease diseussed in PI, § 265. Even before reaching a "final justifieation", it is eertainly "legitimate" in Wittgenstein's sense to eompare two different memories with eaeh other. And this may lead to varying results. Two possibilities ean be provided for. Tbe first is that M2 eoineides with MI. Then it ean be said that Mt provides a eonfirmation ofM2 (even ifnot in a definitive sense); but it may equaHy be said that this eoineidenee provides a eonfirmation of Mt. The second possibility is that Mt and M2 do not eoineide. One eould then 266 GUIDO FRONGIA say that MI contradicts M2, as one could also state the opposite, that M2 contradicts M I. In each case, such a contrast does not provide in itself any conclusive indication on which ofthe two memories is to be considered more reliable. By what can be deduced solelyon the basis ofthis comparison, Mt and M2 might both (totally or in part) be inaccurate or false, and be revealed as such in the light of subsequent verifications. In general, the simple fact that Mt is nearer in time than M2 to the event or object which is (presumably) at the origin ofboth, cannot be held in itselfto be a criterion for greater reliability. Many other interesting considerations about this and other aspects ofthe grammar ofmemory could be entered into, and significant philosophical implications could be drawn from them. However, it is not these which we propose to discuss here. 4. From the standpoint we are taking, on the other hand, it is a basic requirement to return to the observation that in PI, § 265 Wittgenstein lays the conceptual basis for a fairly clear distinction within the grammar of memory. And this distinction makes it possible to define two major categories of cases: a first made up of the uses of memorising which are fully justifiable, and a second made up of those uses for which such a justification is not possible, even in principle. The error of reducing the first to the second is likened in PI, § 265 to the meaningless claim ofa person who buys "several copies ofthe morning paper to assure himseifthat what it said was true." There are two implications, one fairly clear and the other somewhat problematic of the meaning which this well-known image can assurne in relation to the specific problem of memory. The first is fully in agreement with what we have observed so far. The absurdity of someone who compares two copies ofthe same newspaper is similar to that ofthe person who claims to compare two memories ofthe "same" event or object which cannot be "justified" conclusively by recourse to a definitive text, in order to· find out which of them is true. In both cases, this operation is useless because it makes a comparison between entities which are pre-supposed to be homogeneous, and for this reason have the effect of proving nothing in relation to each other. On the more general plane, (and this is certainly the most important philosophical element, which above all links our theme directly to that of private language), the reductio ad absurdum, within which the image of the newspaper which we have already mentioned appears, tends to show, as we noted previously, how wrong it is to reduce one area of the grammar of memory to another, for example that related to the cases for which it is not possible to WITTGENSTEIN AND MEMORY 267 refer to objective justifications to that which concerns cases for which such a possibility does in fact exist. It can be observed that a similar reductionist tendency mayaiso be found in the opposite direction, that is, from the second category of cases to the first. And this naturally produces quite a few paradoxes, even though they be of reverse tendency. Taking account of this, we shall use the term "reductionism" from here onwards (without any further specification) to denote the theory which tends to explain the rules of one of the two areas of the grammar of memory in terms of the other, independently of the direction in which such an attempt is headed. Conversely, we shall use further specification in cases where we wish to refer to one ofthe two possible variants ofthis type oftheoretical approach. 5. It may in any case be useful to point out that, however brilliant on the plane of the argument and effective in literary terms it may be, the analogy between memory ofthe "same" event on the one hand, and copies ofthe same newspaper on the other, may weH be quite inappropriate in other ways, and can create a number of uncertainties, as has in fact been the case. The first is easier to locate, and for that reason theoretically less dangerous. In contrast to what happens with copies ofthe same newspaper, (always assuming that it only appeared in one edition), different memories of a "same" event or object (which are bound to occur at different times), mayaiso be very different to each other, as Wittgenstein in fact admits on a number of occasion. Hence, at least from this standpoint, the metaphor of the newspaper does not work. The second possible observation concerns grammatical aspect which is more difficult to grasp, but which - if it remains unc1arified - may prove to be somewhat underhand. The use of the analogy of the copies of the same newspaper may be inappropriate in that it could easily provide water for the mill of the reductionist, in this case the one who seeks to reduce the whole grammar of memory to the objectivity of comparison. The reference to the copies of the newspaper could in fact suggest that, at least in the case in which it makes sense "to appeal from one memory to another" (because a particular link in the chain of memories can be checked against objective evidence), this takes place in the same way in which two physical objects are compared, for example, two editions of the "same" document written at different times, or even two copies taken from two differing editions of the same newspaper. NaturaHy, Wittgenstein has shown, through innumerable analyses of linguistic cases, that he does not in fact believe in this second type of reduction of the psychological functions to the grammatical state which is proper to observable physical processes or events. Indeed, he always pointed out the 268 GUIDO FRONGIA serious conceptual difficulties to which such an attempt would lead. But this image ofthe newspaper could effectively induce some distracted commentator to forget this evidence. 6. The issues considered uJ> to this point have concerned directly only one of the two major areas of the grammar of memory - that constituted by the memories for which it is possible to provide objective "justifications". So far, the other area (the other half ofthe apple) has only been considered for purposes of confrontation, as having certain grammatical characteristics opposed to the former area. We must therefore ask ourselves at this point, in more positive terms, what happens to those memories which are not "justifiable" in the sense defined above. Could they be considered (at least in a weaker sense than that implied by a "justification") more or less "reliable" - according to each individual case? And iftwo memories ofthe "same" event or object do not agree, how is one to decide between them? On what grammatical basis could such a comparison be made? In order to deal with such a complex set of questions more closely and systematically, we would need to take into consideration a very broad series of linguistic cases discussed by Wittgenstein, which all seem to fall within the grammatical area of memory which we have now moved on to consider. To mentionjust a few examples: memories ofthoughts which we might have had in the past, memories of particular sensations and states of mind, of mental images, etc. Moreover, there is a whole complex of psychological phenomena to which the problem of memorising is intrinsically connected. And apart from offering, certain more general grammatical aspects common to other cases, this has a specific interest in itself. This is the subject of dreams. Dreams are the result of a psychic activity which, inasmuch as it takes place in the state of sleep, is not stimulated in direct, unequivocal and regular fashion by objects or phenomena of the external world. For this reason, the memories which we have of dreams when we pass into the state of wakefulness cannot be in any way ')ustified", in the sense in which Wittgenstein uses the term in PI, § 265. Therefore, we are here faced with the problem of the memory of dreams as an extremely significant case which makes it possible for us to gain a better understanding of some more general aspects of that very broad area of the grammar of memory which we are now considering. 7. As is weIl-known, the notes by Rhees dated 1942-46, and published under the tide Conversations with Freud (CF)2 are dedicated to the specific theme of dreams, as weIl as a large number of scattered notes contained principally WITTGENSTEIN AND MEMORY 269 in Wittgenstein's late unpublished writings. Although not systematic in the conventional sense, this material nevertheless enables us to glean certain indications which are consistent in themselves and which can effectively help us to respond to the questions which we are now facing. In Conversations with Freud, reference is often made to dream activity as characterised by a certain regularity; for example images which occur according to certain constant sequences, or which seem to follow particular internal logics, or which are arranged according to plots with recurrent schemes which sometimes make it possible even to make predietions about their future recurrence, ete. All this would seem justify, as Wittgenstein sometimes acknowledges, the commonly accepted idea that dream activity is "a way of saying something or a way of symbolising something" (CF, p.48); that there is, in fact, "something like a dream language" (CF, p.43). However, to what degree is it justified to use the term "language" to refer to these regular formal and structural aspects of dream-perception? Does this use reflect a genuine analogy with the ordinary concept of "language", or is it only a way of showing so me extrinsic similarity, but one which tends to ignore or hide substantial differences? These questions obviously have very complex implications, which cannot be considered here. Certainly, in support of the notion of a language of dreams there is the undeniable evidence, even historical evidence, of the contributions made by a large areas of contemporary psychology, which, like psychoanalysis itself, have made this notion a key factor in the elaboration of theory and analysis. Wittgenstein does not seem to show any preconceived hostility to these trends of contemporary psyehology, despite what has often been maintained. He certainly sees certain way of understanding and using their results as potentially "dangerous"; just as he sees danger in the fact that they tend to affirm the existence of a "powerful mythology" in contemporary eulture (CF, p.52). But in overall terms his remarks do not have the aim of rejecting the bases of Freudian theory and psycho-analysis en bloc; they are tended rather to provoke critical reflection directed mainly at forestalling certain easily predictable misunderstandings, and at combating the mechanical, repetitive and dogmatic applications which can be made of certain of their theoretical principles. In any case, the subjects dealt with in the conversations with Rhees have a fairly wide theoretical and philosophical range, which goes beyond the more circumscribed polemical concems and particular historical points of reference of eontemporary psychology. 8. Wittgenstein directs a particularly significant critique to the tendency to 270 GUIDO FRONGIA see the exercise of memory, when this is applied to past dream-material, as a function aimed at explaining or describing in the most faithful possible way what has been dreamed. The reason for this criticism is that, in his opinion, this way of understanding attributes an exclusively passive, reproductive, figurative task to memory. Naturally, Wittgenstein is wen aware of the degree to which this aspect has been studied in the context of depth psychology. Almost an these studies, independently of their specific orientation, have revealed the various kinds of interferences which we meet with in the correct memorising of dreams. And this image of a distortion which memory imposes on its own objects is commonly aimed in this context not only at localising the nature of such interference, but also - and above an - at reconstructing the "real" development of the dream processes. The objective of this attempt to go beyond the appearances is obviously not aimed exc1usively at the observance of a disinterested and abstract principle of truth. Such an exercise in this case has rather the aim of intercepting, as far as is possible in their immediate and integral state, the messages which the more profound layers of the Ego send to the conscious mind through the "dream language", to be deciphered later through "interpretation" . The most important aspect of Wittgenstein's reflection on this subject is therefore aimed at a critical review ofthe various moments ofthis theoretical and procedural scheme. His polemic is aimed above all at the idea of a supposed "objective" existence, with an identity in itself, of a dream-material which the memory must tend to reproduce faithfully. To this notion he counterposes the thesis that "the criteria for the truth" of an account of a private event, such as a thought which has arisen in the past, or a dream, cannot be "the criteria for a true description of a process"; and that the importance of such truth "does not reside in its being a correct and certain report of a process".3 As can also be seen quite clearly from Conversations with Freud, Wittgenstein sees the attempt to assume such a criterion for a "faithful" reproduction of a past and irrecoverable reality, such as a dream, as more a reflection ofthe aspiration ofpsychology to adapt itselfto a model ofknowledge and scientific explanation which is completely extraneous to it. This model, in his opinion, is moreover derived from a nineteenth century positivist conception of physical sciences (cf. CF, p.48), and hence not even capable of adaptation to many of the physical phenomena which are more frequently studied today. A double misunderstanding, this, which makes such an interpretation of the problem of memorisation of dreams quite impracticable. We WITTGENSTEIN AND MEMORY 271 should note that, if we wish to simplify matters, the model criticised here could fairly well be adapted to the case of a memory which is objectively "justifiable", as considered in PI, § 265, and discussed above. 