The Struggle against Dogmatism: Wittgenstein and the Concept of Philosophy 9780674033856

Citation preview

THE STRUGGLE AGAINST DOGMATISM

THE STRUGGLE AGAINST DOGMATISM

Wittgenstein and the Concept of Philosophy

OSKARI KUUSELA

HARVARD UNIVERSITY PRESS

Cambridge, Massachusetts London, England 2008

Copyright © 2008 by the President and Fellows of Harvard College All rights reserved Printed in the United States of America Library of Congress Cataloging-in-Publication Data Kuusela, Oskari. The struggle against dogmatism : Wittgenstein and the concept of philosophy / Oskari Kuusela. p. cm. Includes bibliographical references and index. ISBN-13: 978-0-674-02771-8 ISBN-10: 0-674-02771-X 1. Wittgenstein, Ludwig, 1889–1951. 2. Philosophy. 3. Dogmatism. I. Title. B3376.W564K89 2008 192—dc22 2007031578

For Venla, Urho, and Alma

Contents

Acknowledgments

xi

Abbreviations

xv

Introduction

1

1 Wittgenstein on Philosophical Problems: From One Fundamental Problem to Particular Problems 1.1 The Tractatus on philosophical problems

17 18

1.2 Wittgenstein’s later conception of philosophical problems 27 1.3 Examples of philosophical problems as based on misunderstandings 30 1.4 Tendencies and inclinations of thinking: philosophy as therapy 43 1.5 Wittgenstein’s notion of peace in philosophy: the contrast with the Tractatus 46

2 Two Conceptions of Clarification

54

2.1 The Tractatus’s conception of philosophy as logical analysis 55 2.2 Wittgenstein’s later critique of the Tractatus’s notion of logical analysis 65 2.3 Clarification in Wittgenstein’s later philosophy

74

viii

Contents

3 From Metaphysics and Philosophical Theses to Grammar: Wittgenstein’s Turn

96

3.1 Philosophical theses, metaphysical philosophy, and the Tractatus 97 3.2 Metaphysics and conceptual investigation: the problem with metaphysics 102 3.3 Conceptual investigation and the problem of dogmatism 3.4 Wittgenstein’s turn

111

120

3.5 The turn and the role of rules 3.6 Rules as objects of comparison

132 140

3.7 Rules, metaphysical projection, and the logic of language 145

4 Grammar, Meaning, and Language

149

4.1 Grammar, use, and meaning: the problem of the status of Wittgenstein’s remarks 150 4.2 Wittgenstein’s formulation of his conception of meaning 158 4.3 The concept of language: comparisons with instruments and games 163 4.4 Wittgenstein’s development and the advantages of his mature view 168 4.5 Examples as centers of variation and the conception of language as a family 171 4.6 Avoiding dogmatism about meaning

176

4.7 Wittgenstein’s methodological shift and analyses in terms of necessary conditions 180

5 The Concepts of Essence and Necessity

184

5.1 Constructivist readings and the arbitrariness/ nonarbitrariness of grammar 185 5.2 Problems with constructivism

188

5.3 The methodological dimension of Wittgenstein’s conception of essence 192 5.4 The nontemporality of grammatical statements

195

5.5 Explanations of necessity in terms of factual regularities 5.6 Wittgenstein’s account of essence and necessity 5.7 Beyond theses about the source of necessity

204 208

198

Contents

ix

6 Philosophical Hierarchies and the Status of Clarificatory Statements

215

6.1 Philosophical hierarchies and Wittgenstein’s “leading principle” 216 6.2 The concept of perspicuous presentation

228

6.3 The (alleged) necessity of accepting philosophical statements 238 6.4 The concept of agreement and the problem of injustice 6.5 The criteria of the correctness of grammatical remarks

247 252

6.6 Multidimensional descriptions and the new use of old dogmatic claims 258

7 Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics 7.1 Metaphysics disguised as methodology 7.2 The historicity of philosophy

271

7.3 Philosophy and the everyday

275

Notes

287

Index

347

266

265

Acknowledgments

IN THE COURSE of my research on Wittgenstein I have become indebted to many people for their generous help, without which this book would not exist. First, I would like to thank Gordon Baker, Peter Hacker, and Bill Child, who each in turn acted as my supervisor at the University of Oxford. In particular, I am grateful to Baker for his encouragement, to Hacker for his detailed criticisms, and to Child for a helpful “external perspective” on interpretational disputes. One of my greatest regrets is that now that this work is completed, Baker is no longer there to discuss it. I have had the additional pleasure and privilege of discussing Wittgenstein’s philosophy and related matters with a number of people who have also read parts of this work in its various stages. I am grateful to Maija Aalto-Heinilä for correspondence on the Tractatus and to Gisela Bengtsson for discussions and comments on versions of some of the chapters. I am indebted to James Conant for comments on draft chapters, for a chance to discuss Wittgenstein’s philosophy in general, and for encouragement and support. Similarly, I am thankful to Cora Diamond for discussions and correspondence on Wittgenstein. I would like to thank Logi Gunnarsson for comments on an early version of the third chapter, Martin Gustafsson for correspondence and discussions, and Juliet Floyd for correspondence. To Wolfgang Freitag I am grateful for our innumerable discussions on Wittgenstein at Oxford and to Lars Hertzberg for thought-provoking questions and comments on the third chapter. I am indebted to Heikki Kannisto for discussions in the earlier stages of my work on Wittgenstein and to

xii

Acknowledgments

André Maury for his help and the suggestion to go to Oxford. To Vasso Kindi, I owe the opportunity to discuss relevant matters in seminars at the University of Athens. I am grateful to Denis McManus and Katherine Morris for their comments on the fifth chapter. To Marie McGinn and Stephen Mulhall I am thankful, first of all, for an extremely enjoyable viva voce. In addition, both have read and commented on subsequent versions of some of the chapters. Marie McGinn read the whole penultimate draft of the book, for which I am extremely grateful. I would like to thank Alois Pichler for reading parts of the book, for his comments, and for help with matters relating to Wittgenstein’s Nachlass. Rupert Read is to be thanked for discussions, as well as for advice on matters of terminology. I am thankful to Gabriel Sandu for comments on an early draft of (parts of ) the second chapter. Richard Sørli has read chapters and sections, for which I am thankful. Thomas Wallgren has been an inspiring senior colleague, and I am grateful for discussions with him and for his comments on my work. Thanks are due to Risto Vilkko for his help and to Cato Wittusen for discussions of draft chapters. I would also like to express my gratitude to the late Georg Henrik von Wright for his encouragement and for introducing me to Wittgenstein’s Nachlass. Parts of this work have been presented as talks to various audiences in Athens, Bergen, Chicago, Charlottesville, Helsinki, Oslo, Oxford, Åbo, Skjølden, Tampere, and Uppsala. I am grateful for all questions and comments as well as to the organizers of these events. Finally, I would also like to thank the anonymous reviewers for Harvard University Press for their extremely helpful comments and suggestions, Lindsay Waters and Phoebe Kosman at Harvard University Press for their help throughout the publication process, as well as the copy editorial team at Westchester Book Services. My very last debt is to Alun Davies, Tamara Dobler, Simon Summers, and Ben Walker, who came to my aid in the proofreading process. At a rather different level, I would like to thank my partner, Angela Dimitrakaki, who has helped me in innumerable (also academically relevant) ways throughout the process of writing this book. Besides debts of an academic nature, I am extremely grateful to my parents, Tuula and Timo Kuusela, for all their help and support. I am indebted to Anu Kuusela for a million things and more. The final revisions of this book would not have been possible without the help of Despina Tsiskaki. Similarly, I wish to express my gratitude to Christina Dimi-

Acknowledgments

xiii

trakaki. My love goes to my children, Venla, Urho, and Alma, to whom this book is dedicated, as well as to Aleksi, Julius, and Salla Kuusela. I also extend my thanks to all my friends in Athens, Chicago, Helsinki, Oxford, Norwich (where this book was written), and elsewhere, with special mention of Miltos Tsiantis. My work has been supported financially, at various stages, by the Academy of Finland (including the project with decision number 212577), the Kone Foundation, the Alfred Kordelin Foundation, Oskar Öflund’s Foundation, the Jenny and Antti Wihuri Foundation, the AHRB, and ASLA-Fulbright. I wish to thank all these institutions for making it possible for me to engage in this study. Earlier versions of Chapters 3 and 4 have been previously published in Philosophical Investigations under the titles “From Metaphysics and Philosophical Theses to Grammar: Wittgenstein’s Turn” (vol. 28, no. 2 [2005]: 95–133) and “Do the Concepts of Grammar and Use in Wittgenstein Articulate a Theory of Language or Meaning?” (vol. 29, no. 4 [2006]: 309–341), respectively. I am grateful for the permission to use this material here.

Abbreviations

WORKS BY WITTGENSTEIN

The following abbreviations are used to refer to Wittgenstein’s published works. BB

Preliminary Studies for the “Philosophical Investigations,” Generally Known as the Blue and Brown Books, ed. R. Rhees (Oxford: Blackwell, 1958).

CV

Culture and Value, rev. ed., ed. G. H. von Wright in collaboration with H. Nyman, rev. ed. A. Pichler, trans. P. Winch (Oxford: Blackwell, 1998).

LWi

Last Writings on the Philosophy of Psychology: Preliminary Studies for Part II of Philosophical Investigations, ed. G. H. von Wright and H. Nyman, trans. C. G. Luckhardt and M. A. E. Aue (Oxford: Blackwell, 1982).

NB

Notebooks, 1914–1916, ed. G. E. M. Anscombe and G. H. von Wright, trans. G. E. M. Anscombe (Oxford: Blackwell, 1961).

OC

On Certainty, ed. G. E. M. Anscombe and G. H. von Wright, trans. G. E. M. Anscombe and D. Paul (Oxford: Blackwell, 1993).

PG

Philosophical Grammar, ed. R. Rhees, trans. A. Kenny (Oxford: Blackwell, 1974).

PI

Philosophical Investigations, 2nd ed., ed. G. E. M Anscombe and R. Rhees, trans. G. E. M. Anscombe (Oxford: Blackwell, 1997).

PO

Philosophical Occasions, 1912–1951, ed. J. Klagge and A. Nordmann (Indianapolis: Hackett, 1993).

xvi

Abbreviations

PPO

Public and Private Occasions, ed. James C. Klagge and Alfred Nordmann (New York: Rowman & Littlefield, 2003).

PR

Philosophical Remarks, ed. R. Rhees, trans. R. Hargreaves and R. White (Oxford: Blackwell, 1975).

RFM

Remarks on the Foundations of Mathematics, rev. ed., ed. G. E. M. Anscombe, R. Rhees, and G. H. von Wright, trans. G. E. M. Anscombe (Oxford: Blackwell, 1978).

RLF

“Some Remarks on Logical Form,” in J. Klagge and A. Nordmann, eds., Philosophical Occasions, 1912–1951 (Indianapolis: Hackett, 1993).

ROC

Remarks on Colour, ed. G. E. M. Anscombe, trans. L. McAlister and M. Schättle (Oxford: Blackwell, 1978).

RPPi

Remarks on the Philosophy of Psychology, vol. 1., ed. G. E. M. Anscombe and G. H. von Wright, trans. G. E. M. Anscombe (Oxford: Blackwell, 1980).

TLP

Tractatus Logico-Philosophicus, trans. C. K. Ogden (London: Routledge & Kegan Paul, 1951); trans. D. F. Pears and B. F. McGuinness (London: Routledge & Kegan Paul, 1961).

Z

Zettel, ed. G. E. M. Anscombe and G. H. von Wright, trans. G. E. M. Anscombe (Oxford: Blackwell, 1967). WITTGENSTEIN’S NACHLASS

Wittgenstein’s Nachlass: The Bergen Electronic Edition, ed. The Wittgenstein Archives at the University of Bergen (Oxford: Oxford University Press, 2000). References to the Nachlass are by manuscript or typescript number followed by page number (as cited in the von Wright catalogue; see PO). LECTURE NOTES BY OTHERS, CONVERSATIONS, AND CORRESPONDENCE

AWL

Alice Ambrose, ed., Wittgenstein’s Lectures, Cambridge, 1932–35 (Oxford: Blackwell, 1979).

CL

Ludwig Wittgenstein, Cambridge Letters: Correspondence with Russell, Keynes, Moore, Ramsey and Straffa, ed. B. McGuinness and G. H. von Wright (Oxford: Blackwell, 1997).

DC

M. O. C. Drury, “Conversations with Wittgenstein,” in The Danger of Words and Writings on Wittgenstein, ed. D. Berman, M. Fitzgerald, and J. Hayes (Bristol: Thoemmes Press, 1996).

Abbreviations

xvii

LFM

Cora Diamond, ed., Wittgenstein’s Lectures on the Foundations of Mathematics, Cambridge, 1939: from the notes of R. G. Bosanquet, Norman Malcolm, Rush Rhees, and Yorick Smythies (London: Harvester, 1976; Chicago: University of Chicago Press, 1989).

LO

G. H. von Wright, ed., Letters to C. K. Ogden with Comments on the English Translation of the Tractatus Logico-Philosophicus (Oxford: Blackwell, 1973).

LR

G. H. von Wright, ed., Letters to Russell, Keynes and Moore (Oxford: Blackwell, 1974).

VW

Ludwig Wittgenstein and Friedrich Waismann, The Voices of Wittgenstein: The Vienna Circle, ed. G. Baker (London: Routledge, 2003).

WVC

Brian McGuinness, ed., Wittgenstein and the Vienna Circle: Conversations Recorded by Friedrich Waismann, trans. Joachim Schulte and Brian McGuinness (Oxford: Blackwell, 1979).

In order to see more clearly, here as in countless similar cases, we must focus on the details of what goes on; must look at them from close to. —Wittgenstein, Philosophical Investigations

Introduction

PHILOSOPHY IS HABITUALLY IDENTIFIED with philosophical doctrines, theses, or theories put forward by philosophers. This is also how an introduction to the subject often looks: one is provided with a list of names of philosophers and doctrines associated with them. When one begins to study philosophy in a university context, emphasis is soon placed on the ability to philosophize as opposed to the knowledge of philosophy. One learns to construct and analyze arguments, and it is not sufficient simply to know which views a philosopher held. Rather, one is expected to be able to understand why anyone would wish to hold such and such a view. That is, one learns to perceive philosophical doctrines and theses as proposed (perhaps competing) solutions to philosophical problems and becomes able to discern a point behind the things that philosophers say. But even if the activity of philosophizing thus takes the foreground, this activity is presumed to have a particular form. Philosophy is conceived as the search for answers to questions in the manner of the sciences; it is a research program with an unusually long history. A philosopher is expected to solve problems, many of which were already articulated some 2,500 years ago. Against such a background it may then appear alternately incomprehensible and outrageous when someone comes along, as Ludwig Wittgenstein did in the first half of the twentieth century, suggesting that there are no doctrines, theses, or theories in philosophy, and that the questions that philosophers are trying to answer are based on misunderstandings. For if a philosopher’s task is to present theses or theo-

2

Introduction

ries, does denying the existence of theses and theories not amount to abandoning philosophy? If having something to say in philosophy means having a thesis to present and defend, does the rejection of theses mean that one has nothing to say, that one holds no views about philosophical issues and has nothing positive to contribute? Or even worse: if taken seriously, does the view that there are no theses (and so on) in philosophy imply that we should abandon the fruits of the 2,500 and more years of labor by philosophers? But what could be more arrogant and foolish than to declare all these achievements worthless, as based on misunderstandings? Or alternatively, perhaps the claim that there are no theses is just a smoke screen for smuggling in another set of philosophical doctrines. Maybe this assertion really amounts to a declaration of the alleged superiority of its asserter’s philosophical perspective, a contention that his views are somehow beyond dispute, not to be debated but simply to be accepted? This would make the claim that there are no doctrines, theses, or theories not only arrogant but dishonest. Such suspicions may be natural given the background from which they arise. Nevertheless, the conception that the only—or indeed the best—way to positively contribute to philosophy is to present doctrines, theses, or theories cannot be taken for granted. The fact that philosophers, unlike scientists, have not been able to reach any agreement about their doctrines, theses, or theories is already sufficient grounds to question this assumption. The purpose of this book is thus to examine, in the context of Wittgenstein’s work, the possibility, meaning, and implications of an alternative approach to philosophy, one that does not involve philosophical doctrines, theses, or theories. The overall aim of the study might be described as that of rendering comprehensible Wittgenstein’s approach to philosophy by shedding light on what motivates him to abandon philosophical doctrines, theses, and theories. For only when one sees the kinds of problems pertaining to the practice of philosophy to which Wittgenstein is trying to respond can one come to understand the nature of his response, i.e., the point of his conception of philosophy and the methods he articulates. And if, indeed, he is responding to certain identifiable problems, perhaps his work bears some kind of understandable relation to the philosophical tradition after all, even if his response does not constitute a philosophical thesis or theory.

Introduction

3

More specifically, the roots of Wittgenstein’s conception of philosophy might be said to lie in his emphasis on the difference between true or false factual statements and expressions of exceptionless necessity. The failure to distinguish between these two types of statement constitutes, according to Wittgenstein, a fundamental confusion in philosophy, which gives rise to metaphysics as a study of necessary truths pertaining to reality.1 Wittgenstein’s early work, culminating in the Tractatus Logico-Philosophicus, may be conceived as an attempt to develop a philosophical method that respects this distinction between factual statements and expressions of exceptionless necessity. What would be left of philosophy, if the Tractatus managed to accomplish what it seeks to accomplish, would be a method of clarification, but no doctrines, theses, or theories. As Wittgenstein comes to realize later on, however, the method of the Tractatus involves a commitment to philosophical or metaphysical theses about the nature of language and philosophy. Thus his early work fails to achieve the aim of a philosophy devoid of doctrines, theses, or theories. This leads him once more to reconsider the status and function of philosophical statements, and what philosophizing without doctrines, theses, or theories would amount to. Here, a central goal is to avoid the dogmatism or injustice to which Wittgenstein thinks philosophical doctrines, theses, or theories lead, and to which he fell into in his early philosophy. Or to put it more positively, the goal is to obtain flexibility in philosophical thought without loss of rigor. Notably, this new approach also makes possible a novel use of philosophical theses. This means that although Wittgenstein’s shift away from philosophical theses does constitute a shift away from metaphysical philosophy, his philosophy is not antimetaphysical, in the sense of being hostile to the metaphysical tradition of philosophy. It does not imply the outright rejection of what has been said in the course of the history of philosophy. But the differences between metaphysical philosophy and Wittgenstein’s later philosophy remain significant. The latter constitutes a philosophy without theoretical presuppositions and without hierarchies of more and less fundamental concepts—or so I will argue. Another way of characterizing the aim of this book is to say that it purports to clarify the relation between Wittgenstein’s and more traditional approaches to philosophy by presenting the former as a modification, or a series of modifications, of the latter. (Again, it is important

4

Introduction

that these modifications are motivated by certain specific problems relating to the practice of philosophy.) For instance, given that the search for definitions is one of the defining characteristics of the philosophical tradition in which we live, it should be possible to make Wittgenstein’s approach comprehensible by focusing on his novel conception of the function of definitions, and on how it differs from more traditional conceptions of this function. Partly for such strategic reasons, I will emphasize the employment of rules among Wittgenstein’s methods of clarification. Another point where Wittgenstein’s methods make direct contact with “old ways of philosophizing” is the question of the role of examples in philosophy. At this point too, Wittgenstein is making adjustments to the deployment of a familiar philosophical tool. These points of contact create a space for clarifications that could help one see more clearly what it would mean (and not mean) to switch from the more traditional modes of philosophizing to Wittgenstein’s mode of philosophizing. Consequently, one could decide for oneself the potential of Wittgenstein’s new approach to philosophy. So far philosophers have not been very keen to take up the challenge that Wittgenstein’s conception of philosophy poses to philosophy as it has been and continues to be practiced. There may be various reasons for this, but the result is that the potential his methods might have remains to a great extent unexplored. Yet it seems to me that there is much to learn from what Wittgenstein says about philosophy and its methods, and that this could help one improve one’s own philosophizing. As a result, he might be able to help one deal with one’s own philosophical problems, regardless of whether these are problems that Wittgenstein discussed. Nevertheless, my purpose in this book is not to try to convince anyone that Wittgenstein is actually able to help improve one’s philosophical skills. Rather, I wish to bring the reader to a position where she is able to take the possibility of such an improvement seriously.2 To make a brief remark about what kind of wider relevance the discussion of the methodology of philosophy in this book might possess outside philosophy, the novel conception of the status of philosophical statements Wittgenstein develops in his later philosophy could also prove helpful in thinking about certain methodological issues in the humanities, or perhaps even more widely in the sciences. In particular, this conception of philosophical statements might be applied in describing

Introduction

5

the status of methodological statements that constitute frameworks in the context of which research is carried out. The suggestion, then, is that questions relating to the choice and justification of different methodologies and the relations between them could be addressed with enhanced clarity in the light of this conception of methodological statements. This would clarify what type of claims or assumptions are made about the objects of investigation when, for example, a Marxist or psychoanalytic methodology is adopted in art history, or what it means to adopt a particular methodology. (Does that mean making a true/false claim about the nature of the object of investigation? How does that relate to other sets of methodological claims? What does it mean for the adoption of a methodology to involve a political choice? And so on.) Having said this, I must note that I will rarely be painting with such a broad brush as I have done up to now. (Nor will I explicitly address the issue of the status of methodological statements in the humanities or the sciences.) The focus will be on certain specific issues and questions. These include matters relating to the explication of various relevant Wittgensteinian notions, the justification of my interpretation, and the discussion and critique of interpretations put forward by other readers of Wittgenstein. Philosophy requires attention to detail, but this should not lead to the disappearance of the bigger issues from sight. Next, let me briefly provide a more specific context for the interpretation developed here.

Approaches to Reading Wittgenstein and the Interpretational Framework If any one thing is characteristic of Wittgenstein’s philosophy, it is his insistence throughout his philosophical career on not putting forward doctrines, theses, or theories. But for all the centrality of this insistence, what Wittgenstein means by not having theses in philosophy, or by philosophizing without theses, remains unclear. Some commentators and scholars contest Wittgenstein’s selfunderstanding. An example is George Pitcher in The Philosophy of Wittgenstein: “when Wittgenstein says . . . that there are no philosophical theses or theories, he is overstating his case. . . . He does not mean certain statements about language and meaning; for about these, he himself most certainly puts forward theses . . . .”3 Similarly, Hintikka and Hintikka

6

Introduction

maintain that the “crucial role” or “primary function” of the notion of a language-game in Wittgenstein’s philosophy “is theoretical not therapeutical,”4 its purpose being the articulation of a theory of meaning. Anthony Kenny also exemplifies this approach, attributing to Wittgenstein a “theory of meaning as use.”5 Here Kenny’s reactions seem particularly illustrative. For it is apparently not mere conviction or lack of effort to understand Wittgenstein that leads Kenny to attribute to him a theory. After more than thirty years of work on Wittgenstein, he confesses: “Though I have tried my best to do so I do not believe that it is, in the end, possible to reconcile Wittgenstein’s account of philosophy with the entirety of his philosophical activity in the Investigations.”6 A problem with interpretations of this kind, as Kenny recognizes, is that in maintaining that Wittgenstein puts forward theses, or in attributing theories to him, commentators and scholars describe Wittgenstein’s philosophy as inconsistent. From their point of view, there is a conflict between Wittgenstein’s denial that he puts forward theses and his actual practice. Accordingly, Kenny maintains, whoever wants to follow Wittgenstein is forced to choose between Wittgenstein’s theories and his practice.7 Now it might be, of course, that Wittgenstein’s philosophy is inconsistent. But this cannot be accepted without further argument. That it is tempting to read Wittgenstein’s statements as theses, and that it may be hard to see how else to construe them, does not mean that they should be read as theses—especially if this makes his philosophy contradictory. Some other commentators seem, at least at first glance, more cautious. For instance, Norman Malcolm consistently avoids saying that Wittgenstein has a thesis or a theory but instead talks about Wittgenstein’s views and conceptions.8 In a somewhat similar vein, Saul Kripke substitutes “picture” for “theory” in his discussion of Wittgenstein’s view of language and meaning, suggesting that this way, his points would conform better to Wittgenstein.9 Nevertheless, Malcolm and Kripke do not seek to clarify what would distinguish views, conceptions, or pictures from theses and theories. Thus even if these expressions, appropriately explained, might capture Wittgenstein’s thought, it still needs to be spelled out in what sense this would be so. The view that it can be left undecided in what sense Wittgenstein’s statements are not theses or theories suggests that the issue is not crucial or urgent. In effect, commentators who adopt this view are saying:

Introduction

7

“A philosophical statement is a philosophical statement; whether or not it is a thesis, and what it would mean for it not to be a thesis can be decided later.” Hereby the implication is that Wittgenstein’s statements can be grasped without understanding their status. The problem with this view seems obvious. It is comparable to saying that whether the statement “You can count on my support” expresses a genuine commitment or is a joke is irrelevant to understanding it. Hence, one might also assert, contrary to the above types of interpretation, that Wittgenstein’s conception of philosophy is the key to his writings on particular philosophical topics—for instance, language and meaning. Understanding what purpose his statements serve, and that this purpose is not the articulation of theories or theses, is, one could say, a requirement for grasping what he says about particular issues. Only thus is it possible to avoid certain problems to which his statements, when read implicitly or explicitly as theses, will lead. (In this book, Kripke’s interpretation is used here and there to illustrate such problems.) Adopting a point of view of this kind and emphasizing Wittgenstein’s rejection of philosophical theses and doctrines, Cora Diamond writes: “I think that there is almost nothing in Wittgenstein which is of value and which can be grasped if it is pulled away from that view of philosophy.”10 Similarly, Marie McGinn proposes that the question of how Wittgenstein could be understood as not putting forward a substantive philosophical theory, while still achieving something positive that recognizes the depth of philosophical problems, is “the central interpretative issue for the whole of Wittgenstein’s philosophy.”11 I agree with Diamond and McGinn on this point. But the agreement only makes it more urgent to answer the question, How should one understand Wittgenstein’s shift away from theses? Notably, within this third group of scholars who explicitly address the issue of Wittgenstein’s rejection of theses and try to explain what he means by this, there are also some who arguably relapse into philosophical theses in their explanations. Thus they end up covertly attributing such theses to Wittgenstein. This, as I will argue, is the case with Baker and Hacker and with Hans-Johann Glock, for instance. Given that Baker and Hacker’s very carefully thought-out interpretation of Wittgenstein constitutes the most detailed attempt to explain his philosophy, I will mostly take issue with them in this book—not because I particularly enjoy polemics, but because this contrast seems quite illuminating.

8

Introduction

A scholar whose interpretation of the later Wittgenstein comes closest to mine is the later Gordon Baker, who from the late 1980s onward was developing an interpretation that departs radically from the one presented in his work with Peter Hacker. Differences between Baker’s interpretation and mine are probably a result of our focusing on different issues or reading Wittgenstein with slightly different concerns in mind. The significance of this should not be underestimated, however. As a consequence, we offer different elucidations of Wittgenstein’s conception of philosophy. (Sometimes there may also be genuine disagreements about details.) Another reader of Wittgenstein whose interpretation (or aspects thereof) has affinities with the present work is Stanley Cavell. For example, when read in the interpretative framework constituted by my discussion of Wittgenstein’s method, Wittgenstein seems to articulate a conception of language that largely agrees with the interpretation developed by Cavell from the 1960s on. Characteristically, in this conception the possibility of language is not seen as dependent on rules, contrary to many other interpretations of Wittgenstein, including Baker and Hacker’s or Kripke’s. But this is not merely a matter of someone being right and someone else wrong. Taking into account the interpretation of Wittgenstein’s method presented in this book, it is not that these others are wrong to emphasize the rule-governed nature of language. Rather, their view is one-sided. Moreover, as I will explain, being one-sided or simplifying is not necessarily bad. Simplifications are part of Wittgenstein’s method, but it is important to recognize simplifications for what they are. In recent years Wittgenstein’s conception of philosophy has been discussed especially in connection with his early philosophy, as presented in the Tractatus.12 Although some commentators have suggested that Wittgenstein’s conception of philosophy remains constant throughout his career, it is a moot point to what extent this is so.13 Accordingly, it is unclear what Wittgenstein’s statements about the concept of philosophy in the Tractatus can tell us about his later views. The focus of this book is on Wittgenstein’s later work, from the early 1930s onward, although I will also discuss his early philosophy as it finds expression in the Tractatus.14 Roughly, I understand Wittgenstein’s later philosophy as a novel attempt to achieve the aims of the

Introduction

9

Tractatus. Already in his early work he strives to articulate a conception of philosophy as devoid of doctrines or theories.15 But the early philosophy fails to reach this goal. Thus the early and later philosophies are united by their aim while divided by how they try to reach it. Naturally, however, one can talk about such shared aims only at a very abstract level. I do not wish to suggest that Wittgenstein’s view of what exactly is the aim of his philosophy would remain constant amid all the changes in his methodology. Most of all, the Tractatus’s failure is of great interest, because it can alert one to the kinds of difficulties that lie in the way of attempts to break free from philosophical theses. In particular, there seems to be a risk of repeating the Tractatus’s mistakes in a new form in interpreting Wittgenstein’s later philosophy (or in trying to apply Wittgenstein’s ideas in the context of other philosophical debates). Consequently, Wittgenstein’s later philosophy becomes a new set of theses about language, meaning, philosophy, and so on. Not unrelated to this is yet another risk. Knowing that according to Wittgenstein there are “grave mistakes” in the Tractatus,16 one might be inclined to make its mistakes too obvious. This approach is represented in a radical form by Newton Garver, according to whom the Tractatus’s failure is simply that it is nonsense, as Wittgenstein himself acknowledges at the end of the book.17 This interpretation appears to make it unnecessary to pose any further questions about what Wittgenstein might have wished to achieve by writing a nonsensical book, and whether this had anything to do with his attempt to break away from philosophical doctrines. But by making the Tractatus’s mistakes too obvious, one is also in danger of trivializing Wittgenstein’s later philosophy, insofar as one seeks to understand it by way of contrast with the Tractatus, as Wittgenstein recommends. From the point of view of Garver’s reading, practically anything would seem to count as an improvement on the Tractatus. Thus, there is nothing, or very little, to learn from the Tractatus’s failure. A way to try to avert the risk of making the Tractatus’s mistake too obvious is to distinguish between its aims (Wittgenstein’s intentions) and actual achievements. One might say that an attempt to understand the Tractatus’s conception of philosophy is essentially an attempt to understand how Wittgenstein thought he could achieve its aim, how the Tractatus could be read as an articulation of a philosophy without doctrines.18 In this respect, Wittgenstein’s later critique of his early work

10

Introduction

might then not be the best guide to the reconstruction of the Tractatus’s conception of philosophy. The later philosophy probably gives an accurate account of the Tractatus’s shortcomings, and therefore of its actual achievements. But in his later philosophy Wittgenstein seems not particularly concerned with the Tractatus’s attempt to break away from philosophical doctrines, to the extent that he thinks it was already on the right track. Rather, whatever was correct in the Tractatus seems incorporated in Wittgenstein’s later philosophy. Thus, the articulation of what the Tractatus got right is now a part of the articulation of the later philosophy, and is not discussed separately. Wittgenstein hardly discusses the Tractatus’s philosophy explicitly except when concerned with its failure. Accordingly, although one may accept Wittgenstein’s statement that his early philosophy did involve a metaphysics of language, this does not seem a good reason for thinking that the Tractatus’s aim was simply to state a doctrine of language—which then, paradoxically, is nonsense by its own lights. Rather, one must read from the Tractatus (and perhaps other sources from that period) what it was trying to do. This way one can also set the stakes high for the interpretation of the later philosophy. One can arrive at an interesting and strong interpretation of the later philosophy through a reading of the Tractatus that challenges philosophical thought. The cycle is then completed when one comes to see what is wrong with the Tractatus even according to the most favorable account. This way one might force oneself to think more and harder by trying not to downplay the early philosophy. As regards the Tractatus, my objective is to arrive at an interpretation that would allow one to understand how Wittgenstein might have thought of the Tractatus as an articulation of a philosophy without doctrines. This gives more depth to his later account of how the Tractatus nevertheless fails in cutting itself loose from philosophical doctrines. Consequently, this account of the Tractatus’s failure also allows me to shed light on how certain interpretations of Wittgenstein’s later philosophy still remain tied to philosophical doctrines, despite appearances. Insofar as Wittgenstein’s later philosophy enables one to grasp more clearly the Tractatus’s failure, it seems the discussion of the Tractatus’s conception of philosophy would benefit from being connected with the discussion of Wittgenstein’s later conception of philosophy. This book aspires to extend the discussion on Wittgenstein’s early conception of philosophy in this direction and to clarify the relation between his early

Introduction

11

and later thought. For although several volumes devoted to either Wittgenstein’s early or later conceptions of philosophy have been published in recent years, little has been published on the relation between them.19 This means that Wittgenstein’s later conception of philosophy has not yet received the kind of book-length treatment that Wittgenstein suggests would be optimal; i.e., it has not been the subject of a detailed study that seeks to understand it by way of contrast with the Tractatus.