9. The alternative way indicated by Wittgenstein is in line with his general theoretical approach, aimed at demonstrating the great variety of linguistic games, and at pointing out the impossibility of reducing certain areas of the deep grammar of the language to other areas which present apparent similarities with the first. In this particular case, he induces us to recognise the specific nature ofthe rules which govern the various linguistic games concerning dreams:for example, teaching the use ofthe expression "I dreamed that...", recounting one's own dreams to others, recounting them to one's self in a low voice, describing them in writing, listening to the accounts of others, recalling to one's own mind visible images which have appeared in dreams, etc., etc. The exercise of memory, inasmuch as it is aimed at reconstructing the content of the dreams, must be considered in all this variety of contexts of use. It could be said that in general memory is not here subject to limitations of justifiability of the same type as those to which it is subject when it refers to physical events or to objects of perception. The reason for this is that, in the case of a dream, in contrast to what applies in the latter cases, the memory itself necessarily (and hence not in some occasional fashion) contributes to the identity of what has been dreamed. Thus if, after return to the state of wakefulness, there are doubts as to the contents of a dream, the possibility of recourse to any incontestable protocol which has an identity independent of the memory itself is precluded. Similarly, in the case in which two differing memories (and thus two memories which have arisen at different moments) ofthe "same" dream differ from each other, it will be impossible to "justify" conclusively the validity of one or the other. The actual question: "How do I know if I really dreamed that?" (RPP I, §363) thus loses the meaning which an apparently similar demand might otherwise have, if it were aimed at reference to the objective and external source of a fact which is presumed to be reliable: for example: "How do I know that the train leaves at 7.30?" (cf. PI I, §265). The reply, in the case of the dreams, is necessarily univocal, and internal to the question itself: "This is what I remember". In short, memory seems here to be playing a productive and constitutive role in its own object - a role which has no parallel in other areas of the grammar of memory. Now, must we hold that such a productive function, which memory exercises in these contexts, is deprived of restrictions? That it has unlimited power to 272 GUIDO FRONGIA give an identity to its own object? For instance, am I obliged to take everything I remember of a certain dream as equally valid? And if someone recounts his dream to me, and I assume his total sincerity, must I accept as definitive everything which he teIls me? And what happens to any doubts I may have about certain particulars of my dream? And how should I decide, in the case of discordant memories, some ofthem recorded with great care in different moments? In short, are we forced to admit that memory becomes anarchic when it is applied to dreams? But then, how would it be possible to distinguish the memories of dreams from illusions or the creations of fantasy? (cf. RPP I, §201). 10. The replies that Wittgenstein provides for these and other similar questions are often not very clear, but show a certain indecision in assuming a precise theoretical position. Even considering the growing frequency of references to the theme of dreams contained in his later works, it is not impossible that he himself, as he progressively matured a more systematic response to this issue, did not then have the time to develop it completely. Certainly if there was evolution on this specific theme, it is fairly difficult to follow it with continuity through the labyrinth of the writings left unpublished. But if we exclude concerns of a genetic type, at least two rather different solutions can be discerned in his later work. According to the first of these two solutions, in a certain sense, memory becomes a "eriterion for the past" (cfLRPP I, §837), and thus the source of a sort of "subjective truth" (LRPP I, §897). While it is true that these expressions have a rather uncertain meaning, and recur only infrequently in his tater writings, it is perhaps possibte to hypothesise what Wittgenstein intended to state by them. For dreams, and for other analogous psychic phenomena, memory is the true source of information grammatically permitted to act as a reference for their content. And if it does not succeed in resotving any "doubts" which may arise, or fails to cover its own lacunae, the uncertainties which all this may produce will be destined to remain such, for the most part. However, it is interesting to observe that both the expression here mentioned disappear in the corresponding passages in Philosophical Investigations - passages which probably reflect a more advanced stage of his work. On the contrary, there is a section in Last Writings on the Philosophy 0/ Psychology which provides a rather different, and quite indicative, point of reference for our subject: WITTGENSTEIN AND MEMORY 273 Assuming that peoples' dreams can yield important information about a dreamer, what yielded the information would be truthful accounts of dreams. The question whether the dreamer's memory sometimes, often, or always deceives hirn cannot even arise, unless indeed we introduce a completely new criterion for the 'correctness' of the account of a dream (LWPP I, § 898) As we can see, Wittgenstein seems here to be quite ready to admit a certain regulatory limitation in the reporting of dreams. In Philosophicallnvestigations (11, p.222) this indication receives further reinforcement. At the end of the paragraph corresponding to the one cited above, the "completely new criterion" which is required in order to decide about the "report's 'agreeing' with the dream", is given the specific tide of Wahrhaftigkeit. So, something more than a genuine appeal to "correctness" is implied here. And yet, Wittgenstein does not provide us with any specific indication of the significance which should be attributed to such a criterion. He only says that it is quite distinct from that of the "truth" which must be valid for those accounts of memory which, according to his contention in PI, §265, are in proper sense "justifiable". 11. We may gather some indications, even if only indirect ones, which might help to c1arify this crucial point from the Conversations on Freud, which will perhaps enable us to venture on a hypothetical response, tendential and schematic though it may be. Here in general, as previously observed, Wittgenstein does not appear to be preconceptually resistant to the idea that dreams can provide important information on the person who is dreaming. As we have seen, his reservations about the theory of Freud and psycho-analysis do not exc1ude the possibility that the instrument of dreams could be used for a better knowledge of the human mind. More often, his doubt concerns the way in which this possibility is frequently interpreted and translated into action. In this way, we may fee1 that it is justified to remove the reservation ofthe conditional at the beginning of the passage cited earlier. Thus Wittgenstein would now really be searching for a "special" criterion of "truthfulness" for the memorising of dreams. But since the memory, when it is applied to this field of experience, remains the only effective source of information, the desired criterion of correctness must in some way be held to be within the actual construction which memory imposes on its own objects. And this would introduce an important reason for agreement with a fundamental principle of the theory and practice of psycho-analysis: in its productive function of dreammaterial, the memory does not proceed in chaotic and unpredictable fashion, but follows certain regular procedures. It is these latter which have a primary interest for the analysis, even if they have to be "interpreted" appropriate1y. 274 GUIDO FRONGIA This very concept of"interpretation" provides a further indication ofwhere a criterion of "correctness" might be found for the use of memory as applied to dreams. In the light of Rhees' notes, Wittgenstein does not actually seem disposed to set aside the problem ofthe interpretation of dreams from that of the reconstruction carried out in the light of the memory which one has of them. In other words, the relationship between these two moments is not one of succession. It is not a matter ofjirst the memory and then the interpretation of what is remembered of a certain dream. Instead, it is the actual record which constitutes an interpretation (CF, p.46). Or, if we wish to put it in another way, interpretations and memories follow similar procedures. 12. We still need to clarify the meaning to be attributed to such a substantial identification between interpretation and memory. In Conversations on Freud, at least one significant pointer is given about this question. Dream material is presented to the memory in a way not different from the way one might present a complex mass of fragments of a figure which is indecipherable in itself (CF, pp.45-46). The memory tends to unify these fragments according to a coherent design, for example by attributing a unified plot to images and fragments of episodes which otherwise would seem incomprehensible in themselves. Naturally, this unifying function can be called into action by the need to make the dream known to others by recounting it, or may be influenced by the interaction of the person recounting it with the listener (the analyst, for example) (cf. CF, pp.43-44). What Wittgenstein seems to be seeking to indicate here is the need not to consider these "plots" which the memory uses to unify the material which it has succeeded in retainjng, as a "strange disturbance" (cf. LWPP I, §§656, 292). Instead, a "plot" is an integral part ofthis function ofmemory, aimed at gathering and mingling things remembered which may be derived from the most varied sourees, and "turning them into the memory of an event which took place while a person was sleeping" (LWPP I, §656). It is this type of constructive intervention which confers singularity the linguistic game constituted by remembering dreams. It seems clear that here Wittgenstein is thinking of certain characteristics of remembering dreams which are not dissimilar to that of a literary construction suited to certain canonical models of a structural or aesthetic kindfor instance, a particular pattern of recounting, widely accepted and easily discernible logical connections and oppositions, particular stylistic choices, and so on. Both these two activities, just because they adopt formal procedures and structurally regular typologies would succeed in transferring on to WITTGENSTEIN AND MEMORY 275 the plane of language contents which otherwise would remain relegated to the private world ofthe conscience, in the amorphous and non-communicable form of unique and unrepeatable impressions. In other words, here the "plot", far from being "a strange disturbance of memory", in its function of expounding dreams constitutes the essential go-between which gives them their identity. Without the medium of the "plot", it could be said, a dream would not exist, since memory could not confer any identity on it. As can be seen, from this point of view, the problem of the correctness of the memory (its truthfulness, or Wahrhaftigkeit, PI 11, p.223) comes to be identified with the problem of the correctness of the procedure of interpretation, such as is realised for instance in the reconstruction of a dream on the basis of a plot. And it is highly significant that in Conversations with Freud the most severe criticism which Wittgenstein directs at psychoanalysis concerns this very point. From these annotations some very illuminating indications not to be found elsewhere can be gathered. The idea that Wittgenstein views with the greatest suspicion and hostility is that the interpretation of dream constitutes a kind of mechanical and univocal translation from one "symbolism" (specifically that of dreams) to another (that represented by our language of discourse), according to certain rigid mIes of correspondence (cf. CF, p.42). The origin of this mi stake is sited by Wittgenstein in the claim of Freud (and perhaps by psychoanalysis in general) to "be scientific" (CF, p. 44) and in this way to protect the interpretation (and the memory) of dreams from arbitrary interference. The intention of this is to confer objective security, uniformity of results, and hence reliability and prestige, on analysis. In fact the "scientific" model seems to require that the interpretation must either be "complete" or otherwise have no authentic value at an; and in this second case it must be substituted by a more "complete" interpretation (CF, p.49). 13. What we have seen so far provides clear indications on the critical plane, and on the objective polemics which inspire the analysis ofWittgestein. These aspects have been generally accepted by the critics, both for their historical interest and for their effectiveness in discussion. But it has been far less often noted that Wittgestein's observations also provide positive theoretical indications, though often not explicit ones. The interpretation of a dream, he seems to be saying, must be treated as such, i.e. as an attempt to pin down a fleeting object which cannot do other than escape. Such an attempt is made by assuming a particular standpoint or hermeneutic attitude which may change for very varied reason: "What is done in interpreting dreams is not all of one 276 GUIDO FRONGIA sort" (CF, p.46). Thus, to summarise Rhees' testimony, reflecting on a dream and recording certain details of it "will give it a different aspect"; and remembering in different moments and context will produce different dreams (CF, p.46). Given the analogy between recording and interpreting: "there is a work of interpretation which, so to speak, still belongs to the dream itself'; and remembering the dream is a little like "dreaming the dream over again" (ibid.) Therefore, what characterises an interpretation (for instance, in contrast to the certainty which can be attributed to the observation of a physical object or to the "justifiable" memory of it) is the fact of not being conclusive. An interpretation, in a certain sense, tends to produce a "myth" (CF, p.51). A negative connotation need not necessarily be given to this term. It is conferred on it by a comparison of an emulatory type with a (nineteenth century) image of science, tied to the concept of"exactitude" and to the principle ofunivocality of explanations and solutions. But there are extremely significant pages in Wittgenstein' s last works4 where the latter term does not have such a meaning. Here a "mythology" constitutes "the inherited background against which we distinguish between true and false" (OC, § 94) - a knowledge, then, that we accept, in more or less provisional fashion, with the aim of allowing other sectors of enquiry (interpretations) to proceed in relatively freer and speedier fashion with their proper exploratory tasks. Thus a "mythology" may itself become a "norm of description" (OC, § 167) and hence a paradigm of interpretation. If (as may be assumed) a critical intention is present in the observation that psychoanalysis tends to produce a "new myth", this is solely because Wittgenstein (rightly or wrongly) sees in it a certain propensity to undervalue the unavoidable complexity and indeterminacy of the examination of dreams and similar psychological phenomena - a propensity which he believes has its origin in a misconceived intention of "being scientific", which leads that discipline to aspire to an identity which is both alien and inconvenient to it. It is, then, against this unauthentic attitude rather than against particular theoretical and therapeutic advances that Wittgenstein's critique is directed. An attitude which may make for reluctance to accept what probably today many analytical schools would be ready to recognise: that a dream is destined to remain "an enigma", which "doesn't have to have a solution. It intrigues us" (LWPP I, § 195). Universita di Roma 'Tor Vergata' WITTGENSTEIN AND MEMORY 277 NOTES I Philosophische Untersuchungen. Philosophicallnvestigations (PI), ed. by G. E. M.Anscombe & R. Rhees. Trans. by G. E. M. Anscombe, B. Blackwell, Oxford 1953. 2 See in L. Wittgenstein, Lectures & Conversations onAesthetics, Psychology and Religious Belief, ed. by C. Barrett, B. Blackwell, Oxford 1966, pp. 41-52. 3 See Letzte Schriften über die Philosophie der Psychologie. Last Writings on the Philosophy o[ Psychology, Vol. I (LWPP I), ed. G. H. von Wright & H. Nyman. Trans. by C. G. Luckhardt & M. A. E. Aue, B. Blackwell, Oxford 1982, § 897; PI, H, p. 222. 4 Cf. Über Gewissheit. On Certainty (OC), ed. by G. E. M.Anscombe & G. H. von Wright. Trans. by D. Paul & G. E. M. Anscombe, B. Blackwell, Oxford 1969, §§ 94 ff. FRITZ G. WALLNER ASSISTED BY MARKUS COSTAZZA HOW LUDWIG WITTGENSTEIN WOULD HAVE REACTED TO RECENT CHANGES IN PSYCHOLOGY In this paper we want to examine whether L. Wittgenstein's ideas can be made fruitful for basic discussion in cognitive psychology, as the development of methodological discussion in the field of psychology is not as remarkable as in physics. The first main point in this discussion is the confrontation between psychology of experience and cognitive psychology. After this confrontation the perspectives in psychology changed drastically; this was due to the so called cognitive turn. At this point the contrast between behaviorism and psychology of experience is not only overcome, but abolished (in the philosophical sense that we owe to HegeI). I.e., it is solved at a higher level, without sacrifice of the claims underlying those two approaches. One could characterize those claims as the primacy of immediateness and the primacy of probability. We shall see that both play an important role in Wittgenstein's confrontation with psychology, especially when he distances hirnself from Camap's approach. Cognitive psychology, however, in contrast to psychology of experience and behaviorism al ready assurnes a concept of "self' adapted to psychology. The "self' is included in the system's structure in this case. We shall find that in this case a deficiency appearing in Wittgenstein's writings is solved by a single discipline. Wittgenstein's position conceming behaviorism is often misunderstood and seems to be strange; he was often seen as a behaviorist. For example, Rorty (1977, p. 169) points out that Wittgensteinians have entered into an alliance with traditional behaviorism. As everybody knows, there are indeed some places in Wittgenstein' s works, which could be interpreted in a behavioristic way. This becomes evident in particular if we cite the distinction of the first and third person, as it occurs in Zettel, § 472: Plan for the treatment of psychological concepts. Psychological verbs characterized by the fact that the third person of the present is to be verified by observation, the first person not. Statements like the last quotation seduced Robert Fogelin to talk of a form ofbehaviorism in the third person in Wittgenstein's work, which makes hirn accuse Wittgenstein of a basic weakness. I For Wittgenstein the point is that one can apply verification to psychological 279 R. Egidi (ed.), Wittgenstein: Mind and Language. 279-288. © 1995 Kluwer Academic Publishers. 280 FRITZ G. WALLNER sentences in the third person; applying it to sentences in the first person, however, would be senseless. He draws, however, another conclusion than R. Carnap. In our opinion Luckhardt (1983) is right in interpreting Wittgenstein's concept ofpsychological sentences as reaction to Carnap's approach in this context. We can also find one of the few places in which Wittgenstein explicitly quotes Camap: One may also say: 'He made this face' or 'His face altered like this', imitating it-and again one can't describe it in any other way. ((There just are many more language-games that are dreamt of in the philosophy of Camap and others.W Camap was influenced by the principle of verification when solving the problem of the psychological sentence in the first person. He rejected introspection, since, given the lack of verification, it. cannot lead to sentences being intersubjectively meaningful. He offers another explanation for psychological sentences in the first person; it runs as follows: if A leams to use psychological sentences about hirnself, then he is leaming an intersubjective physicallanguage which not only describes the relevant state ofhis body, but also has to describe hirn in such a way as another person could describe it in principle.3 It can be recognized that Camap is defending the c1assical concept of behaviorism, as it was formulated by Watson (1913). It seems that the following is valid for Camap as weH as for Watson: Although consciousness cannot be topic of psychology, it has the "not unimportant function of being the instrument of scientists".4 This implies however a strange ambivalence: it is a metaphysical fiction and yet has scientific function. s In contrast, Wittgenstein foHows another path. He interprets only psychological sentences in the third person as sentences having representative function; sentences in the first person are expressive . . . . The word 'pain' is the name of asensation is equivalent to 'I've got a pain' is an expression of sensation. 6 Having these ideas in mind, Wittgenstein rejects introspection as weH as traditional behaviorism. As a consequence we have to ask, if Wittgenstein creates a new kind ofbehaviorism: behaviorism oflinguistic behavior which we could call the "linguistic behaviorism". Wittgenstein considers this in his Philosophicallnvestigations I, §307: 'Are you not really a behaviourist in disguise? Aren't you at bottom really saying that everything except human behaviour is a fiction?' - If I do speak of a fiction, then it is of a grammatical fiction. HOW LUDWIG WITTGENSTEIN WOULD HAVE REACTED 281 It becomes clear that Wittgenstein is not concerned with empirical statements in such as occur in particular discipline. We shall realize the point of this position only if we look at the behaviorism in psychology. E. C. Tolman (1932) already attacked Watson's position. His argument against Watson's physiological reductionism was that referring to physiological knowledge which has not yet been attained me ans running the risk of a pseudoscientific behaviorism. One should not refer to Watson's "muscletwitching" - but to behavioral acts as conceptual units. This means that intentions become more important for psychological investigations while the intended states of affairs also are determined in regard to the relevant behavior. 7 As in phenomenological psychology the facts have a different meaning, depending on whether the agent is an American, a child or an animal. 8 The logic of this idea led Tolman to his propositionalism (Tolman 1959). By propositional behavior he understands behavior cognitively directed towards a goal. The psychological analysis founded on these ideas is not aimend at describing experience, but rather behavior. Following Tolman's example a chair could have the meaning of"sit upon ableness", of"climb upon ableness" as weIl as of "throw about ableness" (according to the behavior of the user). Things may be comprehended intentionally in a means-end-relation without having to make use of introspective data of experience. By these distinctions Tolman (1959) also introduces distinctions within manifestations which have been normally attributed to subjective soul life. He differentiates between means-end readiness, beliefs and expectancies. On this basis he determines intentional phenomena by behavioral description (Graumann 1965, pp. 248t). In this way he anticipates a position labeled by Miller, Galanter and Pribram (1960, p.211 ft) as subjective behaviorism. Using this approach Tolman could hope to fulfill the tasks of psychology, as conceived, for instance, by phenomenological psychology, without having to recourse to descriptions of experience. Applying these arguments we have made a big development compared to Watson. Watson's reductionism expects too much from the concept of organism. One cannot explain the differentiated structures of behavior, as for instance the bearing in mi nd of a certain goal, from the organism alone, unless one makes pseudoscientific assumptions and constructions. Subjective behaviorism overcomes this deficiency without having to make the assumption of the hypothesis of two domains of reality as psychology of experience does. Let us return to the different views of Wittgenstein and Carnap. Carnap 282 FRITZ G. WALLNER conforms with Watson. He succumbs to a bewitchment of language in his desire to make language a language of science. This bewitchment lies in the demand that even psychological sentences in the first person should be a descriptive language game. This leads to an ontological change in his approach in order to fulfiH the claims ofthe principle ofverification. The change consists in the banishment of fictions mental objects out of the body; Carnap succumbs to a misunderstanding of language (as Heidegger did in his ontologisation of "the nothing"). While Heidegger postulated meaning for a word, Carnap undertakes the effort of deciding between possibility and impossibility of meaning of words. It was Wittgenstein's achievement to unmask introspectionism as weH as behaviorism as a grammatical fiction. His position is quite near to Miller, Galanter and Pribram's approach of subjective behaviorism, but, unlike theirs, his approach permits the overcoming of the dualistic gap. Due to its ontological undecidedness subjective behaviorism runs the risk of being subject to metaphysical fictions, as the literature on its phenomenological interpretation shows. 