Outline of the Discussion The following chapters attempt to approach a set of problems relating to the concept and methods of philosophy from slightly different angles. Thus, although the chapters develop a continuous line of argument, they can be seen as complementing each other in other ways too. What results is the description of a particular philosophical position, which I attribute to the later Wittgenstein. The first chapter addresses Wittgenstein’s early and later conceptions of philosophical problems. The crucial difference between these conceptions (although other differences will be discussed) is that whereas the early Wittgenstein sees philosophy as centered around a fundamental problem, the solution to which contains the solution to all problems, the later Wittgenstein abandons the idea of the hierarchical organization of philosophy assumed in his early approach. This hierarchical organization, he maintains, makes it impossible to get rid of philosophical problems. I introduce here, but can only discuss preliminarily, Wittgenstein’s shift to a conception of philosophy according to which the task of philosophy is to solve particular philosophical problems. The second chapter examines Wittgenstein’s early conception of philosophical clarification as logical analysis, as well as his later critique of the view that there is only one complete analysis of a proposition, or more generally, an ultimate determination of the essential features of an object of investigation. This critique of the Tractatus’s conception of logical analysis then provides the basis for a more detailed discussion of Wittgenstein’s conception of philosophy as the clarification of particular philosophical problems and the wide-reaching implications of this conception for logical or philosophical investigation. The chapter also addresses Wittgenstein’s notion of ordering knowledge of language use, which comes up in various connections in subsequent chapters.

12

Introduction

The third chapter constitutes the center of the argument of the book. It begins with a discussion of the concept of metaphysics and the concept of a philosophical thesis, as well as the Tractatus’s relapse into such theses. Against the background of the Tractatus’s failure and the problem of dogmatism, I propose an interpretation of Wittgenstein’s later conception of the status of philosophical statements that aims to make plain the sense in which his statements do not constitute philosophical theses. I also explain how this conception of the status of philosophical statements is motivated by a desire to avoid dogmatism in philosophy. Special attention is paid to the role of examples, rules, and idealized logical models in philosophical clarification. The interpretation of the status of Wittgenstein’s statements presented in Chapter 3 is further elucidated in Chapter 4 in connection with his remarks on the concepts of language and meaning. My purpose is to show how my interpretation of Wittgenstein’s method fits together with these remarks and receives further support from them. In this context, I also try to clarify the advantages of reading Wittgenstein along the lines I suggest. I try to show that it not only dissolves certain anomalies regarding the interpretation of his writings, but also makes possible a more nuanced understanding of the concepts of language and meaning than seems available for someone committed to philosophical theses. Wittgenstein’s method would thus increase the flexibility of philosophical thought without reducing its rigor. The fifth chapter plays a similar role to the fourth. It employs the results reached in Chapter 3 as a framework for the interpretation of Wittgenstein’s remarks on grammar, essence, and necessity. My interpretation contrasts with widely accepted conventionalist or constructivist readings that attribute to Wittgenstein a conception of essences and relevant kinds of exceptionless necessities as grammatical constructions. I argue that rather than putting forward a philosophical thesis about grammar or linguistic practices as the source of necessity, Wittgenstein avoids any commitment to such theses. Two key issues addressed are the relation of statements about essences and necessities to factual statements, and Wittgenstein’s conception of the nontemporality of statements of the former kind. This discussion also allows me to clarify the relation between Wittgenstein’s and certain more traditional views of essence and necessity. As in Chapter 4, I aim to show how Wittgenstein’s method can lead beyond oppositions between

Introduction

13

philosophical parties (in this case Aristotelianism and Kantianism) to richer philosophical views. In Chapter 6, I seek to further elucidate my interpretation by examining questions relating to the status of grammatical, clarificatory statements. Issues to be addressed include Wittgenstein’s rejection of philosophical hierarchies, his concept of perspicuous presentation, whether one may be forced to accept a philosophical clarificatory statement, what the correctness of grammatical remarks means, and Wittgenstein’s concept of agreement in philosophy. The notion of what I will call “multidimensional descriptions” of language use, which Wittgenstein’s conception of the status of philosophical statements makes possible, is discussed. I also explain how Wittgenstein’s later approach makes possible a new employment of philosophical or metaphysical theses and why worries about relativism are misplaced in the context of this account of philosophy’s tasks. The seventh chapter concludes the book by addressing the question of whether Wittgenstein’s approach to philosophy amounts to a doctrine about the essence of philosophy and examining his conception of the historicity of philosophy. I also discuss Wittgenstein’s view of the relation of philosophical language use to everyday language and seek to elucidate the sense in which there is, according to him, an ethical dimension to philosophy that pervades it in its entirety.

Notes on Source Materials and the Principles of Interpretation In this book I make wide use of Wittgenstein’s Nachlass, i.e., his notebooks, manuscripts, typescripts, and dictations (published as the Bergen Electronic Edition), as well as lecture notes taken by others. As Hilmy notes, this material constitutes an indispensable aid for reading the Philosophical Investigations: The main justification for our appeal to the Nachlaß is that one can best cut through the dense aphoristic quality of such a work as Philosophical Investigations by coming to grips with its fragments in their historical context (and genesis) within Wittgenstein’s own development and by viewing these fragments not only within the context of the bed of remarks where they were first conceived, but also within whatever other

14

Introduction

contexts Wittgenstein saw fit to place them during subsequent stages of revision.20

As well as explaining its justification, Hilmy here outlines the principles of the method of investigation that I will also employ. Ultimately, that is, the Nachlass is only an aid in the interpretation of the Philosophical Investigations, the latter constituting the most polished and authoritative presentation of Wittgenstein’s philosophy. Whatever views are attributed to the later Wittgenstein, they should, therefore, not be in conflict with the Investigations—although what counts as being in conflict with the Investigations is, of course, a matter of interpretation. As regards the use of the Nachlass, it is important that its employment not be a matter of “passage-hunting”21 or simply picking out remarks that best suit one’s purposes without any attention to their context and other factors affecting their interpretation. In this regard it is important to note that there is an order to the Nachlass. Because Wittgenstein often wrote in notebooks while working at the same time on a manuscript or a typescript, it is possible to find in these notebooks different versions of a remark, or attempts to make a particular point that can also be found in a manuscript or typescript. These different versions may greatly facilitate the interpretation of Wittgenstein’s views.22 Similarly, it is also sometimes possible to find in the notebooks remarks that reflect discussions in his lectures of the same time. These may help in deciding between different interpretations of the lecture notes, as well as confirming what is recorded in these notes. In addition to such parallel versions of a remark, there are also successive versions that follow each other in time as Wittgenstein writes “his book.” (Most manuscripts and typescripts may be characterized as work toward what became the Philosophical Investigations, although not all of them are drafts of the book or its parts, but may serve other purposes.) It may be revealing to follow these successive formulations of a remark, its wording sometimes changing almost completely as it makes its journey through the manuscripts and typescripts. Thus the first draft and the last version may be quite different from each other while still connected through various intermediary versions. Finally, as Hilmy notes, the different contexts in which Wittgenstein places a remark also offer useful hints for interpretation. For instance, the study of such different contexts allows one to come to understand connections between different

Introduction

15

remarks and ideas expressed in them. Again, this may help one to decide between different interpretations of the remarks in the Investigations. Where no translations have been published for remarks from the Nachlass, I provide my own. Occasionally, I have modified existing translations. Insofar as I think the modification is significant, this is indicated. In the case of the Tractatus, my translations are sometimes formed by combining existing translations by C. K. Ogden and by D. F. Pears and B. F. McGuinness. Sometimes I use square brackets inside quotations to mark ways Wittgenstein has edited his text, for instance, to give an alternative wording. The same convention is adopted to indicate words that have no straightforword English translation or to alert the reader to the original German when relevant.

TWO

Two Conceptions of Clarification

WHAT PRECISELY IS IT to clarify misunderstandings concerning language use with the purpose of dissolving philosophical problems? In this chapter I discuss Wittgenstein’s early and later conceptions of clarification as well as his later critique of his early approach. This will shed light on his shift away from great problems to the clarification of particular problems and prepare the ground for a discussion Chapter 3 of his turn away from philosophical theses. I begin (in 2.1) by discussing the Tractatus’s conception of clarification as logical analysis, its aspiration to introduce a logical notation— a concept-script—to be used to analyze language use, and Wittgenstein’s method of introducing this notation. (This may be read as a more detailed explication of the notion of the critique of language discussed in 1.1.) I then turn to his later critique of the Tractatus’s approach to clarification, examining first (in 2.21) what he says about misunderstandings pertaining to the Tractatus’s notion of the completeness of analysis and complete exactness. This is followed (in 2.22) by a discussion of problems connected with the granting of a privileged, fundamental status to a particular notation such as a concept-script, or in other words, with the idea of providing final once-and-for-all determinations of the essential logical features of language or particular concepts. The full significance of Wittgenstein’s critique of his early conception of clarification, however, cannot be revealed without turning to his positive characterizations of his later approach to clarification. To

Two Conceptions of Clarification

55

illuminate specific aspects of this critique, I first discuss (in 2.31) his later conception of clarification as the description of language and his shift from once-and-for-all determinations of concepts to the clarification of particular philosophical problems. Subchapter 2.32 then examines his characterization of clarification as bringing an order to the knowledge of language use. Here I also begin to discuss the contrast between my interpretation of Wittgenstein and that of Baker and Hacker by addressing whether there are, according to Wittgenstein, many such philosophical orderings of language or concepts, or only one. I will argue that the views Baker and Hacker attribute to the later Wittgenstein fail in certain important respects to move beyond views on language and philosophy that Wittgenstein already held in the Tractatus. Consequently, they remain in the target area of Wittgenstein’s later critique of his early work, according to my interpretation of this critique. Finally, the discussions in 2.31 and 2.32 also allow me to answer certain questions raised in Chapter 1.

2.1 The Tractatus’s Conception of Philosophy as Logical Analysis In the Tractatus, Wittgenstein characterizes the task of philosophy as follows: Philosophy aims at the logical clarification of thoughts. Philosophy is not a body of doctrine but an activity. A philosophical work consists essentially of elucidations. Philosophy does not result in “philosophical propositions,” but rather in the propositions becoming clear. The task of philosophy is to clarify and delimit sharply thoughts which otherwise are, as it were, opaque and blurred.1

At first sight, Wittgenstein’s characterization might not look very illuminating. What else would the task of the clarification of thoughts be than to make opaque thoughts clear? This is a readily available but not very enlightening metaphor. However, something characteristic of the Tractatus’s conception of clarification comes out in the last clause of this passage: philosophy delimits thoughts sharply. To grasp what Wittgenstein means by this, we must look at other things he says about philosophy, logic, and language.

56

Two Conceptions of Clarification

As explained in 1.1, philosophical problems, according to the Tractatus, arise from a failure to use language according to its logical syntax. One fails to observe some logical distinctions and is thus led to speak nonsense. As Wittgenstein explains, however, confusions of this kind can be avoided by using a sign-language or a notation that is governed by logical syntax or logical grammar and excludes logical errors—a concept-script (Begriffsschrift). He writes about avoiding logical errors (of the type discussed in 1.1): In order to escape such errors we must make use of a sign-language that excludes them by not using the same sign for different symbols and by not using in a superficially similar way signs that have different modes of signification: that is to say, a sign-language that is governed by logical grammar—by logical syntax. (The conceptual notation [Begriffsschrift] of Frege and Russell is such a language, which, however, does still not exclude all errors.)2

But what does Wittgenstein mean by a notation or concept-script governed by logical syntax? His claim is not that ordinary English, for example, does not conform to logical syntax, but that his notation does. A language is able to express something only insofar as it is logical, in accordance with logical syntax. In Wittgenstein’s view, as we saw in 1.1, there is no such thing as an impossible or illegitimate symbol, although there may be signs that do not symbolize. He says of everyday language: “All propositions of our colloquial language are actually, just as they are, logically completely in order.”3 The concept-script, then, is governed by logical grammar in the sense of having been designed in such a way that logico-syntactical distinctions are readily discernible in it. It is a logically perspicuous notation that has logical distinctions embodied, so to speak, in its signs. In the conceptscript, a sign is used to symbolize in only one way (unlike, for instance, the word “is” in ordinary English) and different modes of signification are kept clearly distinct. As a result, what makes sense and what does not is immediately recognizable in the concept-script. By being completely transparent to the logic underlying the use of signs, this notation excludes nonsensical sign formations and prevents logical misunderstandings. There is therefore no such thing as a grammatically well-formed but nonsensical proposition in the concept-script. As Wittgenstein notes, the concept-script would not allow, for instance, the formulation of the question, “Are there simple objects?” which, from the Tractatus’s

Two Conceptions of Clarification

57

point of view, is nonsense.4 Notably, given that according to section 3.325 of the Tractatus the criterion of adequacy for the concept-script is that it excludes all logical mistakes, an adequate concept-script must be able to make perspicuous every possible logical distinction that one might wish to draw—or that one could be unclear about.5 But let us examine more closely the idea that a language or a system of signs can be governed or not governed by logical grammar, i.e., reflect logic with more or less faithfulness. It is characteristic of this idea that logic is not considered dependent on any particular language. Logical investigation, as conceived in the Tractatus, has no interest in the particularities of languages as such. It is concerned with the essential only, this being something common to all different expressions capable of having a certain meaning or a sense. Consequently, particular languages or expressions are of interest to the logical investigation only insofar as the essential—that which makes signification possible— is discernible in them. Such essential features are also universal in the sense that insofar as they are necessary to express a meaning, any expression that is capable of expressing a certain meaning must have these characteristics. As Wittgenstein writes: A proposition possesses essential and accidental features. Accidental features are those that result from the particular way in which the propositional sign is produced. Essential features are those without which the proposition could not express its sense.6 So what is essential in a proposition is what all propositions that can express the same sense have in common. And similarly, in general, what is essential in a symbol is what all symbols that can serve the same purpose have in common.7 So one could say that the real name of an object was what all the symbols that signified it had in common. . . .8 A particular mode of signifying may be unimportant but it is always important that it is a possible mode of signifying. And that is generally so in philosophy: again and again the individual case turns out to be unimportant, but the possibility of each individual case discloses something about the essence of the world.9

Logic for the Tractatus is thus something very abstract. The logic of language is determined by the mutual translatability of linguistic expressions into each other, the essential logical features of an expression

58

Two Conceptions of Clarification

being those it shares with the other members of a class of mutually translatable expressions, as spelled out by a rule of translation.10 The logic of language, in this sense, is the essential core of all possible languages, leaving out everything accidental. Accordingly, the conceptscript, the notation that is governed by logical syntax or grammar, in contrast to the particular grammars of languages, is meant to exhibit the essential but not the accidental features of different languages. It brings to light the common core of all possible languages, revealing the logic of language, where “language” refers to languages in general. Or, one might say, this notation reveals the logic of thought, a particular thought being understood here as something that can find expression in different languages.11 The concept-script reveals concepts in their pure form, as they are behind their impure expression in different languages. Given that the concept-script contains in itself only what is essential to the expression of thought, it is free from all the features of language that might make one misunderstand logic. Thus, it would indeed be the ultimate instrument for the philosopher (and the scientist)—as Frege had conceived and Leibniz dreamt.12 The creation and employment of the concept-script would mean reaching a new ultimate level of clarity in language use: by deploying this notation, all possible misunderstandings could be clarified and thereby every possible philosophical problem resolved! Moreover, although the concept-script would make logical forms, and thereby the essential features of language and reality, readily discernible, it would not constitute a doctrine about them. Insofar as one is able to use a language at all, for instance to make claims, one must already have a grasp—if only an implicit one—of its logic. The same applies to the concept-script. Crucially, however, because the conceptscript does not disguise logic, logical forms would be completely open to view in its employment. By deploying the concept-script one would not be making any statements about the logic of language—such statements could not even be formulated in the concept-script—but the design of this notation would simply allow one to have a perspicuous view of the logic of one’s statements, as one intended them.13 In this sense the employment of the concept-script would result not in a doctrine, but in our propositions becoming clear, just as Wittgenstein characterizes the aim of philosophical clarification.

Two Conceptions of Clarification

59

* * * But given this sketch of the general idea of a concept-script and the philosophical significance of its employment, does the Tractatus actually provide us with a concept-script? As I understand it, it provides at least an outline of such a notation.14 I identify this outline as the Tractatus’s conception that propositions are to be analyzed into truthfunctional combinations of elementary propositions, the latter being composed of simple names.15 The Tractatus’s conception of clarification as logical analysis and the concept-script it projects can be characterized in more detail as follows. Commentators often describe clarification with the help of the concept-script as the translation of the expressions of everyday language into the formulae of the concept-script.16 More precisely, such translations are analyses of complex expressions in terms of definitions given in simpler expressions that determine the modes of signification of the complex expressions. As Wittgenstein says: “Every defined sign signifies via those signs by which it is defined, and the definitions show the way.”17 Insofar as one wants to track down the ways in which a complex expression signifies, this is to be done by analysis, by definitions given in simpler expressions. As for the term “translation,” this movement from complex expressions to simpler ones via definitions is, one might say, a special case of translation. For as Wittgenstein says of such definitions: “Definitions are rules for the translation of one language into another. Every correct sign language must be translatable into every other according to such rules.”18 Analysis, then, is translation in a specific direction: from complex to simple, where complexity and simplicity are the complexity or simplicity of the use of individual signs in terms of which analyses are given, not of signs (formulae) or sign systems. (Certainly, results of analyses in terms of a concept-script are likely to look more complex than the expressions of everyday language.) Given that finer distinctions are drawn in the concept-script than in everyday language, translations from the latter to the former seem appropriately called “logical analyses.” More specifically, as the conceptscript is meant to embody in itself all logical distinctions (in the sense that all logical distinctions can be captured in its formulae), translation in the concept-script is ultimately analysis into what Wittgenstein calls “simple names,” which constitute the end of analysis.19 Once the simple names are reached, the analysis is complete in the sense that

60

Two Conceptions of Clarification

these symbols allow for no further logical distinctions.20 Notably, although unclarities might sometimes be settled already before we reach the level of simple names,21 in other cases settling them might require further analysis. Ultimately there is, according to Wittgenstein, “one and only one complete analysis of the proposition.”22 This is then expected to contain the solution to every possible unclarity that might arise with regard to this proposition, including the problems whose solution does not require complete analysis. In this sense the completely analyzed proposition is thought to contain everything essential for the comprehension of the logic of the proposition in question.23 In line with this characterization of logical analysis, I suggest interpreting what Wittgenstein writes about the truth-functional analysis of propositions and the so-called picture theory of propositions as the outline of his scheme of analysis or his concept-script. Thus, rather than a speculative doctrine of the nature of propositions,24 Wittgenstein’s conception of propositions as pictures constitutes a part of the Tractatus’s concept-script. It is characteristic of Tractarian logical analysis, that is to say, to treat propositions as analyzable into truthfunctional combinations of elementary propositions and ultimately into pictures of states of affairs, where such pictures consist of names that stand for the simple objects in the world and present a state of affairs through their arrangement in the proposition.25 While the truthfunctional analysis of propositions then constitutes the “top level” of the Tractatus’s scheme of analysis, the conception of propositions as pictures constitutes the scheme’s “bottom level.” (The conception of propositions as pictures is by no means an arbitrary addition to the idea of the truth-functional analysis of propositions. The analysis of complex propositions into truth functions of elementary propositions presupposes the notion of a proposition as something true or false, explicated by Wittgenstein with the conception of propositions as pictures, i.e., (re)presentations of states of affairs.) Under the interpretation I propose, the Tractatus then does not merely conceive the task of clarification as something to be undertaken in the future. The carrying out of the task of clarification is not to be identified with the future employments of the concept-script, as if Wittgenstein’s book constituted merely a (metaphysical) prologue to this future philosophy. Rather, given that his concept-script is intended to be governed by logical syntax and, therefore, to reflect the logic of

Two Conceptions of Clarification

61

language without distortion, the introduction of the principles of this notation already counts as clarifying the logic of language. Certain logical problems may therefore be considered resolved simply through Wittgenstein’s introduction of his concept-script.26 The Tractatus is thus engaged in a double task: (1) the introduction of the principles of a notation designed to help us avoid logical mistakes in the future; and (2) the clarification of the logic of language through the introduction of the principles governing this notation.27 The central impetus for the interpretation I propose concerning the role of the truth-functional conception of propositions and the “picture theory” is that this interpretation makes it possible to understand Wittgenstein’s intention not to put forward any doctrines, while still allowing one to say that the Tractatus offers insights into the logic of language. Consequently, the interpretation avoids difficulties with both traditional and “purely therapeutic” resolute readings (for the notion of the purely therapeutic, see below). To begin with, to introduce a notation to be used for the purpose of logical clarification is certainly not to make a statement with a truthvalue. On the interpretation I propose, Wittgenstein therefore is not engaged in theoretical assertion in the sense of advancing a true/false theory of language, world, and so on. Consequently this reading avoids problems arising in the context of traditional interpretations that take the nonsensicality of the statements of the Tractatus to be an implication of the theory of language that the book puts forward.28 The principal problem with such interpretations is that ultimately they result in a sterile paradox. Either the Tractatus’s assertions articulate a theory, in which case they are not nonsense—or they are nonsense and do not articulate anything. The problem here is that rather than even appearing to have succeeded in articulating a conception of philosophy as devoid of theories, such traditional readings result in a stalemate. Either the Tractatus has a theory or, if theoretical assertion is the only possible form of philosophizing, the book is of no more philosophical importance than any other piece of nonsense. The book’s resolute readers, on the other hand, since they do not attribute a theory to Wittgenstein, avoid this paradox.29 Nevertheless, insofar as they take the purpose of the Tractatus’s nonsense to be “purely therapeutic,” i.e., to cure the reader of the temptation to put forward

62

Two Conceptions of Clarification

philosophical theories by making apparent their nonsensical nature, they face a difficulty in explaining how the Tractatus could contain—or give expression to—any insights into the logic of language (for example, concerning the non-representational character of logical constants). Accordingly, for example, Hacker blames Conant and Diamond (rightly or wrongly) for throwing away the logical insights Wittgenstein expresses in connection with his critique of Frege and Russell: “Throwing away the ladder is one thing, throwing away the baby together with the bathwater is another.”30 Moreover, purely therapeutic readings also leave it unclear why Wittgenstein should have spent almost ten years developing the “picture theory,” and so on, if it was meant merely as an example of a nonsensical theory, and why he did not simply discuss nonsense produced by other philosophers.31 The solution I propose to these difficulties with traditional and purely therapeutic resolute readings is essentially that the Tractatus’s logical insights are to be seen not as “expressed” in terms of a paradoxically nonsensical theory, but as embodied in the notation Wittgenstein seeks to introduce. Thus there is indeed a way to read the Tractatus as an expression of logical insights, though not in terms of true/false theoretical statements. This alternative way to express philosophical insights is outlined by Wittgenstein in a notebook from 1929: “R[amsey] does not comprehend the value I place on a particular notation any more than the value I place on a particular word because he does not see that in it an entire way of looking at the object is expressed; the angle from which I now regard the thing. The notation is the last expression of a philosophical view.”32 Although I do not cite this remark as evidence for Wittgenstein’s views in the Tractatus (it would be at best inconclusive), it does, nevertheless, present us with a possible way of thinking about the expression of philosophical insights—one provided by Wittgenstein himself—that seems capable of solving the problems with both the traditional and therapeutic readings. This problem-solving capacity (and not an appeal to uncertain evidence from 1929) is the reason for adopting the interpretation I propose.33 Consequently, if the Tractatus is read as suggesting the adoption of a notation to be used for purposes of clarification, the role of nonsense can be understood differently from what the traditional interpretation has assumed. Nonsense is not used to articulate a theory that provides a foundation for Wittgenstein’s notation, providing it with a “nonsensi-

Two Conceptions of Clarification

63

cal justification,” whatever that might be. Rather, although Wittgenstein does indeed deploy nonsense to introduce the Tractatus’s notation, in so doing he relies solely on the reader’s pretheoretical understanding of language and the distinction between sense and nonsense. The Tractatus, that is to say, seeks to draw the reader’s attention to certain logical distinctions that she already recognizes in her use of language but that are not clearly reflected in the ordinary symbolism, this symbolism therefore being likely to lead to confusion. In this way, the book endeavors to guide its reader to mastery of the notation in which these distinctions are clearly reflected: the sign-language governed by logical syntax or the concept-script. To ease this transition to the concept-script, Wittgenstein also employs nonsense, i.e., phrases that point toward crucial distinctions and their expression in the conceptscript but are in the end to be recognized simply as nonsense. Hence nonsense plays, as Diamond puts it, a transitory role. Its purpose is to lead one to the mastery of Wittgenstein’s notation.34 A central logical distinction to which Wittgenstein seeks to draw his reader’s attention is the distinction between statements of fact and the expression of what is logically necessary and possible, which he describes in the transitory vocabulary as the distinction between saying and showing.35 An important goal of the introduction of his notation, that is to say, is to render perspicuous the difference between the factual and contingent on the one hand and the essential, i.e., the necessary and possible, on the other.36 In the concept-script, no statements can be made about the essential, or even about the distinction between the essential and the contingent. Rather, by employing this notation one would simply be saying “what can be said.”37 I will not go into the details of how exactly Wittgenstein seeks to make manifest the distinction between the factual and the essential. Nevertheless, from the difficulties of the “nonsensical theory interpretation” we may conclude that the distinction has to be elucidated by using examples, that is, with the help of particular cases of language use where the distinction comes clearly into view. For, crucially, any general pronouncements about logic—that is, about logical necessities pertaining to language use—seem doomed to nonsense. Such statements treat the logical characteristics of language as if they were the object of contingent factual statements, thereby muddling the very distinction that Wittgenstein aims to clarify. As for examples used to explicate the

64

Two Conceptions of Clarification

factual-essential distinction, these might be of various kinds: cases where language is employed with sense, cases the Tractatus calls “senseless” (tautologies and contradictions), and, importantly, cases in which we are brought to see how something that seemed to make sense reveals itself as nonsense.38 An advantage of the present interpretation is that it renders readily understandable what Wittgenstein means by saying that the Tractatus constitutes a ladder to be thrown away.39 To throw away the ladder is not to throw away a theory and with it whatever logical insights the theory might provide. Rather, the Tractatus’s introductory nonsense is talk that Wittgenstein’s concept-script excludes. Accordingly, the apparent statements of the Tractatus concerning the essence of the world and language will be discarded once one begins to look at things from the perspective of Wittgenstein’s notation, or—what amounts to the same—adopts “the correct method of philosophy,” as described in Tractatus 6.53. This rejection of the Tractatus’s statements, then, is part of Wittgenstein’s project of the critique of language, aiming to draw limits to the expression of thoughts, which, as Wittgenstein says, ends the exclusion from language and characterization as nonsense of certain statements previously regarded as significant. It is also interesting to observe that once one begins to examine Wittgenstein’s text in the light of the reading proposed here, there is no shortage of forms of expression characteristic of someone introducing a notation rather than making metaphysical statements. (See for example the introduction of the notion of complete analysis of propositions beginning in Tractatus 3.2.) Despite having the appearance of a (paradoxically nonsensical) assertion about the essence of propositions, in the end the general propositional form itself is supposed to be nothing but “the description of the one and only general primitive sign of logic.”40 It is the most general formal characteristic of language expressed in terms of a variable of which all propositions are values, that is, a logical constant, something common to every possible proposition.41 Hence, what Wittgenstein says about the concept of general propositional form may be understood as an attempt to explicate a formal characteristic of language—with which language users are already familiar by virtue of being language users—by introducing a notation that makes this characteristic perspicuous. Nevertheless, I am not suggesting that the above interpretation could ultimately release the Tractatus from philosophical doctrines. Arguably,

Two Conceptions of Clarification

65

Wittgenstein’s claim of the universal applicability of his scheme of analysis—the contention that all philosophical problems can be dissolved by analyzing propositions according to this scheme, or have been dissolved “in essentials” when the scheme is introduced—involves an implicit claim about the nature of language. One might therefore say that in the Tractatus, metaphysics dresses in the gown of methodology and takes refuge in the philosophical method that Wittgenstein spells out. As I explain in Chapter 3, there is a problem relating to the Tractatus’s use of examples and its attempt to establish a method with the help of examples. Regardless of questions concerning the Tractatus’s doctrinal involvements, its conception of philosophy as clarification can now be summed up. Assuming the completion of the initial task of the clarification of the logic of language, through which the proper method of philosophy is established to be logical analysis,42 philosophy clarifies, by providing logical analyses, the logic of expressions that trouble us and entangle us in difficulties. Such analyses make perspicuous the possible ways of using language by excluding nonsensical sign-formations. This is to show either how a proposition can be dissected into simple names or that it cannot be so dissected, being merely an illusory proposition. To return now to section 4.112 of the Tractatus, from which I started, this activity of analysis “aims at the logical clarification of thoughts” in the sense that in a completely analyzed proposition, the “elements of the propositional sign correspond to the objects of thought.”43 Philosophy, therefore, aims to express thoughts clearly by revealing their simple elements, thereby clarifying opaque and blurred thoughts and delimiting them sharply. In this way, philosophical clarification releases thinking from the misleading influence of an everyday language not designed with logical perspicuity in mind. I revisit the Tractatus once more in Chapter 3 to discuss its doctrinal or metaphysical commitments. Let us now turn to Wittgenstein’s later critique of the Tractatus’s conception of clarification.

2.2 Wittgenstein’s Later Critique of the Tractatus’s Notion of Logical Analysis In the Investigations, Wittgenstein introduces and explains his conception of philosophy as clarification by contrasting it with his early

66

Two Conceptions of Clarification

approach. This involves detailed discussions of the unclarities and misunderstandings in the Tractatus’s view of clarification. I will proceed to the “positive” attributes of Wittgenstein’s later approach through a discussion of what he says about problems with the Tractatus. Wittgenstein writes in the Investigations about philosophy as the clarification of “misunderstandings concerning the use of words”: “Many of them can be removed by substituting one form of expression for another; this may be called an ‘analysis’ of our forms of expression, for the process is sometimes like one of taking a thing apart.”44 However, there is a danger of misunderstandings with regard to how we conceive the task of clarification: But now it may come to look as if there were something like a final analysis of our forms of language, and so a single completely resolved form of expression. That is, as if our usual forms of expression were, essentially, unanalysed; as if there were something hidden in them that had to be brought to light. When this is done the expression is completely clarified and our problem solved. It can also be put like this: we eliminate misunderstandings by making our expressions more exact; but now it may look as if we were moving towards a particular state, a state of complete exactness; and as if this were the real goal of our investigation.45

More specifically, according to Wittgenstein, the pursuit of this state of complete clarity and exactness assumes here the form of a great problem concerning the essence of language to be determined once and for all.46 As he says of the search for complete exactness in the immediately following remark: “This finds expression in questions as to the essence of language, of propositions, of thought. . . . We ask: ‘What is language?,’ ‘What is a proposition?’ And the answer to these questions is to be given once and for all; and independently of any future experience.”47 The question of how the Tractatus’s philosophy comes to assume the form of a great problem is discussed in Chapter 3. Here I wish to discuss what Wittgenstein says about misunderstandings pertaining to the Tractatus’s notion of the completeness of analysis and complete exactness, and then to take up problems associated with giving a privileged, fundamental status to a particular notation such as Wittgenstein’s concept-script.48

Two Conceptions of Clarification

67

2.21 The Tractatus’s Notion of the Completeness of Analysis and Complete Exactness The idea of a complete analysis of an expression is the idea that there is an analysis that lays bare all the logical relations of an expression to other expressions, thus resolving all possible misunderstandings and philosophical problems relating to the expression. In the case of a Tractarian name, for instance, this is to lay bare all possible combinations of the name with other names. The result of such an analysis is a “completely resolved form of expression”49 in which the essential features of symbols are completely divorced from their accidental features. But as Wittgenstein explains, the notion of completeness assumed here is problematic: “Formerly I myself spoke about the ‘complete analysis,’ the idea being that philosophy should decompose all propositions once and for all, thus laying down clearly every connection and removing every possibility of misunderstanding. As if there were a calculus in which this decomposition were possible. . . . All of this was based on a mistakenly idealised picture of language and its use.”50 Analysis, as conceived in the Tractatus, may be understood as the codification (translation) of expressions in the formulae of the conceptscript with the purpose of determining the logico-syntactical rules according to which language is used. The problem, then, is that although one can tabulate rules for the use of an expression in order to prevent particular misunderstandings, there is no such thing as a complete list of rules that would remove every possible misunderstanding. Suggestive as the picture of such a complete list may be, it turns out to be devoid of sense upon closer examination. Or as Wittgenstein says in his lectures of 1932/33: “We might feel that a complete logical analysis would give the complete grammar of a word. But there is no such thing as a completed grammar. However, giving a rule has a use if someone makes an opposite rule which we do not wish to follow.”51 The problem encountered in the notion of an ideally complete list of rules is explained in the 1937 version of the Investigations: How should we have to imagine a complete list of rules for the employment of a word?—What do we mean by a complete list of rules for the employment of a piece in chess? Couldn’t we always construct doubtful cases, in which the normal list of rules does not decide? . . .