9 In contrast Wittgenstein does not succeed in accommodating subjectivity and sense in psychology. If cognitive psychology follows hirn, it has to face an undecidedness in principle. Ontologizing subjectivity represents a similar danger for cognitive psychology as previously for introspectionism and physiologism. The accusation that Wittgenstein forfeits subjectivity represents a similar misunderstanding to the accusation of his being a behaviorist. The ontological undecidedness ofmodern psychology causes its uncertainty about its object of research. This is shown in Wittgenstein's Zettel, § 468: 'Man thinks, is afraid etc. etc.': that is the reply one might give to someone who asked what chapters a book on psychology should contain. This implies the special position of psychology and its different character compared to exact natural sciences. Wittgenstein writes in his Remarks on Philosophy 0/ Psychology: Psychologieal concepts are just everyday concepts. They are not concepts newly fashioned by science for its own purpose, as are the concepts ofphysies and chemistry. Psychological concepts are related to those ofthe exact sciences as the concepts ofthe science ofmedicine are to those of old women who spend their time nursing the siek. IO Would it be a possible suggestion to use the method of explicative definition in order to state these concepts more precisely or to delimit them from their everyday use. We will see that Wittgenstein's insight into epistemological interrelations keeps hirn away from this naive cul-de-sac. HOW LUDWIG WITTGENSTEIN WOULD HAVE REACTED 283 Wittgenstein's investigations ofpsychological terms can, I believe, be used for the development of a new approach to psychology. In this respect he has gone beyond the dispute between natural scientific and humanistic psychology. In this respect Stromberg (1987, p. 28) is correct when claiming that Wittgenstein's critical remarks do not play an important role from the point ofview ofpresent-day trends in psychology. To my mind, however, one cannot accept the self-understanding (i.e. the unification of application and theory) of some ofthese trends. Humanistic psychology is an example ofthis. The joint development oftheory and application is represented in psychology in those approaches which suppose that the creation of psychological terms and specific human selfinterpretation go hand in hand. One could talk ofa pragmatic turn in psychology, when its motto is the "self-making" ofthe human being. We have to differentiate this from the action-theoretical turn in psychology. It can be understood as the re action to behaviorism, in which the human being was seen as a stimulus-response mechanism. Wittgenstein's ideas must also be correlated with the approach of action theory. This relation is quite similar to the relations between the late Wittgenstein philosophy of language and Searle's speech act theory (cf. Wallner 1981). There is a twofold relationship: in contrast to approaches in theory of action Wittgenstein is not in search of an explanation of human behavior. The aim, again, is to gain clarity. Wittgenstein's view ofbehavior represents an essential difference to the theory of action. The representation of mental pro ces ses against the background of behavior reveals structures which are not taken into consideration in theory of action. My own behaviour is sometimes, but rarely the object of my own observation. And this hangs together with the fact that I intend my behaviour. Even if an actor observes the expressions ofhis own face in a glass, or the musician attends c10sely to every note in playing, and judges it, this happens after all so that he shall direct his action accordingly. [Zettel, § 591]" As Wittgenstein sees it, two aspects of behaving are of special interest: its deliberateness and the question of the role the observer plays in this context. Because of its deliberateness it is different from mentallife (this is also the argument against the misinterpretation ofWittgenstein as a behaviorist). This determines the role of behavior in the argument: undoubtedly it represents the first-person point of view; hence its expressivity, but its deliberateness makes behavior an object of observation. This happens quite rarely, usually when it serves some purpose of action. This becomes clearer when Wittgenstein talks about the role of self-observation, as in § 839 of Remar/,s 284 FRITZ G. WALLNER on philosophy ojpsychology. Two aspects become c1ear: on the one hand the strange role of self-observation which does in "first person" what is done to the other by means ofthe "third person" and, on the other hand, the possibility of self-observation is made c1ear. We have to focus on the conceptual relation between action and behavior. This relation is not easy to realize and we cannot make use of tradition al categories to comprehend it. The criterion of immediate planning represents an important distinctive feature. Action is immediately determined by planning, whereas behavior is planned seldom and only indirectly. After an I know that when I am angry, I simply don 't need to learn this from my behaviour. But do I draw a conclusion from my anger to my probable action? One might also put the matter, I think, like this: my relation to my actions is not one of observation. t2 An argument for this approach to the concept of action is evident in RPP I, § 561 Tbe expression, the behaviour, of considering. Ofwhat do we say: It is considering something? Ofa human being, sometimes ofa beast. (Not ofa tree or a stone.) One sign ofconsidering is hesitating in what you do (Köhler). (Not just any hesitation.) Acting is planned immediately if reflection disturbs it. The function of observation is different in the context of behavior and in that of action. Action is not observed because it is planned, behavior is observed in order to be planned. The relation between anger and angry actions could be observed in principle; it would not correspond, however, to normal human behavior. Hence the methodological argument against such an observation is not lack of clearity in the observed object. Let us turn to RPP I, § 705 Why do I never conclude from my words to my probable actions? For the same reason as I don't conclude from my facial expression to my probable behaviour, - for the interesting thing isn't that I don't conclude from my expression of emotion to my emotion, but rather than I don't conclude from that expression to my later behaviour either, as others do, who observe me. [Zettel, § 576] The conceptual relation between actions and words corresponds to the relation between behavior and facial expression, it is that of specification. The relation between the two forms of behavior is not a relation of causality: as regards behavior we rather have to differentiate the first-person point ofview from the third-person point ofview. If one infers from behavior (1) to behavior (2), one looks at behavior (1) from a third-person point ofview and interprets it as action. Language clarifies the situation in the phrase "expression of HOW LUDWIG WITTGENSTEIN WOULD HAVE REACTED 285 emotion". Behavior (1) is an expression of emotions in so far as I make my emotions public. When I am expressing my emotions, the conclusion from this to my emotions is trivial. The conclusion from such expression to future behavior is not trivial, however. In my opinion, this is the crucial point of modem cognitive psychology. Tuming to the investigation of human action-plans implied tuming away from the incorrect anthropological concept of behaviorism. But a category mistake was committed. The unsatisfactoriness of interchanging the firstperson point of view with the third-person observation was seen; it was realized that this unsatisfactoriness could not be rectified by addition a data from introspective psychology. It was been seen that in the concept of action-plans there arises a similar problem. One could describe it as follows: If actionplans are made from the perspective of the third person, they will become relatively unimportant and trivial for the individual, and for his emotional life. 13 One can try to draw up action-plans from the first-person point of view. However, these would not be action-plans in the same sense as discussed above. The problems may be covered up by a concept of "participating observation", but they will not be surmounted. Neisser (1976, p.7) describes the problems in the context of history of psychology: Indeed, the assumptions that underlie most contemporary work on information processing are surprisingly like those of nineteenth-century introspective psychology, though without introspection itself. His suggestions for improvement, however, do not go beyond methodological corrections: First, cognitive psychologists must make a greater eifort to understand cognition as it occurs in the ordinary environment and in the context of natural purposeful activity. This would not mean an end to laboratory experiments, but a commitment to the study ofvariables that are ecologically important rather than those that are easily manageable. Second, it will be necessary to pay more attention to the details ofthe real world in which perceivers and thinkers live, and the fine structure ofinformation which that world makes available to them. 14 A similar judgment underlies Wittgenstein's thesis conceming the role of experiments in psychology, however without hope of improvement: Rather, in psychology there is what is problematic and there are experiments which are regarded as methods of solving the problems, even though they quite by-pass the thing that is worrying us. 1S In the view of the misunderstanding of the conceptual role of "action" in cognitive psychology, tuming to experiments, leads, only to apparent pseudosolutions. 16 286 FRITZ G. WALLNER Wittgenstein's thoughts, as mentioned and reconstructed above, show us the reasons why cognitive psychology was not capable of fulfilling the hopes which had been placed in it. We must ask whether Wittgenstein can offer us suggestions for finding a way out of the present crisis in cognitive psychology. This crisis can be briefly characterized as folIows: one group of approaches does not reach our emotionallife, the other group overrides it. The computational approach of simulating cognitive processes is based on a formal assumption, i.e. that emotional life follows strict and explicable mIes, and on the assumption that psychological performances can be understood criterially; i.e. they have to be interpreted as given results. Artificial lntelligence (the symbolic paradigm) abandons all claim to explain emotionallife, but rather tries to reproduce cognitive processes under the general aspect of information processing but with results drawn from artificial models. 17 Finally we must estimate in what respect we can leam from Wittgenstein in this situation. It seems that Allan Turing's successful approach in Artificiallntelligence research narrows the horizon of Wittgenstein's thoughts interpretation by criteria. It seems that Allan Turing's approach, so succesful in Artificial Intelligence, narrows the horizons of Wittgenstein' s thought by a purely criterial interpretation. It puts in the shade the "psychological propositions" whose conceptual elucidation is so ticklish, but which could be important even for present-day psychology. We observed this, for instance, in differentiating between acting and behaving. Universität Wien NOTES J. R. Fogelin, Wittgenstein, p. 176. Wittgenstein (1980), § 920. 3 Camap (1959), p. 192. 4 Graumann (1965), p. 244. 5 Ibid. 6 RPP I, § 313. 7 Tolman poses 1922 in his m paper the following question: "By what single common set of concepts can we possibly take care both of the facts of gross behavior and of those of consciousness and imagery?" (p. 3). And he opines that a "non-physiological behaviorism", which nevertheless does also investigate behavior, has to distinguish at least four aspects: "stimulating agency, behavior cue, behavior object and behavior act" (p. 4). "Behavior object" represents in this case the meaning (depending on the situation) of a bundle of possibilities of behaving, "because of his particular training and past experience and state ofbehavior readiness ... arouse in him a very specific group of behavior tendencies; e.g., those of sitting upon, getting up from, kneeling on, moving up to 1 2 HOW LUDWIG WITTGENSTEIN WOULD HAVE REACTED 287 the table, etc. This group of aroused tendencies defines his behavior object." (p. 6). Tolman (J 959), p. 6. 9 Cf. e.g. Wann (1964). 10 RPP H, § 62. 11 RPP I, § 838. 12 RPP I, § 712. 13 Zitterbarth and Werbik (1987) already locate this methodological weakness in the beginnings of cognitivistic explanations of behavior: "Es kommt also, wenn Miller, Galanter & Pribram von 'Bildern' und 'Plänen' reden, nicht auf den 'subjektiven Sinn' an, den Handelnde mit ihrem Tun verbinden ... 'Pläne' und 'Bilder' meinen also nicht die faktischen Sinngehalte, mit denen ein Subjekt operiert, sondern 'theoretische' Konstrukte, die an die Stelle des physiologischen Reflexbogens im Behaviorismus treten." (When Miller, Galanter and Pribram speak of"images" and "plans" they are not referring to the "subjective meaning", which acting people connect with their actions ... therefore "plans" and "images" don 't mean the actual meaning content present to an acting subject, but "theoretical constructs" replacing the physiological reflex arc assumed by behaviorism.) (p. 231). 14 Neisser (J 976), pp. 7--8. 15 Wittgenstein (J 967), § 1039. 16 This opinion, saying that shortcomings arise when unclear basic concepts coincide with unproved empirical methodology, is shared by Neisser (1976) too, who writes: "Perhaps as a result, the actual development of cognitive psychology in the last few years has been disappointingly narrow, focusing inward on the analysis of specific experimental situations rather than outward toward the world beyound the laboratory." (Preface, p. XI). 17 Compare Schiffner (1989), especially p. 23 ff. 8 REFERENCES Carnap R.: 1959, 'Psychology in Physical Language'. In: Logical Positivism, ed. by A. J. Ayer, New York, p. 165 ff. Fogelin J. R.: 1976, Wittgenstein, Routledge & Kegan Paul, London. Graumann C. -F.: 1965, 'Subjektiver Behaviorismus?', Archiv./Ur die gesamte Psychologie, Bd. 117, pp. 240-251 Luckhardt C. G.: 1983, 'Wittgenstein and Behaviorism', Synthese, 56, pp. 319-338. Miller G. A., Galanter E. & Pribram K. H.: 1960, Plans and the Structure 01 Behavior, Holt, Rinehart & Winston, New York. Neisser U.: 1976, Cognition and Reality. Principles and Implications olCognitive Psychology, W. H. Freeman, New York. Rorty R.: 1977, 'Wittgensteinian Philosophy and Empirical Psychology', Philosophical Studies. 31, pp. 151-172. Schiffner E.: 1989, Verstehende Maschinen versus Verstehensmaschinen. Eine philosophische Untersuchung über Bedingungen der Möglichkeit von Maschinenintentionalität, Phi!. Diss., Wien. Stromberg W: 1987, 'Wittgenstein. Theoretical Psychology and the Classification ofPsychological Concepts', Philosophicallnvestigations, 10, pp. 11-30. Tolman E. C.: 1966, 'A New Formula for Behaviorism', in E. C.Tolman, Behavior and Psychological Man. Essays in Motivation and Learning, Univ. ofCalifornia Press, Berkeley, 288 FRITZ G. WALLNER Los Angeles, pp. 1--8. Tolman E. C.: 1932, Purposive Behavior in Animals and Men, Century, New York. Tolman E. C.: 1959, 'Principles ofPurposive Behavior', in Psychology: A Study ofa Science, ed. by S. Koch, Vol. 2., McGraw-HiIl, NewYork, pp. 92-157. Wallner F.: 1983, 'Läßt sich die Sprechakttheorie als eine präzisierende Fortführung und systematische Ausarbeitung von Wittgensteins späterer Philosophie verstehen?' ,Der Mensch und die Wissenschaft vom Menschen. Die Beiträge des XII. Deutschen Kongresses für Philosophie, ed. by G. Frey, Innsbruck, Vol. 2, p. 1033 ff. Wann, T. W. (ed.): 1964, Behaviorism and Phenomenology. Contrasting Bases for Modern Psychology, Univ. ofChicago Press, Chicago. Watson J. B.: 1913, 'Psychology as the Behaviorist views it', Psychological Review, 20, pp. 158-177. Wittgenstein L.: 1953, Philosophische Untersuchungen. Philosophical Investigations, trans. by G. E. M. Anscombe, B. Blackwell, Oxford. Wittgenstein L.: 1980, Bemerkungen über die Philosophie der Psychologie. Remar/cs on the Philosophy ofPsychology, Vol. I, ed. by G. E. M.Anscombe & G. H. v. Wright. Oxford: Basil Blackwell, Oxford (Quoted as RPP I). Wittgenstein L.: 1967, Zettel, ed. by G. E. M.Anscombe & G. H. v. Wright. B. Blackwell, Oxford Zitterbarth w., Werbik H.: 1987, 'Subjektivität als methodisches Prinzip. Argumente und Verfahrensweisen einer dialogisch-verstehenden Psychologie', in Wissenschaftstheorie und Erkenntnistheorie. Eine Einfiihrung für Psychologen und Humanwissenschaftier, ed. by J. Kriz, H. E. Lück, H. Heidbrink, Leske & Budrich. INDEX OF NAMES Ambrose A., 81 n, 136n, 183n, 231 n Anderson C., 169n Anderson T., 56n AngeleIli l., 81n Anscombe G. E., 35n, 36n, 46n, 56n, 81n, 91n, 105n, 114n, 135n, 168n, 182n, 183n, 205n, 210n, 222n, 231n, 247n,262n, 277n AristotIe, 33,37,42,43,45,47,121 Armstrong 0., 203,205n Aue M.A.E., 81n, 182n, 205n, 247n, 262n, 277n Augustine,121 Austin 1.L., 26 Ayer AJ., 247 181, 183n, 221, 222n, 225, 247n, 279282, 286n Casati R., 185 Cellucci c., 83,90n Chisholm R.M., 53,54, 56n Chomsky N., 77 Church P.A., 247n Claesges U., 192n ClarkR.W., 135n ClarkeT., 216 Coleridge S.T., 39 ComteA., 1 Copi 1., 247n Corsi G., 90n Costazza M., 279 BakerG.P., 93-95, 105n, 136 Beard R., 247n Bell J., 90n Bergmann G., 137, 144-146, 157, 167, 169n,247 Beth E.W., 88 Black M., 11, 17 Blackmann L., 169n Barrett C., 136n,277 Boltzmann L., 4 Boricic B.R., 90n Bosanquet R.G., 81n Bouveresse 1., 59 BowerG., 247n Bradley H., 152,154, 170n Brandl1., 47 BrentanoF., 3,131,136,171,181 BrochH., 9 Burnyeat M., 222n Burkhardt H., 169n D'AgostinoM., 90n DarwinC., I Davidson 0., 166-168, 170n, 221 Dell'Utri M., 117 Descartes R., 38, 42, 44, 118, 122n, 211-216,218,222 Diamond c., 81n,91 Dostoyevsky F., 4,6 Dummett M., 84, 90n Eames E., 169n Egidi R., 171,231 n Einstein A., 3, 17, 225, 230 ElieH., 81n Euclide., 68 FeyerabendP.K., 224,231n FichteJ.G., 7 Fodor J.A., 107-115,117,120,121, 123n Fogelin R., 279,286n Frank Ph., 225 Frazer J.G., 7 FregeG., 8,11,68,69, 75,8In, 170n, 171, 175,181,177,182,207,208 Cantor G., 69,70,81 Carl w., 211 Camap R., 2,3,4, 5, 7, 14, 16n, 17n, 175, 289 290 INDEX OF NAMES Freud S., 9 Frongia G., 263 Gabbay D., 91n Galanter E., 281,282, 287n Galilei G., 18n Gargani A.G., 127 Geach P.T.164, 169n Gentzen G., 84, 86-90 Gödel K., 71 Goethe J. w., 256, 261 Goodman N., 231n Graumann C.-F., 281,286n Grillparzer F., 6 GrimmJ., I Guenther F., 91n Hacker P.M.S., 93-95, 105n Hahn H., 3, 17n, 225 Haller R., xi, 17n, 223, 247n Hanson N.R., 231n Hargreaves R., 135n, 183n, 191 Hegel G.w.F., 1,4 Heidegger M., 282 Hilmy S.S., 7,8, 17n, 18n, 230, 231n, 235, 247n, 248n Hintikka J., 90n, 228, 231n Hintikka M.B. 231n Hochberg H., 137, 168n, 169n HollisM., 231n Holtzman S.H., 136n Hume D., 44,49, 142,240 Husserl E., 77-79, 81n, 192n James w., 80, 249-262 Janik A., 6-8, 17n Kaila E., 2 KantI., 12,218,225 Kelkel L., 81n KeimR.,56n Kenny A.J.P., 37,47, 56n, 91n, 135n, 169n, 183n, 222n Kierkegaard S., 4 King J., 81n, 135n, l83n Kneale M. & w., 90n Köhler w., 284 Kraus K., 9,10 Krausz M., 231n Kripke S., 50,93,94,97,99, 100, 113,225, 226 Kuhn T.S., 231n Lee D., 81n, 135n, 183n Leibniz G.w., 16,66 Leich C.M., 136n Leinfellner E., 247n Lewis H.D., 90n LoosA.,9,10 Luckhardt C.G., 81n, 135n, 182n, 205n, 247n, 262n, 277n, 280n Lukes S., 231 n Lyell C., I Macdonald M., 81n, 136n, I 83n, 231n MachE., 3 Machover M., 90n Malcolm N., 81n, 94, 192n, 195,207-210 Marconi D., 107,117-122 Martinich AP., 105n Marty A., 184n Mauthner F., 10 McAlister L.L., 136n, 183n, 231n McGinn C., 228,231n McGinn M., 219-220, 222n McGuinness B.F., xi, 35n, 81n, 9In,135n, 183n, 192n, 247n Meiland J., 231n Miller G.A., 281,283, 287n Monk R., 170n Mondadori M., 88,89, 90n, 91n Moore G.E., xi, 18n, 26, 72, 137-147, 156, 168n, 169n, 174, 183n, 195-205,211, 213,215,219 Mulligan K. 170n Mundle C., 247n Musil R., 9 Nagel T., 101,213 Neisser U., 285, 287n Neurath 0., 3,7,225 Nietzsche F., 4, 6 INDEX OF NAMES Nyiri C. 5,6, 17n-19n Nyman H., 18n, 35n, 81n, 136n, 182n, 205n, 247n, 262n, 277n Paul D., 183n, 222n, 231n, 277n Pears D., 1, 23, 35n, 135n, 148, 155-164, 170n, 183n Picardi E., 207 Pitcher G., 122n Plato., 33,34,39,160 Prawitz D., 84,85,87,88, 91n Pribram K.H., 281,282, 287n Prior A., 55,56n Protagoras., 223 PutnamH., 117-119,121, 122n, 123n,231n Quine W.O.V., 56n,231n RaskR.K., 1 Reichenbach H., 225 Rhees R., 35n, 36n, 46n, 56n, 81n, 105n, 114n, 122n, 123n, 136n, 182n, 183n, 191n, 211n, 222n, 223n, 231n, 247n, 248n, 262n, 268n, 277n, 279n Russell B., 2,3,8-11, 17n, 83, 127-129, 135n, 137, 139-170, 173,182, 184n, 191, 235, 236,240, 241, 247n Sambin G., 90n Sauer W., 231 n Savage C. 169n Schättle M. 136n, 183n, 231 n Schther R., 81n Schiffner E., 287n Schilpp A.N., 205n Schlick M., 3,191,225 Schönberg A., 9 Schubert F., 68 Schulte J., 249, 135n Searle J., 283 Shanker S., 192n 291 Shoesmith D.J., 91n SikB Z., 91n SlaterJ.G., 169n Smiley T.J., 91n Smith B., 56n,169n Smullyan R.M., 91n Smythies Y., 8ln Spencer H., 1 Spengler 0., 4, 5 Spiegelberg H., 192n Spinoza B., 16,42 SraffaP., 4 StrollA., 93 Stromberg W., 383 Stroud B., 212,216-218, 222n Sundholm G., 91n Szabo M.E., 90n Tarski A., 167 Tolman E.C., 281, 286n Tolstoy L., 8-10 Toulmin S., 23ln Travis C., 56n Turing A., 286 Waismann F., 91n, 174 Wallner F., 279,283 Wann T. w., 287n Watson J.B., 280,281 Werbik H., 287n White R., 135n. 183n, 191n Whitehead A.N., 169n Williams B., 222n Winch P., 35n, 136n, 247n Wright C., 231 n Wright G.H. (von), 1,16,35,46, 81n, 90n, 91n, 135n, 136, 168n, 182n, 205n, 222n, 231n, 247n, 262n, 277n Zermelo E., 81n Zitterbach W., 287 INDEX OF S UBJECTS Behaviorism, 37,44,45,239,240,241,245, 279-287 Iinguistic, 280 non-physiological, 286 subjective, 281,282 Believe/belief, 137-144, 168, 176-178, 195-205, 207-210, 280 'Bewegungsgefühlraum', 190 Bodily states, 48. See also Mental states Body, 48 Brain, 39 'brain in the vat', 117-119,212 Abilities, 43, 44,47 Absolute concepts, 230 Acquaint"!lnce, 162 Action/s, 54, 201, 202, 209 and certainty, 215,216 and facts, 176-178 individual, 179 intentional/voluntary, 175, 176,178 rule-following 22. See also Explanation, Intentional explanation Action theory, 283-286, 288 Actualisations, 43 Actuality, 37,43 and potentiality, 47. See also Potentiality Analytic truth, 214 Analyticity, 169 Artificial intelligence, 286 Aspects, theory of, xi Assertion, 207, 208, 210 and supposition, 197-199,203 indirect, 200 tentative or hesitant, 197, 202, 204. See also Assumption; Force indicator Associating and thinking, 29 Associationist principle Russell's, 130 Asimmetry, See Firstlthird person Assumption ('Annahme'), 208,209 , Attente', 60, 61. See also Expectation Attitudeis, 207 non epistemic, 210,222, See Dispositional states, Judgmentls 'Bedeutung', 182 'Bedeutungskörper' , See 'Corps de signification' Behavior, 179, 180,279-287 cognitivistic explanations of, 287 Behavioral data, 224 Care ('Sorge'), 256-261 Cartesian ego, 9 Cartesianism, 37,47 Causal explanation, See Explanation Causallaws, 29 Causality 284. See also Intentionality Causation and empirical concomitance, 228 mnemic, See Associationist principle Causes, See Motives Certainty/ties, xi, 211-222 Chemistry, 282 'Classes', 'theorie de', 71 Cognitive psychology, 279,282,285,286 computational approach to, 286 Colours, 172, 185, 186 phenomenological theory of, 174 Complexes, 156, 157 'Compositionalite, 'principe de', 75 Computer, 119,212. See also Machine language 'Conceptual analysis', 173 Conditionalls, 208, 209 Consequentialist theory of meaning, See Meaning Conservatism, 6 'Contradiction' 293 294 INDEX OF SUBJECTS logical, 200 'loi de', 80 Conventionalism, 24, 224, 225 'Corps de signification', 74,75,79,80 'Couleur', 72, 73. See also Colours Criteria, 177,181 pragmatic, 176 semantic, 218 Custom, 227 Deduction system of natural, 85-87, 90 Definition ostensive, 38 'Demonstration', 65,66,68,71,76 'Denotation', 75 Descriptions definite, 139, 140, 162, 165 'Desir', 60,61 Desires, 176 Dispositional states/dispositions, 43, 203 Doubt, xi, 177,212-214 and certainty, 221 cartesian, 211,212,215 existential, 14 internal, 215-217 sceptical, 211, 213, 214 Dream language, 27&-276 Dreaming, argument from, 212,217,218 Drearns, interpretation of, 270, 274-276 Emotions, xi, 249-262, 285 and bodily sensations, 249,250, 252-256 and care ('Sorge'), 256,261 and imagination, 257 and one's own mind, 250 causes of, 256-261 expressions of, 252-260, 285 symptons of, 257-259 'Energeia', 43 'Ensembles', 'theorie de', 69-71 'Entites'/'etres' geometriques, 68, 69 intermediaires/fantömatiques, 59-81 Epistemic preference, 53-55 Epistemic states, 8 Error, possibility of, 216, 217 'Espace' 'physique', 73 'visuel', 73 Essentialism, 11, 55 Excluded middle, principle of, 80, 90 Expectation/expecting, expect, 129-133, 176, 178, 243-245 Explanation, 29,30 Carnap's causal, 247 causal/intentional, 175, 178, 180 mentalistic, 329,240 non-causal, 177 of actions, 175, 178 of behavior, 283. See Intentional explanation Extensionality, 47-49, 54 External relations, See Relationls Factls atomic, 170n empirical, 214 . !,xistential, 153 existentially general, 153,160,161,163, ·"164 grammatical, 208 intentional, 153 non-existent, 138, 139, 141, 145, 151 possible, 144, 145, 166 potential, 145 relational, 149. See also Possibility/ies 'Fait' 'negatif', 64 'possible', 61 Falsity/false, See Truthltrue Firstlthird person, 49,51,9&-101,196,202, 205,209,279,280,282-285 Force indicator, 207, 208 Forms of life, 228, 229 Freudian theory, See Psychoanalysis 'Gedankeninhalte' . See SenseIs 'Generalite', 71 Generality empiricallinductive, 172, 181 INDEX OF SUBJECTS high, 172,180,181 'Geometrie', 68 'Grammaire' 'regles de', 63, 72, 73, 75, 76 'philosophique', 62, 63, 72 'pure logique', 77,79 Grammar and psychology, 171-184 philosophical, 171-184,225 surface/depth grammar, 208, 242, 243, 246. See also Rules 'Hexis' , 43, 44 'Homunculus fallacy', 39 Hopes, 176, 178 I (pronoun), 48, 98-100, 104 Idealism ontological, 221 Imagination, 257 Immediate experience, 173, 180 Implication, 88 'Impossibilite' les', See 'Possibilite' les' I 'Impossibilite' les' Impossibility, degrees of, 59,61 Incommensurability, thesis of, 224 Incomplete symbols, See Descriptions Indeterminacy of translation, 224 Individuals, 188, 189,225 Induction/inductive procedures, 178 'Infini' 'potentiel' I'actuel', 69-72 Infinity axiom, 70 Inner, See Mental Inner eye, 47 Inner experience, See Internal experience Inner life, xi, 117, 236 Inner/outer, 145, 246 Intensions, 75 Intentionls, 176-178, 182 sense and content of, 146 Intentionality, xi, 127-136, 153, 175 Bergmann'sandMoore'sanalysisof, 144 and causality, 178. See also Relational analysis Intentional 295 action, See Actionls contexts/propositions, 158, 159, 163, 175,177-179,181 element of language, 182 acts, 128, 129, 131, 132 Intentional explanation, 178 Intentional reference, 137-170 Internal experience/s, 174, 175, 182 phenomenological analysis of, 171. See also Mental processes Internal properties, 143-145, 152, 155 Internal relations, See Relationls Internal statements, 221 Introspection/introspectionism, 249, 254, 280,282 Intuition in Husserl, 78 Irrationalism, xi Judgmentls agreement in, 227,228 concept of, 219 relation al analysis of, 149 synthetic a prior, 225 value, 178 'Kinesis', 43 Knowledge and action, 215,216 and certainty, 221 empirical, 214 factual, 176 foundations of, 220 holistic, 218 objective, 213 theory of, 171 trascendental conditions of, 218 'Langage des valences', 67 Language, 24, 25, 185 and mind, 235-248 and world, 26 as a caJculus, 129 as a system of internal relations, 128129 critique of, 10, 12 296 INDEX OF SUBJECTS ideal, 10 intentional use of, 181 mental, 31 mythology of, See also Mythology natural state of, 11, 12 'never.;tever', II non-public, 114 of data, 173 ofthought, 107-114, 117, 120, 121 philosophy of, xi, 10, 28, 181 pragmatic nature of, 181 primary, See Physicallanguage public, 122 secondary, See Phenomenological language sodal dimension of, 117, 123 thing, 221. See also Dream language, Machine language, Private language Language-game/s, 191,216,218,219,228 verbal fallacy of, 228, 229 Law of mnemic causation, See Associationist principle Laws, 174 'Lebensformen', See Forms oflife 'Lingua mentalis', See Private language Linguistics, 181 Logic, 171-173 modal, 55 philosophy of, 10 truth-functional/extensional, 172, 177, 181 Logical constants, 128 form/s, 127 product, 127 relation, See Intentional reference Logical atomism, 24 Logical positivism, 2, 15, 191 Logical symbols, meaning of, 83-91 Logical relation, See Intentional reference 'Mathernatiques', 65, 66, 68-71 Machine-Ianguage, 112, 113, 133 computer'sandprogrammer's, 112,113. See also Computer Materialism, 44,45,240,241,245 Materialist