68

Two Conceptions of Clarification

The regulation of traffic in the streets permits and forbids certain actions on the part of the drivers and pedestrians; but it does not attempt to guide the totality of their movements by prescription. And it would be senseless to talk of an “ideal” ordering of traffic which should do that; in the first place we should have no idea what to imagine as this ideal. If someone wants to make traffic regulations stricter on some point or other, that does not mean that he wants to approximate such an ideal.52

The analogy between the rules of language and the rules of traffic that Wittgenstein deploys here has certain limitations. If we assume the Tractatus’s idea that what is logically possible coincides with what makes sense, there is no such thing as prescription in logic.53 Nevertheless, there are also important similarities between the regulation of traffic and the clarification with which Wittgenstein is concerned. No ideal order of traffic or of language—or as one might well say, no ideal, complete state of clarity—can be approximated by laying down rules, because no ideal has been determined. Just as in the case of traffic we have no conception of the totality of movements to be regulated, in the case of language we have no conception of the totality of misunderstandings to be removed. Hence the notion of an ideal system of rules lacks a definite sense. No criteria have been specified for what it would mean for a system of rules to meet this ideal. More specifically, assuming that its task would be to remove particular misunderstandings, a list of rules would be complete insofar as it removed those misunderstandings it was intended to remove. In this case, the criteria for the completeness of the list are specifiable, and the task of tabulating rules is a sensible one. What constitutes the class of every possible misunderstanding, however, is a different question. As Wittgenstein observes, there seems to be no limit to misunderstandings that can be imagined. One can always imagine misunderstandings that are not covered by the rules one provides. Moreover, these rules themselves can be misunderstood. Consequently, the class of every possible misunderstanding is impossible to specify. But if it is impossible to determine which cases fall into the class of every possible misunderstanding, no criteria can be specified for the completeness of the list of rules needed to remove them. Thus in this case, the notion of a complete list of rules is devoid of sense, and there are no criteria for saying that the task of tabulating rules has been accomplished.54 If language were a calculus operating according to clear and precise

Two Conceptions of Clarification

69

rules, it would be possible to talk about the totality of all rules and about a complete list of rules for the use of an expression. Given a clearly defined system, such as a calculus, it is possible to list the rules that determine the possible employments of signs. This seems to be the sense in which the Tractatus assumes an idealized picture of language, as Wittgenstein remarks above in the quotation from manuscript 116. (As we have seen, the Tractatus conceives language as presentable in the form of a calculus of simple names. Accordingly, it is plausible that the word “formerly” in the passage above refers to the Tractatus.) Nevertheless, even in this case misunderstandings concerning these rules of language are possible. Constitutive rules such as those defining a calculus cannot remove every possible misunderstanding. There is no notation that cannot be misunderstood, and the concept-script is no exception. The Tractatus’s idea that logical analysis should aim to eliminate every possible unclarity is therefore confused. Conceived in this way, clarification becomes an impossible task. There is no such thing as an ultimate, final clarification that could remove every possible misunderstanding relating to an expression at once. Instead, clarifications are given when unclarities arise. They concern particular unclarities, and clarification requires no absolutely complete specification of the rules of language in the Tractatus’s sense. Or as Wittgenstein says about the concept of explanation: “One might say: an explanation serves to remove or to avert a misunderstanding—one, that is, that would occur but for the explanation; not every one that I can imagine.”55 Similar considerations apply to the notion of complete exactness. What it means for something to be exact as opposed to inexact is only determined in particular cases. For instance, what counts as exactness in measuring time in a laboratory and in the case of train arrivals is not the same. Accordingly, there is no general concept of exactness that would allow one to say that a complete exactness has been reached for every possible case. Just as the idea of completeness of clarification in an absolute sense is confused, so is the idea of complete exactness in an absolute sense.56 2.22 The Notion of a Fundamental Form of Expression In the Tractatus, the complete analysis of a proposition is envisaged as terminating in elementary propositions consisting of simple names that allow for no further distinctions, allegedly bringing to light all logical

70

Two Conceptions of Clarification

distinctions that could be drawn. But as Wittgenstein later explains, the conception of a fundamental form of expression that contains in itself everything that might be considered essential for the clarification of an expression is problematic. Not that some expressions could not be called “simple” (indefinable, not allowing for further distinctions, and so on) in contrast to “complex expressions.” The problem lies in the idea that everything essential in our expressions could be captured through a reduction to a single particular form of expression. This can be elucidated as follows: As Wittgenstein points out, unanalyzability or simplicity is not characteristic of an object of investigation as such. For instance, given that logic investigates the uses of language, no use of a sign is simple per se. What counts as simple or unanalyzable is relative to the assumed framework of analysis, or as one could say, to the mode of presentation employed. He writes: “ ‘Indefinable,’ here one imagines something unanalysable; and in such a way as if there was an object here that is unanalysable (like a chemical element). . . . —But the impossibility of analysis corresponds to a mode of presentation which we assume (lay down).”57 There are thus no simples in an absolute sense, as Wittgenstein says in the Investigations §§47–48, where he seeks to clarify this point by discussing a few concrete examples, illustrating in this way the concept of simplicity. He asks what we would call the simple parts of a tree or a chair, for instance, and discusses a (fictional) language-game of describing certain combinations of colored squares, where the individual squares are called “simple” in contrast to their combinations (“language-game (48)”). These examples make manifest that what is called “simple” (in the sense of not consisting of parts) has to be specified in a particular case or context. The criteria for simplicity and complexity need to be laid down, i.e., the way of conceiving things (Betrachtungsweise) or the mode of presentation has to be specified before it makes sense to talk about the “simple”(PI §47). That simplicity is relative to ways of conceiving things is exemplified, for instance, by the difference in what would count as the simple parts of a chair from the points of view of molecular physics and furniture design. Moreover, it seems that as regards logic, no mode of presentation (and no particular notion of simplicity) can be given a privileged status. Wittgenstein explains the point with the help of his invented example:

Two Conceptions of Clarification

71

Let us imagine language-game (48) altered so that names signify not monochrome squares but rectangles each consisting of two such squares. Let such a rectangle, which is half red half green, be called “U”; a half white a half green one “V”; and so on. Could we not imagine people who had names for such combinations of color, but not for the individual colors? Think of the cases where we say: “This arrangement of colors (say the French tricolor) has a quite special character.” In what sense do the symbols of this language-game stand in need of analysis? How far is it even possible to replace this language-game by (48)?—It is just another language-game; even though it is related to (48).58

In this variation of language-game (48)—let us call it (48)*—what is taken as “simple” is different than in (48). In this example it would appear most natural to analyze (48)* in terms of (48), as the names U, V, and so on in (48)* stand for combinations of the monochrome squares of (48). But as Wittgenstein points out, it is not necessary to understand (48)* in terms of (48). One might treat names in (48)* as not analyzable into and replaceable by names in (48). It is perfectly possible, logically, to treat certain color combinations as simple and basic instead of individual colors, and to have a notation in which color combinations are simple, that is, in which the particular character of a combination is treated as not further analyzable in terms of simple parts. And although one might be tempted to think otherwise, there are readily available examples of something like this, such as the French tricolor or the combination of red, blue, and white in the U.S. flag. Both games (48) and (48)*, one might say, have the same freedom, and neither is superior to the other. But to get a firmer grasp of Wittgenstein’s point and its relevance to the notion of analysis, let us look at how he arrives at it through the remarks that precede §64. Here it is important to bear in mind that logical analysis is not, as it were, the analysis of mere signs, but of symbols—or signs with logicosyntactical employment, as one might say from the Tractatus’s perspective. As regards philosophical problems with the concept of analysis, therefore, the phenomenon of ambiguity (i.e., that a sign can be used in many ways and would be analyzed differently with respect to these different uses) is of no great interest. Rather, the interesting question concerns the analysis of a sign with a particular use, whether analysis in such a case could be said to have some one terminus—a definitive account of

72

Two Conceptions of Clarification

the use of the sign. That Wittgenstein is not concerned with mere ambiguity is evident from his example. In language-games (48) and (48)* different signs are used. Accordingly, the question concerning the analyzability of (48)* in terms of (48) is the question of whether the use of the signs in (48)* can and must be explained in terms of the use of the signs in (48). Insofar as the task of logical analysis is to distinguish the essential (logical) features of a symbol from its accidental features, the above question can be formulated as the question of whether there is only one possible way of sorting out the essential from the accidental in the use of a sign. Here Wittgenstein’s answer is that it need not always be so. There need not necessarily be only one generally valid answer to the question, what is the point of a particular language-game with a sign? There need not be only one correct answer to the question, what are the important and central features of a game? and consequently to the question, what is essential in the game? Rather, what is essential in the game depends on one’s interests or concerns and may vary according to the point of view from which one looks at it. Put in another way, there need not be one single answer to the question of whether the uses of two particular signs are the same or different—that is, whether one could be translated in terms of the other (or replaced by it) or whether something important gets lost in translation. (Sometimes a translation is appropriate in some respects but problematic in others.) If there is no definitive and exclusive answer to these questions, there is no single answer to the question, what is essential in the uses of these signs? Wittgenstein discusses the issue with the help of a few examples. We could treat a proposition (a), which names the different parts of a chair and talks about chairs in this way, as an analyzed version of a proposition (b), which talks about chairs. Would we then say that ordering someone to fetch the parts of a chair is the same as ordering someone to fetch the chair? There are grounds for both an affirmative and a negative answer. Wittgenstein, imagining an interlocutor, writes: “But all the same you will not deny that a particular order in (a) means the same as one in (b); and what would you call the second one, if not an analysed form of the first?”—Certainly I too should say that an order in (a) had the same meaning as one in (b); or . . . they achieve the same. And this means that if I were shewn an order in (a) and asked: “Which order in (b) means the same as this?” or again “Which order in

Two Conceptions of Clarification

73

(b) does this contradict?” I should give such-and-such an answer. But that is not to say that we have come to a general agreement about the use of the expression “to have the same meaning” or “to achieve the same.” For it can be asked in what cases we say: “These are merely two forms of the same game.”59 Suppose for instance that the person who is given the orders in (a) and (b) has to look up a table co-ordinating names and pictures before bringing what is required. Does he do the same when he carries out an order in (a) and the corresponding one in (b)?—Yes and no. You may say: “The point of the orders is the same.” I should say so too.—But it is not everywhere clear what should be called the “point” of an order. (Similarly one may say of certain objects that they have this or that purpose. The essential thing is that this is a lamp, that it serves to give light;—that it is an ornament to the room, fills an empty space, etc., is not essential. But there is not always a clear distinction between essential and inessential.)60

Depending on one’s interests or point of view, one might adopt various conceptions of what is essential in the use of a sign. (Consider the contrasting viewpoints: “The function of a lamp is to give light” and “Only an extremely unsophisticated person thinks that lamps are only about light.”) Depending on which aspects of the use of the sign one focuses on—for instance, the issue of which orders a particular order contradicts, or what exactly is involved in carrying it out—one may have different views of what is essential in it. Consequently, different analyses of the use of the sign—that is, explanations of what is essential in it—are possible.61 Similarly, it might be precisely the complexity of the expression that is regarded as essential to it. This might be the case, for instance, in language-game (48)*. Here the combination of colors might be treated as essential, not analyzable in terms of names for monochrome squares, as this analysis treats the combinations as merely accidental. Wittgenstein writes about the idea of the fundamentality of the analyzed form: “To say, however, that a sentence in (b) is an ‘analysed’ form of one in (a) readily seduces us into thinking that the former is the more fundamental form; that it alone shews what is meant by the other, and so on. For example, we think: If you have only the unanalysed form you miss the analysis; but if you know the analysed form that gives you everything.—But can I not say that an aspect of the matter is lost on you in the latter case as well as the former?”62

74

Two Conceptions of Clarification

Although an analysis of an expression may bring out something interesting that is not discernible in the unanalyzed form, it may at the same time conceal something else. The analysis emphasizes certain aspects of the expression at the cost of others, and something may get lost in the process of analysis. (The analyzed and the unanalyzed form are different, which, after all, is the point of the analysis. If there were no differences between the analyzed and unanalyzed expressions, what would be the point of analysis?) And although the features of the analyzed form may be important to us for certain purposes, this may be true of the unanalyzed form as well. Hence, one simply begs the question if one says: “Some features may be lost in the analysis, but what is important is that the essential features remain.” Consequently, according to the later Wittgenstein, it would be misplaced to maintain that one must regard some form of expression as the fundamental one, in the manner of translations into the conceptscript. There might not be any final once-and-for-all answer to the question, which features of the expression are essential? and it may be that no single particular form of expression encompasses all the features that could be considered essential to the expression. Rather, the same point we ended up with in the previous section emerges from Wittgenstein’s discussion: clarifications—or logical analyses—are to be considered as relative to particular problems and questions. There are no ultimate analyses, no final renderings of what we say in a fundamental form that could clear away all misunderstandings concerning language use. Therefore, although translations in a concept-script such as the Tractatus proposes may be helpful sometimes, they may be unhelpful or misleading at other times. The formulae of a conceptscript cannot be treated as the fundamental form of expression. Arguably (to refer back to 2.21), the same applies to any tabulation of the rules of language: no particular determination of the rules for the use of an expression can be treated as the description of its use so far as the clarification of misunderstandings is concerned. But this last point requires further elucidation, and I return to it later.63

2.3 Clarification in Wittgenstein’s Later Philosophy Read against the backdrop of the Tractatus, Wittgenstein’s later conception of philosophy as clarification can be seen as designed to overcome

Two Conceptions of Clarification

75

problems that ultimately prevent the Tractatus from reaching its aim: philosophy without doctrines. (His early and later philosophy are, therefore, continuous, constituting a continuous attempt to spell out a certain kind of philosophical approach.) More specifically, as we have seen, Wittgenstein’s critique of the Tractatus’s conception of analysis is informed by the idea that a shift from once-and-for-all determinations of concepts to the clarification of particular unclarities is required to eliminate the philosophical difficulties that plague the Tractatus. The following two subchapters seek to clarify further the idea of this shift by examining some of Wittgenstein’s positive characterizations of his approach. 2.31 Description of Language Use and the Clarification of Particular Problems The construction of a concept-script and the analysis of language in its terms might be characterized as an attempt to bring to the fore the logic of language—the essential logical properties obscured by the grammatical features of particular languages, which are considered accidental.64 By contrast, Wittgenstein’s later philosophy understands logical or philosophical investigation as concerned with particular languages. The object of investigation of logic or philosophy, according to him, is not language in an abstract sense, but actual natural languages such as German and English or specialist languages such as scientific languages, for instance, “the chemist’s language.”65 He writes: “Philosophy deals with existing languages and should not pretend that it must discuss language in an abstract sense.”66 And in the Investigations: “The philosophy of logic speaks of sentences and words in exactly the same sense in which we speak of them in ordinary life when we say e.g. ‘Here is a Chinese sentence,’ or ‘No, that only looks like writing; it is actually just an ornament’ and so on. [New paragraph] We are talking about the spatial phenomenon of language, not about some non-spatial, nontemporal phantasm.”67 More specifically, philosophy seeks to dissolve philosophical problems by describing language use.68 As Wittgenstein says: “Philosophical problems are not solved by explanation but by description.”69 Accordingly, he now understands logical analysis, or clarification more generally, as an activity of describing language use: “Logical analysis. Who describes the use of a word clarifies the concept to us.”70

76

Two Conceptions of Clarification

But at this point one might want to ask, does Wittgenstein’s conception not make philosophy quite uninteresting? Insofar as philosophy is concerned with the description of languages in an ordinary, nonabstract sense, is it not concerned with the contingent and perhaps merely accidental? Of what philosophical interest could it be to investigate the contingent concepts or structures of a particular language? Perhaps it is this fear that has led commentators to insist that Wittgenstein is not concerned with mere particularities of actual languages but with something common to different human languages. For instance, Rush Rhees writes about Wittgenstein: “He is still interested in ‘human language’ rather than in the language or languages that people speak.”71 Similarly Oswald Hanfling: “He was not concerned about the grammar of particular languages . . . , but about logical features of what may be called human language (or human thought).”72 Arguably, however, this worry signals a misunderstanding of the point of Wittgenstein’s descriptions of language. If so, Rhees’s and Hanfling’s attempts to respond to the worry are misplaced. Wittgenstein is not moving in his later philosophy from abstract philosophical or logical theses allegedly holding for every possible language73 to observations concerning actually existing languages that claim a universal validity in Rhees’s and Hanfling’s sense. That is to say, the significance of Wittgenstein’s descriptions of particular languages does not depend on his remarks somehow reaching beyond the merely particular to what is universally true of all languages. But this does not mean that, according to Wittgenstein, we must accept a weaker form of philosophy, i.e., that in his view the only available answers to philosophical questions about essences are historical statements about contingent conceptual determinations.74 Rather than thinking about the significance of philosophy in these terms—and, consequently, either affirming or rejecting the claim about the universal significance of philosophy in a strong ahistorical or transhistorical sense—Wittgenstein sees the matter differently.75 The significance of philosophy is as great as the importance of having a clear comprehension of what one is saying. In this sense philosophy’s significance depends on the importance of language for human beings and on the importance of the misunderstandings to be cleared away. As he remarks: “The problems arising through a misinterpretation of our forms of language have the character of depth. They are deep disquietudes; their roots are as deep in us

Two Conceptions of Clarification

77

as the forms of our language and their significance is as great as the importance of our language.” And: “The results of philosophy are the uncovering of one or another piece of plain nonsense and of bumps that the understanding has got by running its head up against the limits of language. These bumps make us see the value of the discovery.”76 Apparently anticipating the above worries, Wittgenstein also writes: That I have to use a full-blown language (and not some sort of preparatory, provisional one) in giving my explanations already shows that I can only adduce something external about it. Yes, but how can these accounts then satisfy us?—Well, your very questions were framed in this language; they had to be expressed in this language, if there was anything to ask! And your scruples are misunderstandings. Your questions refer to words; so I have to talk about words.77

This remark may be read as seeking to eliminate the troubling asymmetry, i.e., the appearance that Wittgenstein is trying to answer questions that concern something transhistorical and universal— immutable essences and structures—by investigating historically contingent concepts. The solution to the problem of historicism is not to claim a strong universal status for Wittgenstein’s statements, as Rhees and Hanfling construe his remarks about language.78 Rather, the solution consists in recalling that the questions that express philosophical problems are themselves articulated in some particular language or another. Philosophical questions are posed in the language or languages spoken by those involved in the philosophical discussion, and these languages have a historically contingent existence. Accordingly, insofar as clarification is needed to get clear about such questions and problems, it will concern the expressions of the language(s) in which they were articulated. “Your questions refer to words; so I have to talk about words.” Or more specifically, to the extent that philosophical questions, claims, and hopes are formulated in everyday or ordinary language, it is everyday language (as used by relevant persons) that must be examined. Consequently, one can bracket and put aside the question of whether philosophical statements concern something universal in the strong, transhistorical or ahistorical sense or instead have more limited, perhaps merely “local” interest. The point is, first, that although

78

Two Conceptions of Clarification

one cannot assume philosophical concerns that find their expression in some particular language to be of universal and transhistorical relevance, this does not mean that they could not have such relevance either. This issue is not decided one way or the other by taking particular languages as the object of philosophical investigation. Second, it is important to notice that the idea that philosophy deals with immutable, universally relevant matters, and that this is what gives philosophy its importance, is itself an idea expressed in particular languages in particular historical contexts. This idea, therefore, is subject to logical or grammatical investigation just like anything else. It does not identify a requirement that philosophy must meet but that is beyond the reach of philosophical investigation as conceived by Wittgenstein. Accordingly, that Wittgensteinian clarifications might (sometimes) lack universal relevance does not yet constitute an objection to his approach. A point related to the one expressed in the Investigations §120 concerning Wittgenstein’s approach is expressed in the following remark: “What is it that is repulsive in the idea that we study the use of a word, point to mistakes in the description of this use and so on? First and foremost one asks oneself: How could that be so important to us? It depends on whether what one calls a ‘wrong description’ is a description which does not accord with established usage—or one which does not accord with the practice of the person giving the description. Only in the second case does a philosophical conflict arise.”79 As this remark makes clear, Wittgenstein’s descriptions of language are not motivated by the assumption that some established linguistic practices—or a philosopher’s description of them—should be accepted as a standard for how one is to speak and think. Deviation from established ways of using language—for instance, the practices adopted by the clarifier—is not philosophically problematic. Rather, philosophical problems arise from conflicts within a person’s language use. The problem of other minds provides an example. By conceiving sensation-talk according to the model of statements about inner objects of knowledge, the philosopher is confronted with the difficulty of not being able to attribute sensations to others.80 Here the root of philosophical problems is not that the philosopher’s description of language use—her conception of sensation-talk—does not accord with the normal use of language. Rather, the roots of the problems lie in the conflicts that arise between the ways in which the philosopher herself

Two Conceptions of Clarification

79

wants to use language. On the one hand, she wishes to be able to attribute sensations such as pain to others; on the other hand, her conception of sensation-talk makes it impossible for her to account for the possibility of such attributions in a way satisfactory to her. It is in this sense that philosophical problems arise from unclarities about what one is saying and from conflicts within one’s language use.81 Thus, a philosophical problem must be cleared up, so to speak, from the inside. Clarification, as Wittgenstein conceives it, is not a matter of imposing an alleged standard of correct language use on the interlocutor from the outside, but of clarifying the interlocutor’s language use to her on the basis of her own criteria for what makes sense. In this sense, clarification is essentially a dialogue between the philosopher and her interlocutor.82 Hence it would also be wrong to think that ordinary or everyday language, or a certain description of it, constitutes for Wittgenstein a standard of sense that he urges one to abide by, even though this is how he has often been interpreted.83 As Savickey notes, Wittgenstein’s philosophy would thus become “a form of intellectual constraint or censorship.”84 In other words, Wittgenstein’s point is not that philosophical questions must be answered in terms of everyday language, but that the intelligibility of the questions must be examined in the language in which the questions were formulated.85 Let us now turn to another aspect of Wittgenstein’s later conception of clarification as the description of language use. As explained in 2.21 and 2.22, Wittgenstein in his later philosophy problematizes the Tractatus’s conception of clarification as an attempt to find final, once-andfor-all analyses of concepts. He thus departs from the idea of philosophy as an attempt to find great answers, i.e., to settle once and for all every possible unclarity relating to a concept through its final analysis. This means that in his later philosophy, Wittgenstein is not only dealing with particular, actually existing languages but with particular, actual philosophical problems. As he says: “The task of philosophy (in my sense) is to point out actual mistakes.”86 This change in Wittgenstein’s comprehension of the task of clarification has very important consequences for his conception of the role of descriptions of language use in philosophy, as I will explain. Wittgenstein characterizes the point of his descriptions as follows: “The game with these words, the use which is made of them is more

80

Two Conceptions of Clarification

involved—the role of these words in our language other than we are inclined to believe. [New paragraph] It is this role which we must understand in order to dissolve philosophical paradoxes. . . . [New paragraph] We seek to find the meaning of a word only insofar as we describe this role. And we describe it only as far as is necessary for dissolving philosophical problems.”87 According to this, dissolving philosophical paradoxes requires comprehending the uses or roles of the relevant words in the language in question, the purpose of Wittgenstein’s descriptions being to help one understand such uses or roles. With respect to present concerns, the last paragraph of the quotation is particularly interesting. Insofar as the task of philosophy is the clarification of particular, actual philosophical problems, the statement that the roles of words are described only as far as is required for dissolving philosophical problems may be read as a rejection of the idea that philosophical descriptions should provide once-and-for-all accounts of the words’ roles. Rather than seeking to cover all uses of a word in the relevant meaning, and to settle every possible problem relating to it, the descriptions only capture its uses to the extent required by the clarification of certain particular, actual problems. Wittgenstein makes this point in his lectures as well: “The point of examining the way a word is used is not at all to provide another method of giving its meaning. When we ask on what occasion people use a word, what they say about it, what they are right to substitute for it, and in reply try to describe its use, we do so only insofar as it seems helpful in getting rid of certain philosophical troubles.”88 Thus, rather than examining the use of a word in the abstract with the purpose of providing an account of its meaning, Wittgenstein always focuses in his descriptions on the clarification of specific philosophical problems. Unlike the Tractatus’s completed analyses, descriptions in Wittgenstein’s later philosophy are not intended to cover all possible uses of the relevant terms in relevant contexts of their employment and to contain an answer to every possible unclarity that might arise. Rather, the completeness of a philosophical description is relative to particular philosophical problems. A description is complete insofar as it dissolves the problems it was intended to dissolve.89 This point is expressed in the Investigations as follows: “It is not our aim to refine or complete the system of rules for the use of our words in unheard-of ways. [New paragraph] For the clarity we are aiming at is

Two Conceptions of Clarification

81

indeed complete clarity. But this simply means that philosophical problems should completely disappear. . . .”90 No doubt complete analyses, as the Tractatus envisages them, would count as unheard-of refinements of the system of rules for the use of our words. The first sentence of the quotation can therefore be plausibly interpreted as referring to the Tractatus.91 Moreover, as regards the interpretation of this sentence, it is important that in the 1936 and 1937 versions of the Investigations it occurs as the first paragraph of the remark (quoted in 2.21) comparing traffic rules and the rules of language, thus problematizing the notion of a complete list of rules for the use of a word.92 As explained in 2.21, this remark can be read as spelling out Wittgenstein’s grounds for rejecting the aim of the completeness of philosophical accounts as it is understood in the Tractatus. There is, according to him, no such thing as a complete account of the logical grammar of a word that excludes all possible misunderstandings relating to it. The remark quoted in 2.21 can, therefore, help us to understand the point in the Investigations §133 about not refining rules in unheard-of ways and Wittgenstein’s later conception of the completeness of descriptions of language. Although complete clarity—rather than lesser clarity of some sort— continues to be the goal of Wittgenstein’s later philosophy, his conception of this goal has been transformed. Now complete clarity is understood as the complete disappearance of particular philosophical problems relating to problematic concepts, which does not require once-and-for-all accounts or “completed grammars” of the relevant concepts. It is also significant that in the Investigations the paragraphs of §133 quoted above are followed by the part of this remark discussed in 1.5: “Problems are solved (difficulties eliminated), not a single problem.” This confirms that Wittgenstein’s shift away from great once-andfor-all solutions to the clarification of particular philosophical problems is, indeed, motivated by problems relating to the notion of the completeness of philosophical analyses or accounts. The aim of philosophical clarification, according to the later Wittgenstein, is not completeness in an abstract sense, as he conceived it in the Tractatus.93 But let us try to focus even more sharply on the contrast between the Tractatus’s and Wittgenstein’s later approaches and the precise sense in which his turn to the discussion of particular problems constitutes a turn away from once-and-for-all determinations of concepts. The quotations

82

Two Conceptions of Clarification

above from manuscript 121 and the lectures might seem to say merely that philosophical descriptions need not cover all uses of a word in a particular meaning and that it is enough for such descriptions to extend only as far as is necessary for the elimination of certain particular problems. According to this reading, Wittgenstein’s point would be that in philosophy one can often do with less than a once-and-for-all account. But this interpretation is problematic. The problem is that the view thus attributed to the later Wittgenstein is also compatible with the Tractatus and thus obscures the difference between his early and later conceptions of the completeness of clarifications. In particular, the interpretation creates a conflict between Wittgenstein’s critique of the Tractatus and his positive characterizations of his later approach, bringing his later thought into the target area of the critique aimed at the Tractatus. As noted in 2.1, the Tractatus too allows for the possibility of dissolving philosophical problems before reaching the ultimate end of analysis. The idea of interrupting an analysis before its ultimate end, therefore, is perfectly compatible with the idea that there are ultimate analyses of concepts that contain solutions to all philosophical problems relating to them, including those problems that do not require taking analyses to their ultimate end. Similarly with other descriptions of language use: the idea that one does not always need to describe the use of a word in all respects in order to solve a philosophical problem connected with the word is compatible with the idea of an ultimate description that provides the solution to all problems with the word. Interpreted in the way just outlined, Wittgenstein’s later approach therefore does not constitute a decisive shift away from philosophy as a search for great once-and-for-all solutions. It still leaves open the possibility of such solutions. As explained in 2.22, however, Wittgenstein’s critique of the idea that there is a single, ultimate end to logical analysis rejects the whole notion of such once-and-for-all solutions. As he argues, it is possible to analyze the use of an expression in more than one way, and what counts as an appropriate analysis is relative to one’s interests, where such interests may be specified by reference to the particular philosophical problems that are one’s concern. What should then be said about the uses of a word—which logical distinctions must be drawn— depends on which problems more specifically one aims to resolve. Accordingly, there might be different analyses of a concept or different

Two Conceptions of Clarification

83

descriptions of a word’s use relative to different problems. The notion of a “description of the essential features of a concept” has no absolute sense, and there need not be a unifying, fundamental, once-and-for-all account that brings together all different analyses or descriptions. Taking these considerations into account, clarification emerges as essentially concerned with particular philosophical problems. That this interpretation, rather than the one mentioned before, captures Wittgenstein’s later conception of the description of language is confirmed by the following remark, from a 1933 notebook:94 “As my students set other examples against examples of the use of a word so as to demonstrate that the word is not used the way I think, the answer is always that these counter-examples are very useful but they do not demonstrate that I have not described the use correctly; for I did not want to say at all that my examples show the use of the word but only one way of using it. The mistake is the assumption that we wanted to illustrate the essence of, say, understanding with these examples and the counter-examples demonstrate that this has not been grasped correctly. As if our aim were to give a theory of understanding that would then have to explain all cases of understanding.”95 Thus, according to Wittgenstein, his students misunderstand his purposes when they assume that his characterization of the use of a word—“understanding,” in this example—with the help of certain examples is meant to account for all its uses (in the relevant meaning). Consequently, the students’ counterexamples are beside the point. They may bring to view matters of interest about the use of the word in question. But they do not make Wittgenstein’s description incorrect, because it was not meant to cover all the word’s uses (in the relevant meaning) in the first place. Or as Wittgenstein says, his purpose is not to put forward a theory of the essence of understanding, where such a theory would provide an account of what all instances of understanding must be.96 Rather, he only wants to describe, or clarify with his examples, a certain way or ways in which the word is used. Insofar as Wittgenstein’s students’ examples really are genuine counterexamples to Wittgenstein’s account, they are not compatible with his description in the sense that when used to provide the basis for an account that is intended to cover all cases of the word “understanding,” the students’ and Wittgenstein’s examples exclude each other. This means that insofar as the cases that Wittgenstein and his students

84

Two Conceptions of Clarification

bring up are all agreed to be genuine cases of understanding, and appropriately described (as Wittgenstein accepts in the remark), they cannot form a unified account. They cannot be consolidated into a once-and-for-all account of the use of “understanding” or the essence of understanding. But this lack of unity by no means makes these descriptions useless from the point of view of philosophical clarification, as conceived by the later Wittgenstein. Descriptions of the use of a word that focus on the different ways in which it is used, or on different aspects of its use, may be employed as reminders about these uses or aspects. Accordingly, such descriptions can enable one to dissolve particular philosophical problems that arise from confusions relating to those uses or aspects, or from ignoring them. As regards the interpretation of Wittgenstein’s shift away from philosophical once-and-for-all accounts, the significance of the passage above then is that it provides us with a characterization of his method in which Wittgenstein explicitly denies that the aim of his descriptions is the composition of once-and-for-all accounts of the uses of words.97 What I have said about the uniformity of philosophical accounts raises other questions about the status of Wittgenstein’s later descriptions of language. If there are no once-and-for-all answers to philosophical problems, philosophy’s task being the clarification of particular problems and the description of particular cases of language use, does this mean that philosophy becomes an empirical study of language? Does the task now become that of describing a multitude of different instances of language use, apparently an empirical task? Wittgenstein’s reply to this question may be found in the following passage on the notion of the description of language, which rejects the idea of philosophy as an empirical study of language: “And this description gets its light, that is to say its purpose, from the philosophical problems. These are, of course, not empirical problems; they are solved, rather, by looking into the workings of our language, and that in such a way as to make us recognize those workings: in despite of an urge to misunderstand them. The problems are solved, not by giving new information, but by arranging what we have always known.”98 The point is that philosophical clarification is concerned with dissolving misunderstandings, not with informing anyone about language use as if

Two Conceptions of Clarification

85

she did not know how to use language or as if the purpose were to collect data about different ways of using language. Whereas an empirical study informs one about something one did not know (perhaps confirming a hypothesis about how things might be), philosophical descriptions are intended to dissolve confusions relating to something one already knows but has difficulties in understanding and explaining to oneself.99 This difference comes to view in that whereas the correctness of an empirical investigation is judged on the basis of evidence for it (reports about language use need to be well documented, and so on), the criterion of correctness for a philosophical account, insofar as philosophical problems are based on misunderstandings, is the disappearance of the problem: the recognition that it has been resolved or dissolved.100 Thus although Wittgenstein’s shift away from unified once-and-forall accounts of concepts to the description of actual language use may create the impression that philosophy now becomes an empirical, case-by-case study of language, this impression is misleading. Philosophy is concerned with describing language use because such descriptions can be used to dissolve misunderstandings and unclarities relating to language. But such descriptions concern something one is already familiar with as a competent user of the relevant expressions. They only serve to remind one of what one already knew.101 In a similar vein, it would be misleading to think that descriptions of particular uses or aspects of language might be employed as a basis for aggregative descriptions combining depictions of individual cases into more comprehensive accounts—unless the aim of comprehensiveness is given a more precise characterization by reference to problems to be resolved. The aim of philosophical clarification cannot be the comprehensiveness of philosophical descriptions as such, in the sense that the descriptions should cover the different uses of words as widely as possible. Rather, comprehensiveness is to be understood in terms of the completeness of a philosophical account, as described above, an account being complete if it dissolves particular actual problems.102 The foregoing discussion seems also to supply us with an answer to the question of what constitutes Wittgenstein’s shift away from questions that are impossible to answer to questions that are easy to answer, as discussed in 1.2 and 1.5. Wittgenstein’s “easy questions” concern the