theory of mind, 205 Mathematical and physical language, 225 concepts, xi, 177, 179 formulas, See Rules Mathematics, 180,224 foundations of, 171 Meaning as use, 128 causal theory of, 118 concept of, 235 consequentialist theory of, 84, 89, 90 criteria of, 176 non-Fregean theory of, 171 ofthe word, 214 theories of, 171, 175 'Means that', 167 Medicine, 282 Memory, xi, 263-277 and knowledge, 263, 270 and language, 263, 270 as 'criterion for the past', 271,272 concept of, 29 justification of, 264-268 two notions of, 263,264 Mental and bodily states, 48 and physical distinction, 211 causes, 30 concepts, 179 images, 244 involvement, 53-55 objects, 181,282 phenomena, 30, 175, 176, process/es, 39, 181,243 states, 177,199-201,203,207,209,244. See also Psychological/mental Mentallife, grammar of, 127 Mentalese/mentalesians, 117-123 Mentalism, 119, 121, 122,237,238,240, 245 Metaphysical claims three types of, 47-53 Metaphysics, 14-16,37-38 dynamic, 43,46 foundationalist, 42, 45 INDEX OF SUBJECTS of mind, 47-56 speculative, 54 spiritualistic, 38,41,42,44-46 trascendental, 52, 53 Mind, 44,45 and brain, 28, 29, 44, metaphysical concept of, 41 metaphysical theory of, 37 metaphysics of, 37,47-56 philosophy of, 15,43,44,171,175 Russell's theory of, 235 Modal involvement, 55 Modemity, 1-7, 16 Moore's paradox, See Paradox Motives and causes, 178, 278. See also Reasons 'Myth'/s', 238,246,276 Mythology, 235-248 of mental processes, 237 of symbolism, 237 Natural history, knowledge of, 213 Necessary propositions and non-necessary, 225 and probabilistic, 173 Necessity, 55 nomological, 110, 115 'Negation', 64 Neural explanation, 31 Neurology, 28 Neurophysiology, 29 'Nombres' 'entiers', 70 'transfinis " 69 'Ombre' ,See 'Possibilite'/es' /'Impossibilite' / es' Objectivity, 213, 221-224 Objects material, 186 visual, 187 Optimism rationalist, 7 Other minds/people/persons, 44-49,54,55, 201,209,221,259,260 297 Pain, 49, 52, 255, 280 Paradox Kripke's, 226 Moore's, xi, 195-205, 207-210 Russell's, 169n Parts visual, 187. See also Whole/parts Perception, problem of, 192 Person/s, 48, 54, 55 "personal experiences", 175. See also Firstlthird person, Other minds Persuasion, 229 Phenomenalism, 240,241,245 Phenomenalistic sensualism, 3 Phenomenological-not-phenomenological, 172, 178 Phenomenological/phenomenal analysis, 173-175 language, 191 propositions/sentences, 174 problems/"problems", 173, 192 Phenomenology, 171, 172-178, 183n, 185-192 Philosophicalpsychology, xi, 174, 175, 181, 235-288 Philosophy 'first', 42 of psychology, See Philosophical psychology Physical language, 280 objects, 174 phenomena, 270 sentences/propositions, 174. See also Phenomenological Physics, 173, 174,224,235,282 Physiologism, 281, 282 Platonism, 31-33 'Possibilite' /es' /'Impossibilite' /es' , 61, 64, 66,67,70,72,73,75-79 'chimique', 63, 66, 67 'et existence', 66 'et realite' /'reel' , 59-81 'et sens', 59--81 'geometrique', 67 'Iogique'/'grammaticale', 62 298 INDEX OF SUBJECTS Possibility/ies and facts, 25-28 in Russen, 143, 145 logical, 195, 197, 198, 202 Possible/'Possible' See Fact/s, 'Possibilite'les' I' ImpossibilitC' I es' IPossibility/ties Possible worlds, 56n, 175 Potentiality, 37,43,44 and actuality, 47 Powers active and passive, 43, 44 Prediction, 177, 178 Presentlpasttense, 196, 197, 199, 205, 207, 209 Primary phenomena, See Phenomenological language Private language, xi, 51, 52, 53, 93, 263 argument, 107-115, 117-123 as 'lingua mentalis', 230 as non public language, 109, 110 Proof, 225 method of, 180 Propositional attitudes, 173, 175, 177 attributes, 144-147 entities, 137, 140, 149, 150-152, 158 Propositions/sentences/statements external existential, 221 'grammaticales' I'mathematiques', 67, 68, 70 normative, 239 not-extensional, 182 systematic/empirical, 178, 179, 225 Propositionalism Tolman's, 281 Psychoanalysis, 269, 273, 275 Psychological/mental concepts/propositions, xi, 171, 172, 175, 180, 181,202,236,238-243,245, 246,279,280,282,286 events, 178 expressions, 245 phenomenalprocesses, 177 taxonomicallcausal view of, 236, 240 verbs, 173 Psychologism, xi, 174-177, 182 Psychology, 137,171-184,237,279-287 and grammar, 171-184 and philosophy, 29 and physics, 279 cognitive, See Cognitive psychology contemporary, 269 depth, 270 empirical, 236, 238, 246 history of, 285 humanistic, 283 introspective, 285 of mental states, 185 phenomenological, 281 philosophical, See Philosophical psychology Quantification/quantifier existential, 86 universal, 83,84,88 Quantities and measures, problem of, 185, 186 Realism, 11,213, 'Realite', See 'Possibilite'/es'I'Impossibilite' les' Reasons, 216, 219 Reductionism, 44 in the theory of memory, 267 physiological, 281 vehic1e, 44 Watson's, 281 'Reel' 'et son ombre', 59-81. See also 'PossibilitC' les' I' Impossibilite' les' Reference, 119, 120. See also Intentional reference Relation/s external, 128, 182 internal, 128, 184n Relational analysis Russell's, 139,141, 142, 145, 147, 148, 151, 156,164-166 Relational contexts non-symmetrical, 158, 159, 163 Re1ativism, xi, 180,223-231 INDEX OF SUBJECTS epistemological, 223-231 ontological, 224 Protagoras thesis, 223 Relativity, theory of, 230 Representation/s, 201, 212 internal, 107, 11 0, 111 mental, 117, 119, 121 physical, 119 'Representation', 'systeme de', 62,63 Representational relation Bergmann's, 145 Representational system, 122 Rule following, 28,31-35,93-105,226 Kripke's argument of, 50,51 Rule/s, 178,179,181,182,218 and agreement, 227,228 and mathematical formulas, 95,104 and postulates, 225, 226 and trespass, 103, 104 as edicts, decrees, etc., 94-97 constative, 96,97,99--103 disjunction, 86 elimination, 84-87,89,90 inference, 83, 85 introduction, 84, 88, 90 of agame, 162 of gramrnar/grammatical, 162,225 of induction, 87 system of, 176,178-180 universal, 83 Scepticism, xi, 26,47,49,211-222 Science/s, 173 and philosophy, 23-36 empirical, 172 mental, 179 'naturelle', 70 of science, 23 Scientific theories, 235, 236 Seeing concept of, 237 Self, concept of, 279 Self-contradictory sentence, 196,202,203 Self-interpretation/self-understanding, 283 Self-observation, 202, 282-284 Semantics of possible worids, 56n 299 Sensation/s, xi terms, 117 expression of, 280 'Sens', 66, 73-75, 80 'determination de' ('Sinnbestimmung'), 68 et 'contresens' , 77 et 'non-sens', 59,61-66,70,73,74,79, 80 'remplissant', 76 'signification', 76-78 'utilisation de' ('Sinnverwendung'), 68 Sense/s, 174-176 as 'Gedankeninhalte' , 181 as 'Sinne', 181,182 Sense datum, language of, 132 Sense-impressions, 174, 209 Sense--perception, 217 Set theory 180. See also 'Ensembles', 'theorie des' Shadow of the expected fact, 131 Shadowy beings, See 'Entites'/'etres' 'Signification', 76-80 Similarity, notion of, 130-132 'Sinn'/'Sinne', See Sense/s, 'Sens' Solipsism, 25-28 Soul-body, 39,46. See also Body Space, 189 absolute, 186-188,191 'Gesichtsraum' , 192 orientation ('Gefühlsraum'), 189-191, 192n visuaI185-192. See also 'Espace' Speech act theory, 283 Subjectivism, xi, 180 Supposition, 197-199 T-sentences, 167 Tautologies, 225 Thinking without looking, 203 Thought, 208 language of, See Assumption, Language ofthought Thought-experiment 300 INDEX OF SUBJECTS James', 249,259,251 True 'for me', 224 'per se', 223,224 Truth-conditions, 166 Truth/true, 176, 223, 228-230 ana1ytica/conceptual, 253 subjective, 272 Truth-functions, II Truth-makers, 151, l70n Truth values, 174 Understanding, 39,40, 41, 127, 182, 214, 218,228 'Übersichtliche Darstellung' ('perspicuous presentation '), 12 Validity, 85,87,88 Value judgments, See Judgmentls Verification, principle of, 280, 282 Verificationism, 107, \08 'View from nowhere', 101 Vienna Circle/'Wiener Kreis', 3,8 Visual coordinates system of 188-190. See also Space Whole/parts, 186, 187, 190 Will, xi 'Wissenschaftliche Weltauffassung' , 3 World and representation, 212 extema1, 211, 217 Wor1d picture/,We1tbild', 214,218-221 SYNTHESE LIBRARY Studies in Epistemology, Logic, Methodology, and Philosophy of Science 1. J. M. Bochenski, APrecis of Mathematical Logic. Translated from French and German by O. Bird. 1959 ISBN 90-277-0073-7 2. P. Guiraud, Problemes et meehodes de la statistique linguistique. 1959 ISBN 90-277-0025-7 3. H. Freudenthai (ed.), The Concept and the Role of the Model in Mathematics and Natural and Social Sciences. 1961 ISBN 90-277-0017-6 4. E. W. Beth, Formal Methods. An Introduction to Symbolic Logic and to the Study ofEffective Operations in Arithmetic and Logic. 1962 ISBN 90-277-0069-9 5. B. H. Kazemier and D. Vuysje (eds.), Logic and Language. Studies dedicated to Professor Rudolf Carnap on the Occasion of His 70th Birthday. 1962 ISBN 90-277-0019-2 6. M. W. Wartofsky (ed.), Proceedings ofthe Boston Colloquiumfor the Philosophy of Science, 1961-1962. [Boston Studies in the Philosophy of Science, Vol. I] 1963 ISBN 90-277-0021-4 7. A. A. Zinov'ev, Philosophical Problems of Many-valued Logic. A revised edition, edited and translated (from Russian) by G. Küng and D.D. Comey. 1963 ISBN 90-277-0091-5 8. G. Gurvitch, The Spectrum of So cia I Time. Translated from French and edited by M. Korenbaum and P. Bosserman. 1964 ISBN 90-277-0006-0 9. P. Lorenzen, Formal Logic. Translated from German by F.J. Crosson. 1965 ISBN 90-277-0080-X 10. R. S. Cohen and M. W. Wartofsky (eds.), Proceedings ofthe Boston Colloquiumfor the Philosophy of Science, 1962-1964. In Honor of Philipp Frank. [Boston Studies in the Philosophy of Science, Vol. 11] 1965 ISBN 90-277-9004-0 11. E. W. Beth, Mathematical Thought. An Introduction to the Philosophy of Mathematics. 1965 ISBN 90-277-0070-2 12. E. W. Beth and J. Piaget, Mathematical Epistemology and Psychology. Translated from French by W. Mays. 1966 ISBN 90-277-0071-0 13. G. Küng, Ontology and the Logistic Analysis of Language. An Enquiry into the Contemporary Views on Uni versals. Revised ed., translated from German. 1967 ISBN 90-277-0028-1 14. R. S. Cohen and M. W. Wartofsky (eds.), Proceedings ofthe Boston Colloquiumfor the Philosophy of Sciences, 1964-1966. In Memory of Norwood Russell Hanson. [Boston Studies in the Philosophy of Science, Vol. III] 1967 ISBN 90-277-0013-3 15. C. D. Broad, Induction, Probability, and Causation. Selected Papers. 1968 ISBN 90-277-0012-5 16. G. Patzig, Aristotle's Theory of the Syllogism. A Logical-philosophical Study of BookA of the Prior Analytics. Translated from German by J. Barnes. 1968 ISBN 90-277-0030-3 17. N. Rescher, Topics in Philosophical Logic. 1968 ISBN 90-277-0084-2 18. R. S. Cohen and M. W. Wartofsky (eds.), Proceedings ofthe Boston Colloquiumfor the Philosophy of Science, 1966-1968, Part I. [Boston Studies in the Philosophy of Science, Vol. IV] 1969 ISBN 90-277-0014-1 19. R. S. Cohen and M. W. Wartofsky (eds.), Proceedings ofthe Boston Colloquiumfor the Philosophy of Science, 1966-1968, Part ll. [Boston Studies in the Philosophy of Science, Vol. V] 1969 ISBN 90-277-0015-X SYNTHESE LffiRARY 20. J. W. Davis, D. J. Hockney and W. K. Wilson (eds.), Philosophical Logic. 1969 ISBN 90-277-0075-3 21. D. Davidson and J. Hintikka (eds.), Words and Objections. Essays on the Work of W. V. Quine. 1969, rev. ed. 1975 ISBN 90-277-0074-5; Pb 90-277-0602-6 22. P. Suppes, Studies in the Methodology and Foundations of Science. Selected Papers ISBN 90-277-0020-6 from 1951 to 1969. 1969 23. J. Hintikka, Modelsfor Modalities. Selected Essays. 1969 ISBN 90-277-0078-8; Pb 90-277-0598-4 24. N. Rescher et al. (eds.), Essays in Honor of Carl G. Hempel. A Tribute on the ISBN 90-277-0085-0 Occasion ofHis 65th Birthday. 1969 25. P. V. Tavanec (ed.), Problems ofthe Logic of Scientijic Knowledge. Translated from Russian.1970 ISBN 90-277-0087-7 26. M. Swain (ed.), Induction, Acceptance, and Rational Belief. 1970 ISBN 90-277-0086-9 27. R. S. Cohen and R. J. Seeger (eds.), Ernst Mach: Physicist and Philosopher. [Boston Studies in the Philosophy of Science, Vol. VI]. 1970 ISBN 90-277-0016-8 ISBN 90-277-0155-5 28. J. Hintikka and P. Suppes, Information and Inference. 1970 29. K. Lambert, Philosophical Problems in Logic. Some Recent Developments. 1970 ISBN 90-277-0079-6 30. R. A. Eberle, Nominalistic Systems. 1970 ISBN 90-277-0161-X 31. P. Weingartner and G. Zecha (eds.), Induction, Physics, and Ethics. 1970 ISBN 90-277-0158-X 32. E. W. Beth, Aspects of Modern Logic. Translated from Dutch. 1970 ISBN 90-277-0173-3 33. R. Hilpinen (ed.), Deontic Logic. Introductory and Systematic Readings. 1971 See also No. 152. ISBN Pb (1981 rev.) 90-277-1302-2 34. J.-L. Krivine, Introduction to Axiomatic Set Theory. Translated from French. 1971 ISBN 90-277-0169-5; Pb 90-277-0411-2 35. J. D. Sneed, The Logical Structure ofMathematical Physics. 2nd rev. ed., 1979 ISBN 90-277-1056-2; Pb 90-277-1059-7 36. C. R. Kordig, The Justijication of Scientific Change. 1971 ISBN 90-277-0181-4; Pb 90-277-0475-9 37. M. Capek, Bergson and Modern Physics. AReinterpretation and Re-evaluation. [Boston Studies in the Philosophy of Science, Vol. VII] 1971 ISBN 90-277-0186-5 38. N. R. Hanson, What I Do Not Believe, and Other Essays. Ed. by S. Toulmin and H. Woolf. 1971 ISBN 90-277-0191-1 39. R. C. Buck and R. S. Cohen (eds.), PSA 1970. Proceedings of the Second Biennial Meeting of the Philosophy of Science Association, Boston, Fall 1970. In Memory of Rudolf Carnap. [Boston Studies in the Philosophy of Science, Vol. VIII] 1971 ISBN 90-277-0187-3; Pb 90-277-0309-4 40. D. Davidson and G. Harman (eds.), Semantics ofNatural Language. 1972 ISBN 90-277-0304-3; Pb 90-277-0310-8 41. Y. Bar-Hillel (ed.), Pragmatics ofNatural Languages. 1971 ISBN 90-277-0194-6; Pb 90-277-0599-2 42. S. Stenlund, Combinators, Ä-Terms and ProofTheory. 1972 ISBN 90-277-0305-1 43. M. Strauss, Modern Physics and Its Philosophy. Selected Paper in the Logic, History, and Philosophy of Science. 1972 ISBN 90-277-0230-6 44. M. Bunge, Method, Model and Matter. 1973 ISBN 90-277-0252-7 ISBN 90-277-0253-5 45. M. Bunge, Philosophy of Physics. 