86

Two Conceptions of Clarification

clarification of particular problems, in contrast to attempts to resolve every possible unclarity that might arise about a concept. They are questions concerning particular usages of words (or aspects of words’ uses), a comprehension of which is required for the resolution of certain particular problems. The difficulty with this mode of clarification is in arranging what we already know about language use in such a way that problems are eliminated. As G. E. Moore reports from Wittgenstein’s lectures, although Wittgenstein said “that he would only tell us ‘trivial’ things—‘things which we all know already’ . . . the difficult thing was to get a ‘synopsis’ of these trivialities, and that our ‘intellectual discomfort’ can only be removed by a synopsis of many trivialities—that ‘if we leave out any, we still have the feeling that something is wrong.’”103 This is a genuine difficulty not to be underestimated. Nevertheless, in contrast to the impossibility of reaching definite, once-and-for-all accounts of the grammar of the relevant words—“complete presentations of their grammar”—this difficulty is manageable in principle. Chapters 3 and 4 examine in more detail the method of clarification of particular problems outlined here. Now, however, I will discuss Wittgenstein’s characterization of the task of philosophy as that of arranging or ordering knowledge of language use. This provides us with another way of explaining the difference between his early and later conception of philosophical clarification as well as an opportunity to further elucidate his conception of philosophy as the clarification of particular problems. Here I also begin to discuss the contrast between the interpretation of Wittgenstein’s approach articulated in this book and the reading of Baker and Hacker. 2.32 Clarification as Arranging and Ordering As well as depicting philosophy as the description of language use, Wittgenstein characterizes its task as that of arranging or ordering. The task of philosophy is to bring an order to “our concepts,” “the things,” or “our knowledge of the use of language,” as he alternately formulates his conception. He writes: “The philosophical problem is an awareness of disorder in our concepts, and can be solved by ordering them.”104 And in the Investigations: “The problems are solved, not by giving new information, but by arranging what we have always known.”105 A more detailed characterization is given in manuscript 117:

Two Conceptions of Clarification

87

The expression of a philosophical confusion: we don’t know what to say about it. I don’t know how to arrange things, what order to give to the concepts. I don’t know, for example, whether I should e.g. classify proofs as experiments, mathematics as a game, contradictions as confusions. Or whether I should say that the difference between mathematical and experimental truths is that of a degree, whether I should say a new proof gives the proposition a new sense. I don’t know my way about in the human actions, the techniques of using words, mathematical propositions, proofs. When I should describe them, I’m unable to have a perspicuous view of them.106

One can say that arranging and ordering, as Wittgenstein conceives it, aims at making concepts (things or human activities) perspicuous. This is achieved by pointing out interconnections, similarities, and differences between different uses of a word—for example, different uses of “proposition”107—or also between the uses of different words. For example, one might ask with this purpose in mind how mathematical truths relate to empirical truths discovered by way of experiment, whether mathematical proofs are to be understood as demonstrations of the truth of something, or are perhaps better understood as determinations of the sense of mathematical propositions, and so on.108 As an activity of pointing out similarities and differences, arranging and ordering involves an important element of comparison: pointing out analogies and so on.109 Or as Wittgenstein puts it: “The investigation of language is a description and comparing of concepts, also with ad hoc constructed concepts.”110 Similarly, he writes about philosophy as conceptual investigation, contrasting this with natural history: “What is it, however, that a conceptual investigation does? Does it belong in the natural history of human concepts?—Well, natural history, we say, describes plants and beasts. But might it not be that plants had been described in full detail, and then for the first time someone realized the analogies in their structure, analogies which had never been seen before? And so, that he establishes a new order among these descriptions. He says e.g. ‘compare this part, not with this one, but rather with that!’. . . . He is saying ‘Look at it like this’ and that may have advantages and consequences of various kinds.”111

88

Two Conceptions of Clarification

The only example given in this remark (in a part that I have not quoted) of the advantages and consequences of the activity of ordering is the possibility that it might give a new direction to scientific investigation. Nevertheless, it is clear that Wittgenstein regards the activity of ordering as also of philosophical significance. Establishing an order among things (by pointing out analogies, for instance, or by stating a rule) can turn a set of phenomena that is difficult to make sense of— say, the different cases of the use of a word—into an orderly, perspicuous, and comprehensible whole. Similarly, it may be necessary to introduce a new order in the place of an old one if the old one creates philosophical problems, for instance, because it is based on misleading analogies. In this way, establishing an order may allow one to understand what was not understandable before. As an example of this procedure of reordering, consider Wittgenstein’s suggestion that we regard utterances of pain as extensions of primitive, natural behavior such as cries of pain, rather than as statements about inner objects of knowledge.112 As a consequence of reclassifying expressions of pain together with such primitive behaviors, or comparing them with such forms of behavior, problems arising from the traditional conception of first-person utterances of pain as knowledge seem to dissolve. A cry or a moan is not a statement about anything, but an instance of expressive behavior. Through this reconceptualization of the relation of pain to its expression, the gap between them is eliminated and knowledge of others’ pain becomes comprehensible. Now my knowledge of another person’s pain is not seen as involving a problematic claim about an inner object of knowledge to which I have no access. Rather, my knowledge of the other’s pain is based on the perception of her pain through its immediate expression.113 Thus, what figures in the old account as an “inner object” inaccessible to others is reclaimed and made accessible by Wittgenstein’s account with the help of the notion of expressive behavior. For notably, although I may indeed doubt the genuineness of your moaning in individual cases—this uncertainty is characteristic of the “languagegame”114—it is a different thing to say that expressions of pain are therefore always doubtful. This inference presupposes an arguably problematic transition from each (individual) case to every case.115 In the passage quoted above, the activity of ordering leads to a suggestion to look at things in a new way. Wittgenstein similarly describes or-

Two Conceptions of Clarification

89

dering as a matter of introducing new pictures: “Who brings about an order to where there was no order before, introduces a new picture.”116 Accordingly, he regards it as characteristic of philosophy to articulate and propose pictures or ways of looking at things with the purpose of dissolving philosophical problems. He writes: “The philosopher says: ‘Look at it like this—.’ ‘Are you still puzzled by it?’”117 Such a suggestion to look at things in a particular way may, for instance, be employed to release someone from the grip of a misleading picture and a thought-cramp.118 Importantly, however, the goal of philosophy is not, according to Wittgenstein, to establish anything like the only correct order of concepts. He writes in the Investigations: “We want to establish an order in our knowledge of the use of language: an order with a particular end in view; one out of many possible orders; not the order.”119 Discussions earlier in this chapter can be seen as suggesting that, according to Wittgenstein, there may be different philosophical orderings of concepts depending on which of their aspects one is concerned with. In order to solve certain philosophical problems, one might establish an order that highlights certain conceptual connections—for instance, certain analogies and non-analogies between concepts—whereas in the case of other problems, one might concentrate on other features of the concepts in question, establishing a different order. But there need not be an ultimate order, or the order, that brings together all such different orderings. The orders are established, as Wittgenstein says in the passage above, “with a particular end in view,” this being the solving of particular philosophical problems. However, Wittgenstein’s wording in §132 leaves room for different interpretations. According to an interpretation put forward by Baker and Hacker, the different possible orders that Wittgenstein talks about are not different philosophical orders. Rather, Wittgenstein is contrasting his philosophical order with nonphilosophical orders—for instance, with an order established by a philologist. Baker and Hacker write about philosophy’s task of ordering: “The order it seeks is not the only proper one. The grammarian or philologist arranges linguistic data quite differently, for quite different purposes. The order philosophy seeks to establish is guided by the purpose of resolving philosophical, or conceptual problems.”120 Hence, although philosophy does not seek to establish the order of language for all possible purposes, it seeks

90

Two Conceptions of Clarification

to establish the philosophical order of language. Let us examine this suggestion in more detail. Baker and Hacker, of course, are right in saying that what distinguishes Wittgenstein’s ordering of “linguistic data” from that of, say, the philologist’s is its purpose—even if the borderline between these two undertakings might be unclear in some individual cases. Unlike the philologist’s, Wittgenstein’s investigations aim at resolving philosophical problems.121 But that there are different kinds of orderings of language for different purposes does not yet decide the issue of whether there would be one or many specifically philosophical orders. The existence of different kinds of orders is compatible with both possibilities. Importantly, however, when considered against the background of Wittgenstein’s critique of his early thought, problems begin to emerge with Baker and Hacker’s view that there is a certain single, ultimate philosophical order of language. Given that the purpose of establishing a philosophical order among concepts or in the knowledge of language use is to resolve philosophical problems, Baker and Hacker’s interpretation, in effect, amounts to the claim that there is a certain order of language that contains the solution to all philosophical problems in the context of a relevant language, a region of language, or particular concepts.122 On the basis of discussions earlier in this chapter, however, one should immediately be alarmed by how closely reminiscent this is of the Tractatus. Baker and Hacker’s interpretation brings one right back to the idea that there is a certain definite logical order implicit in language, which logical investigation aims to make perspicuous and the comprehension of which leads to the dissolution of all philosophical problems. But then Baker and Hacker’s interpretation appears to be in the target area of Wittgenstein’s critique of his early work. Wittgenstein writes in the Tractatus about the notion of an implicit logical order of language and the task of determining it: “All propositions of our colloquial language are actually, just as they are, logically completely in order. That simple thing which we ought to give here is not the model of the truth but the complete truth in itself.”123 More specifically, as explained in 2.1, the Tractatus’s aim was not to state truths about logic or to represent the logic of language in propositions. Rather, it was to allow the logical order of language to reveal itself through the construction and employment of a notation governed by logical syntax, in which logic was not obscured by misleading sur-

Two Conceptions of Clarification

91

face grammatical forms. Now, compare this with what Wittgenstein says in the early 1930s about the notion of a logical order of language and about the remark just quoted: “Was it not a mistake from me (for that is how it seems to me now) to assume that anyone who uses language plays always a definite game? For was that not the meaning of my remark that everything in a proposition ‘is in order’—however casually it is expressed? . . . But there is nothing either in order or disorder in that—it would be in order if one could say: this person too plays a game according to a definite fixed set of rules.”124 A definite set of rules in the relevant sense is one through which all possible uses of a concept, for example, are determined. Rather than leaving anything open, such a set of rules provides a once-and-for-all definition or a complete grammar of the concept in the sense discussed in 2.21, charting all logical relations into which the concept may enter.125 Here it is then crucial that although there may be important differences between Baker and Hacker’s view of Wittgenstein’s later conception of the philosophical order of language and that of the Tractatus,126 insofar as the purpose of the philosophical order envisaged by Baker and Hacker is to provide the solution to all philosophical problems or unclarities relating to a concept or a set of concepts, this order clearly must amount to a complete grammar of the concept or concepts in the relevant sense. Problems with the notion of such a complete grammar were already discussed in connection with the Tractatus’s notion of the completeness of analysis. How they reemerge in the context of Baker and Hacker’s reading may be explained as follows. Given that new philosophical problems may always arise, the totality made up of such problems is indeterminate not only in practice, but in principle. This, however, means that the philosophical order of language postulated by Baker and Hacker is unspecifiable, too. As long as no determinate sense is given to the locution “all philosophical problems,” no criteria have been determined for what counts as establishing the ultimate philosophical order of language that solves them. Hence, just like the Tractatus’s complete analyses of concepts, the ultimate philosophical order postulated by Baker and Hacker is an unattainable ideal. By contrast, as I argue in Chapter 3, it is central to Wittgenstein’s later outlook and his approach to philosophy that, rather than postulating a structure of rules of language use which constitutes philosophy’s object of description, he conceives rules as a means of description and the clarification of particular philosophical problems.127 This

92

Two Conceptions of Clarification

methodological move is motivated, on the one hand, by problems relating to the concept of necessity that arise when it is assumed that rules of language constitute philosophy’s object of description. On the other hand, as Wittgenstein comes to realize, the logic or the possible uses of language are too complex to be captured in any particular set of rules or a formal notation such as the Tractatus’s concept-script. As I explain in Chapter 3, here one is threatened by a fall into dogmatism, and this threat leads Wittgenstein to rethink the issue of the role of rules in philosophical clarification. In postulating an underlying order of language that philosophy should describe, Baker and Hacker therefore fail to take into account what is arguably a key aspect of the development of Wittgenstein’s conception of language and philosophy. In effect, they attribute to him a conception he abandoned in the early 1930s. Another problem with Baker and Hacker’s interpretation is that from its point of view, Wittgenstein’s introduction of his method of the description of language emerges effectively as a renewed version of the Tractatus’s claim that philosophical problems have been solved “in essentials.” For insofar as there is an order implicit in language that contains the solution to philosophical problems, and a method has been established for rendering this order perspicuous, then apparently all philosophical problems are already settled in principle. What remains is to work out the details. Thus Baker and Hacker’s interpretation seems to involve a return to the kind of great programmatic claim the Tractatus makes about philosophy and its method, now attributed to the later Wittgenstein. But as I explain in 3.1, this programmatic aspect of the Tractatus is ultimately the source of its failure, i.e., its relapse into philosophical theses.128 These problems with Baker and Hacker’s interpretation suggest that it does not correctly capture Wittgenstein’s notion of establishing an order of language. His view is not that there is just one philosophical ordering of the knowledge of language use, but many such orderings. But leaving aside for the moment the rather complex philosophical questions touched upon here, let us next examine textual evidence for this alternative reading. In the quotation from manuscript 117 in the beginning of this subchapter, Wittgenstein mentions as an example of a philosophical

Two Conceptions of Clarification

93

unclarity the problem of whether mathematics should be classified among games or characterized as a game. Insofar as his purpose is to establish a definite conceptual order, one should apparently expect him to try to give a definite answer to this problem. Wittgenstein, however, rejects this aim in his lectures on the foundations of mathematics: “The thing is not to take sides, but to investigate. It is sometimes useful to compare mathematics to a game and sometimes misleading.”129 His approach is similar in the case of whether a proof gives a new sense to a proposition. According to Wittgenstein, although one may sometimes characterize proofs in this way, the characterization does not fit all cases. And as he emphasizes, the important thing is to avoid dogmatism about this issue.130 Although I cannot discuss Wittgenstein’s philosophy of mathematics here at any length, this suggests that the purpose of his comparison between mathematics and games is not at all to establish the conceptual order and to provide a once-and-for-all answer to the question, what is mathematics? i.e., to determine the essence of mathematics or what all cases falling under the concept of mathematics must be. Rather, the purpose is to highlight a particular aspect of mathematics. The same applies, mutatis mutandis, to the concept of a mathematical proof. Correspondingly, Wittgenstein’s conception of the utterances of pain as instances of expressive behavior rather than statements about inner objects of knowledge is by no means meant to hold for all utterances of pain. There are also descriptions of pain in the first-person case, as he is well aware, and the conception of utterances of pain as an extension of primitive behavior does not apply to such cases. (A cry of pain is not a description.)131 The conceptual order established with the help of the comparison between utterances of pain and primitive expressions, therefore, is not meant as the order that allegedly determines how one must always think about the expressions of pain. Instead, it is a possible way of ordering the concepts designed, more specifically, to dissolve philosophical problems that arise when talk about sensations is understood on the model of knowledge claims. In his later work on Wittgenstein, Baker abandons the interpretation he propounded earlier with Hacker. Here he seems exactly right when he writes about Wittgenstein’s descriptions of grammar: “A comparison which is illuminating for one purpose may be unhelpful or even obfuscating in another context, and a pair of seemingly inconsistent

94

Two Conceptions of Clarification

analogies may facilitate grasping different aspects of a single thing.”132 And as Baker also notes in this context, such characterizations with analogies and comparisons should not be assumed to be “additive.” This ought to be readily understandable against the background of the preceding discussions in this chapter. There is no reason to think that descriptions of language in Wittgenstein’s sense should be additive, insofar as they serve the purpose of clarifying particular problems. Arguably, this is also the case with the orderings of language use. They do not add up to a Great Order of Language, but serve to dissolve particular problems. There is also more direct textual evidence in support of the view that the contrast Wittgenstein draws in §132 is between different philosophical orders rather than between philosophical and nonphilosophical orders. Consider the 1937 version of the same remark: “As our aim is to break the bewitchment in which certain forms of language hold us, we want to establish an order in our knowledge of the use of language, which makes this possible. I.e. an order with a particular end in view; one out of many possible orders.”133 According to this formulation, the purpose of the ordering is to break the bewitchment that certain forms of language exert over us. But to say that the purpose is to defeat bewitchment by certain—rather than by some other—forms of language is to say that the purpose of ordering is to dissolve certain specific philosophical problems whose roots are in specific misunderstandings concerning the forms of language. The order, then, is one out of many possible ones in the sense that it is designed to make possible the dissolution of these particular problems. Correspondingly, it is plausible that Wittgenstein’s contrast in §132 is a contrast between different orderings for philosophical purposes, not between the philosophical order and different kinds of nonphilosophical orders. Interestingly, in the Investigations Wittgenstein also makes the following comment on classifying the expressions of a particular languagegame into different kinds: “how we group words into kinds will depend on the aim of the classification . . . . Think of the different points of view from which one can classify tools or chess-men.”134 Apparently, this point about classification applies quite generally. How one should order concepts depends on the purposes the order serves. Thus, what is an appropriate philosophical order of concepts depends on the unclarities one is trying to clarify and the problems one is trying to resolve.

Two Conceptions of Clarification

95

This seems to settle the dispute concerning Wittgenstein’s notion of ordering. Consequently, we also have an answer to the question raised in 1.4 as to whether by achieving a “cramp-free position” Wittgenstein means some specific state, a particular regrouping of the expressions of language, or only “freedom from thought-cramps.” Insofar as his aim is not to establish the unique philosophical order of language, the “crampless position” cannot be identified with any particular ordering of our knowledge of language use. Philosophical therapy for Wittgenstein does not set out to establish any specific order for language use and thought but strives to release thought-cramps. By doing this it aims to enable one’s thoughts to move more freely and allow one to “philosophize well.”135 Next, we need to examine more closely Wittgenstein’s conception of philosophical descriptions of language in order to get a clearer grasp of his shift away from great problems and philosophical theories and theses. In the following chapter, I also seek to define the notion of a philosophical theory or thesis more precisely by discussing the concept of metaphysics or metaphysical philosophy, which is a mode of philosophizing that aspires to establish philosophical theses in the relevant sense, and with which Wittgenstein explicitly contrasts his approach. One key problem with metaphysical philosophy for him is the danger of dogmatism. Accordingly, his later approach to philosophy constitutes a strategy for avoiding dogmatism in philosophy.

THREE

From Metaphysics and Philosophical Theses to Grammar: Wittgenstein’s Turn

in a notebook from 1938: “The greatest danger that threatens the mind in philosophizing comes from the metaphysical tendency that takes over it and completely topples the grammatical.”1 Evidently he wishes to distinguish his approach from metaphysical philosophy, but the question is, how exactly is this distinction to be understood? The Investigations employ the phrase “turning our whole examination around” to characterize a methodological shift in Wittgenstein’s philosophy. I argue that this shift, Wittgenstein’s turn, marks his break away from metaphysical philosophy and explains what he means by not having theses, doctrines, or theories in his later philosophy. The turn may also be characterized as a passage from his early philosophy to his later philosophy, that is, to philosophy as conceptual or grammatical investigation.2 Accordingly, an examination of Wittgenstein’s turn can provide us with a more precise account of the shift outlined in earlier chapters from great problems to particular problems. I begin by discussing the Tractatus’s attempt to disengage itself from metaphysical philosophy and its relapse into metaphysics and philosophical theses (3.1). I then proceed through what Wittgenstein says about metaphysical philosophy and his analysis of the Tractatus’s failure (3.2) to an examination of the notion of conceptual investigation as the description of language use. Having discussed the problem of dogmatism in connection with such descriptions, and how this problem emerges in the context of Baker and Hacker’s interpretation of the later Wittgenstein (3.3), I examine the idea of the turn first in more

WITTGENSTEIN WRITES

From Metaphysics and Philosophical Theses to Grammar

97

general terms (3.4) and then specifically with regard to the role of rules in philosophical clarification (3.5 and 3.6). Finally, I address (in 3.7) some possible misunderstandings relating to the interpretation developed in this chapter, among these a question about the notion of the logic of language. In this chapter, Wittgenstein’s turn is described as a response to the problem of dogmatism in philosophy as well as to problems relating to the concept of necessity. As explained in Chapter 6, the turn also leads to the dissolution of philosophical (conceptual) hierarchies.

3.1 Philosophical Theses, Metaphysical Philosophy, and the Tractatus A philosophical thesis in the traditional sense is a thesis concerning an essence. Just as traditionally, essences have been the concern of metaphysical philosophy or metaphysics. Therefore, it seems possible to clarify the notion of a philosophical thesis by examining the concept of metaphysical philosophy. I discuss here two characteristics of metaphysical philosophy and how they manifest themselves in connection with the Tractatus’s program for clarification. My discussion of these two characteristics, however, is not meant to provide a definition of the concept of metaphysics valid for every possible case in which we might classify something as an instance of metaphysical philosophy. The characterization I offer is not meant as exclusive in this sense. Rather, I aim simply to draw attention to certain characteristics that seem central to the metaphysical mode of thinking, while leaving open the possibility that there might be other grounds for describing someone’s thinking as metaphysical. (The characterization offered here is a definition to be used as an object of comparison in a sense to be explained later.) Metaphysical philosophy can be characterized as the pursuit of knowledge or understanding of necessary truths or principles that govern what there is, or rather, what there could be.3 Metaphysics, that is to say, is not just an inventory of empirically, contingently existing things. It puts forward theses—doctrines or theories4—about the essence or nature of its objects of investigation. Such theses are intended to capture the necessary characteristics of these objects in contrast to what is accidental to them. Metaphysical theses, that is to say, state what something must be in order to be (or count as) whatever it is. Put in another way, such theses bring to

98

From Metaphysics and Philosophical Theses to Grammar

view what things really are in contrast to what they might happen to be, or—from the point of view of epistemological concerns—might appear to be. As regards their generality, given that metaphysical theses concern characteristics necessary to the objects falling under a concept, all objects falling under that concept must have the relevant characteristics. Hence metaphysical theses, if correct, are universally valid for those objects. (If y is a necessary characteristic of an x, then anything, in order for it to count as an x, must have the characteristic y.) Although the concepts of universality and generality are not identical,5 one might say that a metaphysical thesis applies to objects falling under a relevant concept with complete generality. The generality of such a thesis is qualitatively different from that of an empirical generalization. As for the Tractatus, as explained in 2.1, it might be understood as an attempt to move away from statements concerning essences and in this sense from philosophical or metaphysical theses. Instead, the necessary is to be elucidated in terms of the concept-script. Making plain the logic of expressions, this notation brings to the fore what is necessary and possible, thereby also excluding statements about the necessary and the possible.6 Consequently, the concept-script would make clear the logical distinction between the elucidation of what is essential on the one hand and factual statements on the other. Any attempt to make a statement about the essential would be revealed as nonsensical through logical analysis, that is, translation into the concept-script.7 The motives of Wittgenstein’s attempt to clarify the distinction between the essential and the factual are evident. He is concerned with dissolving “the confusion, very widely spread among philosophers, between internal relations and (proper) external relations,” where an internal relation is one that it is inconceivable for an object not to possess.8 At the level of logic, the purpose is to get rid of “the confusion of formal concepts with proper concepts which runs through the whole of the old logic.”9 Similarly, in a letter to Russell after the completion of the Tractatus manuscript, Wittgenstein characterizes the distinction between the factual and the necessary—or, expressed in the Tractatus’s transitional vocabulary, the distinction between what a proposition says and what it shows—as his “main point” and “the cardinal problem of philosophy.”10 Wittgenstein’s distinction implies a shift in the methodology of philosophy. His idea that a concern with the essential is not a matter of making true/false statements about anything suggests an elemental difference

From Metaphysics and Philosophical Theses to Grammar

99

between philosophy and the sciences. Philosophy with its statements about essences is not a superscience that reaches an even higher level of generality than other sciences. Nor is it a more fundamental science that investigates what other sciences presuppose. Rather, philosophy is not engaged in theoretical assertion at all. Its task consists solely in the clarification or clear expression of what is said.11 But although the Tractatus’s distinction between a concern with truths and a concern with essences seems to make obsolete the central metaphysical notion of necessary truth, to draw this distinction as the Tractatus seeks to do arguably does not suffice to constitute a departure from metaphysical philosophy and philosophical theses. The problem pertains, in the first instance, to what Wittgenstein thinks it means to introduce a philosophical method in general (what sort of a claim this involves), and consequently to the Tractatus’s conception of philosophy as logical analysis. More specifically, the Tractatus aspires to put forward a program for philosophy as logical analysis that is universally applicable to all philosophical problems. Any philosophical problem, Wittgenstein maintains, can be dissolved by means of the method introduced in the Tractatus—if it has not already been dissolved just by setting up this method. Accordingly, Wittgenstein seems to be in a position to claim that the Tractatus has solved all philosophical problems “in essentials.”12 In effect, however, these methodological views constitute a thesis about the essence of philosophy, that is, about how one must philosophize or about the correct method of philosophy. That is, any method that can do the same job as the Tractatus’s would apparently count as equivalent to it. With respect to their clarificatory power, such methods would be essentially identical, differing only in their accidental characteristics. But if so, then Wittgenstein’s method is not just one among others, that is, a way in which one can solve philosophical problems. His method is how one must philosophize. This is the first sense in which the Tractatus puts forward a philosophical thesis. Moreover, by presenting the Tractatus’s scheme of analysis or concept-script as one that can be applied to any philosophical problem whatsoever, Wittgenstein commits himself to a particular metaphysics of language. For insofar as he maintains that (1) there is only one complete analysis of a proposition that terminates in simple names, that (2) all propositions can be analyzed in this way, and that (3) all logical confusions can be clarified through such an analysis, he

100

From Metaphysics and Philosophical Theses to Grammar

thereby commits himself to a thesis about the nature of propositions. According to this thesis, all propositions are analyzable into concatenations of simple names without any kind of “logical remainder,” i.e., distinctions that would not be captured and could therefore cause confusion. In this sense every possible proposition is expected to fit the Tractatus’s model for analysis, and the model is assumed to be universally applicable to propositions. One might therefore say that the Tractatus has a metaphysics of language built into its conception of the method of logical analysis. Analysis, as conceived in the Tractatus, makes use of a particular notion of the general form of a proposition, assuming as its modus operandi that every proposition has this form, i.e., is a (re)presentation of a state of affairs.13 This means that even if all formulations of the Tractatus’s “theory of language” are nonsense, insofar as the purpose of the book is to introduce a universally applicable scheme of analysis for philosophy as clarification, the metaphysics of language is inevitably there too. It is embodied in the activity of clarification as the form of this activity.14 Consequently, the Tractatus’s attempt to jettison metaphysical doctrines is unsuccessful. Even though metaphysics seemingly disappears as a body of statements, it finds a refuge in methodology, as a thesis about the correct method of philosophy and as an implicit assumption about the essence of propositions and language. To fully appreciate the Tractatus’s metaphysical commitments, however, it is important not to construe Wittgenstein’s commitment to a thesis about language too narrowly, i.e., to take the problem to be his adoption of a certain very special thesis about language. One might say more generally about the type of approach the Tractatus promotes that even if one abandoned the Tractatus’s concept-script for another one, insofar as one nevertheless continued to maintain that there is some such notation that “excludes all logical mistakes” and that philosophy must become logical analysis in terms of some particular notation of this kind, one would still remain committed to a thesis about the nature of philosophy and language.15 For even if one thus abstracts from Wittgenstein’s commitment to the conception of propositions as pictures (and so on), language continues to be seen as possessing an essence that can be captured once and for all in some logical notation or another that shows what meaningful expressions must be. That is, regardless of which alternative regimentation of language one substitutes for the

From Metaphysics and Philosophical Theses to Grammar

101

Tractatus’s, this would still count as a commitment to a thesis about the essence of language.16 The notion of the universal applicability of the Tractatus’s method brings us to the second characteristic of metaphysical philosophy: the construction of philosophical hierarchies. As noted in 1.5, this line of thought is exemplified by the Aristotelian idea that since everything that there is, is, philosophy could grasp everything insofar as it could grasp the nature of being, which now emerges as the fundamental question. In the Tractatus, the question concerning the method of philosophy, and therefore also the question of the essence of a proposition, plays a similar role. Given Wittgenstein’s view of the correct method of philosophy as a method of logical analysis, and given that developing such a method, i.e., a scheme for the analysis of language, seems to require grasping the essence of language, the determination of the essence of language becomes the fundamental task. Thus the problem of the essence of a proposition assumes the role of a fundamental problem upon which the solution to all philosophical questions depends insofar as they are to be resolved by logical analysis.17 The way in which Wittgenstein’s requirement that his method of analysis should be universally applicable to any logical unclarity leads him to the creation of a philosophical hierarchy can be explained as follows. The method’s universal applicability, of course, is not guaranteed by its success in solving any particular, actual problems. This would still leave open the possibility that the method might not be applicable in further cases that may arise. Consequently, the task of establishing the method assumes the form of determining the essence of any meaningful linguistic expression once and for all. For assuming that there are certain essential features shared by all meaningful linguistic expressions, insofar as the method of analysis can be known to focus on such essential features, then its applicability to any meaningful expression and to any unclarity seems ascertained. In this way a division is created between problems regarding the applicability of the method in general and problems to be solved by applying the method, whereby the former emerge as more fundamental in that their solution seems to be a requirement for the justified adoption of the method of analysis. Accordingly, solving problems relating to the applicability of method have been solved, other philosophical problems are thereby resolved in principle or “in essentials.”18

102

From Metaphysics and Philosophical Theses to Grammar

In the Tractatus, this division between more and less fundamental issues comes to light in Wittgenstein’s distinction between a “strictly correct method” that consists in the application of the scheme of analysis developed in the Tractatus and the not strictly correct method that the book itself employs to introduce the method of analysis.19 As is well known, the Tractatus itself offers no analyses in terms of the method of analysis it articulates, leaving the method’s application to others. Instead, the Tractatus is concerned with the question of methodlanguage, and Wittgenstein devotes himself to matters concerning the foundations of philosophical or logical analysis—including the question of the essence of language.20 Accordingly, the Tractatus’s conception of logical analysis might be characterized as ultimately resting on a theoretical foundation in the sense that this conception of analysis involves a determination of the essence of language arrived at independently of any actual applications of the method of analysis itself. Rather, this determination is supposed to be able to ascertain the legitimacy of analyses to be undertaken in the future, thus anticipating and fixing the form of all possible analyses. Despite its aspiration to a philosophy without philosophical theses and doctrines, and a highly interesting attempt at one, the Tractatus is thus not free from philosophical theses and theoretical presuppositions. Although it may appear at first sight to have discovered how to avoid putting forward metaphysical theses as necessary truths about philosophy’s objects of investigation, the philosophical program proposed by the Tractatus involves a commitment to a thesis about philosophy and the nature of language. There is a metaphysical residue in the Tractatus’s conception of philosophy as clarification. Let us now turn to Wittgenstein’s later philosophy to see how it seeks to resolve the problem of the Tractatus’s relapse into theses. I begin with Wittgenstein’s later conception of metaphysical philosophy and the confusion that, according to him, characterizes metaphysical thinking.

3.2 Metaphysics and Conceptual Investigation: The Problem with Metaphysics In his later philosophy, Wittgenstein continues to identify the problem with metaphysics as an unclarity about the status of philosophical statements and about what kind of investigation a philosophical investigation

From Metaphysics and Philosophical Theses to Grammar

103

is. In this respect, he preserves the Tractatus’s distinction between philosophy and the sciences.21 Whereas metaphysics is conceived as an investigation of necessary truths pertaining to philosophy’s objects of investigation, and therefore as a factual or “superfactual” investigation, philosophy for Wittgenstein is a conceptual investigation. He writes: “Philosophical investigations: conceptual investigations. The essential thing about metaphysics: that the difference between factual and conceptual investigation is not clear to it. A metaphysical question is always in appearance a factual one, although the problem is a conceptual one.”22 Wittgenstein characterizes metaphysical assertions as follows: “we feel that we have said something about the nature of pain when we say that one person can’t have another person’s pain. . . . as though it would be not false but nonsense to say ‘I feel his pains,’ but as though this were because of the nature of pain, of the person etc. as though, therefore, this statement were ultimately a statement about the nature of things.”23 Thus, even though he describes metaphysical statements as having the appearance of factual ones, he is not portraying them as empirical statements. By making a statement about the essence of pain—“I cannot feel another person’s pain”—I am not saying that perhaps tomorrow I will be able to feel it. It is not that this statement is contingently true now but might be false later. Rather, it seems inconceivable that I should feel somebody else’s pain. And this is taken to be so because of the nature of pain. The statement “I cannot feel another person’s pain,” therefore, seems to express a necessary truth. According to Wittgenstein, however, rather than describing necessary features of our objects of investigation, the “musts” and “cans” of metaphysical statements indicate rules for the use of expressions. Even though the metaphysician thinks she is making a statement about the nature of things, she is better understood as putting forward a rule for the use of language disguised as a factual statement. As Wittgenstein says, summing up a discussion in the Blue Book: “What we did in these discussions was what we always do when we meet the word ‘can’ in a metaphysical proposition. We show that this proposition hides a grammatical rule.” And: “The man who says ‘only my pains are real’ . . . objects to using the word in the particular way in which it is commonly used. On the other hand, he is not aware that he is objecting to a convention.”24 So although it might appear that the metaphysician has made a discovery about pain, she is in effect making a suggestion concerning language

104

From Metaphysics and Philosophical Theses to Grammar

use—as her rule denying the reality of other’s pain incidentally deviates from the common use of the word “pain.” This contrast between factual statements and expressions of a rule can be further elucidated with the help of what Wittgenstein says about the notion of a priori propositions, which he treats as examples of metaphysical statements:25 “The avowal of adherence to a form of expression, if it is formulated in the guise of a proposition dealing with objects (instead of signs) must be ‘a priori.’ For its opposite will really be unthinkable, inasmuch as there corresponds to it a form of thought, a form of expression that we have excluded.”26 As above, the remark concerns a statement expressing a linguistic commitment (“an avowal of adherence to a form of expression”) disguised as a proposition about objects. This conflation of two kinds of statement results in what looks like a statement about objects, but one with which no experience can conflict. As the possibility of things being otherwise is excluded, the statement is clearly not a contingent, empirical one. Hence one may feel tempted to conclude that the statement is “a priori”. It concerns something known to be true independently of the contingent facts of which experience informs us, expressing a necessary truth that cannot be in conflict with reality.27 But again, rather than being due to a necessity in (our or the objects’) nature, the exclusion of the possibility of things being otherwise, Wittgenstein suggests, is due to a form of expression that the speaker assumes and that excludes other possibilities from consideration. The apparent “a priori” statement is not a statement about anything but expresses commitment to a form of expression and thought, a rule of language. Consequently, one might sum up the metaphysician’s confusion by saying that she projects a way of using language onto reality, as exemplified by the “a priori” proposition. Projected onto reality, the necessities characteristic of this way of using language seem as if they were necessary features of reality. The metaphysician, therefore, is caught up in an illusion, mistaking the reflection of her concepts on reality for a truth about reality. Or as Wittgenstein says: “The ‘order of things’ the idea of the form(s) of imagination/phenomena [Vorstellung], that is, of the a priori is a grammatical illusion itself.”28 But given the confusion that, according to Wittgenstein, underlies metaphysics, how should one understand the notion of conceptual investigation or Wittgenstein’s shift from metaphysics to conceptual in-

From Metaphysics and Philosophical Theses to Grammar

105

vestigation and the difference between these two modes of thinking? Certainly the shift to conceptual investigation cannot be merely a matter of turning to investigate language instead of reality. Wittgenstein’s distinction between metaphysics and conceptual investigation is not based on a claim about the proper object of philosophical investigation. Although the Tractatus too regards language as its object of investigation, this does not suffice to divorce it from metaphysics. Rather, as an object of investigation language is at the same level as any other part of reality, and as likely a target of metaphysical projections as anything else. This is evident in the fact that the Tractatus is the later Wittgenstein’s primary example of the confusion of metaphysics. At this point, it is instructive to look more closely at Wittgenstein’s remarks on his relapse into a metaphysical use of words in the Tractatus and to try to apprehend the different aspects of this relapse as he describes it. Wittgenstein writes in the 1937 typescript of the Investigations: “Every proposition says: This is how things stand.” Here we have the kind of form that can mislead us. (Misled me.) . . . This is the kind of proposition that one repeats to oneself countless times. One thinks that one is tracing the outline of the thing’s nature over and over again, and one is merely tracing round the frame through which we look at it. . . . Again and again we trace out the form of expression and think we have depicted the thing.—Due to an optical illusion we appear to see within the thing what is marked on our spectacles. . . . Only when this illusion has been removed can we simply see language, as it is. The expression of this confusion is the metaphysical use of our words. For now we predicate of the thing what lies in our mode of presenting it. Impressed by the possibility of a comparison, we think we are perceiving a state of affairs of the highest generality.29

What Wittgenstein describes in these passages is an example of a philosopher projecting a mode of presentation (form of expression) that she is employing onto the object of investigation and mistaking the characteristics of the mode of presentation for the characteristics of the object. Thus in the Tractatus, a particular conception of propositions was turned into a metaphysical thesis about what propositions must be. Let us look at this more closely.

106

From Metaphysics and Philosophical Theses to Grammar

For Wittgenstein’s above characterization to apply to the Tractatus, one need not maintain that the book intended to put forward a thesis about propositions. Rather, the Tractatus may be taken to fall in the target area of the above description because it assumes that the formula “Every proposition says: This is how things stand” spells out a universally applicable scheme for the analysis of propositions. For although the proposal that we adopt a notation or a particular scheme of analysis does not yet constitute a thesis about anything, the claim that this scheme of analysis is universally applicable to any proposition does constitute, in effect, an implicit thesis about propositions.30 The following remark also suggests that Wittgenstein did not think he was putting forward a theory, but this was what actually happened: “We have a theory . . . of the proposition; of language, but it does not seem to us a theory. For it is characteristic of such a theory that it looks at a special, clearly intuitive case and says: ‘That shows how things are in every case. This case is the exemplar of all cases.’—‘Of course! It has to be like that’ we say, and are satisfied. We have arrived at a form of expression that enlightens us. . . .”31 This remark outlines the anatomy of the Tractatus’s mistake, so to speak, both its phenomenology and the logic driving it. At the former level, it did not seem to Wittgenstein that he was putting forward a theory or a thesis about anything. Rather, he thought that he had caught a glimpse of something: that he could perceive the essence of propositions in the examples he was contemplating. Or, as he says in the first long quotation from typescript 220: “we think we are perceiving a state of affairs of the highest generality.” It is crucial that he was not aware of entertaining a particular conception, presenting things in a particular way, or employing a mode of presentation. It is characteristic of the mistake that the involvement of a mode of presentation completely escapes one’s attention. One does not notice that the “enlightening example” assumes the role of a mode of presentation functioning as a model for the things it exemplifies. Instead, the situation assumes the appearance that something is directly perceived, as if one simply saw in the example the innermost essence of the things it exemplifies and did not use the example as a mode of presentation. As Wittgenstein says above, employing the metaphor of glasses to signify a form of presentation: “we appear to see within the thing what is marked on our spectacles.” And in the 1936 manuscript of the Investigations: “The idea . . . is a form of presentation, but we

From Metaphysics and Philosophical Theses to Grammar

107

have moved far away from recognising it as a form of presentation. It is like a pair of glasses on our nose through which we see everything we look at.”32 But behind this phenomenology is a particular “logic of thought” that is the driving force of the mistake. The error is the result of certain unexamined presuppositions of thinking. According to Wittgenstein, he lost sight of the “enlightening example” functioning as a mode of presentation because of a particular conception of concepts that he assumed. According to this conception, the unity of a concept is determined by the presence of a common feature or features that all objects falling under the concept must share. This problematic assumption then prevented Wittgenstein from recognizing that the Tractatus’s conception of propositions as pictures was a comparison rather than an insight into the nature of all propositions that he could read out of his examples. He writes: “I had used a comparison; but through the grammatical illusion that a certain one thing, something common to all its objects, corresponds to a concept it did not seem like a comparison.”33 The point is that against the background of the assumed conception of the unity of concepts, it seems one may perfectly legitimately take the enlightening example to show what all individual instances of propositions are. It is not merely a particular case that other cases might be compared with and that could be used as a mode of presentation in this sense, i.e., as a model for the other cases. Rather, the example really seems to bring to light the common essence, what all cases really are. The point is put as follows in the continuation of the passage from typescript 220 §93 quoted above: “The tendency to generalize the case seems to have a strict justification in logic: here one seems completely justified in inferring: ‘If one proposition is a picture, then any proposition must be a picture, for they must all be of the same nature.’ For we are under the illusion that what is sublime, essential about our investigation consists in grasping one comprehensive essence.”34 As Wittgenstein says, the above assumption concerning the unity of concepts constitutes an illusion about philosophy or logic, making it seem that their nature lies in grasping such unified comprehensive essences.35 Indeed, as the above diagnosis of the Tractatus’s mistake shows, the conception of concepts as unified through universally shared characteristics is a crucial component of the idea that philosophy’s task is to provide once-and-for-all determinations of concepts or

108

From Metaphysics and Philosophical Theses to Grammar

comprehensive theses about essences, i.e., great philosophical answers. It is against the background of this conception of the unity of concepts that philosophy’s search for once-and-for-all accounts of concepts seems possible and desirable. In this sense, what Wittgenstein finds problematic in his early work, and in metaphysical philosophy more generally, can be traced to a remarkable extent to this conception of the unity of concepts. This conception facilitates the metaphysical projection by making it look unproblematic to use examples to exhibit what all cases falling under a concept must be. Consequently, it becomes hard or impossible to notice that anything is being projected onto the objects of investigation. Now we are in a position to see how this problematic view of concepts, which forms the background for the Tractatus’s program for philosophy, causes it to collapse, even if the Tractatus avoids problems arising with interpretations that attribute to it a “nonsensical theory.” Let us therefore assume that an interpretation along the lines suggested in 2.1 is able to release the Tractatus from the paradox of the nonsensical theory.36 Regardless, a fatal problem remains: there is still a gap between what the examples deployed by the Tractatus in introducing its notation can illustrate and the grounding of a universal program allegedly applicable to any proposition whatsoever. Observations concerning particular examples cannot guarantee the applicability of the Tractatus’s scheme for logical analysis to all propositions. Hence they cannot justify Wittgenstein’s claim of the universal applicability of his method. The justification of this claim would require the assumption that “all propositions are the same by their nature” and that their common essence comes to view in the Tractatus’s examples. Granted this assumption, the Tractatus would apparently be on safe ground. But this assumption is precisely what Wittgenstein later problematizes, bringing into focus the problem of the justification of the Tractatus’s program for philosophy. Ultimately, the problem with the Tractatus’s attempt to establish its method with the help of examples, therefore, lies in its use of those examples, or the significance it attaches to them. Let us approach the problem with the Tractatus from another angle. It is not that the Tractatus’s conception of propositions as pictures is inherently problematic. It may be illuminating for certain purposes to characterize propositions as pictures or (re)presentations of states of affairs,

From Metaphysics and Philosophical Theses to Grammar

109

to define or explain the meaning of “proposition” in this way.37 This possibility, however, is eclipsed when the conception is not recognized as a particular way of conceiving propositions or a particular mode of presentation but is turned into a thesis about their essence.38 Conceived as a truth about the philosopher’s object of investigation, the conception must—on pain of being reduced to a mere empirical statement about some propositions—be taken as valid for every possible proposition. The projection of the philosopher’s mode of presentation onto the object of investigation gives rise to dogmatism. When the mode of presentation is (mis)understood as a truth, the philosopher is, as it were, driven by the dynamics of the situation to state dogmatically that the characterization must hold for all propositions. Thus what is a defining feature of her mode of presentation is claimed to be a necessary feature of her object of investigation. Wittgenstein writes: “now everything which holds of the model will be asserted of the object of the examination; & asserted: it must always be. . . . This is the origin of a kind of dogmatism.”39 I will return to this passage in greater detail in 3.4. As regards my present point, however, the conception of propositions as pictures may be taken as an example of the kind of model referred to in the first sentence of the quotation. The problem of dogmatism then arises when one claims that every proposition must fit the model of propositions as pictures, that for anything to count as a proposition it must be a picture of a state of affairs. This way one comes to impose a dogma upon the phenomena, i.e., onto propositions as one’s object of investigation. No one, of course, is in a position to justify such a dogma by actually examining every possible proposition and showing that they do indeed correspond to it. Rather, one comes to accept the correctness of the model—so it seems—on the basis of certain examples and the model’s ability to clarify some unclarities relating to the concept of the proposition.40 But however illuminating the model might be with respect to such particular problems, there may be others it cannot resolve. Arguably, the model of propositions as pictures does not fit all propositions equally well and cannot do equally good work in the case of every philosophical problem relating to the concept of the proposition. Indeed, when turned into a dogma that every proposition must allegedly fit, the model may come to positively prevent one from finding a satisfactory account of other examples of propositions. Thus, the idea of what propositions must be may come to cramp one’s understanding,

110

From Metaphysics and Philosophical Theses to Grammar

holding one’s mind captive in the way described in 1.3. Moreover, such difficulties may throw the model in doubt in general, undermining the already achieved results. For insofar as the model is supposed to capture something that holds universally for every proposition, anomalous cases presented as counterexamples apparently show that it does not correctly capture a universal characteristic or characteristics. Perhaps, it is then reasoned, the model’s success in explaining whatever it was able to explain was merely accidental. Hence, turning the model into a dogma may erode the original reasons for its acceptance. Not only is a dogma, therefore, unjustifiable and potentially unjust; it is inherently unstable. The assumption of a common essence is also the basis of what Wittgenstein calls “the sublimation of logic,” i.e., the idea that philosophy or logic is not concerned with signs in an ordinary sense but with something more abstract and pure that lies behind ordinary signs.41 For the ordinary signs do not quite seem to satisfy the philosopher’s idea of what they must be. Not every proposition appears to be a presentation of a state of affairs, for instance. But if this is their essence, then apparently there must be a deeper level at which they can be revealed to be such.42 This leads one to the (in)famous philosophical postulation of a realm of the “really real” behind the veil of appearances that everyday thought (wrongly) assumes to be real. (An example is the sublimation of our concepts through the assumption of the Platonic forms.) With the Tractatus as his example, Wittgenstein writes: “When we believe that we must find that order, the ideal in actual language, we are easily led to speak of the ‘real’ sign (sentence or word), to look for the real sign, so to speak, behind what is called that in customary language use. For we aspire after something more pure than the sign in the sense of a written or printed word etc. We are in search of a sublime essence.”43 In the Tractatus, he was led to think that there are clear and precise rules or an ideal “crystalline logic” to be revealed behind the ordinary signs through logical analysis. The reasoning can be briefly described as follows: Insofar as every proposition is to be understood as a true/false (re)presentation of a state of affairs, it seems that propositions must have a definite sense. Otherwise they do not have a definite truth-value. And if everyday propositions on the surface level do not seem to have a definite sense, then apparently it must be possible to re-

From Metaphysics and Philosophical Theses to Grammar

111

veal such a sense through analysis. Thus the conception of propositions as (re)presentations of states of affairs, when turned into a dogma, leads one to assume and look for “ideally pure rules” and “the real signs” that this conception of language seems to require. In this way logic’s aspiration for exactness and clarity gets turned into a thesis about the underlying clear and precise rules of language, a hidden structure below the surface of language.44 Given these problems with the Tractatus, the question is, if not simply a shift to the investigation of language, what is the way out of metaphysics to conceptual investigation? Here we come to what Wittgenstein characterizes as his turn, explaining it by reference to the Tractatus’s view that language use must be governed by a “crystalline logic”: “The preconceived idea of crystalline purity can only be removed by turning our whole examination around.”45 Evidently, he believes that turning the examination around provides a solution to the problem of the projection of the Tractatus’s ideal of crystalline logic onto language in the form of a metaphysical thesis. But what exactly is it to turn the examination around? I spell out my answer in 3.4. First, however, I discuss Wittgenstein’s notion of conceptual investigation as the description of language and certain problems with Baker and Hacker’s interpretation of this notion.46 The purpose of this is to have a clearer grasp of the problem before discussing the solution.

3.3 Conceptual Investigation and the Problem of Dogmatism Although Wittgenstein explains the idea of the turn in connection with the Tractatus, it is arguably meant to be of more general methodological significance. The turn constitutes a strategy for avoiding metaphysical projections of forms of presentation onto the objects of investigation in the guise of statements about necessary truths. Thus, the turn is also a strategy for avoiding dogmatism. One might say that the Tractatus functions as an example of metaphysical philosophy in Wittgenstein’s explanation of the turn. A turn of the kind he recommends is then to be performed wherever the Tractatus’s example applies, i.e., whenever one encounters problems of the type described above.47 However, to characterize the turn only in a negative way—as a turn

112

From Metaphysics and Philosophical Theses to Grammar

away from metaphysical projections—does not suffice to explain what it would be to engage in a conceptual investigation instead of metaphysics. We must examine more closely what Wittgenstein means by turning to conceptual investigation. Contrasting conceptual investigations with scientific ones and in particular with metaphysical ones understood according to the model of science, Hacker characterizes the idea of conceptual investigation as follows: “There is a sharp distinction between using a (token) sentence to express a rule of grammar and using one to state a fact. These distinct uses of sentences are as exclusive as using a rod to measure lengths of objects and measuring the length of the rod. Conceptual investigations are investigations into our measuring rods and their uses, not into what is measured.”48 Hacker’s point is that conceptual investigation is concerned with modes of presentation (as exemplified by the measuring rod) rather than with what is presented (making true/false statements about objects, as exemplified by measurements). But what is it for conceptual investigation to investigate measuring rods and their uses, or our modes of presentation, rather than what is presented? (I return to the distinction between factual statements and statements of a rule shortly.) Hacker’s answer to this question can be found in his characterization of Wittgenstein’s view of the task of philosophy as that of the perspicuous presentation of language. (Hacker uses the term “surview.” I discuss the concept of perspicuous presentation in more detail in 6.2.) Hacker writes: “A surview enables us to grasp the structure of our mode of representation, or whichever segment of it is relevant to a given philosophical problem.” And: “a surview is to be obtained by a careful description of our ordinary uses of language. . . . Language is the means of representation. Its inner structure, constituted by the rules which determine the use of sentences and their constituents is the form of representation, the web of conceptual connections by means of which we conceive of the world.”49 And finally: “Philosophy is a conceptual investigation, it describes our conceptual structure from within.”50 According to Hacker, therefore, Wittgenstein’s turn to conceptual investigation is to be understood as a turn to the description of our forms of representation or conceptual structures as they are given in ordinary uses of language—where “ordinary” contrasts with ideal

From Metaphysics and Philosophical Theses to Grammar

113

structures, allegedly discovered below the surface of language.51 This view is also stated explicitly as an interpretation of Wittgenstein’s turn by Baker and Hacker in their discussion of the Investigations §108: “We must examine logic from the vantage point of an unclouded vision of ordinary language, rather than looking at ordinary language through the sharp lattices of a preconceived crystalline logic. This is, indeed, to rotate the axis of the investigation around the fixed point of our need for an Übersicht, a correct ‘logical point of view.’”52 Insofar as this means that instead of projecting an idea of clear and precise rules onto language as a requirement, one should turn to examine language use free of prejudices, this certainly seems a correct characterization of what Wittgenstein aims at with his turn. But to characterize an aim is not the same as to say how it can be reached. That conceptual investigation is or involves the description of language use is also evident. Wittgenstein says, for instance: “Grammar describes the use of words in language. So it has a similar relation to the language as the description of a game, the rules of a game, have to the game.”53 Apparently, “grammar” in this passage means grammatical rules, as Wittgenstein’s comparison between grammar and the rules of a game indicates. But what exactly is to be understood by the description of language use, and what is the role of rules of grammar in such descriptions?54 A clear comprehension of these issues is a requirement for a grasp of Wittgenstein’s turn. Let us begin by examining more closely the notions of a statement of a rule and a statement of a fact, so as to get these two types of statement into focus.55 To begin with, it is not immediately obvious in what sense a rule can be said to be a description. This is best elucidated with the help of an example. I may define a unit of measurement, let us call it a “unit,” by picking up a stick from the ground and saying, “this is one unit long.” By doing this I am stating a rule that determines what it is to be one unit long and defining this particular stick as a standard of the length of one unit. However, it is important to note that with this definition I am not giving a correct or incorrect description or making a true/false statement about anything, in particular the length of the stick. This comes to light in that had I picked up another stick with a different length and defined it as one “unit,” I would not have made a mistake or said something false but merely defined “unit” differently.

114

From Metaphysics and Philosophical Theses to Grammar

Hence, the logical role of my definition is not that of a true/false statement (description) about any particular stick. What might confuse one here is that exactly the same sentence, “This is one unit long,” can also function as a description of a stick or as a statement about the length of a stick. But one must attend to the use of the sentence and not be misled by its form. Put forward as a description of a stick, the sentence is a contingent, true/false statement. In the case of my making such a statement, it might turn out that I was mistaken about the length of the stick and misrepresented it. Importantly, the possibility of error arises here because in this second case the statement represents or describes a possibility (a state of affairs) of which we have a conception independently of my statement. A certain stick is assigned a particular property, whereby it is presupposed that we understand what it is for it to have or lack this property. By contrast, in the case of my definition we have no conception of how long one unit is independently of my statement and therefore no criterion for judging whether the stick is one unit long or not. Rather, my definition determines the meaning of “unit.” In this sense the definition, unlike a factual statement, is not a statement about an object of discourse and has no object about which it is true or false. Rather, it only defines an object of discourse, making possible certain kinds of true/false statements, i.e., statements about objects in terms of the measure of length “unit.”56 Here it is also important to note that the roles of a statement of a fact and a statement of a rule are exclusive, as Hacker also notes in the first passage above.57 One may, of course, measure a rod that is normally used for measuring—and might therefore be called a “measuring rod” for “historical” reasons. In this case, however, the rod is not functioning as the means and standard of measurement, but its logical role is that of an object of measurement. That something should take up the two roles simultaneously is as impossible as using a ruler to measure itself. The means of measurement must be (logically although not historically) independent of the object of measurement. In this sense it would also be misleading to say that my definition of “unit” states a necessary truth about the length of the stick that I use in the definition. As explained, the definition does not state a truth about anything. Consequently, although one might perhaps say that it is true by definition that the standard stick for “unit” is one unit long, it is crucial to keep in

From Metaphysics and Philosophical Theses to Grammar

115

mind how such truths by definition differ from true statements concerning objects.58 But if a statement of a rule does not describe anything in the sense of making a true/false statement about an object of discourse that exists independently of it, how can one describe language use with rules? This brings us to the point of Wittgenstein’s reference to games in the passage from manuscript 140 quoted above. As he observes in another notebook, the word “description” is not only used in cases where there is an object of description given independently of the description and of which the description is true or false. We also talk about descriptions in connection with, for instance, geometrical constructions or designs whereby what is described does not exist independently of the description.59 Thus, one may say that by defining rules for a calculus, or stating rules that constitute a calculus, I describe a system for the use of signs. Similarly, by stating rules for a game, I describe a game. (Likewise, one might talk about my definition of “unit” as a description of a unit of measurement.) Statements of a rule are then to be understood as descriptive of language in this or a similar sense. For notably, Wittgenstein only talks about similarity in the passage from manuscript 140. The word “similar” is significant. The notion of description in the second sense (as in the case of designing a game) cannot alone explain Wittgenstein’s notion of the description of language use. This is because philosophy as conceptual investigation is the clarification of actual confusions and, therefore, the description of actual language use. But as Wittgenstein also remarks (see below), description in the second sense does not inform us about actual practices of playing games or using language. That is, although there is a sense in which rules of a game are descriptive of that game, they only describe it, so to speak, as a possible game. Rules that define or constitute a game do not say or indicate anything about whether this game is ever actually played. They only tell us how it is to be played insofar as it is ever played. Similarly, although I may describe the use of a word by stating a rule and thus determine how the word is to be used, this rule as such does not yet tell us anything about whether and how this word is used in any actual language or by any actual speaker. It is a further matter to determine whether or not language is actually used according to this rule.60 Admittedly, the ambiguity of the word “description” noted above may easily obstruct a clear view of matters at this point. Not observing

116

From Metaphysics and Philosophical Theses to Grammar

this ambiguity and not being clear about the sense in which one is talking about description in a particular case puts one at risk of serious confusion with respect to the notion of the description of language use. For the fact that one and the same sentence may take either role— i.e., function as a description in either of the above senses—may create the impression that a sentence might be both a statement of a rule and a description of actual language use. But this is an illusion created by wavering between the two uses of “description” without noticing it. As explained, a sentence cannot take both roles at once. Accordingly, it is crucial to see that even if the rules one states for the use of a word coincide perfectly with its actual use, they are not descriptive of its actual use. The logical roles of statements of rules and statements of facts are not brought any closer by such coincidences. Wittgenstein writes about the two senses of “description” and the danger of being misled in the directly relevant context of a discussion of statements of rules and a perspicuous notation to be used for philosophical clarification: “What seems to mislead here is a double-meaning of the word ‘description’ when one sometimes talks about the description of a real house or a tree etc., and sometimes about the description of a shape, construction etc., of a notation, a game. Whereby, however, one does not mean a sentence that says that such a game will actually be played somewhere or such a notation actually used; rather the description stands for the words used here ‘such a game’ and ‘such a notation.’”61 This remark summarizes the key points above: the ambiguity of the word “description” and the fact that statements of a rule do not describe actual practices of playing games or using language. But given that to state a rule is not yet to provide a description of actual language, what sort of a statement is the philosopher making when she employs a rule for the purposes of clarification, describing actual language use? This question may be instructively discussed in connection with Baker and Hacker’s interpretation, which seems problematically unaware of the distinction between the two senses of “description.” According to Hacker (and consistent with the second sense of “description,” as in the case of designs), grammatical rules are descriptive of language in the sense of constitutive rules.62 Importantly, however, Hacker takes such constitutive rules as also descriptive of actual language use, attempting— so it seems—to tacitly combine the idea of grammatical rules having the

From Metaphysics and Philosophical Theses to Grammar

117

status of constitutive rules with the idea of their being descriptions of actual language use. This can be seen, for instance, in the following statement: “the descriptions are rules we actually employ as standards for the correct use of expressions.”63 But assuming that the standards of correct use spoken of here are constitutive rules, and granted that such rules do not inform one about the actualities of language use, on what grounds are these rules claimed to be descriptive of actual language? There seems to be an additional claim hidden in the philosopher’s statement as conceived by Hacker. The philosopher is not only stating a rule, but also asserting that this rule is one we actually follow or one actually constitutive of language. This claim requires closer scrutiny. In the interest of keeping things perspicuous, we may distinguish two components in the philosopher’s statement: a rule for the use of language, and a proposition that this rule governs actual use. The second component can be understood in more than one way. First, the philosopher might be understood as making an empirical statement about language, i.e., that language users in fact follow such and such rules. Given that this interpretation of the notion of philosophical descriptions of language is explicitly excluded by both Wittgenstein and Hacker, however, I will not discuss it further.64 Second, the philosopher might state that language is used according to such and such rules, asserting that this is how language must be used unless one wants to deviate from its actual and normal use. The philosopher says, for example, that one must use a word in such and such a way in order for it to mean such and such or to express a certain concept, whereby “such and such a way” is specified by a statement of a rule. Breaking the rule means talking nonsense or using an expression in a sense different from the ordinary. In this case, the philosopher is making not merely an empirical claim about language use but a statement about the possible ways of using language, a normative statement. A rule of this type by which language is asserted to be governed is thought to reveal the “bounds of sense.”65 This way of conceiving philosophical descriptions of language, however, brings us back to a setting analogous in important respects to the Tractatus, as I will explain. Of course, there are important differences too. Unlike the Tractatus, which ends up putting forward a thesis about every possible language, the philosopher is now concerned with a particular language or languages. The philosopher’s claim is a “localized” one.66 Accordingly, the philosopher’s statement is no longer conceived

118

From Metaphysics and Philosophical Theses to Grammar

as concerning hidden essential characteristics of language, or a “crystalline logic.” Significantly, however, these differences do not release the philosopher from the problem of dogmatism, which arises anew in the context of the “localized” descriptions of language. Thus, the philosopher is still in danger of coming to project and impose her conceptions and characterizations onto language as claims about what it must be. For even if the philosopher’s statement is now restricted to a particular language, it is still a statement about a necessity pertaining to language use: about how language must be used or its use described. Clearly, to say that using a word in such and such a way (as specified by a particular rule) is a requirement for the word to have a particular meaning is to make a statement about how a word must be used. The necessity of using the word this way is now conditional or relative to one’s wish to use a word in the relevant meaning. Nevertheless, this does not change the fact that the philosopher is still making a claim about how things must be in the context defined by the conditional. In important respects, her claim is therefore exactly of the metaphysical type. It is an assertion about a necessity pertaining to her object of investigation as if this were a fact about the object. This is clearly exemplified by the kind of statements that Hacker attributes to Wittgenstein, specifically by Hacker’s idea that Wittgenstein’s descriptions ascribe certain rules the status of being actually constitutive of language. Here a claim is made about actual or factual necessity, while the alleged necessity is still assumed to be logical, not empirical. Thus the philosopher’s statement of a rule assumes the role of a thesis or a dogma, a metaphysical projection being performed in the context of describing ordinary language use. As an example, compare the Tractatus’s rule for propositions, according to which they are all presentations of states of affairs, with the rule “The meaning of a word is its use in a language.”67 Baker and Hacker comprehend this rule or remark of Wittgenstein’s as stating that having a use (in the relevant sense of the term “use”) is necessary for a word to have meaning. Only cases in which a word has a relevant kind of use, or its employment is parasitic on such uses, count as instances of the meaningful use of words.68 But if this is how one should read Wittgenstein’s remark about meaning, there is a perfect analogy between this remark and the Tractatus’s statement about propositions as (re)presentations of states of affairs. Where the Tractatus comes in effect to claim (in its

From Metaphysics and Philosophical Theses to Grammar

119

relapse into metaphysics) that being a picture is a necessary condition for being a proposition, the remark about meaning as use makes a parallel claim about words’ meanings. Consequently, it remains completely unclear how one would be better protected against dogmatism in the case of the remark about meaning. Construed as Baker and Hacker suggest, Wittgenstein’s later approach seems as dogmatic as that of the Tractatus and is shadowed by an imminent threat of a relapse into theses. Accordingly, one might very well ask: is it really the case that the remark as interpreted by Baker and Hacker is uncontroversial, as Wittgenstein expects philosophical statements to be, in contrast to theses?69 Here Hacker’s suggestion that one could say of the rules stated by the philosopher that “their aptness is guaranteed by the fact that they are elicited from the person whose bewilderment is in question”70 offers no solution. For being a competent speaker, I know how to use language. But to be able to use a word and to describe its use are two different things, and one’s ability to use language in no way guarantees one’s ability to describe its uses. As Wittgenstein says: “We are not at all prepared for the task of describing the use of e.g. the word ‘to think’ (And why should we be? What is such a description useful for?)” and “One learns the word ‘think,’ i.e. its use, under certain circumstances, which, however, one does not learn to describe.”71 Accordingly, for instance, the ability to teach the use of a word does not require the ability to describe it. Indeed, as numerous instances in philosophy and elsewhere testify, it is certainly possible to mistake for a correct definition or a description of a concept something that is not. Therefore, my acknowledgment that a rule is descriptive of my language use or the fact that a rule is elicited from me cannot guarantee its aptness. (Correspondingly, I can be dogmatic, blinded by misleading pictures, and so on, in attempting to describe my own language use.) Hence Baker and Hacker’s interpretation cannot explain how conceptual investigation would constitute a shift away from metaphysical philosophy and philosophical theses. The argument that brought us here and its conclusion are worth recapitulating. As explained, statements of a rule are not true or false about anything and do not constitute theses. Such statements, however, do not tell us anything about actual language use, and as soon as a rule is claimed to be descriptive of actual language use, the problem about the status of the philosopher’s statements arises. The question is, what allows one to say that the

120

From Metaphysics and Philosophical Theses to Grammar

philosopher’s statement according to which language must be used in such and such a way (or the use of a word is governed by such and such a rule, or such and such is a rule of language) is not a philosophical or metaphysical thesis about language? Baker and Hacker’s interpretation does not seem to offer a satisfactory answer to this question. As an attempt to characterize the problem with Baker and Hacker’s interpretation, one might say that they only exchange one object of description for another. Instead of a hidden logic of language, we are to describe the logic of language as it is given in the ordinary practices of language use. But although it is correct that Wittgenstein turns in his later philosophy to the description of ordinary linguistic practices, this observation does not yet do anything to clarify his notion of the description of language. By discarding the prejudice of crystalline logic, one has gotten rid of one prejudice, but there seems to be nothing to guard against other prejudices and dogmas to which one might fall prey as one turns to describe ordinary practices. Hence characterizing conceptual investigation in the manner of Baker and Hacker as a turn to the description of forms of (re)presentation given in ordinary language, and as a concern with the rules of language, cannot account for Wittgenstein’s shift away from philosophical theses.72 The remainder of this chapter seeks to articulate an interpretation of Wittgenstein’s shift away from philosophical theses. I argue that an extra step is required to complete what emerges as a half-turn by Baker and Hacker in order to solve the problems left unsolved by their interpretation. This step consists of the description of a third way of conceiving what I characterized above as the philosopher’s hidden additional claim (or the second component of her statement). Next, I discuss Wittgenstein’s turn at a more general level by reference to remarks intended to explain it in the Nachlass, after which I return to the question concerning the role of clarificatory rules.