1973 SYNTHESE LIBRARY 46. A. A. Zinov'ev, Foundations 0/ the Logical Theory 0/ Scientific Knowledge (Complex Logic). Revised and enlarged English edition with an appendix by G. A. Smirnov, E. A. Sidorenka, A. M. Fedina and L. A. Bobrova. [Boston Studies in the Philosophy of Science, Vol. IX] 1973 ISBN 90-277-0193-8; Pb 90-277-0324-8 47. L. Tondi, Scientific Procedures. A Contribution conceming the Methodological Problems of Scientific Concepts and Scientific Explanation. Translated from Czech by D. Short. Edited by R.S. Cohen and M.W. Wartofsky. [Boston Studies in the Philosophy of Science, Vol. X] 1973 ISBN 90-277-0147-4; Pb 90-277-0323-X ISBN 90-277-0192-X 48. N. R. Hanson, Constellations and Conjectures. 1973 49. K. J. J. Hintikka, J. M. E. Moravcsik and P. Suppes (eds.), Approaches to Natural Language. 1973 ISBN 90-277-0220-9; Pb 90-277-0233-0 50. M. Bunge (ed.), Exact Philosophy. Problems, Tools and Goals. 1973 ISBN 90-277-0251-9 51. R. J. Bogdan and I. Niiniluoto (eds.), Logic, Language and Probability. 1973 ISBN 90-277-0312-4 52. G. Pearce and P. Maynard (eds.), Conceptual Change. 1973 ISBN 90-277-0287-X; Pb 90-277-0339-6 53. I. Niiniluoto and R. Tuomela, Theoretical Concepts and Hypothetico-inductive In/erence. 1973 ISBN 90-277-0343-4 54. R. Fraisse, Course oJ Mathematical Logic - Volume 1: Relation and Logical Formula. Translated from French. 1973 ISBN 90-277-0268-3; Pb 90-277-0403-1 (For Volume 2 see under No. 69). 55. A. Grünbaum, Philosophical Problems 0/ Space and Time. Edited by R.S. Cohen and M.W. Wartofsky. 2nd enlarged ed. [Boston Studies in the Philosophy of Science, Vol. XII] 1973 ISBN 90-277-0357-4; Pb 90-277-0358-2 56. P. Suppes (ed.), Space, Time and Geometry. 1973 ISBN 90-277-0386-8; Pb 90-277-0442-2 57. H. Kelsen, Essays in Legal and Moral Philosophy. Selected and introduced by O. Weinberger. Translated from German by P. Heath. 1973 ISBN 90-277-0388-4 58. R. J. Seeger and R. S. Cohen (eds.), Philosophical Foundations 0/ Science. [Boston Studies in the Philosophy of Science, Vol. XI] 1974 ISBN 90-277-0390-6; Pb 90-277-0376-0 59. R. S. Cohen and M. W. Wartofsky (eds.), Logical and Epistemological Studies in Contemporary Physics. [Boston Studies in the Philosophy of Science, Vol. XIII] 1973 ISBN 90-277-0391-4; Pb 90-277-0377-9 60. R. S. Cohen and M. W. Wartofsky (eds.), Methodological and Historical Essays in the Natural and Social Sciences. Proceedings 0/ the Boston Colloquium Jor the Philosophy 0/ Science, 1969-1972. [Boston Studies in the Philosophy of Science, Vol. XIV] 1974 ISBN 90-277-0392-2; Pb 90-277-0378-7 61. R. S. Cohen, J. J. Stachel and M. W. Wartofsky (eds.), For Dirk Struik. Scientific, Historical and Political Essays. [Boston Studies in the Philosophy of Science, Vol. XV] 1974 ISBN 90-277-0393-0; Pb 90-277-0379-5 62. K. Ajdukiewicz, Pragmatic Logic. Translated from Polish by O. Wojtasiewicz. 1974 ISBN 90-277-0326-4 63. S. Stenlund (ed.), Logical Theory and Semantic Analysis. Essays dedicated to Stig Kanger on His 50th Birthday. 1974 ISBN 90-277-0438-4 64. K. F. Schaffner and R. S. Cohen (eds.), PSA 1972. Proceedings 0/ the Third Biennial Meeting 0/ the Philosophy 0/ Science Association. [Boston Studies in the Philosophy of Science, Vol. XX] 1974 ISBN 90-277-0408-2; Pb 90-277-0409-0 SYNTHESE LIBRARY 65. H. E. Kyburg, Jr., The Logical Foundations 0/ Statisticalln/erence. 1974 ISBN 90-277-0330-2; Pb 90-277-0430-9 66. M. Grene, The Understanding 0/ Nature. Essays in the Philosophy of Biology. [Boston Studies in the Philosophy of Science, Vol. XXIII] 1974 ISBN 90-277-0462-7; Pb 90-277-0463-5 67. J. M. Broekman, Structuralism: Moscow, Prague, Paris. Translated from German. 1974 ISBN 90-277-0478-3 68. N. Geschwind, Selected Papers on Language and the Brain. [Boston Studies in the ISBN 90-277-0262-4; Pb 90-277-0263-2 Philosophy of Science, Vol. XVI] 1974 69. R. Fraisse, Course 0/ Mathematical Logic - Volume 2: Model Theory. Translated from French. 1974 ISBN 90-277-0269-1; Pb 90-277-0510-0 (For Volume 1 see under No. 54) 70. A. Grzegorczyk, An Outline 0/ Mathematical Logic. Fundamental Results and Notions explained with all Details. Translated from Polish. 1974 ISBN 90-277-0359-0; Pb 90-277-0447-3 71. F. von Kutschera, Philosophy o/Language. 1975 ISBN 90-277-0591-7 72. J. Manninen and R. Tuomela (eds.), Essays on Explanation and Understanding. Studies in the Foundations ofHumanities and Social Sciences. 1976 ISBN 90-277-0592-5 73. J. Hintikka (ed.), Rudolj Carnap, Logical Empiricist. Materials and Perspectives. 1975 ISBN 90-277-0583-6 74. M. Capek (ed.), The Concepts 0/ Space and Time. Their Structure and Their Development. [Boston Studies in the Philosophy of Science, Vol. XXII] 1976 ISBN 90-277-0355-8; Pb 90-277-0375-2 75. J. Hintikka and U. Remes, The Method 0/ Analysis. Its Geometrical Origin and Its General Significance. [Boston Studies in the Philosophy of Science, Vol. XXV] 1974 ISBN 90-277-0532-1; Pb 90-277-0543-7 76. J. E. Murdoch and E. D. Sylla (eds.), The Cultural Context 0/ Medieval Learning. [Boston Studies in the Philosophy of Science, Vol. XXVI] 1975 ISBN 90-277-0560-7; Pb 90-277-0587-9 77. S. Amsterdamski, Between Experience and Metaphysics. Philosophical Problems of the Evolution of Science. [Boston Studies in the Philosophy of Science, Vol. XXXV] 1975 ISBN 90-277-0568-2; Pb 90-277-0580-1 78. P. Suppes (ed.), Logic and Probability in Quantum Mechanics. 1976 ISBN 90-277-0570-4; Pb 90-277-1200-X 79. H. von Helmholtz: Epistemological Writings. The Paul Hertz / Moritz Schlick Centenary Edition 0/ 1921 with Notes and Commentary by the Editors. Newly translated from German by M. F. Lowe. Edited, with an Introduction and Bibliography, by R. S. Cohen and Y. Elkana. [Boston Studies in the Philosophy of Science, Vol. XXXVII] 1975 ISBN 90-277-0290-X; Pb 90-277-0582-8 80. J. Agassi, Science in Flux. [Boston Studies in the Philosophy of Science, Vol. XXVIII] 1975 ISBN 90-277-0584-4; Pb 90-277-0612-2 81. S. G. Harding (ed.), Can Theories Be Refuted? Essays on the Duhem-Quine Thesis. 1976 ISBN 90-277-0629-8; Pb 90-277-0630-1 82. S. Nowak, Methodology 0/ Sociological Research. General Problems. 1977 ISBN 90-277-0486-4 83. J. Piaget, J.-B. Grize, A. Szeminska and V. Bang, Epistemology and Psychology 0/ Functions. Translated from French. 1977 ISBN 90-277-0804-5 SYNTHESE LIBRARY 84. M. Grene and E. Mendelsohn (eds.), Topics in the Philosophy of Biology. [Boston Studies in the Philosophy of Science, Vol. XXVß] 1976 ISBN 90-277-0595-X; Pb 90-277-0596-8 85. E. Fischbein, The Intuitive Sources of Probabilistic Thinking in Children. 1975 ISBN 90-277-0626-3; Pb 90-277-1190-9 86. E. W. Adams, The Logic of Conditionals. An Application of Probability to Deductive Logic. 1975 ISBN 90-277-0631-X 87. M. ｐｲｺ･ｬｾｫｩ＠ and R. W6jcicki (eds.), Twenty-Five Years of Logical Methodology in Po land. Translated from Polish. 1976 iSBN 90-277-0601-8 88. J. Topolski, The Methodology of History. Translated from Polish by O. Wojtasiewicz. 1976 ISBN 90-277-0550-X 89. A. Kasher (ed.), Language in Focus: Foundations, Methods and Systems. Essays dedicated to Yehoshua Bar-Hillel. [Boston Studies in the Philosophy of Science, Vol. XLIß] 1976 ISBN 90-277-0644-1; Pb 90-277-0645-X 90. 1. Hintikka, The Intentions of Intentionality and Other New Models for Modalities. 1975 ISBN 90-277-0633-6; Pb 90-277-0634-4 91. W. Stegmüller, Collected Papers on Epistemology, Philosophy of Science and History ofPhilosophy. 2 Volumes. 1977 Set ISBN 90-277-0767-7 92. D. M. Gabbay, Investigations in Modal and Tense Logics with Applications to Problems in Philosophy and Linguistics. 1976 ISBN 90-277-0656-5 93. R. J. Bogdan, Local Induction. 1976 ISBN 90-277-0649-2 94. S. Nowak, Understanding and Prediction. Essays in the Methodology of Social and Behavioral Theories. 1976 ISBN 90-277-0558-5; Pb 90-277-1199-2 95. P. Mittelstaedt, Philosophical Problems of Modern Physics. [Boston Studies in the Philosophy of Science, Vol. XVIII] 1976 ISBN 90-277-0285-3; Pb 90-277-0506-2 96. G. Holton and W. A. Blanpied (eds.), Science and Its Public: The Changing Relationship. [Boston Studies in the Philosophy of Science, Vol. XXXIß] 1976 ISBN 90-277-0657-3; Pb 90-277-0658-1 97. M. Brand and D. Walton (eds.), Action, Theory. 1976 ISBN 90-277-0671-9 98. P. Gochet, Outline of a Nominalist Theory of Propositions. An Essay in the Theory of Meaning and in the Philosophy of Logic. 1980 ISBN 90-277-1031-7 99. R. S. Cohen, P. K. Feyerabend, and M. W. Wartofsky (eds.), Essays in Memory of Imre Lakatos. [Boston Studies in the Philosophy of Science, Vol. XXXIX] 1976 ISBN 90-277-0654-9; Pb 90-277-0655-7 100. R. S. Cohen and 1. J. Stachel (eds.), Selected Papers of Leon Rosenjield. [Boston Studies in the Philosophy of Science, Vol. XXI] 1979 ISBN 90-277-0651-4; Pb 90-277-0652-2 101. R. S. Cohen, C. A. Hooker, A. C. Michalos and J. W. van Evra (eds.), PSA 1974. Proceedings ofthe 1974 Biennial Meeting ofthe Philosophy of Science Association. [Boston Studies in the Philosophy of Science, Vol. XXXß] 1976 ISBN 90-277-0647-6; Pb 90-277-0648-4 102. Y. Fried and J. Agassi, Paranoia. A Study in Diagnosis. [Boston Studies in the Philosophy of Science, Vol. L] 1976 ISBN 90-277-0704-9; Pb 90-277-0705-7 103. M. Przelec;;ki, K. Szaniawski and R. W6jcicki (eds.), Formal Methods in the Methodology ofEmpirical Sciences. 1976 ISBN 90-277-0698-0 104. J. M. Vickers, Belief and Probability. 1976 ISBN 90-277-0744-8 105. K. H. Wolff, Surrender and Catch. Experience and Inquiry Today. [Boston Studies in the Philosophy of Science, Vol. LI] 1976 ISBN 90-277-0758-8; Pb 90-277-0765-0 SYNTHESE LIBRARY 106. K. Kosik, Dialectics of the Concrete. A Study on Problems of Man and World. [Boston Studies in the Philosophy of Science, Vol. LII] 1976 ISBN 90-277-0761-8; Pb 90-277-0764-2 107. N. Goodman, The Structure of Appearance. 3rd ed. with an Introduction by G. Hellman. [Boston Studies in the Philosophy of Science, Vol. LIII] 1977 ISBN 90-277-0773-1; Pb 90-277-0774-X 108. K. Ajdukiewicz, The Scientijic World-Perspective and Other Essays, 1931-1963. Translated from Polish. Edited and with an Introduction by J. Giedymin. 1978 ISBN 90-277-0527-5 109. R. L. Causey, Unity ofScience. 1977 ISBN 90-277-0779-0 ISBN 90-277-0781-2 110. R. E. Grandy, Advanced Logicfor Applications. 1977 111. R. P. McArthur, Tense Logic. 1976 ISBN 90-277-0697-2 112. L. Lindahl, Position and Change. A Study in Law and Logic. Translated from Swedish by P. Needham. 1977 ISBN 90-277-0787-1 113. R. Tuomela, Dispositions. 1978 ISBN 90-277-0810-X 114. H. A. Simon, Models of Discovery and Other Topics in the Methods of Science. [Boston Studies in the Philosophy of Science, Vol. LIV] 1977 ISBN 90-277-0812-6; Pb 90-277-0858-4 115. R. D. Rosenkrantz, Inference, Method and Decision. Towards a Bayesian ISBN 90-277-0817-7; Pb 90-277-0818-5 Philosophy of Science. 1977 116. R. Tuomela, Human Action and Its Explanation. A Study on the Philosophical Foundations of Psychology. 1977 ISBN 90-277-0824-X 117. M. Lazerowitz, The Language of Philosophy. Freud and Wittgenstein. [Boston Studies in the Philosophy of Science, Vol. LV] 1977 ISBN 90-277-0826-6; Pb 90-277-0862-2 118. Not published 119. 1. PeIe (ed.), Semiotics in Poland, 1894-1969. Translated from Polish. 1979 ISBN 90-277-0811-8 120. I. Pöm, Action Theory and Social Science. Some Formal Models. 1977 ISBN 90-277-0846-0 121. J. Margolis, Persons and Mind. The Prospects of Nonreductive Materialism. [Boston Studies in the Philosophy of Science, Vol. LVII] 1977 ISBN 90-277-0854-1; Pb 90-277-0863-0 122. J. Hintikka, I. Niiniluoto, and E. Saarinen (eds.), Essays on Mathematical and Philosophical Logic. 1979 ISBN 90-277-0879-7 123. T. A. F. Kuipers, Studies in Inductive Probability and Rational Expectation. 1978 ISBN 90-277-0882-7 124. E. Saarinen, R. Hilpinen, I. Niiniluoto and M. P. Hintikka (eds.), Essays in Honour ofJaakko Hintikka on the Occasion ofHis 50th Birthday. 1979 ISBN 90-277-0916-5 125. G. Radnitzky and G. Andersson (eds.), Progress and Rationality in Science. [Boston Studies in the Philosophy of Science, Vol. LVIII] 1978 ISBN 90-277-0921-1; Pb 90-277-0922-X 126. P. Mittelstaedt, Quantum Logic. 1978 ISBN 90-277-0925-4 127. K. A. Bowen, Model Theory for Modal Logic. Kripke Models for Modal Predicate Ca1culi. 1979 ISBN 90-277-0929-7 128. H. A. Bursen, Dismantling the Memory Machine. A Philosophical Investigation of Machine Theories of Memory. 1978 ISBN 90-277-0933-5 SYNTHESE LIBRARY 129. M. W. Wartofsky, Models. Representation and the Scientific Understanding. [Boston Studies in the Philosophy of Science, Vol. XLVIII] 1979 ISBN 90-277-0736-7; Pb 90-277-0947-5 130. D. Ihde, Technics and Praxis. A Philosophy of Technology. [Boston Studies in the Philosophy of Science, Vol. XXIV] 1979 ISBN 90-277-0953-X; Pb 90-277-0954-8 131. J. J. Wiatr (ed.), Polish Essays in the Methodology olthe Social Sciences. [Boston Studies in the Philosophy of Science, Vol. XXIX] 1979 ISBN 90-277-0723-5; Pb 90-277-0956-4 132. W. C. Salmon (ed.), Hans Reichenbach: Logical Empiricist. 1979 ISBN 90-277-0958-0 133. P. Bieri, R.-P. Horstrnann and L. Krüger (eds.), Transcendental Arguments in Science. Essays in Epistemology. 1979 ISBN 90-277-0963-7; Pb 90-277-0964-5 134. M. Markovic and G. Petrovic (eds.), Praxis. Yugoslav Essays in the Philosophy and Methodology of the Social Sciences. [Boston Studies in the Philosophy of Science, Vol. XXXVI] 1979 ISBN 90-277-0727-8; Pb 90-277-0968-8 135. R. W6jcicki, Topics in the Formal Methodology 01 Empirical Sciences. Translated from Polish. 1979 ISBN 90-277-1004-X 136. G. Radnitzky and G. Andersson (eds.), The Structure and Development 01 Science. [Boston Studies in the Philosophy of Science, Vol. LIX] 1979 ISBN 90-277-0994-7; Pb 90-277-0995-5 137. J. C. Webb, Mechanism, Mentalism and Metamathematics. An Essay on Finitism. 1980 ISBN 90-277-1046-5 138. D. F. Gustafson and B. L. Tapscott (eds.), Body, Mind and Method. Essays in Honor of Virgil C. Aldrich. 1979 ISBN 90-277-1013-9 139. L. Nowak, The Structure olIdealization. Towards a Systematic Interpretation of the ISBN 90-277-1014-7 Marxian Idea of Science. 1980 140. C. Perelman, The New Rhetoric and the Humanities. Essays on Rhetoric and Its Applications. Translated from French and German. With an Introduction by H. Zyskind. 