3.4 Wittgenstein’s Turn I have taken the expression “Wittgenstein’s turn” to signify a methodological shift. Many commentators agree that a methodological repositioning occurs in his philosophy in the early 1930s (around 1930–32). This is manifested, for example, by his adoption of the language-game method, the examination of language with the help of “primitive

From Metaphysics and Philosophical Theses to Grammar

121

forms of language” in which the uses of expressions are readily surveyable.73 Accordingly, in the preface to the Investigations Wittgenstein contrasts his “new thoughts” with his “old way of thinking,” indicating that a change has occurred in his philosophical thought.74 But what exactly this shift amounts to is less clear. How the shift in the early 1930s relates to what Wittgenstein calls “turning our whole examination around” in the Investigations §108 also requires explanation, because Wittgenstein only comes to talk about the reorientation of his thought in these terms in the late 1930s (around 1936–37). What is evident is that Wittgenstein’s discussion of the turn in 1936–37 concerns the proper role of the ideal of crystalline purity of logic that the Tractatus projected onto language, claiming that such clear and precise rules must be found in language. This concern is particularly clear in the notebooks Ms157a and Ms157b, where Wittgenstein makes numerous attempts to find the core of the problem with the Tractatus and state the crucial difference between his old and new approaches.75 The discussion of the turn in the late 1930s, therefore, is a discussion of how to avoid dogmatism, prejudices, and metaphysical projections in philosophy. That this is indeed connected with the methodological shift in the early 1930s can be established as follows. A number of earlier remarks also address the theme of dogmatism and projections, suggesting a certain kind of methodological shift as a solution to the problems arising in this context. Furthermore, because the shift articulated in both the earlier and later remarks is essentially the same—I will argue that the relevant remarks form a continuous string of attempts to formulate the idea of this methodological reorientation—we are able to identify this shift as Wittgenstein’s turn. Hence the discussion of turning the examination around in 1936–37 can be understood as an attempt to further clarify the idea of the methodological repositioning that in fact occurred in the early 1930s. This shift involves a novel conception of the role of examples and rules (for instance, definitions) in philosophy. Let us now look more closely at what Wittgenstein says about the problem of metaphysical projections and dogmatism. Consider the following remark from 1937: “When you are tempted to make general metaphysical statements, ask yourself (always): What cases am I actually thinking of?—What sort of case, which conception do I have in mind

122

From Metaphysics and Philosophical Theses to Grammar

here? Now something in us resists this question for we seem to jeopardize the ideal through it: whereas we are doing it only in order to put it in the place where it belongs. For it is supposed to be a picture with which we compare reality, through which we represent how things are. Not a picture by which we falsify reality.”76 The immediate context of this remark is a discussion of the notion of the “ideal name,” which stands for an indestructible, simple element of reality. Wittgenstein characterizes such an ideal name as “a picture, a form of presentation to which we are inclined” when philosophizing about language.77 Not conceiving the ideal name as a picture “which we compare reality with” and a form of presentation “with the help of which we present how things are,” however, one is led to claim that the ideal shows us what language must be, that there must be such ideal names in language.78 What Wittgenstein describes here is clearly an example of the metaphysical projection. One projects a form of presentation that one employs—a certain conception of what being a name amounts to— onto the object of investigation (language), and comes to put forward a dogmatic claim about the nature of names. But how is the question, “What cases am I actually thinking of?” meant to help? This can be clarified by reference to an earlier remark from 1934: “The object of comparison [insertion: model], the object from which a way of conceiving things is derived, should be announced so that the examination does not become unjust. For now everything which holds of the model will be asserted of the object of the examination; & asserted: it must always be . . . This is the origin of a kind of dogmatism. One forgets the role of the prototype in the examination: it is as it were the unit of measurement with which we measure the object of examination. Dogmatism, however, claims that every measured object must be a whole number of the units of measurement.”79 In accordance with Wittgenstein’s analysis of the Tractatus’s mistake, which relates the error of his early philosophy to its employment of examples (see 3.2), the metaphysical projection may in some cases be described as the conferral of the characteristics of a prototype upon the object of investigation. A prototype here is a paradigmatic or exemplary case on which other cases may be modeled, and which, in this capacity, forms the basis for a way of conceiving those cases and a mode of presenting them. The metaphysical projection, however, constitutes a more ambitious use of the prototype as a mode of presentation. Now

From Metaphysics and Philosophical Theses to Grammar

123

the prototype is claimed to be something that the objects of investigation must match; it is seen as bringing to light a necessary truth about the objects of investigation. The question, “What cases am I really thinking of?” is then meant to counteract this temptation to employ the prototype as the basis of a metaphysical thesis. More specifically, the question is meant to achieve this goal by offering a way of depriving metaphysical statements of their charm. One can bring a statement down to earth from the heights of metaphysical abstraction by reminding oneself of its actual origins: the cases that one really had in mind and that inspired the statement. As Wittgenstein puts the point in the early versions of the Investigations by reference to the Tractatus’s ideal of language: “We are released thus also from the enchantment of the ideal . . . by recognising it as a picture & announcing its origin. How did you come to this ideal; of which material did you form it? Which concrete idea was its real prototype? We must ask this, otherwise we will not be able to get rid of its misleading aspect.”80 That Wittgenstein talks about a “misleading aspect” is important. As explained in 3.2, it is not that the Tractatus’s conception of propositions as pictures, for instance, is misleading in and of itself. Rather, the conception takes up a misleading aspect when put forward as something to which all propositions must correspond. This misleading aspect of the employment of the model is to be avoided, not the model as such, which may have an illuminating use in specific contexts. Accordingly, Wittgenstein’s question regarding how one arrived at the philosophical statement is intended to enable one’s release from this dogmatic aspect of the employment of a philosophical conception. It seeks to accomplish this by (re)directing attention to the actual examples and other more specific considerations that functioned as the basis of the conception.81 It would not be correct, however, to think that Wittgenstein is recommending, as a solution to the problem of metaphysical projections and dogmatism, a descent from universal theses to empirical statements concerning particular cases. Consider the following earlier version (from 1931) of the remark from manuscript 115 above: The object of comparison, the object from which this way of conceiving things is derived must be announced so that prejudices do not constantly slip into discussion. Because then we shall willy nilly ascribe what

124

From Metaphysics and Philosophical Theses to Grammar

is true of the prototype of the comparison also to the object to which we are applying the examination; & we claim “it must always be. . .” This comes about because we want to give the prototype’s characteristics a foothold in the examination. But since we confuse prototype & object we find ourselves dogmatically conferring on the object properties which only the prototype necessarily possesses. On the other hand we think the examination will lack the generality we want to give it if it really holds of the one case. But the prototype must be presented for what it is; as characterizing the whole examination and determining its form. In this way it stands at the head & is generally valid by virtue of determining the form of examination, not by virtue of a claim that everything which is true only of it holds for all the objects to which the examination is applied.82

In this remark Wittgenstein is again concerned with the role of exemplary, prototypical cases in philosophical examination and the metaphysical projection of their properties onto the objects of investigation. As above, he is making the point that the prototype should be explicitly presented as a case “from which [a] way of conceiving things is derived.” But, as he also notes, his drawing attention to the actual cases used as examples may cause worries about the generality of philosophical statements. If what the philosopher says applies only to the exemplary cases, the philosophical account seems to lose its generality. Crucially, however, rather than recommending that we comprehend philosophical statements as empirical statements about particular cases, Wittgenstein is proposing a shift away from any theses, whether metaphysical or empirical.83 This can be explained as follows. The point of an example, rather tautologically, is to exemplify something. An example is used to make a point that applies more generally, i.e., also to other cases. But although what an example brings to light may apply generally to many cases, problems arise when its applicability is claimed to be greater than it actually is, in particular when the example is claimed to show something that is necessarily true of the cases it concerns. This brings us to Wittgenstein’s suggestion that instead of being used as the basis for a philosophical thesis, an example should be comprehended as an object of comparison.84 In this role the example is employed to characterize the objects of investigation by way of comparison, noting both similarities and differences between the example

From Metaphysics and Philosophical Theses to Grammar

125

and the cases modeled on it. The example, that is to say, is used to draw attention to certain characteristics of the objects of investigation, but to what extent the latter actually correspond to the former is left open. Or as Wittgenstein remarks: “We give up general dogmas concerning our object,—the particular examples throw as much general light on their surroundings as is cast on them.”85 The metaphor of the amount of reflection of light may be explained by saying that the generality of a philosophical model is a function of its successful application to whatever it is used to (re)present.86 More specifically, a model employed as an object of comparison is not used to make an empirical statement about any particular objects in the sense of being valid of only those objects, though perhaps inductively generalizable. Neither is the model used as a basis for a thesis that states what all objects falling under a concept must be. Rather, the model is the articulation of a way of conceiving things or a mode of presenting the objects of investigation. As Wittgenstein says about prototypical examples in the passage from manuscript 111, such an example determines a form of examination. Thus, for instance, the conception of propositions as pictures articulates a model for what it is to be a proposition, but the necessity expressed by the words “what it is to be” is to be seen as a characteristic of the model, not to be conferred upon the objects of investigation as if stating a necessary fact about them. To be a proposition, according to this model, is to be a picture. But the scope of this characterization is limited to specific contexts of philosophical examination where this model is employed. The statement “to be a . . . is to be a . . .” pertains to the form of an examination or the mode of presentation and does not constitute a claim about a necessity in reality. Consequently, the model’s generality is as great as its successful application in the clarification of philosophical problems. Used as an object of comparison, the model thus loses its dogmatic aspect. In words of the passage from manuscript 183 quoted above, the model becomes “a picture with which we compare reality, through which we represent how things are; not a picture by which we falsify reality,” where “reality” includes anything one might want to take as one’s object of investigation. The philosophical ideal, however, is not “jeopardized” in the sense of a collapse to empiricism but “put in the place where it belongs.” Or as Wittgenstein says in another notebook

126

From Metaphysics and Philosophical Theses to Grammar

from the same period: “The ideal loses none of its dignity if it is posited as the principle determining the form of one’s approach.”87 This conception of the status of philosophical statements is summarized as follows in the 1936 manuscript of the Investigations: “For we can avoid the injustice or emptiness of our statements only by presenting the model as what it is, as an object of comparison—as, so to speak, a measuring rod—and not as a preconceived idea to which reality must correspond. (I am thinking of Spengler’s mode of examination.) Here namely lies a certain dogmatism which our philosophy can easily fall into.”88 This remark outlines the very strategy for avoiding dogmatism discussed above. (Below I discuss in more detail the relation between Wittgenstein’s remarks on this theme in the 1930s and 1940s.) Very interestingly, however, Wittgenstein now specifies the problem of dogmatism as presenting us with a dilemma. Putting forward philosophical statements as theses about what things must be, one runs the risk of either injustice or emptiness. Let us look at this more closely. By injustice, Wittgenstein apparently means the failure of philosophical theses to capture the manifoldness of phenomena they seek to describe, or the tendency of philosophers’ descriptions (definitions and so on) to misleadingly simplify the concepts they are meant to clarify.89 Accordingly, he characterizes the problem of injustice as arising when one focuses on similarities and analogies and ignores differences.90 At other times he associates the problem of injustice with “philosophical parties.”91 Presumably, this association refers to such parties urging one to accept, as the exclusive account of a matter, some simplified conception articulated in the form of a philosophical thesis. It is also noteworthy that, insofar as the philosopher’s task is to clarify people’s thinking for them, injustice is not merely, metaphorical injustice toward the phenomena, but injustice toward those people whose thinking one sets out to clarify. What Wittgenstein means by emptiness is less obvious, but might be explained in the following way.92 The problem of emptiness arises when one tries to avoid conflicts with the ordinary comprehension of things, or tries to avoid doing injustice to the manifoldness of phenomena by subliming the philosophical conception and stating that it really applies to certain ideal cases, not to the phenomena of the everyday world. Thus, for instance, the Tractatus’s conception of propositions was not

From Metaphysics and Philosophical Theses to Grammar

127

intended to be valid for the unanalyzed propositions of everyday language in their manifoldness but to describe what propositions really are (or must be) under the surface of language. Only such analyzed propositions, not ordinary propositions, are expected to meet the standard of exactness that the Tractatus lays down.93 This, however, leads to a peculiar emptiness of the philosophical conception. Applied at an ideal level and to ideal things, the philosophical conception is indeed universally valid. At the level of completely analyzed propositions, no proposition fails to meet the standards set by the Tractatus. (Anything that does not meet these standards is not a proposition.) The price of attaining this universality, however, is a loss in the conception’s power to mark any distinctions. For example, if all language use is perfectly exact, as it is supposed to be at the Tractatus’s ideal level, no room remains for vagueness. But a consequence of thus making vagueness disappear is that the adjective “exact” no longer has any use either. It does not distinguish one kind of proposition from another, as there is nothing to distinguish, no contrast to draw. Hence, in its application to the ideal propositions, the adjective “exact” fails to convey anything. Employed to express a requirement to which language must conform, the term “exact” is used, as Wittgenstein says, “in a typically metaphysical way, namely without an antithesis; whereas in their correct and everyday use vagueness is opposed to clearness.”94 On the other hand, if the exactness of the ideal propositions is contrasted with the vagueness of ordinary unanalyzed propositions, we are brought back to the problem of injustice. For it does not seem right to claim that unless ordinary propositions meet the Tractatus’s standard of exactness, they do not qualify as propositions. Hence, by simply presenting one’s theses as theses about everyday phenomena, one is not released from the difficulties. Attempting to avoid the horn of injustice by resorting to a sublime ideal level, one faces the horn of emptiness, and vice versa.95 The conception of philosophical conceptions as objects of comparison offers a way out of the dilemma. Exaggerated claims about what things must be are eliminated by comprehending the modality expressed by the philosophical statement as characterizing the form of approach or as a characteristic of the philosophical model. Consequently, the need to postulate ideal entities of which philosophical statements are really true is also eliminated.

128

From Metaphysics and Philosophical Theses to Grammar

* * * To ascertain the justification of the above interpretation, let us address in more detail certain exegetical questions relating to the developmental history of the remark just quoted from manuscript 142, and therefore the genesis of the Investigations §131 on how to avoid the dilemma of injustice and emptiness. Is it justified to use the remarks quoted above from manuscripts 111 and 115 to explain Wittgenstein’s methodological response to the problem of dogmatism and his turn? Can these earlier remarks be used to explain Wittgenstein’s view in the Investigations? As becomes evident, these questions are very much about how to read §131 itself. It is not that the interpretation of this remark is somehow obvious and one only needs to understand how the earlier remarks relate to it. This becomes apparent once one starts asking questions about the relation of §131 to its close predecessors in the manuscripts and typescripts of the Investigations. Read in its context in the Investigations, §131 might create the impression that it only concerns the role of simple language-games, which are characterized as objects of comparison in the preceding remark. Notably, however, the context of this remark (the one we know as Investigations §131) is different in the 1936 manuscript and the 1937 typescript of Wittgenstein’s book. In manuscript 142 as well as typescript 220 (§107), the remark occurs in the context of a discussion concerning the role of logical calculi and the idealization of concepts in philosophy. More specifically, in these two texts the remark constitutes the penultimate paragraph of a longer remark the beginning of which was also quoted above, according to which we can be released from the enchantment of the ideal by announcing its origin.96 This longer remark contains a paragraph on the role of logical calculi and a paragraph that provides another example of idealization in philosophy. Similarly, in the notebook Ms157b, where §131 is drafted in a form similar to those in manuscript 142 and typescript 220 (quoted in 3.6), the remark has the word “ideal” in the place of the word “model.” How are these changes to be explained? Insofar as the discussion of the proper role of the ideal—and of turning the examination around—concerns the question of how to avoid metaphysical projections and dogmatism, an answer is readily available. Presumably Wittgenstein’s method of using simple languagegames for purposes of clarification is also meant to accord with the aim

From Metaphysics and Philosophical Theses to Grammar

129

of avoiding dogmatism. But then the remark on models as objects of comparison could very well be presented as concerning at one time the proper role of the ideal and at another time the role of languagegames. In both cases the problem is the same, i.e., that of coming to project a mode of presentation onto the object of investigation in the form of a thesis. Similarly, the solution in both cases is the same. We are to take the model—the ideal or the language-game; the term “model” applies to both—as an object of comparison. This “multifunction reading” of the Investigations §131 seems especially natural in typescript 238 (§144), the 1942 revision of (roughly) the philosophy sections of typescript 220, where the remark is positioned between remarks concerning the role of the ideal and the role of language-games. (Typescript 238 constitutes, as it were, an intermediate link between the prewar versions and the final version of the Investigations.) Here §130 (concerning the role of language-games) occurs directly after §131 (on models as objects of comparison), beginning with the words: “Also our exact//clear//simple language-games.” This seems as explicit a statement as one can hope for to the effect that “model” is indeed intended to cover both logical calculi and simple language-games.97 It is also important to note that in the Investigations, questions concerning the regularization of language (§130) and the refining and completing of the rules of language (§133) are also raised in the immediate context of §131. Hence although the notion of the ideal of crystal rules does not occur here, questions relating to logicians’ aspirations to describe and regulate language by laying down rules are very much present. These observations are also important for understanding the relation of the Investigations §131 to the earlier remarks on the role of prototypes and models quoted above from manuscript 111 (typescripts 211–213) and manuscript 115. Remark §131 also shares an origin with these remarks, as is evident in certain textual facts relating to the way Wittgenstein explains his methodological idea, that is, the examples and comparisons he uses. More significantly, the context for the occurrence of the relevant remarks from manuscripts 111 and 115 is the very same discussion of the role of logical calculi that also provides the context for §131 in manuscript 142 and typescripts 220 and 238. Let us look at these issues more closely. First, as regards the examples and comparisons employed, it is notable that in all relevant remarks up

130

From Metaphysics and Philosophical Theses to Grammar

to the 1942–43 revisions of the Investigations (typescripts 238 and 239), Wittgenstein connects the idea of how to avoid dogmatism with dogmatism in Spengler. He uses Spengler’s theory of the development and decline of cultures as an example of a dogmatic way of putting one’s point and also explains how his methodological amendment allows one to avoid the dogmatism of the claim that this is how cultures must evolve. (I have edited out the discussion of Spengler from the quotations from manuscripts 111 and 115 above, but these remarks can be found in print in Culture and Value.) Another recurring feature is the comparison of a philosophical statement with a measuring rod or a measuring unit, which first occurs in manuscript 115 and is present also in the Investigations.98 (On the other hand, there is no doubt that the passage from manuscript 111 is an earlier version of this remark from manuscript 115.) Up to typescript 220, the relevant remarks are also immediately followed by remarks concerning units of measurement originally drafted in manuscript 115. Second, regarding the context of the remarks on the role of prototypes or models, every such occurrence (from manuscript 111 to typescript 239, except manuscript 157b) is in the immediate context of another remark (originally drafted in manuscript 111) discussing the role of examples in logic. According to this remark, it is mistaken to conceive logical calculi as applicable to ideal cases instead of ordinary, concrete cases of language use, an example being the idea that the Tractatus’s concept-script is really applicable to ideal signs to be revealed by analysis, examples of which are promised only in the future. Rather, Wittgenstein says, insofar as a calculus has application at all it is to actual language use, regardless of whether the calculus can capture actual uses in all respects. Such concrete cases are thus also to be recognized as the prototype of the calculus.99 Examination of the contexts of Wittgenstein’s remarks on prototypes and models as objects of comparison reveals therefore that in each case the relevant remarks (up to typescript 239) occur as a response to a particular problem relating to the status and use of logical calculi. More specifically, these remarks may be read as attempts to articulate a positive conception of logical calculi as applicable to actual cases of language use, and as attempts to generalize this conception of the role of logical calculi so as to cover philosophical statements more widely. (Spengler’s ambitions do not relate to logic and, as mentioned, other

From Metaphysics and Philosophical Theses to Grammar

131

examples of philosophical idealization are discussed in manuscript 142 and typescript 220.) Thus, the remarks in question seem appropriately characterized as venturing to describe a strategy for avoiding dogmatism in logic and philosophy in general. The differences in wording between the remarks in manuscripts 111 and 115 on the one hand, and the post-1936 versions on the other, should not be allowed to obscure the fact that these remarks all try to articulate what is ultimately one and the same point. In the Investigations the generality of this point might, at first sight, seem reduced, as §131 appears to concern only the role of simple language-games. But as already explained, there are good grounds for not reading this section as a remark on the role of languagegames alone. The word “model” in the remark is intended to apply to philosophical conceptions more generally. The relevant remarks from Wittgenstein’s manuscripts on prototypes and models as objects of comparison, or on the problem of dogmatism, may therefore be taken to form a continuous string of attempts to articulate the idea of a methodological shift in philosophy. This methodological shift is then what one might call “Wittgenstein’s turn.” Apparently, we can now also date the emergence of the central idea of the turn to 1931, when manuscript 111 is written. The 1936–37 remarks on turning the examination around are further attempts to find a clear formulation for the idea of this methodological repositioning. But as Wittgenstein’s discussion of this topic in manuscript 115 (1934) shows, he is concerned with the formulation of this methodological insight in the intervening period as well. As regards the significance of the metaphor of turning the examination around, while discussing the turn in 1937 Wittgenstein remarks that he is looking for a “quick passage, with one step” from his old mode of examination to his new mode of examination (Betrachtungsweise), in contrast to “diverse, step by step passages.”100 Such diverse passages are presumably clarifications of the Tractatus’s and other earlier confusions in minute detail.101 In contrast to such diverse passages, however, Wittgenstein now tries to find a quick passage, i.e., a general and more abstract way of articulating the difference between his old and new modes of examination, allowing one to perceive their difference at one glance and independently of questions of detail. This quick passage is then provided by the metaphor of turning the examination around.

132

From Metaphysics and Philosophical Theses to Grammar

* * * The idea of Wittgenstein’s turn might now be summed up as follows. The turn is a shift away from a philosophical conception as a Vorurteil, a preconception to which the object of investigation allegedly must correspond, to the conception as a Vorbild, a model used as a mode of presentation and more specifically an object of comparison. Thus the turn is a move away from theses about the objects of investigation to philosophical statements as the articulation of ways of conceiving the objects of investigation. One way of making the turn consists in the recognition that some actual cases serve as an Urbild, a prototype for the philosophical conception, and the preconception as a projection of the properties of the prototype onto the objects of investigation. With the turn, the prototype is brought back to its role as an illuminating example to be used as an object of comparison. However, there is no need to assume that some actual cases must always serve as the basis of a model. It is also possible to use purely fictional cases as models and to describe language use (or characterize concepts) with the help of such models. This is the role of fictional language-games in Wittgenstein’s philosophy.102 But let us next discuss the turn and the role of rules more specifically to provide further support for the interpretation I propose.

3.5 The Turn and the Role of Rules With the foregoing interpretation of Wittgenstein’s turn we also arrive at a conception of the role of rules in conceptual or grammatical investigation that solves the problem of relapse into theses and dogmatism that threatens conceptual investigation as understood by Baker and Hacker. Comprehended as articulating models to be used as objects of comparison, rules can be used to describe actual language use without falling back to either empirical statements or assertions about how language must be used (in the sense explained in 3.3). But let us first look at what Wittgenstein says about rules in connection with the turn. He writes in the Investigations: “We see that what we call ‘sentence’ and ‘language’ has not the formal unity that I imagined, but is the family of structures more or less related to one another.—But what becomes of logic now? Its rigour seems to be giving way here—But in that case doesn’t logic altogether disappear?—For how can it lose its rigour? Of

From Metaphysics and Philosophical Theses to Grammar

133

course not by our bargaining any of its rigour out of it—The preconceived idea of crystalline purity can only be removed by turning our whole examination round. (One might say: the axis of reference of our examination must be rotated, but about the fixed point of our real need.)”103 What Wittgenstein refers to as “our real need” is the need for clarity, a clear comprehension of what we say.104 Trying to interpret his suggestion as to how to reach this goal, however, we enter a rather complicated and not entirely transparent dialectical situation. Wittgenstein raises the question of whether rigor could be bargained out of logic. But due to the ambiguity of the word “logic,” it is not clear whether he means logic as a discipline or the logic of language and, therefore, what exactly the problem is. Baker and Hacker assume in their discussion that rigor would be bargained out of the logic of language.105 Thus, the problem discussed in the Investigations §108 would be whether language loses its rigor when recognized as not having the formal unity imagined in the Tractatus, according to which every proposition is a picture of a state of affairs with a definite truth-value. This seems a possible interpretation, although it is perhaps odd that Wittgenstein should return to this issue, having already discussed it thoroughly in the Investigations §§65–88. However, insofar as this is the problem, Baker and Hacker are right to observe that the logic of language “cannot lose its rigour as a result of philosophical transactions; it does, because it must, ‘take care of itself.’ The rigour it has it has. All that is lost are our preconceived ideas about what rigour we are inclined to demand of it.”106 Accordingly, they conclude: “We must examine logic from the vantage point of an unclouded vision of ordinary language, rather than looking at ordinary language through the sharp lattices of a preconceived crystalline logic.”107 In Baker and Hacker’s view, therefore, Wittgenstein’s point is that nothing important can be bargained out of the rigor of language, but one can only lose one’s prejudices and philosophical illusions about its rigor. The turn then consists of our discarding the picture of the crystalline logic as a misconceived requirement imposed on language. There is no such crystalline structure of language.108 Naturally, I agree with Baker and Hacker that, according to Wittgenstein, the rigor of language does not depend on what philosophers say about it. Already in the Tractatus he thought that the propositions of everyday language are in order as they are109 and that all philosophy

134

From Metaphysics and Philosophical Theses to Grammar

can do is clarify this logical order. That §108 is concerned with the issue of how to get rid of philosophical prejudices is also clear. Unfittingly for Baker and Hacker’s reading, however, Wittgenstein suggests in the 1936 and 1937 versions of §108 that rather than discarding the notion of crystalline purity, the turn changes things such that “this purity assumes another position.”110 Consider these two formulations of the idea of the turn from Wittgenstein’s notebooks of this time: The complete crystal clarity of logic: it should be a crystal, nothing amorphous in it. (Wherefrom this ideal?) For the knowledge of the amorphous does not interest us here (that is correct). And we also have to do with the crystal systems. I.e. we have to do with exactness. For a greater exactness removes certain misunderstandings. But there are also misunderstandings that easily come about by striving for a greater exactness. But it is not that we could have something bargained out of this crystal clarity! The preconception that lies in it can only be removed by rotating our whole examination; and thereby placing the preconception in another position.111 It looks as if the logic would lose the essential in it; its rigour. As if it had been bargained out of it. But it only now plays another role. It has turned from a preconception concerning reality into a form of presentation. Where has the crystalline clarity come to? It has become a form of presentation and nothing besides that.112

In accordance with the 1936 and 1937 versions of the Investigations, Wittgenstein says that the purity now assumes a new position and a new role. Even more oddly for Baker and Hacker, Wittgenstein states that his investigations still have to do with crystal systems and that he is not interested in the knowledge of the amorphous, i.e., the noncrystalline. And strikingly, according to Wittgenstein, nothing can be bargained out of the crystal clarity of logic. All of this seems to fit poorly with Baker and Hacker’s idea of disposing of the crystal clarity as a misconception about language. These problems suggest that what is at stake might be the rigor of logic as a discipline rather than the rigor of language. Wittgenstein is concerned with the problem of rigor being bargained out of the discipline of logic when it turns out that language does not fit the Tractatus’s

From Metaphysics and Philosophical Theses to Grammar

135

preconception of crystalline purity. From the point of view of this interpretation the above dialectic and problem can be recast as follows. The conclusion that language lacks a formal unity raises concerns about logic as a discipline. Does it, for instance, follow from there being no universal formal criterion for being a proposition (such as the Tractatus assumed) that logic becomes an empirical investigation of the family of cases that forms the concept of language? And even if logic does not become an empirical study, insofar as its task is to investigate such imprecisely bounded families, does the exactness of logic not have to be relaxed so as to match the actual vagueness of language? This interpretation, according to which it is the faith of logical investigation that is at stake, seems to find immediate confirmation in other texts in the Nachlass. Thus in manuscript 114 (from 1932) the question concerning logic in the Investigations §108 is formulated as a question about philosophy: “But if the general concept of language dissolves in this way, doesn’t philosophy dissolve as well?”113 Given Wittgenstein’s conception of philosophy as a logical or grammatical investigation, this might be taken as an alternative formulation of the worry expressed in §108, but one that lacks the ambiguity of the latter. Similarly, in manuscript 152 (from 1936) Wittgenstein discusses the faith of the discipline of logic and the prospect that logic disappears if there is no exactness and clarity in language, an inner, sublime structure with which logic is concerned. For the consequence seems to be that logic becomes an empirical investigation of the uses of language.114 It is also important that after the paragraph quoted from §108, Wittgenstein in fact continues by talking about “the philosophy of logic.”115 Let us turn to examine this alternative, according to which Wittgenstein’s concern is the philosophy of logic, the “nature” of the discipline of logic, philosophy as a logical or grammatical investigation, and the faith of logic as a discipline. Now the bargaining situation looks as follows. The idea of the crystalline clarity of logic is that of rigor and exactness. This is what logic as a discipline aims at. Rigor and exactness are the essence of logic as a discipline.116 But it also seems that if logic is to exist as a rigorous discipline, language must somehow live up to its standards. If our concepts are not exact, then it seems false to present them as such in logic and to ignore the complexities of actual language use. But as Wittgenstein is then led to observe and acknowledge, language does not live up to this demand of exactness. He is unable to

136

From Metaphysics and Philosophical Theses to Grammar

hold on to the claim that there is an exact definition for the essence of a proposition that captures all cases falling under the concept of proposition, or that our expressions are always analyzable into exact ones. And so a very difficult bargaining situation is created. There is something that looks like a justified request to lower the claim as to how exact or rigorous language can be said to be. But if this request is granted and the bid accepted, logic as a rigorous discipline striving for exactness and rigor, i.e., clarity, is destroyed! Nothing can be bargained out of the rigor of the discipline of logic because it thereby loses what is essential to it. To break out of this difficulty, the whole context of the bargain has to be demolished. This is the point of Wittgenstein’s turn. For the bargaining situation is created by the idea that logic involves theses about language use, i.e., that the exactness of logic as a discipline involves a claim about the exactness of actual language use. This brings us to Wittgenstein’s innovation regarding the new position of crystalline clarity. Whereas the Tractatus asserts that language use must be governed by clear and precise rules, Wittgenstein’s novel approach abandons all such claims about language use. The rules are not to be envisaged as characteristic of the object of investigation. Rather, with the turn they are understood as a form or mode of presentation and as characterizing the form of logical examination. Philosophy or logic uses clear and precise rules as a mode of presentation in describing and clarifying the use of language. As soon as it is recognized that the rules and calculi used in logic constitute a form of presentation, not a thesis about what language is (or must be), the bargaining situation collapses. To employ a particular form of presentation is not yet to make any claims about the object of presentation. To say that language can, for certain purposes, be described as being used according to exact rules is very different from saying that it is really used according to such rules. Hence the use of exact rules as a mode of presentation as such does not commit one to any claims about the exactness or inexactness of language. Accordingly, rigor is not bargained out of the discipline of logic. It is given a novel position through rearrangement and by “recognising the ideal of the order as a part of the mode of presentation.”117 Wittgenstein’s novel conception of the status of logical models thus dissolves the problem of whether language can live up to the standards

From Metaphysics and Philosophical Theses to Grammar

137

of exactness of the discipline of logic. That there is no exact universal definition of a concept does not mean that one could not use such a definition for the purposes of a discussion of a certain particular philosophical problem. There is no problem with simplifying matters in this way as long as the definition captures the aspects of the concept that are relevant from the point of view of the problems at hand and one is aware of making a simplification. Or as Wittgenstein says, concluding a discussion of ambiguity in our concepts and having remarked that even though we cannot say that language users must be following exact rules, we may compare language to calculi in logic: “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.”118 Here it is crucial to notice that this is indeed the only way to get rid of the crystalline purity as a prejudice. For if one simply denies the exactness of language, the prejudice will stay with one in the form of the negative thesis that language does not match the requirements of the discipline of logic and this discipline is unable to reach the level of rigor considered essential to it. In this sense the turn—conceived as a turn away from the idea that logic advances any theses about the exactness or inexactness of language—emerges as a move that is necessary (in our particular historical situation; see 7.2) for successfully abandoning metaphysics. Exchanging a thesis about the exactness of language for a thesis about its inexactness does not count as getting rid of philosophical theses. The realization that the Tractatus projected a form of presentation on language, therefore, only constitutes the first half of Wittgenstein’s turn. Having reached this point, one still faces the problem about the rigor of logic as a discipline. This problem is not solved merely by saying that language is not what the Tractatus took it to be. Rather, one must move away from the idea that logic assumes or involves any claims about language and that the rigor of the discipline of logic depends on any theses about the nature of language. This is the point at which Baker and Hacker fall short of the turn.119 For although it is quite apt to say that “[t]he preconceptions of the Tractatus forced an ideal upon language,”120 Hacker still appears to assume that the use of exact rules and calculi must involve a commitment to a claim about the nature of language. Since Wittgenstein is

138

From Metaphysics and Philosophical Theses to Grammar

clearly critical of the claim that language is a calculus, it might seem that he must also reject the idea of using exact rules and calculi in clarification. As Hacker writes: “The idea that the logical structure of a language can be illuminatingly represented in a calculus is chimerical. For natural languages are very unlike logical calculi, and a perspicuous representation of their philosophically puzzling features cannot typically be fruitfully given in the forms of artificial notations. . . . Artificial calculi can contribute nothing to the solution of philosophical problems.”121 The problem with this view as an interpretation of Wittgenstein, however, becomes apparent by posing the analogous question of whether natural languages are like Wittgenstein’s simple languagegames, for instance, the builders’ language-game in the Investigations §2, which consists only of orders. There can be no doubt that the answer is negative. There are no natural languages that consist only of orders, and the shopping situation in §1 is genuinely strange, far from realistic.122 Nevertheless, it is clear at the same time that, according to Wittgenstein, such artificial language-games can be used for the purposes of clarification. Why could this not be the case with exact rules and calculi?123 Here it is also helpful to contrast the view Hacker expresses with how Wittgenstein explains his approach in the Blue Book. For although Wittgenstein too makes clear that he does not think that language is used according to strict rules like a calculus, his conclusion as regards the usefulness of exact rules in clarification differs from Hacker’s. Wittgenstein writes: [I]n general we don’t use language according to strict rules . . . . We, in our discussions on the other hand, constantly compare language with a calculus proceeding according to exact rules. This is a very one-sided way of looking at language. For not only do we not think of the rules of usage—of definitions, etc.—while using language, but when we are asked to give such rules, in most cases we aren’t able to do so. We are unable clearly to circumscribe the concepts we use; not because we don’t know their real definition, but because there is no real “definition” to them. To suppose that there must be would be like supposing that whenever children play with a ball they play a game according to strict rules. . . . Why then do we in philosophizing constantly compare our use of words with one following exact rules? The answer is that the

From Metaphysics and Philosophical Theses to Grammar

139

puzzles which we try to remove always spring from just this attitude towards language.124

That language is not used according to exact rules is therefore, in Wittgenstein’s view, not a reason against using such rules in philosophical clarification.125 Indeed, there may be several reasons for employing exact rules, one of which is mentioned here. It has to do with how philosophical problems come about or, as Wittgenstein says, philosophers’ “attitude towards language,” their keenness on giving definitions of concepts and their constant running into trouble as a result. “The man who is philosophically puzzled sees a law in the way a word is used, and, trying to apply this law consistently, comes up against cases where it leads to paradoxical results.”126 A possible response in such a situation is to provide other rules that may be just as acceptable instead of the philosopher’s original rules. This way one may counteract the idea that one must think about a matter in a particular way and consequently resolve philosophical problems that arise from thinking about it in this way.127 Or as Wittgenstein says about the problem of our not being able to define the concept of knowledge: “We should reply: ‘There is no one exact usage of the word “knowledge”; but we can make up several such usages, which will more or less agree with the ways the word is actually used.’”128 Thus Wittgenstein is not denying that the philosophers’ exact definitions have value. Rather, his point is to move away from the situation in which we look for the right definition (a once-and-for-all characterization), as if philosophy’s task were the description of language as such, in the abstract and independently of the clarification of particular unclarities. On the other hand, an exact definition may well enable the clarification of particular problems provided that the definition adequately captures those aspects of language use that are unclear and that give rise to philosophical problems. But one should not present the definitions as dogmas to which language must correspond. As Wittgenstein also says: “Shall I . . . say that there is no list of rules for our language, and the whole exercise of setting up one is nonsense? But it is clear that it is not nonsense for we lay down rules successfully; & we must only refrain from setting up dogmas.”129 To return to Hacker’s interpretation, however, it would not be fair to say that he allows no room for the use of exact rules and calculi in philosophy. He does note that Wittgenstein attributes a role to exact rules

140

From Metaphysics and Philosophical Theses to Grammar

and calculi as objects of comparison. When Hacker says of the ideal languages that philosophers construct that “their primary role in philosophy is to function as objects of comparison to illuminate what is present in ordinary language,”130 this seems exactly correct. This is what ideal languages or artificial calculi are for Wittgenstein (see below). However, on the basis of Hacker’s account such objects of comparison are doomed to remain strangely useless, as he says in the previous quotation, able to “contribute nothing to the solution of philosophical problems.” Or as he also says of artificial calculi, they are “[a]t best . . . only of use in philosophy as objects of comparison in the course of pursuing the real business of philosophy.”131 Insofar as this means that the real business of philosophy does not involve the use of exact rules as objects of comparison, it seems again misplaced.

3.6 Rules as Objects of Comparison Let us examine more closely what Wittgenstein says about the role of rules and calculi in philosophy. In the passages above from the Blue Book and the dictation to Schlick (in note 125), the idea of comparing language to games or calculi with exact rules was mentioned. This expresses Wittgenstein’s basic insight regarding the role of rules: in order to avoid dogmatism, one should understand the ideal rules of logic as an object of comparison. As he writes in manuscript 157b about the Tractatus’s ideal of crystalline purity (this remark is a draft of the Investigations §131): “For we can avoid injustice—or emptiness, in our assertions only by presenting the ideal in our examination as what it is, namely an object of comparison—as, so to speak, a measuring-rod—[and] not as a preconception, which everything must conform to. For here lies the dogmatism into which we fall so easily in doing philosophy.”132 Having been rotated to its right place, the ideal of crystalline logic thus becomes an object of comparison. Rather than maintaining dogmatically that language must conform to the exact rules that one states, rules are to be comprehended as articulating models with the help of which language use is described by way of comparison. The advantage of this characterization of the role of rules is that it renders plain the function of rules as distinct from statements of fact. To employ an exact rule (such as a definition in terms of necessary conditions) as an object of comparison is not to assert that language use

From Metaphysics and Philosophical Theses to Grammar

141

conforms to the rule, as if stating a fact about a necessity in language use. Rather, the role of the definition is similar to that of a measuring rod used for describing certain characteristics of an object of investigation. The point of a measuring rod, one might say, is that it helps to establish similarities and differences between different objects with respect to their length. But it would be ludicrous to claim that for anything to count as an object with a length, its length must be a whole number of our units of measurement.133 Similarly, one should not confer upon language the properties of the rules one employs as a mode of presentation. Although one criterion for the value of a definition is that it captures many cases of the use of a word in a particular meaning, one falls into dogmatism if one demands that the definition must hold of all the cases of the use of the word in the relevant meaning.134 If one then demands that the artificial calculi and exact rules correspond perfectly to actual language use in order to be useful tools of clarification, one has not moved away from dogmatic requirements. This is to fail to follow Wittgenstein’s turn and to continue making claims about language use. One is still demanding that language must be this in order to be presented thus, for example, that it must be a calculus in order to be presented fruitfully as such. But because the choice of the appropriate mode of presentation (like that of a unit of measurement) is relative to the task at hand, it is confused to claim at an abstract level, independently of the task of clarifying any particular unclarities, that a certain mode of description is either appropriate or inappropriate. Similarly for all general objections against using artificial calculi in clarification, such as expressed by Hacker above. On the other hand, it may be perfectly in order to object to using a particular rule to characterize a concept in the context of a particular discussion. Following Wittgenstein in his turn, one may therefore very well look at language “through the sharp lattice of crystalline logic” and describe language use with the help of exact rules, contrary to Baker and Hacker’s statement (quoted in 3.5). The crucial thing is simply not to confuse this lattice with reality. Although philosophy, to be sure, “talks about language in an ordinary sense,” it may use exact rules and calculi to make its points about language—and this implies no theses about the nature of language. As Wittgenstein writes in the Investigations: “in philosophy we often compare the use of words with games and calculi which have fixed rules, but cannot say that someone who is using language

142

From Metaphysics and Philosophical Theses to Grammar

must be playing such a game.”135 This contains the essential core of Wittgenstein’s conception. The dogmatism of claims about what language must be (how someone must be using it) can be avoided by comprehending logical or philosophical models as objects of comparison. Consider how Wittgenstein characterizes his method of describing language in 1934: If we look at the actual use of a word, what we see is something constantly fluctuating. In our investigations we set over against this fluctuation something more fixed, just as one paints a stationary picture of a constantly altering landscape. When we study language we envisage it as a game with fixed rules. We compare it with, and measure it against, a game of that kind. If for our purposes we wish to regulate the use of a word by definite rules, then alongside its fluctuating use we set up a different one by codifying a characteristic aspect of the normal use of the first one in rules. Thus it could be said that the use of the word “good” (in an ethical sense) is a combination of a very large number of interrelated games, each of them, as it were a facet of the use.136

Although actual language use may be fluctuating, such fluctuating uses can, according to Wittgenstein, be studied by comparing them with something more fixed, by envisaging language as being used according to fixed rules. In this sense the description of language as a game according to fixed rules provides a particular point of view that may be adopted. (A literal translation of the third sentence quoted above is: “We look at language from the point of view of a game according to fixed rules.”)137 The rules thus function as instruments employed to reveal aspects of the actual language use—or according to Wittgenstein’s deleted alternative, normal language use—similarly to how a stationary picture captures aspects of an altering landscape. Or as he says in an earlier manuscript: “I would like to call the rule an instrument. ‘To clarify grammar’ means to produce it in the form of a game with rules.”138 Again, however, this is not to say that language is a calculus or a game governed by rules. Rather, language use may be presented as such a game or calculus with the purpose of clarifying its uses. This purpose is clearly expressed, for example, in the following passage from the previous page in manuscript 140, which provides yet

From Metaphysics and Philosophical Theses to Grammar

143

another characterization of this procedure of clarification: “if we wish to draw sharp boundaries in the use of a word, in order to clear up philosophical paradoxes, then alongside the actual picture of the use, in which as it were different colours flow into one another without sharp boundaries, we may put another picture, which is in certain ways like the first one but is built up of colours with clear boundaries between them.”139 Similarly, Wittgenstein writes in manuscript 114: “If I wish to draw sharp boundaries to clear up or avoid misunderstandings in the area of a particular use of language, these will be related to the fluctuating boundaries of the natural use of language in the same way as sharp contours in a pen-and-ink sketch are related to the gradual transitions between patches of colour in the reality depicted.”140 The significance of these characterizations with respect to present interpretative issues is that a version of these remarks also occurs in the Investigations (§§76, 77). This shows that the method of clarification with rules as objects of comparison described above is not just a transient phase in Wittgenstein’s thinking (in the early to mid-1930s) but an important constituent of his later methodology. As in manuscripts 140 and 114, in the Investigations Wittgenstein compares a sharply defined concept to a color patch with sharp contours and talks about using such sharply contoured patches to characterize color patches with blurred contours.141 Here a striking image is used to counteract the idea that the task of definitions should (always) be that of providing us with something we might call the definition of the concept: “But if the colours in the original merge without a hint of any outline won’t it become a hopeless task to draw a sharp picture corresponding to the blurred one? Won’t you then have to say: ‘Here I might just as well draw a circle or heart as a rectangle, for all the colours merge. Anything— and nothing—is right.’—And this is the position you are in if you look for definitions corresponding to our concepts in aesthetics or ethics.”142 But it would not be correct to conclude from this remark that exact definitions are worthless in cases like this. Rather, as Wittgenstein says in the passage from manuscript 140 above, fixed rules—such as sharp definitions—may be used to capture “characteristic aspects” of a concept or, in the case of a word such as “good” in an ethical sense, certain “facets of the use.” Thus, exact definitions may be deployed to draw distinctions that are important from the point of view of certain particular philosophical problems, even though it would be admitted that the boundary between the classes thus distinguished is in reality blurred

144

From Metaphysics and Philosophical Theses to Grammar

and the classes merge with each other without a clear border. For it may be crucial to see a difference of kind between two classes of cases even if they have no definite border and it remains (partly) open which cases fall into which class. Moreover, the fact that the definition does not capture the use of the expression in every respect should not be considered a defect, as the point is to dissolve particular philosophical problems. As noted, given this purpose, it is enough to focus on the aspects of the concept relevant to the problem. This way rules may be deployed to present blurred linguistic uses in a clear fashion without the problematic consequences of simplifying dogmas about the use of language.143 The foregoing also shows that it would be a misunderstanding to think that in Wittgenstein’s view a family-resemblance concept may never be characterized with the help of a sharp definition. For in both manuscript 140 and the Investigations, when Wittgenstein discusses the role of definitions, his examples are concepts that he takes to be family-resemblance concepts. Nevertheless, he clearly accepts the possibility of characterizing these concepts with the help of sharp definitions, although he does not think that any such definition can adequately capture the concept as a whole. Whether an exact definition is appropriate depends on the particular purposes of the discussion, i.e., on the particular problems addressed. It is dogmatic to maintain that an exact definition must always be given. It is similarly dogmatic, however, to claim that such a definition can never be appropriate. An example of the procedure of clarifying the use of words assisted by a sharply defined concept, without claiming that this concept is an accurate definition of the actual concept, is found in Wittgenstein’s discussion of the concept of reading in the Investigations. Here reading is defined as an “activity of rendering out loud what is written or printed; and also of writing from dictation, writing out something printed, playing from a score, and so on.”144 Clearly this definition does not adequately capture what we mean by reading, but only some part of it (including also something we would normally not call reading). Nevertheless, it can be used to clarify specific aspects of reading as well as the concept of rule-following more generally, as Wittgenstein does in this context. Again, that reading, according to him, is a familyresemblance concept is not considered a problem.145 In Chapter 4 I argue that Wittgenstein’s characterization of meaning as use is yet another, similar example of the use of a sharply defined concept to characterize “fluctuating language use.”

From Metaphysics and Philosophical Theses to Grammar

145

Finally, although Wittgenstein’s point about the role of rules may seem radical, there is also a sense in which he is simply building on something already accepted by many philosophers. Their own attempts to give definitions show that they are happy to talk about concepts and language use in terms of rules. Wittgenstein might then be seen as making only a slight (albeit arguably crucial) methodological adjustment when he says that definitions should not be projected onto language in the form of theses.

3.7 Rules, Metaphysical Projection, and the Logic of Language Wittgenstein’s conception of rules as objects of comparison constitutes a solution to the problem of dogmatism as it arises in Baker and Hacker’s interpretation. Employed as an object of comparison, a rule used to describe language does not constitute a claim about a necessity pertaining to language use. Thus, the philosopher’s statements are clearly distinguished from theses about language and the “bounds of sense.” However, the proposed interpretation may also invite misunderstandings, to which I now turn. The point of Wittgenstein’s introduction of his conception of clarificatory rules as objects of comparison is not that, according to him, it is impossible to describe the actual uses of language as they really are— and that one, therefore, must be satisfied with something less: mere comparisons, and so on. Although the uses of language may often be too complex to capture exhaustively in particular formulae, this does not imply that descriptions given for particular purposes would be inadequate.146 Accordingly, rather than making a claim about what can or cannot be done (that it is impossible to describe language adequately), Wittgenstein’s purpose is to eliminate a source of philosophical problems. After all, philosophical models are supposed to be something “with the help of which we present how things are,” not something “by which we falsify reality.”147 The point is that, when put forward as theses about what language (or the things falling under concepts) must be, the philosopher’s rules are in danger of becoming just such falsifications. Wittgenstein, therefore, is trying to counteract the problematic tendency of our preventing ourselves from seeing how language functions by setting up dogmas, including ones we are not aware of. As he notes about the Tractatus’s projection of its conception of propositions onto

146

From Metaphysics and Philosophical Theses to Grammar

language: “Only when this illusion has been removed can we simply see language, as it is.”148 Here Wittgenstein is not saying that the Tractatus’s dogma needs to be replaced with another one, the correct one. His point is that language can only be seen as what it is when dogmas are no longer imposed upon it. One must not, that is to say, fall back to the position the Tractatus characterized by saying: “That . . . which we have to formulate here, is not a likeness of the truth, but the truth itself in its entirety.”149 This remark, quoted in 2.32, is a comment on the conception that everyday language is in a perfect logical order, which it is philosophy’s purpose to disclose. Problematically, as explained in 3.3, the Tractatus imagined this disclosure as if it did not involve any particular point of view or the employment of modes of presentation as an expression of a point of view. Similarly, a temptation to assume a position of “pure observation without a point of view” may arise in the context of descriptions of language that imitate the later Wittgenstein. Here too one may be misled to imagine that grammatical rules bring to light the order of language.150 But a rule is a mode of presenting language: to state a rule in order to describe uses of language is to employ language for a particular purpose, not to transcend language and such purposes.151 To refrain from presenting rules as theses about language and to deploy them as objects of comparison is then to move from metaphysics to conceptual investigation, as Wittgenstein conceives the latter. Hence conceptual investigation cannot simply be characterized as an investigation that takes modes of representation as its objects, as Baker and Hacker suggest.152 It is essential to this investigation that it be carried out with the awareness that the investigation itself involves the use of modes of presentation and that clarificatory statements themselves possess a grammar of their own. Crucially, this grammar, i.e., the necessities such statements express, is not to be projected onto language as if one would be stating necessary truths about language. Here the distinction between philosophy as a conceptual and as a superfactual investigation is again lost from sight, and a relapse into metaphysical claims about language occurs. Finally, the question might be raised whether Wittgenstein’s conception of rules as objects of comparison amounts to denying that language possesses a logic independent of its descriptions in terms of rules employed as objects of comparison. Or as one might also formulate

From Metaphysics and Philosophical Theses to Grammar

147

this question, Does the view attributed to Wittgenstein amount to an instrumentalism about logic according to which to speak about the rules of language is only a useful fiction? From the point of view of the interpretation developed in this book, Wittgenstein must certainly be seen as rejecting the Tractarian picture of language as governed by an implicit structure of rules through which the possibilities of its use are determined.153 Rather than thought of as dependent on language’s possession of such an implicit structure of rules, the logic of language is understood, in his later philosophy, as something that resides in the use of language: it is embodied in the making of judgments (and so on), that is, in the linguistic practices we engage in.154 Ultimately, therefore, what can and cannot be said is decided in the context and against the background of the form of life that language is part of. What can be said is a matter of what language users can make sense of given their lives with language.155 Moreover, importantly, the fact that such lives are something not purely conventional or arbitrary brings to light a sense in which the logical is fundamentally intertwined with the factual.156 Wittgenstein writes about this intertwinedness: “If language is to be a means of communication there must be agreement not only in definitions but also (queer as this may sound) in judgments. This seems to abolish logic, but does not do so.—It is one thing to describe methods of measurement, and another to obtain and state results of measurement. But what we call ‘measuring’ is partly determined by a certain constancy in results of measurement.”157 As this remark may be interpreted, the use of language for the purpose of communication does not merely presuppose an agreement on definitions in an abstract sense, i.e., an agreement on rules used to determine the meanings and use of terms. Rather, to agree on definitions is also to agree, to a large extent, on judgments, that is, on the actual uses of language. (Ultimately, only agreement in judgments shows that there is an agreement on definitions or the meanings of terms— including the terms used in stating definitions.)158 Crucially, however, this does not mean that logical or grammatical statements (employed to describe language) can be understood as factual statements about language use, i.e., about its users’ agreements (and so on). Such a reduction of logic to the factual would, indeed, abolish logic. The discipline of logic would now become a branch of linguistics, anthropology, or sociology, and no room would be left for logical or grammatical

148

From Metaphysics and Philosophical Theses to Grammar

statements as expressions of exceptionless necessities. Statements of logic or grammar must, therefore, be distinguished from factual statements. As Wittgenstein notes in the remark just quoted, it is one thing to describe a method—or a unit—of measurement, i.e., to state a rule, and another to state results of measurement (see 3.3). However, on the basis of these observations concerning the relation of logic to the factual, we can already say that, although logical or grammatical descriptions are not factual (nor superfactual) statements, what they describe is not a fiction. Logic, that which the logicians try to capture, is there in the use of language to be described correctly or incorrectly, and it is in this sense something real and concrete. (To think that Wittgenstein’s turn, as described in this chapter makes the logic of language something that is merely in the eye of the logician-beholder is to draw the negative conclusion from Wittgenstein’s rejection of logical theses that language lacks the characteristic it was previously claimed to possess. But to move from a thesis concerning language to its negation is to remain within the sphere of theses, not to leave behind this way of thinking about language and logical investigation.) The difficult issues relating to the intertwinedness of the logical and factual touched upon here are discussed and elucidated from various angles in Chapters 4–6.159 Next I will seek to further clarify Wittgenstein’s method of grammatical description by discussing how the use of rules as objects of comparison is exemplified by what he says about the concepts of meaning and language.

FOUR

Grammar, Meaning, and Language

on Wittgenstein’s later remarks on the concepts of grammar, meaning, and language. Attention to the details of these remarks reveals his views to be perfectly congruent with the method of clarification described in Chapter 3. Here we also arrive at an explanation of the sense in which his conceptions of meaning as use and language as a rule-governed practice and his characterization of language as a family-resemblance concept do not constitute philosophical theses about the essence of meaning or language or about the use of the relevant terms. Accordingly, Wittgenstein’s remarks on these concepts can be used to further elucidate and deepen our grasp of his method. To make the interpretative and philosophical problem more concrete, I again contrast my reading with that of Baker and Hacker, who despite their intentions to the contrary arguably end up attributing philosophical theses to Wittgenstein. (See especially 4.1 and 4.4.) In establishing the consistency of Wittgenstein’s remarks on meaning and language with his methodological statements (in 4.2–4.6), my purpose is first of all to throw light on the more general issue of how his remarks on the concept of philosophy and his actual philosophical practice can be interpreted as constituting a coherent whole.1 The methodological considerations in Chapter 3 can be understood as a framework for the interpretation of Wittgenstein’s remarks that allows one to do just that, i.e., to read him as not putting forward philosophical theses. This approach to establishing the coherence of Wittgenstein’s writing can be adopted and applied to his remarks on other topics as

THIS CHAPTER FOCUSES

150

Grammar, Meaning, and Language

well, and parallel nonmetaphysical readings developed in such contexts. (Chapter 5 illustrates this in connection with his remarks on the concepts of necessity, essence, and grammar.) My discussion is also intended as a contribution to the interpretation of Wittgenstein’s views on meaning and language. In particular, I seek to explain how his approach makes possible the articulation of richer and more nuanced views on language and meaning than seems possible in terms of philosophical theses. This way the sense in which language is conventional and the sense in which it is deeply rooted in facts about human beings and their environment can be done equal justice, as I explain in 4.3–4.5. Similarly, as shown in 4.6, dimensions of meaning that are eclipsed by interpreting Wittgenstein’s definition of meaning as use as a thesis about word-meaning now become accessible to philosophical examination. In 4.7 I return to more general methodological questions, clarifying how the discussion of Wittgenstein’s method in this chapter bears on the status of philosophical analyses in terms of necessary conditions. This explicates a contrast between analyses in terms of necessary conditions and analyses in terms of what Wittgenstein calls “centers of variation” that emerges in the context of my discussion of the notion of family resemblance in 4.5. Overall, an important goal of this chapter is to elucidate the way in which Wittgenstein’s method appears to make possible a certain gain in the flexibility of philosophical thought without a loss of rigor, a potential that cannot be grasped as long as his remarks are (in effect) read as theses. If Wittgenstein’s method really is capable of delivering such goods, the advantages of his approach should be plain. But to get to a position to evaluate this, we must first see what can be made of his remarks on language and meaning.

4.1 Grammar, Use, and Meaning: The Problem of the Status of Wittgenstein’s Remarks How should Wittgenstein’s conception of meaning as use and meaning as something determined or constituted by grammatical rules be interpreted? He discusses the concept of meaning particularly in manuscripts 114 and 140,2 from which I consider several remarks together with formulations encountered in other manuscripts, lectures, and the

Grammar, Meaning, and Language

151

Investigations. I begin with meaning and grammar and come later to the conception of meaning as use. Consider the following series of remarks:3 I want to say: the place of a word in grammar is its meaning. But I can also say: the meaning of a word is what the explanation of its meaning explains. . . . The explanation of the meaning explains the use of the word. The use of a word in the language is its meaning. Grammar describes the use of words in the language.4

The first of these remarks establishes a connection between the words “grammar” and “meaning.” Wittgenstein says that he wants to talk about meaning as “the place of a word in grammar.” He then writes that he could also say that the meaning of a word is what the explanation of its meaning explains. The second remark can be taken to specify what it means to say that the meaning of a word is its place in grammar. To see how it can be understood as such an explanation, let us look at the different formulations Wittgenstein gives of this remark. The second remark occurs in different forms in Wittgenstein’s Nachlass. In the Investigations it reads: “ ‘The meaning of a word is what is explained by the explanation of the meaning.’ i.e.: if you want to understand the use of the word ‘meaning,’ see what are called ‘explanations of meaning.’”5 As Wittgenstein spells out this strategy on the opening page of the Blue Book, it is possible to bring the question, What is meaning? “down to earth” by examining what are called “explanations of meaning.” Approaching the question in this way, he says, “will teach you something about the grammar of the word ‘meaning’ and will cure you of the temptation to look about you for some object which you might call ‘the meaning.’”6 An important aspect of Wittgenstein’s remark about the explanations of meaning is, therefore, that it articulates a particular way of approaching the question, what is meaning?7 If one follows Wittgenstein’s advice and asks, “What are called ‘explanations of meaning’?” one will come to see, according to him, that they are characteristically rules. He writes about this in manuscript 140 (some ten pages after the passage above): “We said that by ‘meaning’ we meant what an explanation of meaning explains. And an explanation of meaning is not an empirical proposition and not a causal explanation, but a rule, a convention.”8 Meaning, according to Wittgenstein, can

152

Grammar, Meaning, and Language

therefore be understood as something explainable in terms of rules.9 Various kinds of definitions—ostensive and verbal—provide examples. The sense in which the second remark from manuscript 140 constitutes an explanation of the idea of the meaning of a word as its place in grammar can now be stated as follows. By drawing attention to explanations of meaning, the remark highlights the possibility of talking about meaning in terms of rules. Given this, one may also characterize the meaning of a word as its place in grammar or talk about meanings of words in terms of such places. Grammatical rules that specify the word’s role in a language define a place for a word in grammar.10 Accordingly, Wittgenstein also talks about meaning as the role of a word in a calculus and in a broader sense as something determined or constituted by grammatical rules.11 Finally, as regards Wittgenstein’s remark that the use of a word in the language is its meaning, when the surrounding remarks in manuscript 140 are taken into account, “use” is most naturally understood as “rulegoverned use.” If, as Wittgenstein says, grammatical explanations of meaning in terms of rules are explanations of use, and grammar is the description of the use of words in language, then evidently use can be spelled out in terms of rules.12 These explanations of Wittgenstein’s conception of meaning also make plain its connection with his conception of philosophy as the clarification of language use. Following his observation that explanations of meaning are rules, clarification can be understood as an activity of stating rules for the use of expressions. Here the activity of clarification emerges as an entirely linguistic and therefore logical or grammatical affair. Clarification is a concern with language, not with something else that supposedly determines linguistic meaning—for example, objects in the world or mental states—but cannot itself be taken as an object of logical or grammatical investigation.13 Especially in the early 1930s, Wittgenstein often expresses this point by saying that concepts such as “picturing,” “meaning,” and “understanding” are not metalogical concepts. By this he means that they are not concepts of a different order, which provide a foundation for logical investigation but cannot themselves be taken as the object of this investigation. Rather, these concepts are concepts like any other and are to be investigated by the same means.14 Given Wittgenstein’s conception of philosophical clarification as an activity of stating rules, his remarks on meaning from manuscript 140

Grammar, Meaning, and Language

153

above are presumably also to be seen as such clarificatory rules. Interpreted in this way, these remarks are examples of the method of clarification by stating rules themselves, and explain their own status reflexively. But even if things seem fairly straightforward up to this point, questions now arise concerning the status of Wittgenstein’s explanations or the role of grammatical rules in philosophical clarification. Is the point of his remarks on meaning—for instance, the above set of grammatical rules—to determine what linguistic meaning must be? Is Wittgenstein stating that, given our concept of meaning, a word only has meaning if it has a rule-governed use or is parasitic on such uses? Is his characterization of the concept of meaning intended in this sense to be valid for every possible case falling under the concept of wordmeaning? If so, would this not count as a philosophical thesis about (the possibility of ) meaning? To address these questions, let us first consider what Wittgenstein says about meaning as use in the Investigations: “For a large class of cases—though not for all—in which we employ the word ‘meaning’ it can be explained thus: the meaning of a word is its use in the language.”15 This remark explicitly expresses reservations concerning the generality of the conception of meaning as use. But due to a certain ambiguity, it still seems to allow for different readings. According to the interpretation put forward by Baker and Hacker, examples of the use of “meaning” that are excluded are cases such as “those clouds mean rain” and “you mean so much to me,” but all cases of word-meaning are included in Wittgenstein’s large class of cases.16 According to their reading, Wittgenstein’s statement about meaning therefore expresses something that holds necessarily for all cases of word-meaning. As they write: “There is no such thing as meaning independently of rules which determine how an expression is to be used.”17 Similarly, according to Glock: “A sign becomes meaningful not through being associated with an object, but through having a rule-governed use. Whether a sign is meaningful depends on whether there is an established use . . . ; what meaning it has depends on how it can be used.”18 Taken in this way, Wittgenstein’s explanation of the concept of meaning might then be further characterized in Kantian terms by saying that it determines the limits of this concept, and to this extent the “bounds of sense,” by articulating a necessary condition of word-meaning.19 Put

154

Grammar, Meaning, and Language

another way, Wittgenstein’s explanation is taken to show how one must use the relevant words, unless one wishes to deviate from their normal or actual use—possibly at the price of talking mere nonsense. In this sense, Wittgenstein’s explanation is taken to articulate the rule by which the actual use of “meaning” is governed. Or put yet another way, given that Wittgenstein’s explanation establishes a necessary condition of word-meaning and that there are no cases of word-meaning falling outside the scope of this explanation, the explanation may be characterized as the definition of the concept of word-meaning. The problem with this interpretation, however, is that it seems to turn Wittgenstein’s statement about meaning into a philosophical thesis about the concept of meaning (our actual concept) or the essence of meaning. For in stating that a rule-governed use is necessary for wordmeaning, a philosopher certainly appears to be making a statement about what cases falling under the concept of meaning must be, and in this sense about the concept or the essence of meaning. (According to this, all instances that qualify as meaningful uses of words must be either rule-governed or parasitic on such uses.) Alternatively, one might ask why the philosopher’s statement about how the word “meaning” must be used in order for it to express our concept of meaning—i.e., that it must be used in the way specified by the philosopher’s rule—is not a thesis about what language use must be or about the rules by which our language is governed. The problem of dogmatism arises again. One might at least ask: how can the philosopher be sure that in stating her rule she is not overlooking cases that do fall under our concept of wordmeaning but do not fit the criterion of rule-governedness as a necessary condition of word-meaning? And what guarantees that her statement about what meaning must be is not yet another example of a philosopher projecting a mode of presentation (her definition of the concept of meaning) onto reality?20 Baker and Hacker’s response, I believe, would run along the following lines. Grammatical statements are not theses because they would be acknowledged by everyone who masters the use of the relevant expressions. In this sense they are, unlike theses, wholly uncontroversial. More specifically, as opposed to disputes about matters of fact, there can be no controversies about grammatical rules because deviation from genuine grammatical rules constitutes a fall into nonsense, not a false statement.21 Nevertheless, this response fails to resolve the prob-

Grammar, Meaning, and Language

155

lem. Ultimately, Baker and Hacker’s view that there can be no controversies about grammar is simply a corollary of their interpretation of Wittgenstein’s conception that grammatical rules determine what makes sense. There can be no controversies about grammar on the assumption that Wittgenstein’s statement about meaning as constituted by rules articulates the rule that defines the concept of word-meaning (or the rule we follow in the actual use of the word “meaning”). But given that this account of why there are no controversies in philosophy, and why Wittgenstein’s statement about meaning is not controversial, evidently depends on a suspected thesis about meaning, it cannot clarify the issue of why Wittgenstein’s conception of meaning is not a philosophical thesis. In effect, Baker and Hacker’s response constitutes a restatement of the suspected thesis and begs the very questions at stake. A second possible way of unpacking Baker and Hacker’s response results in a corresponding outcome. Ultimately, their response seems to boil down to the claim that there are statements in language that language users would necessarily agree upon, these being the grammatical rules. Since Baker and Hacker presumably do not intend to make an empirical claim that there are such necessarily agreed-upon statements, they seem to be claiming that there must be such statements in language. But now the original question concerning the role of grammatical statements arises again. Why is this attempt to account for the status of grammatical statements not a philosophical thesis about language? As the status of Baker and Hacker’s response overall is as unclear as that of the statement about meaning that it was meant to explain, the problem remains unresolved. These problems can be avoided, however, if Wittgenstein’s definition of the concept of meaning (as articulated through the rules that he states) is taken as a model to be used as an object of comparison, following what he says about the role of “our assertions” in philosophy in the Investigations §131. According to this interpretation, rather than stating that the use of “meaning” must accord with the model of meaning as rule-governed use, he is saying that one should compare the actual concept (or the actual use of the word “meaning”) with this model, noting both the similarities and the dissimilarities between the model and actual use. This alternative reading essentially comprehends Wittgenstein’s use of the rule “The meaning of a word is its use in the language” in terms of the description of his methodology provided in

156

Grammar, Meaning, and Language

Chapter 3 and articulated in the following remark (quoted more fully there): “When we study language we envisage it as a game with fixed rules. We compare it with, and measure it against, a game of that kind. [New paragraph] If for our purposes we wish to regulate the use of a word by definite rules, then alongside its fluctuating use we set up a different one by codifying a characteristic aspect of the first one in rules.”22 Here the point of the model—the philosopher’s rule—is to provide something concrete, which is under the clarifier’s control, to be set against the manifoldness and blurriness of language use, so as to bring order into linguistic relations and to make it possible to perceive them more clearly.23 In philosophy, that is to say, one may use rules for the purpose of codifying in them characteristic aspects of the actual, blurred uses of language. Abiding by the method of comparison, however, one refrains from claiming that actual language use must fit the models one puts forward. In this sense no theses are advanced about how things—including language use or concepts—must be. Thus when used as an object of comparison, the conception of meaning as constituted by rules does not amount to a thesis about the essence or the possibility of word-meaning. Rather, it articulates a particular way of conceiving word-meaning, while leaving open the possibility that there might be cases that do not fit this model. Hence the necessity expressed by the rule is now treated as characteristic of the model and not projected onto the objects of description as a thesis about their essence. Consequently, one is in principle safeguarded against dogmatism.24 An important consequence of this way of understanding Wittgenstein’s conception of meaning is that there is no longer any need to insist (with Baker and Hacker) that anyone who does not agree with the view that word-meaning is constituted by rules must be either speaking nonsense or deviating from normal language use. Both claims, indeed, are highly controversial ways to account for the uncontroversiality of Wittgenstein’s conception.25 When understood as an object of comparison instead, the conception of meaning as constituted by rules itself becomes a great deal less controversial. (Its uncontroversiality no longer needs to be backed up by further claims or elaborations on language but should be apparent in its philosophical use.) As a result, the conception might now be acceptable even to those who would otherwise reject the view that, according to Wittgenstein, word-meaning de-

Grammar, Meaning, and Language

157

pends on rules (Cavell, for example).26 For provided that Wittgenstein’s conception is only meant to capture a particular aspect of the concept of word-meaning, and not to constitute the account or the definition of word-meaning, it no longer seems particularly difficult to agree about, assuming that one grants some role to rules in the constitution of word-meaning. And significantly, the cases in which this clarificatory model is to be used are now, in principle, decidable one by one in the sense explained in 3.4. Thus agreement on Wittgenstein’s conception does not imply an acceptance of it as applicable to every case of the meaningful use of words. It also follows from this interpretation that even though Wittgenstein says, for instance, “We are inclined to forget that it is the particular use of a word only which gives the word its meaning,” “The use of a word in practice is its meaning,” and “a label would only have meaning to us insofar as we made a particular use of it,”27 one cannot without further justification extract a universally applicable statement about meaning from such remarks. In the context of these quotations Wittgenstein is arguably making a certain specific point about meaning, namely, that a word is not given meaning simply by pointing at something and saying “This is . . .” The meaning of a word, like that of a label, depends on the use made of it, and labeling something does not yet determine the meaning of the label.28 But by making this point Wittgenstein does not yet commit himself to the view that wordmeaning can (or must) always be understood according to the model of meaning as use. Similar observations apply to the Investigations §43, in the context of which Wittgenstein discusses the idea that the meaning of a name is an object that the name stands for. (The discussion here concerns more or less the same issues as in the Blue Book.) As a response to this referentialist conception of meaning, Wittgenstein then offers the explanation of meaning as use, aiming to eliminate certain problems that arise in the context of referentialism, like the problem of a name becoming meaningless when its bearer ceases to exist.29 Again, however, the suggestion to replace the referentialist conception of meaning implies no commitment on Wittgenstein’s part to the view that meaning can always be explained as use. To attribute to Wittgenstein a thesis about what word-meaning must be on the basis of §43 reads more into the text than it can support. Such a reading fails to take seriously the possibility that

158

Grammar, Meaning, and Language

the statement is only meant to dissolve certain particular philosophical problems. Note also that in his lectures Wittgenstein explicitly offers the conception of meaning as use as a solution to problems arising in the context of referentialism. As he writes, discussing the referentialist problem of having to postulate abstract objects as the meaning of numerals: “As a way out of the difficulty posed by this question I suggest that we do not talk about the meaning of words but rather about the use of words.”30 But is there any textual evidence for the interpretation of Wittgenstein’s conception of meaning that I suggest? Although the interpretation derives initial plausibility from his general methodological remarks as discussed in Chapter 3, such remarks alone cannot settle the interpretation of his remarks on meaning. Perhaps there are conflicts between Wittgenstein’s pronouncements on methodology and his actual practice. We must, therefore, examine more closely the ways in which he formulates his conception of meaning.