1979 ISBN 90-277-1018-X; Pb 90-277-1019-8 141. W. Rabinowicz, Universalizability. A Study in Morals and Metaphysics. 1979 ISBN 90-277-1020-2 142. C. Perelman, lustice, Law and Argument. Essays on Moral and Legal Reasoning. Translated from French and German. With an Introduction by HJ. Berman. 1980 ISBN 90-277-1089-9; Pb 90-277-1090-2 143. S. Kanger and S. Öhman (eds.), Philosophy and Grammar. Papers on the Occasion of the Quincentennial of Uppsala University. 1981 ISBN 90-277-1091-0 144. T. Pawlowski, Concept Formation in the Humanities and the Social Sciences. 1980 ISBN 90-277-1096-1 145. J. Hintikka, D. Gruender and E. Agazzi (eds.), Theory Change, Ancient Axiomatics and Galileo' s Methodology. Proceedings of the 1978 Pisa Conference on the History and Philosophy of Science, Volume I. 1981 ISBN 90-277-1126-7 146. J. Hintikka, D. Gruender and E. Agazzi (eds.), Probabilistic Thinking, Thermodynamics, and the Interaction 01 the History and Philosophy 01 Science. Proceedings of the 1978 Pisa Conference on the History and Philosophy of Science, Volume 11.1981 ISBN 90-277-1127-5 147. U. Mönnich (ed.), Aspects 01 Philosophical Logic. Some Logical Forays into Central Notions ofLinguistics and Philosophy. 1981 ISBN 90-277-1201-8 148. D. M. Gabbay, Semanticallnvestigations in H eyting' s Intuitionistic Logic. 1981 ISBN 90-277-1202-6 SYNTHESE LIBRARY 149. E. Agazzi (ed.), Modern Logic-A Survey. Historical, Philosophical, and Mathematical Aspects ofModern Logic and Its Applications. 1981 ISBN 90-277-1137-2 150. A. F. Parker-Rhodes, The Theory of Indistinguishables. A Search for Explanatory ISBN 90-277-1214-X Principles below the Level ofPhysics. 1981 151. J. C. Pitt, Pictures, Images, and Conceptual Change. An Analysis ofWilfrid Sellars' ISBN 90-277-1276-X; Pb 90-277-1277-8 Philosophy of Science. 1981 152. R. Hilpinen (ed.), New Studies in Deontic Logic. Norms, Actions, and the FoundaISBN 90-277-1278-6; Pb 90-277-1346-4 tions ofEthics. 1981 153. C. Dilworth, Scientific Progress. A Study Conceming the Nature of the Relation between Successive Scientific Theories. 2nd, rev. and augmented ed., 1986. 3rd rev. ed., 1994 ISBN 0-7923-2487-0; Pallas Pb 0-7923-2488-9 154. D. Woodruff Smith and R. Mclntyre, Husserl and Intentionality. A Study of Mind, Meaning, and Language. 1982 ISBN 90-277-1392-8; Pb 90-277-1730-3 155. R. J. Nelson, The Logic ofMind. 2nd. ed., 1989 ISBN 90-277-2819-4; Pb 90-277-2822-4 156. J. F. A. K. van Benthem, The Logic ofTime. A Model-Theoretic Investigation into the Varieties ofTemporal Ontology, and Temporal Discourse. 1983; 2nd ed., 1991 ISBN 0-7923-1081-0 ISBN 90-277-1437-1 157. R. Swinburne (ed.), Space, Time and Causality. 1983 158. E. T. Jaynes, Papers on Probability, Statistics and Statistical Physics. Ed. by R. D. Rozenkrantz. 1983 ISBN 90-277-1448-7; Pb (1989) 0-7923-0213-3 159. T. Chapman, Time: A PhilosophicalAnalysis. 1982 ISBN 90-277-1465-7 160. E. N. Zalta, Abstract Objects. An Introduction to Axiomatic Metaphysics. 1983 ISBN 90-277-1474-6 161. S. Harding and M. B. Hintikka (eds.), Discovering Reality. Feminist Perspectives on Epistemology, Metaphysics, Methodology, and Philosophy of Science. 1983 ISBN 90-277-1496-7; Pb 90-277-1538-6 162. M. A. Stewart (ed.), Law, Morality and Rights. 1983 ISBN 90-277-1519-X 163. D. Mayr and G. Süssmann (eds.), Space, Time, and Mechanics. Basic Structures of a ISBN 90-277-1525-4 Physical Theory. 1983 164. D. Gabbay and F. Guenthner (eds.), Handbook of Philosophical Logic. Vol. I: Elements of Classical Logic. 1983 ISBN 90-277-1542-4 165. D. Gabbay and F. Guenthner (eds.), Handbook of Philosophical Logic. Vol. 11: Extensions of Classical Logic. 1984 ISBN 90-277-1604-8 166. D. Gabbay and F. Guenthner (eds.), Handbook of Philosophical Logic. Vol. III: Alternative to Classical Logic. 1986 ISBN 90-277-1605-6 167. D. Gabbay and F. Guenthner (eds.), Handbook of Philosophical Logic. Vol. IV: Topics in the Philosophy of Language. 1989 ISBN 90-277-1606-4 168. A. J. I. Jones, Communication and Meaning. An Essay in Applied Modal Logic. 1983 ISBN 90-277-1543-2 169. M. Fitting, Proof Methods for Modal and Intuitionistic Logics. 1983 ISBN 90-277-1573-4 170. J. Margolis, Culture and Cultural Entities. Toward a New Unity of Science. 1984 ISBN 90-277-1574-2 171. R. Tuomela, A Theory of Social Action. 1984 ISBN 90-277-1703-6 172. J. J. E. Gracia, E. Rabossi, E. Villanueva and M. Dascal (eds.), Philosophical Analysis in Latin America. 1984 ISBN 90-277-1749-4 173. P. Ziff, Epistemic Analysis. A Coherence Theory of Knowledge. 1984 SYNTHESE LIBRARY ISBN 90-277-1751-7 174. P. Ziff, Antiaesthetics. An Appreciation of the Cow with the Subtile Nose. 1984 ISBN 90-277-1773-7 175. W. Balzer, D. A. Pearce, and H.-J. Schmidt (eds.), Reduction in Science. Structure, Examples, Philosophical Problems. 1984 ISBN 90-277-1811-3 176. A. Peczenik, L. Lindahl and B. van Roermund (eds.), Theory of Legal Science. Proceedings of the Conference on Legal Theory and Philosophy of Science (Lund, Sweden, December 1983). 1984 ISBN 90-277-1834-2 ISBN 90-277-1835-0 177. I. Niiniluoto, Is Science Progressive? 1984 178. B. K. Matilal and J. L. Shaw (eds.), Analytical Philosophy in Comparative Perspective. Exploratory Essays in Current Theories and Classical Indian Theories of Meaning and Reference. 1985 ISBN 90-277-1870-9 179. P. Kroes, Time: Its Structure and Role in Physical Theories. 1985 ISBN 90-277-1894-6 180. J. H. Fetzer, Sociobiology and Epistemology. 1985 ISBN 90-277-2005-3; Pb 90-277-2006-1 181. L. Haaparanta and J. Hintikka (eds.), Frege Synthesized. Essays on the Philosophical and Foundational Work of Gottlob Frege. 1986 ISBN 90-277-2126-2 182. M. Detlefsen, Hilbert's Program. An Essay on Mathematical Instrumentalism. 1986 ISBN 90-277-2151-3 183. J. L. Golden and J. J. Pilotta (eds.), Practical Reasoning in Human Affairs. Studies ISBN 90-277-2255-2 in Honor of Chaim Perelman. 1986 184. H. Zandvoort, Models of Scientific Development and the Case of Nuclear Magnetic Resonance. 1986 ISBN 90-277-2351-6 185. I. Niiniluoto, Truthlikeness. 1987 ISBN 90-277-2354-0 186. W. Balzer, C. U. Moulines and J. D. Sneed, An Architectonic for Science. The Structuralist Program. 1987 ISBN 90-277-2403-2 187. D. Pearce, Roads to Commensurability. 1987 ISBN 90-277-2414-8 188. L. M. Vaina (ed.), Mauers oflntelligence. Conceptual Structures in Cognitive Neuroscience. 1987 ISBN 90-277-2460-1 189. H. Siegel, Relativism Refuted. A Critique of Contemporary Epistemological Relativism. 1987 ISBN 90-277-2469-5 190. W. Callebaut and R. Pinxten, Evolutionary Epistemology. A Multiparadigm Program, with a Complete Evolutionary Epistemology Bibliograph. 1987 ISBN 90-277-2582-9 191. J. Kmita, Problems in Historical Epistemology. 1988 ISBN 90-277-2199-8 192. J. H. Fetzer (ed.), Probability and Causality. Essays in Honor of Wesley C. Salmon, with an Annotated Bibliography. 1988 ISBN 90-277-2607-8; Pb 1-5560-8052-2 193. A. Donovan, L. Laudan and R. Laudan (eds.), Scrutinizing Science. Empirical ISBN 90-277-2608-6 Studies of Scientific Change. 1988 194. H.R. Otto and J.A. Tuedio (eds.), Perspectives on Mind. 1988 ISBN 90-277-2640-X 195. D. Batens and J.P. van Bendegem (eds.), Theory and Experiment. Recent Insights and New Perspectives on Their Relation. 1988 ISBN 90-277-2645-0 196. J. Österberg, Self and Others. A Study of Ethical Egoism. 1988 ISBN 90-277-2648-5 197. D.H. Helman (ed.), Analogical Reasoning. Perspectives of Artificial Intelligence, Cognitive Science, and Philosophy. 1988 ISBN 90-277-2711-2 SYNTHESE LIBRARY 198. I. Wolenski, Logic and Philosophy in the Lvov-Warsaw Schoo!. 1989 ISBN 90-277-2749-X 199. R. W6jcicki, Theory of Logical Calculi. Basic Theory of Consequence Operations. 1988 ISBN 90-277-2785-6 200. J. Hintikka and M.B. Hintikka, The Logic of Epistemology and the Epistemology of Logic. Selected Essays. 1989 ISBN 0-7923-0040-8; Pb 0-7923-0041-6 ISBN 90-277-2808-9 201. E. Agazzi (ed.), Probability in the Sciences. 1988 202. M. Meyer (ed.), From Metaphysics to Rhetoric. 1989 ISBN 90-277-2814-3 203. R.L. Tieszen, Mathematical Intuition. Phenomenology and Mathematical Knowledge.1989 ISBN 0-7923-0131-5 ISBN 0-7923-0135-8 204. A. Melnick, Space, Time, and Thought in Kant. 1989 205. D.W. Smith, The Circle of Acquaintance. Perception, Consciousness, and Empathy. 1989 ISBN 0-7923-0252-4 206. M.H. Salmon (ed.), The Philosophy of Logical Mechanism. Essays in Honor of Arthur W. Burks. With his Responses, and with a Bibliography of Burk's Work. 1990 ISBN 0-7923-0325-3 207. M. Kusch, Language as Calculus vs. Language as Universal Medium. A Study in Husserl, Heidegger, and Gadamer. 1989 ISBN 0-7923-0333-4 208. T.C. Meyering, Historical Roots of Cognitive Science. The Rise of a Cognitive Theory ofPerception from Antiquity to the Nineteenth Century. 1989 ISBN 0-7923-0349-0 209. P. Kosso, Observability and Observation in Physical Science. 1989 ISBN 0-7923-0389-X 210. J. Kmita, Essays on the Theory ofScientific Cognition. 1990 ISBN 0-7923-0441-1 211. W. Sieg (ed.), Acting and Reflecting. The Interdisciplinary Turn in Philosophy. 1990 ISBN 0-7923-0512-4 212. J. Karpiriski, Causality in Sociological Research. 1990 ISBN 0-7923-0546-9 213. H.A. Lewis (ed.), Peter Geach: Philosophical Encounters. 1991 ISBN 0-7923-0823-9 214. M. Ter Hark, Beyond the Inner and the Outer. Wittgenstein's Philosophy of Psychology. 1990 ISBN 0-7923-0850-6 215. M. Gosselin, Nominalism and Contemporary Nominalism. Ontological and EpistemologicalImplications of the Work of W.V.O. Quine and of N. Goodman. 1990 ISBN 0-7923-0904-9 216. J.H. Fetzer, D. Shatz and G. Schlesinger (eds.), Definitions and Definability. Philosophical Perspectives. 1991 ISBN 0-7923-1046-2 217. E. Agazzi and A. Cordero (eds.), Philosophy and the Origin and Evolution of the Universe. 1991 ISBN 0-7923-1322-4 218. M. Kusch, Foucault' s Strata and Fields. An Investigation into Archaeological and Genealogical Science Studies. 1991 ISBN 0-7923-1462-X 219. C.I. Posy, Kant's Philosophy ofMathematics. Modem Essays. 1992 ISBN 0-7923-1495-6 220. G. Van de Vijver, New Perspectives on Cybernetics. Self-Organization, Autonomy and Connectionism. 1992 ISBN 0-7923-1519-7 221. I.C. Nyiri, Tradition and Individuality. Essays. 1992 ISBN 0-7923-1566-9 222. R. Howell, Kant' s Transcendental Deduction. An Analysis of Main Themes in His Critical Philosophy. 1992 ISBN 0-7923-1571-5 SYNTHESE LIBRARY 223. A. Garda de la Sienra, The Logical Foundations of the Marxian Theory of Value. 1992 ISBN 0-7923-1778-5 224. D.S. Shwayder, Statement and Referent. An Inquiry into the Foundations of Our ConceptualOrder.1992 ISBN 0-7923-1803-X 225. M. Rosen, Problems of the Hegelian Dialectic. Dialectic Reconstructed as a Logic of Human Reality. 1993 ISBN 0-7923-2047-6 226. P. Suppes, Models and Methods in the Philosophy of Science: Selected Essays. 1993 ISBN 0-7923-2211-8 227. R. M. Dancy (ed.), Kant and Critique: New Essays in Honor ofW. H. Werkmeister. 1993 ISBN 0-7923-2244-4 228. J. Wolenski (ed.), Philosophical Logic in Poland. 1993 ISBN 0-7923-2293-2 229. M. De Rijke (ed.), Diamonds and Defaults. Studies in Pure and Applied Intensional Logic. 1993 ISBN 0-7923-2342-4 230. B.K. Matilal and A. Chakrabarti (eds.), Knowing from Words. Western and Indian Philosophical Analysis of Understanding and Testimony. 1994 ISBN 0-7923-2345-9 231. S.A. Kleiner, The Logic of Discovery. A Theory of the Rationality of Scientific Research. 1993 ISBN 0-7923-2371-8 232. R. Festa, Optimum Inductive Methods. A Study in Inductive Probability , Bayesian Statistics, and Verisimilitude. 1993 ISBN 0-7923-2460-9 233. P. Humphreys (ed.), Patrick Suppes: Scientific Philosopher. Vol. 1: Probability and Probabilistic Causality. 1994 ISBN 0-7923-2552-4; Set Vols. 1-3 ISBN 0-7923-2554-0 234. P. Humphreys (ed.), Pa trick Suppes: Scientific Philosopher. Vol. 2: Philosophy of Physics, Theory Structure, and Measurement Theory. 1994 ISBN 0-7923-2553-2; Set Vols. 1-3 ISBN 0-7923-2554-0 235. P. Humphreys (ed.), Patrick Suppes: Scientific Philosopher. Vol. 3: Language, Logic, and Psychology. 1994 ISBN 0-7923-2862-0; Set Vo1s. 1-3 ISBN 0-7923-2554-0 236. D. Prawitz and D. Westerstähl (eds.), Logic and Philosophy of Science in Uppsala. Papers from the 9th International Congress of Logic, Methodology, and Philosophy of Science. 1994 ISBN 0-7923-2702-0 237. L. Haaparanta (ed.), Mind, Meaning and Mathematics. Essays on the Philosophical Views ofHusserl and Frege. 1994 ISBN 0-7923-2703-9 238. J. Hintikka (ed.): Aspects of Metaphor. 1994 ISBN 0-7923-2786-1 239. B. McGuinness and G. Oliveri (eds.), The Philosophy of Michael Dummett. With Replies from Michael Dummett. 1994 ISBN 0-7923-2804-3 240. D. Jamieson (ed.), Language, Mind, and Art. Essays in Appreciation and Analysis, In Honor ofPaul Ziff. 1994 ISBN 0-7923-2810-8 241. G. Preyer, F. Siebelt and A. Ulfig (eds.), Language, Mind and Epistemology. On Donald Davidson's Philosophy. 1994 ISBN 0-7923-2811-6 242. P. Ehrlich (ed.), Real Numbers, Generalizations of the Reals, and Theories of Continua.1994 ISBN 0-7923-2689-X 243. G. Debrock and M. Hulswit (eds.): Living Doubt. Essays concerning the epistemology of Charles Sanders Peirce. 1994 ISBN 0-7923-2898-1 244. J. Srzednicki: To Know or Not to Know. Beyond Realism and Anti-Realism. 1994 ISBN 0-7923-2909-0 245. R. Egidi (ed.): Wittgenstein: Mind and Language. 1995 ISBN 0-7923-3171-0 246. A. Hyslop: Other Minds. 1995 ISBN 0-7923-3245-8 SYNTHESE LIBRARY 247. L. P6los and M. Masuch (eds.): Applied Logic: How, What and Why. Logical ISBN 0-7923-3432-9 Approaches to Natural Language. 1995 248. M. Krynicki, M. Mostowski and L.M. Szczerba (eds.): Quantifiers: Logics, Models and Computation. Volume I: Surveys. 1995 ISBN 0-7923-3448-5 249. M. Krynicki, M. Mostowski and L.M. Szczerba (eds.): Quantifiers: Logics, Models and Computation. Volume 11: Contributions. 1995 ISBN 0-7923-3449-3 Set ISBN (volume 248 + 249) 0-7923-3450-7 KLUWER ACADEMIC PUBLISHERS - DORDRECHT / BOSTON / LONDON