4.2 Wittgenstein’s Formulation of His Conception of Meaning In formulating his conception of meaning, Wittgenstein makes numerous qualifications concerning its generality and nature. From the interpretative standpoint that attributes to Wittgenstein a view of what word-meaning must be, such formulations are bound to appear idiosyncratic, odd, and even anomalous.31 But whether or not this is correct, it is worth examining in more detail how Wittgenstein expresses himself. Let us first look at how he explains his point to his students in his lectures of 1933–34. Wittgenstein introduces his characterization of meaning as use with the following words: “in discussing understanding, meaning etc. our greatest difficulty is with the entirely fluid use of words. I shall not proceed by enumerating different meanings of the words “understanding,” “meaning” etc., but instead shall draw ten or twelve pictures that are similar in some ways to the actual use of these words.”32 The first runs: “To begin with, I have suggested substituting for ‘meaning of a word,’ ‘use of a word,’ because use of a word comprises a large part of what is meant by ‘the meaning of a word.’ . . . The use of a word is what is defined by the rules, just as the use of the king of chess is defined by the

Grammar, Meaning, and Language

159

rules.”33 The second of Wittgenstein’s pictures is the characterization of meaning as what is explained by the explanation of meaning. Wittgenstein says (direct continuation of the previous quotation): “I also suggest examining the correlate expression ‘explanation of meaning.’ This will teach us something about the meaning of ‘meaning.’” In parallel to the passage from manuscript 140 quoted in the beginning of 4.1, he then connects this with the conception of meaning as use: “The meaning of a word is explained by describing its use.”34 That is, given that use is something defined by rules, one may describe the use of a word by tabulating rules and in this way explain its meaning. Clearly, Wittgenstein is spelling out here the same conception of meaning as in manuscript 140, although he is more specific about the way he understands the relation between his characterizations and the actual use of words. To say that a characterization of meaning is one of several “pictures that are similar in some ways to the actual use,” and that it “comprises a large part of what is meant by ‘the meaning of a word,’” is not to say that the characterization is intended to capture the actual use or the meaning of “meaning” in every respect and exactly as it is. Similarly in the case of the second picture, Wittgenstein says that attention to the explanation of meaning “will teach us something about the meaning of ‘meaning’” (italics mine). Because of the parallel with the first picture, and the fact that the point is put in exactly the same way in the Blue Books (see quotation in 4.1), this choice of words does not seem accidental. We can therefore conclude that, according to Wittgenstein, focusing on the use of words and the explanations of meaning, i.e., on rules and the rule-governedness of language, makes manifest only a particular aspect of the concept of meaning. It is not meant to be all there is to say about meaning, and there might be other aspects to meaning.35 This, however, is still compatible with different interpretations. The aspect to which Wittgenstein draws attention could be a necessary condition of word-meaning, something presupposed by every instance of word-meaning and determinative of all possible cases falling under the concept of word-meaning. That this condition is only necessary and not also sufficient means that it does not account for everything that is relevant for word-meaning. Nevertheless, according to the interpretation of rule-governedness as a necessary condition of meaning, the aspect that Wittgenstein is referring to is the defining feature of word-meaning in

160

Grammar, Meaning, and Language

the sense that there are no cases of meaningful word usage that are neither instances of nor parasitic on rule-governed uses.36 Accordingly, from this point of view, anyone who talks about word-meaning in the normal sense (the sense in which the word is actually used) presupposes this condition. Nevertheless, the above passages do not support this interpretation in any unambiguous way. Wittgenstein does not actually say in so many words that he conceives rule-governedness as a necessary condition of meaning. In fact, if this is what he really means, it seems strange that he does not say so. At any rate, it is fair to say that if this is his point, he explains himself to his students in a rather unclear way. Alternatively, Wittgenstein might be taken to say that the above pictures of meaning constitute something with which the concept of meaning is to be compared. These pictures are models to be used as objects of comparison and, in this capacity, employed to bring to light a particular aspect of the concept of meaning. Rule-governedness, then, is not to be taken as the essential and defining feature of the concept, a necessary condition that is always presupposed when one talks about word-meaning in the normal sense. Rather, since the pictures are only “similar in some ways” to the actual use of “meaning,” the claim that the actual cases must fit these pictures is mistaken. It is a misleading simplification to say that the word “meaning” is actually used according to the rule “The meaning of a word is its use in language.” To decide between these interpretations, we must examine more textual evidence. Such evidence can be found, for instance, in the context of the above remarks. Indeed, what Wittgenstein says here seems to provide quite strong support for the second interpretation. For although Wittgenstein does not actually use the phrase “object of comparison,” the quoted passages are preceded by a discussion of the idea of clarification with the help of exact rules to which actual language use is compared.37 Wittgenstein says, for instance: “Inasmuch as our language is complex, I shall point out simpler structures which can be set side by side with it to see what light they shed on it.”38 And: “We shall compare the use of language to playing a game according to exact rules.”39 Consequently, it seems most natural to understand the two pictures of meaning he calls “similar in some ways to the actual use” as such “simpler structures” with which the actual “entirely fluid” use of “meaning” is compared. Now the apparent oddity of how Wittgenstein

Grammar, Meaning, and Language

161

expresses himself also disappears. The way he explains himself is an apposite way of making his point.40 We find another interesting characterization in a notebook from the same academic year. Here Wittgenstein makes essentially the same point about the status of his conception of meaning, but in a slightly different way: “If we wanted to bring our explanation that meaning is constituted by the rules for the use of a word closer to the actual use of the word ‘meaning,’ we would state that such rules constitute the meaning which we could easily have an overview of [leicht übersehen könnten], which would give us an easily surveyable [leicht übersehbares] characteristic picture, or something like this.”41 Thus, as Wittgenstein says, the explanation that meaning is constituted by rules for the use of a word is not meant as a statement about how the word “meaning” is actually employed when talking about the meanings of words.42 Rather, constitutive rules of this kind serve the purpose of the perspicuous presentation of the use of words. Such rules provide, as Wittgenstein says, a characteristic picture of a word’s meaning. This explanation applies also to the word “meaning” itself. The definition of meaning as constituted by rules is itself a picture designed for the purpose of the perspicuous presentation of language and intended to capture a characteristic aspect of the use of the word “meaning.” The relation between Wittgenstein’s explanation of meaning and the actual use of “meaning” can therefore be described by saying that the former talks about meaning in a particular, easily manageable sense. The point of Wittgenstein’s conception of meaning as rule-governed use is now clearly delineated. It is intended to bring to view certain aspects of the concept of meaning, but it is not a thesis about the essence of (the concept of ) meaning. Accordingly, one might say that the remarks in the first quotation from manuscript 140 in 4.1 are a set of grammatical rules that Wittgenstein uses to articulate or define a certain conception or model of what having a meaning and describing language use comes to. Crucially, however, he does not claim that this conception or model of meaning captures the meaning of “meaning.” (I will shortly provide more support for this reading.) This interpretation of the remarks in manuscript 140 can be also explained as follows. In a paper on the context-principle in Wittgenstein’s later philosophy, Katherine Morris draws attention to a textual

162

Grammar, Meaning, and Language

device in Wittgenstein that she calls “outer modals” in contrast to “inner modals.” This device is exemplified, for instance, by the following statement from Wittgenstein: “Here we might say . . . that a sign . . . may be sometimes a word and sometimes a proposition.”43 From the perspective of Morris’s distinction, “might” is to be understood in this instance as an outer modal governing “may” as an inner modal and in this way qualifying it. Thus Wittgenstein is not simply making a claim about what a sign may (or may not) be or stating a “grammatical truth” pertaining to the concepts of sign, word, and proposition. Rather, he is articulating a particular conception of what it is for a sign to be a word or a proposition, a possible way of looking at the issue. Or as Morris says about the outer modals: “What these ‘modal’ operators govern . . . is conceptions of certain key concepts: they introduce new possibilities of ways of seeing concepts.”44 Notably, Wittgenstein’s articulation of his conception of meaning as use and as constituted by grammatical rules in manuscript 140 (quoted in 4.1) is also qualified in a corresponding way. He does not simply assert that it is a rule of our language that the meaning of a word is its place in grammar, but explicitly qualifies this with “I want to say.” This may be read as indicating that what he wishes to articulate is a particular conception of meaning rather than the definition of meaning. Accordingly, the chain of remarks about meaning in manuscript 140 is not to be read as a series of dogmatic statements about how one must understand the concept of meaning. It is intended to spell out a particular conception of meaning or a model for the concept of meaning. That is, given what Wittgenstein wants to say (the first line of the quoted passage), one can further say that meaning is what is explained by the explanation of meaning, and so on. Indeed, this is also what Wittgenstein actually says, the second remark being prefixed with “But I can also say.” The point that the conception of meaning as explainable in terms of rules spells out a specific sense of “meaning” but does not constitute a claim about the sense of “meaning” is made repeatedly in Wittgenstein’s manuscripts. For instance: “What interests us in the sign, the meaning which matters for us is what is embodied in the grammar of the sign.”45 And: “Our proposition ‘meaning is what an explanation of meaning explains’ could also be interpreted in the following way: let’s only bother about what’s called the explanation of meaning, and let’s

Grammar, Meaning, and Language

163

not bother about meaning in any other sense.”46 Similar qualifications are also expressed in the immediate context of the first quoted passage in 4.1. Here too Wittgenstein says: “Meaning, in our sense, is embodied in the explanation of meaning.”47 Although there is some ambiguity to these remarks as well, against the background of the preceding discussion there should remain no doubt with regard to their interpretation. As to what “our interest” is or why Wittgenstein characterizes this conception as “meaning in our sense,” this too seems plain. He is interested in rules and in stating rules for language because rules are a good instrument for clarification. This purpose is also stated in manuscript 140, where Wittgenstein writes (having just remarked that meaning “in our sense” is embodied in the explanation): “An explanation of meaning can remove every disagreement with regard to a meaning. It can clear up misunderstandings. The understanding here spoken of is a correlate of explanation.”48 Again, however, the characterization of understanding as a correlate of explanation is not to be taken as a statement about what understanding meaning must be. As Wittgenstein is well aware, understanding meaning does not necessarily entail being able to explain meaning.49 Consequently, it would be erroneous to maintain that Wittgenstein takes the rule-governedness of words to be a necessary condition of their meaningfulness. Rather, it is simply characteristic of his model for the concept of meaning that it presents word-meaning as dependent on rules, and rules as necessary for meaning. This necessity, however, is not to be projected onto the object of investigation, and the model turned into a thesis about the use of the word “meaning” or the concept of meaning. To do so is to relapse into dogmatism and to fail to follow Wittgenstein in his turn away from metaphysics.50

4.3 The Concept of Language: Comparisons with Instruments and Games Next I discuss certain remarks by Wittgenstein on the concept of language. As in the case of meaning, I argue that it would be problematic to interpret him as making a claim about what language must be or what something must be in order to qualify as part of language, or more specifically, claiming that the essence of language can be defined by reference to its rule-governedness. (Subchapters 4.4 and 4.5 seek to

164

Grammar, Meaning, and Language

further elucidate Wittgenstein’s view of language and to make clearer what one gains by following Wittgenstein in abandoning such philosophical theses.) The first thing to notice is that insofar as the purpose of Wittgenstein’s definition of meaning in terms of rule-governed use is not to capture the essence of word-meaning, he can hardly be read as having meant rule-governedness as the essential characteristic of language, either—even if it would be assumed that what is essential in linguistic expressions is their meaning. (But this assumption too might be considered not self-evident, as explained shortly.) Instead, as a closer examination of Wittgenstein’s characterizations of the concept of language readily reveals, he only presents the conception of language as a rule-governed practice as capturing one particular aspect of language. Accordingly, he remarks about philosophy’s concern with the rules of language: “We consider language from a one-sided point of view.”51 Although the aspect of rule-governedness may be of particular interest to philosophy, it is nevertheless just one aspect of the concept of language, not the essential feature of language. To focus on it is one-sided. Let us examine this in more detail. Considering the context of the passage above, we find that having made the remark about the one-sidedness of his point of view, Wittgenstein goes on to cite and elaborate examples of expressions that cannot be explained in terms of rules. His example is the word “boo” used to give someone a fright.52 He concludes: “One may now say that the purpose, the effect of the word ‘boo’ is the important thing about the word; but explaining the purpose or the effect is not what we call explaining the meaning.”53 Thus the word “boo” used with the purpose of giving someone a fright is excluded from the class of words whose meaning is explained in terms of rules (see also the context of the quoted passage). As Wittgenstein says, what is important about this word would be explained in terms of its effect. Clearly, such explanations differ significantly from explanations in terms of rules. Unlike a definition, an explanation in terms of effects states something about empirical facts, i.e., the actual effects of words. (It could not be said that the purpose of the word “boo” is to give a fright if it did not, in fact, in most cases have this effect.)54 Unlike in the case of a rule, where the effects of a word are concerned one might well talk about discovery. Whereas it makes

Grammar, Meaning, and Language

165

sense to say that one has discovered that a word has particular effects on people, it does not make sense to talk about discovering a rule. (Rules are laid down rather than discovered. Apparently, one may speak of the discovery of rules only in the sense of finding out about already-established rules.) Similarly, whereas I might stipulate a rule, I cannot stipulate the effects of a word.55 But can one say that “boo” (in the above capacity) has a meaning at all? There are reasons for saying that “boo” does not have a meaning. One cannot, for instance, really talk about understanding in this connection. If I sneak behind you, shout “boo,” and you jump, we would not say that you understood what I meant by this word. You just reacted to it in a particular way. Similarly, rather than saying that “boo” has a meaning, Wittgenstein talks about the purpose or effect as “the important thing about the word”—or as one could also say, the essential thing about this word. But in the end it is not very important from the point of view of present interpretative questions whether words like “boo” are described as having a meaning or not. For even if “boo,” according to Wittgenstein, does not have meaning, he is clearly willing to talk about it as a word. This willingness to acknowledge it as part of language—a tool of language of some sort—indicates that he does not consider rule-governedness as the essential feature of language, i.e., as a necessary condition for something to qualify as part of language. Nothing, however, depends on this particular choice of words in manuscript 140. The same point is also made in the Investigations: “To invent a language could mean to invent an instrument for a particular purpose on the basis of the laws of nature (or consistently with them); but it also has the other sense, analogous to that in which we speak of the invention of a game. [New paragraph] Here I am stating something about the grammar of the word ‘language,’ by connecting it with the grammar of the word ‘invent.’”56 Wittgenstein distinguishes here between two senses in which one could talk about inventing a language: as comparable to inventing an instrument that serves a particular purpose (i.e., the discovery that something can be used to bring about such and such) and as comparable to inventing a game. What is the contrast he wishes to draw? Obviously games, just as instruments, conform to natural laws, and in the construction of both games and instruments such regularities are presupposed and utilized. However, unlike the design of a game,

166

Grammar, Meaning, and Language

the design of an instrument is ultimately determined by external purposes it serves, i.e., by purposes comprehended independently of the instrument’s invention. Accordingly, in an earlier version of the remark from 1930–31 (Ms110), Wittgenstein mentions the gasoline engine and the sewing machine as examples of instruments. Clearly, these instruments are invented to serve external purposes in the relevant sense. They are designed to perform more effectively tasks that existed before their invention. (I return to this earlier version of the remark in 4.4.) External purposes of this kind differ from internal ones, such as the purpose of checkmating one’s opponent in chess. The purpose of checkmating, of course, is to win the game, but that it constitutes winning the game depends on the rules of the game. That checkmating constitutes winning in chess is a matter of definition, not a discovery about the most effective or entertaining (and so on) way to reach a predetermined goal. This point might also be expressed by saying that the rules of games are, in a certain sense, arbitrary. Of course, games too may serve external purposes. Often they are invented for entertainment, for example, and in this respect their rules might not be arbitrary. (A change in rules may make a game more or less entertaining.) But this only illustrates the point that (systems of ) rules may serve as instruments for particular purposes. It is not to say that for something to count as a rule, it must serve some external purpose. Similarly, for something to count as a game, it is not necessary for it to serve any external purposes. In analogy to games one may also say that the rules of grammar are arbitrary. As Wittgenstein writes (in the immediate context of the previously quoted passage): “The rules of grammar may be called ‘arbitrary,’ if that is to mean that the aim of the grammar is nothing but that of the language.”57 Grammatical rules, therefore, do not have any predetermined purpose that they serve, any more than language use has such a purpose, and to this extent they are arbitrary. Consequently, Wittgenstein’s distinction can be spelled out as follows. The two senses of inventing a language to which he refers make manifest two aspects of the concept of language: (1) language as analogous to games and defined by arbitrary rules and (2) language as analogous to instruments that serve particular external purposes, i.e., purposes that are not arbitrary but determined, for instance, by their effects. The connection with “boo” in manuscript 140 is now clear. This word is an example of the second case. As Wittgenstein says, the role or

Grammar, Meaning, and Language

167

function of this word is defined in terms of its effect rather than by rules. It is also clear that certain qualities of the word “boo”—its sound—are relevant in a way that is different from words whose function is defined by rules. “Boo” can only perform the function of giving a fright insofar as it is uttered loudly and suddenly, which is not a matter of convention but is due to the empirical fact that people react to sudden loud noises in a particular way. Hence “boo” used to frighten someone indeed resembles an instrument designed for a particular purpose on the basis of certain regularities in nature. We can now conclude that although “boo” lacks certain features characteristic of linguistic expressions—its function does not depend on rules, and in this sense rule-governedness is not essential to it—this does not automatically mean, according to Wittgenstein, that it could not be taken as a word and as part of language. For “boo” exhibits certain other features that may be said to be characteristic of language. One may then say on the basis of these features that “boo” is part of language. That is, if we accept Wittgenstein’s characterization of the grammar of “language” through the two analogies in the Investigations §492, then the fact that the function of “boo” is not defined by and dependent on rules is not a sufficient reason to exclude it from language. At this point one might wish to object that it is because of the rules that govern the use of “boo” that it is classified as part of language. For there is, of course, a convention about using this particular sign for giving a fright, where from the instrumental point of view any sudden loud sound would do. Note, however, that to insist that “boo” counts as part of language because of the rules that govern it is to say that it counts as part of language for reasons that are inessential to its functioning as the expression it is, i.e., as a word employed to give a fright. As explained, the capacity of “boo” to serve this purpose depends not on rules but on facts about the expression and about human beings. Hence the explanation in terms of a rule cannot do any real work here. Rather, one might suspect that it only serves to satisfy a demand arising from a commitment to a certain form of account, i.e., the form of a uniform philosophical thesis that explains the essence of a class of objects by reference to a characteristic or characteristics common to them all. In any case, Wittgenstein evidently is not to be read as committed to such a thesis about the essence of language. But although 4.2 and 4.3 seem to allow us to decide the question of the status of Wittgenstein’s characterizations of the concepts of meaning and

168

Grammar, Meaning, and Language

language, it might be less clear what would be gained by adopting his conception of the status of philosophical statements. Why should one understand characterizations and definitions in philosophy in this way, i.e., as clarificatory pictures or models to be used as objects of comparison?58 I have argued that this allows one to avoid dogmatism. But admittedly this is still rather abstract. To further clarify this approach and what kind of advantages can be expected from it, I next consider Wittgenstein’s conception of language from the perspective of its development.59 For at an earlier stage he seems to have thought of language as exclusively definable by reference to grammatical rules.

4.4 Wittgenstein’s Development and the Advantages of His Mature View Wittgenstein writes in an early version of the Investigations §492 in manuscript 110 from 1930–31: What does the one do who constructs (invents) a new language? Which principle does he follow? For this principle is the concept “language.” To invent a language does not mean to invent an instrument for a particular purpose on the basis of the laws of nature (or consistently with them). . . . Also the invention of a game is not an invention in this sense, but comparable to the invention of a language. I don’t need to say that I merely describe further the grammar of the word “language” by connecting it with the grammar of “invent.”60

Here Wittgenstein denies what he later affirms, i.e., that the invention of a language is comparable to that of an instrument. Evidently, the reason for this denial is his rejection of the idea that the grammar of language would be determined by purposes external to it like the design of instruments. As already noted, grammatical rules, according to Wittengstein, are arbitrary.61 Crucially, however, to put the wrong kind of emphasis on this point about the arbitrariness of grammar runs the risk of dogmatism and of injustice in cases where the instrumental aspect of language is more prominent and conventionality present only as an accidental feature, as in the case of “boo.” Looking at Wittgenstein’s development in light of the problem of dogmatism, one might say that what leads Wittgenstein to abandon the dogmatic view of manuscript 110 is his coming to realize and fully

Grammar, Meaning, and Language

169

appreciate that to talk about language in terms of rules is to adopt a particular picture or conception of language, which should not be projected onto language as a thesis about its essence. Instead, in accordance with the idea of the turn (apparently first formulated slightly later in manuscript 111; see Chapter 3), one must comprehend this picture of language as a clarificatory model to be used as an object of comparison. This then allows one to avoid dogmatism and enables one to do justice to other aspects of language besides its rule-governedness. Consistent with this interpretation, the following interesting formulations and deletions can be found in manuscript 114 on the same page where the final version of the Investigations §492 first occurs: “The grammar of language constitutes its essence. The language is particularly for us a calculus; it is characterized by linguistic acts, not their practical purpose.”62 The point of the deletions can be explained as follows. It would be misleading to state without any qualifications that language is a system determined by its rules like a calculus, these rules constituting its essence. Rather, the conception of the use of language as an essentially rule-governed activity brings into focus a particular aspect of language. This conception is a characterization of language from a particular angle, motivated by certain specific philosophical issues, as Wittgenstein’s qualification “for us” indicates. Here my interpretation can be contrasted with Hacker’s reading of the Investigations §492 to illustrate the former’s clarificatory capacity and the more general philosophical advantages of Wittgenstein’s view as interpreted here. Consistent with his attribution of (what I would characterize as) a thesis about the essence of language as rule-governed, Hacker finds it difficult to accept that Wittgenstein would allow “purely instrumental” cases as belonging to language. According to Hacker, this interpretation “seems to concede too much.”63 The problem, as Hacker apparently sees it, is that admitting such instrumental cases as part of language seems to open doors to causal theories of language that seek to explain language in terms of its effects. Hacker is, of course, entirely right that it would be problematic to think of language exclusively according to the model of instruments, i.e., in terms of the effects of expressions, and that Wittgenstein does not accept this. Ultimately, according to such a view, an expression means whatever effect it happens to have. As Wittgenstein notes in the relevant

170

Grammar, Meaning, and Language

context in the Investigations: “When I say that the orders ‘Bring me sugar’ and ‘Bring me milk’ make sense, but not the combination ‘Milk me sugar,’ that does not mean that the utterance of this combination of words has no effect. And if its effect is that the other person stares at me and gapes, I don’t on that account call it the order to stare and gape, even if that was precisely the effect that I wanted to produce.”64 But it is crucial to note that to bar the exaggerated view that a word means whatever effect it happens to have, one does not need to retreat to the thesis that every instance of language use must be rule-governed or else parasitic on such uses.65 For if understood as objects of comparison, the instrumental and game conceptions of language do not imply the possibility—much less the necessity!—of explaining language exclusively from one of these perspectives. Rather, each conception accentuates a particular aspect of language. Accordingly, the worry that allowing for purely instrumental cases concedes too much finds a foothold only if one assumes that such examples should constitute the ground for universal theses about language. But this is precisely not how the role of examples should be understood. To avoid the problem of dogmatism and injustice, individual examples are not to be used as the basis of theses about what other cases must be, but as objects of comparison, as explained in 3.4. When this methodological correction is made, there is no problem with allowing for purely instrumental cases, or indeed with using such examples for the purposes of clarification.66 Therefore, pace Hacker, Wittgenstein’s abandonment of “the hard line” in manuscript 110 is not an anomaly requiring explanation.67 Once we give up the assumption that one should identify Wittgenstein’s conception of language exclusively with the game conception (which assumption appears to be the root of Hacker’s problem in interpreting §492), both the problem of “conceding too much” and the apparent anomaly concerning Wittgenstein’s development are dissolved.68 Hence we seem to fare better in the interpretational task if we incorporate the lessons of Wittgenstein’s turn into the interpretation of his view of language and take the conception of language as a rulegoverned practice as an object of comparison. More generally, a very important feature of Wittgenstein’s method is that it enables one to take into account various features of one’s object of investigation that would, if turned into bases of philosophical theses about the essence of the object, exclude each other from consideration.

Grammar, Meaning, and Language

171

Thus adopting Wittgenstein’s method, one is able, for instance, to do justice to both the instrumentality and the arbitrariness of language and does not need to maintain rigidly that one of these aspects alone should explain the functioning of language or be the essential feature of language. Rather, which features are to be regarded as essential depends on one’s more specific concerns (the particular issues one is discussing). In this sense Wittgenstein’s method has the advantage of leading to a certain increase in the flexibility of philosophical thought.69 Accordingly, taking Wittgenstein’s conception of the arbitrariness of grammar on board, along with whatever clarificatory benefits it might bring, does not presuppose the thesis that language must be rulegoverned. Rather, the conception can be taken to apply to language to the extent that it is rule-governed or comparable to rule-based games. The clarificatory force of the conception of the arbitrariness of linguistic rules, in other words, is relative to the aptness of the model of language as rule-governed in the various contexts in which the model is used. Crucially, therefore, this relativity does not imply the loss of any strength or philosophical momentum Wittgenstein’s conception may have. For wherever language is aptly regarded as rule-governed, the points based on the conception of the arbitrariness of grammar apply with their clarifying force. But where language is not aptly characterized in this way, these points are not justified anyway, instead constituting a potential source of philosophical confusion. Hence nothing of the problem-solving capacity of Wittgenstein’s conception of the arbitrariness of grammar is lost when the picture of language as rulegoverned is understood as an object of comparison. We have only gained in justice and lost in dogmatism.70

4.5 Examples as Centers of Variation and the Conception of Language as a Family The present interpretation of Wittgenstein’s development can also be explained as follows. While manuscript 110 suggests that the comparison of language with games (the conception of the use of language as a rule-governed activity) provides a principle that determines the concept of language, in the early 1930s Wittgenstein gave up the search for such a unifying principle. This means giving up the search for the definition

172

Grammar, Meaning, and Language

of the concept of language or a thesis of the essence of language by reference to certain allegedly necessary characteristics of language. Giving up the search for such a unifying principle, a once-and-for-all determination of the concept of language, Wittgenstein comes to adopt the view of language as a family-resemblance concept. But it is important to understand why this does not constitute an alternative thesis about language, and what it means not to use the concept of family resemblance to articulate theses. As regards the contrast between the definition of language by reference to necessary conditions and the characterization of language as constituting a family of cases, it would be contradictory to maintain both that language necessarily presupposes rules and that it is a family of cases. More precisely, to define language by rule-governedness as a necessary condition while simultaneously characterizing it as a family is contradictory insofar as the two conceptions are put forward as exclusive theses about the nature or the concept of language. For to characterize language as a family in Wittgenstein’s sense is to say that the different cases falling under this concept need not have anything in common that allows the application of the concept of language to them.71 The definition of rule-governedness as a necessary condition of language, however, when employed as a thesis, entails that every instance of language use is either rule-governed or parasitic on such uses. Thus rule-governedness is posited as a principle that defines universally the concept of language.72 Wittgenstein writes about family-resemblance concepts, referring to the idea that a concept must be defined by a common feature: “This notion is, in a way, too primitive. What a concept-word indicates is certainly a kinship between objects, but this kinship need not be the sharing of a common property or a constituent. It may connect the objects like the links of a chain, so that one is linked to another by intermediary links. Two neighbouring members may have common features and be similar to each other, while distant ones belong to the same family without any longer having anything in common.”73 Wittgenstein’s method of using examples as “centers of variation” for the purpose of characterizing family-resemblance concepts can further elucidate the notion of family resemblance. Such centers of variation are essentially examples exhibiting clearly some feature(s) characteristic of the different cases falling under a concept. In this capacity the examples can be used as reference points in arranging the cases under a concept into a perspicuous order.

Grammar, Meaning, and Language

173

Wittgenstein writes about this use of examples, commenting on the philosopher’s inclination to declare one case or one characteristic the essential one: This is the case when we are asked: what is the essence of punishment and now one says that every punishment is really society’s revenge, and another, its essence is deterrence etc. But are there not typical cases of society’s revenge and typical cases of deterrent measures and others of punishment as reform, and countless mixtures and intermediate cases? If we, therefore, were asked about the essence of punishment, essence of revolution, of knowledge, of cultural decline or refined sense for music—we should [insertion: would] not try to give something common to all cases, not what they all really are, that is, an ideal which is contained in them all; but instead of this examples, as it were centres of variation.74

The example of “boo,” for instance, might be used as a center of variation. Admittedly, “boo” is in the periphery of language, a kind of limiting case. However, it is precisely for this reason that it can be used illuminatingly to characterize the concept of language. It exhibits a particular feature of language—its instrumentality—in a purified form. Correspondingly, with §492 in mind, one might present arbitrary stipulations with no real practical use as another center of variation at the opposite extreme of language. For example, I might stipulate that “oob” means the color of my eyes, but only when there are exactly 159 persons within a square mile with the speaker at the center. With such a definition, “oob” is entirely useless, a mere caricature of a linguistic expression. It is empirically impossible for anyone to use it, except in specially arranged circumstances. But because it exhibits the feature of rule-governedness, one may say that “oob” is a possible linguistic expression. “Boo” and “oob” would then appear to be the kind of distant family members with no common features that Wittgenstein talks about in the passage from manuscript 140 above. In any case, whatever features “boo” and “oob” may share, these features are not essential to their functioning as the expressions they are. Nevertheless, “boo” and “oob” can be united under the same concept through other cases—intermediary links—with which they do share features. (Presumably, most linguistic expressions both are governed by rules and possess instrumental value.) Accordingly, “boo” and “oob” can be used as examples to explain what it means to say that language constitutes a family of cases. With “boo” and

174

Grammar, Meaning, and Language

“oob” as centers—or limits—of variation, the concept of language is not defined by a feature common to all cases but, as Wittgenstein puts it, by “similarities overlapping and criss-crossing.”75 The point of using “boo” and “oob” to characterize the concept of language is simply that they exhibit in a clear fashion the two characteristics of language that Wittgenstein identifies in §492. It is important to emphasize, however, that one would again be slipping back to the dogmatic mode of philosophizing if one said that the cases singled out as centers of variation really are the ones that must be regarded as the central or essential cases. For then one would again be making a claim about the objects of investigation, establishing the true philosophical order of language. The only difference would be that instead of declaring one case the central and essential one on which other cases are to be modeled, one would be giving several cases core status. But philosophical orderings of concepts are to be seen as relative to particular contexts of discussion or to particular interests. What is an appropriate order depends on the philosophical issues one seeks to clarify. In other words, which aspects of the concept need to be highlighted by the arrangement of examples is determined by one’s more specific concerns.76 It is therefore crucial to keep in mind that the singled-out cases have the status of a center of variation in the context of a possible way of arranging the material under investigation—where “material” refers to the different cases that fall under the concept of language or punishment, for instance. Arranging cases of punishment by using examples of revenge, deterrence, and so on as points of reference serves the purpose of making perspicuous the whole constituted by the concept of punishment, creating an overall picture of this concept. And although competent users of the concept already possess a “practical” understanding of individual cases that fall under it (they are able to tell punishments from rewards, and so on), such an arrangement of the cases can further clarify the concept for them by reminding them of the different kinds of cases that make up the concept and by explicating the relations between them. But such an arrangement of instances of punishment into a whole must not be projected onto reality as the order that constitutes the essence of punishment or the characterization of the concept. Rather, the cases singled out are central simply to this specific arrangement. The arrangement based on selected examples as

Grammar, Meaning, and Language

175

centers of variation is a picture of the concept of punishment, a means of presentation to be employed as an object of comparison. Why Wittgenstein’s characterization of language as a familyresemblance concept does not constitute a thesis about language (or the concept of language) can be explained in essentially the same way. As noted, the concept of family resemblance spells out a conception of the unity of concepts. The crucial difference between the employment of this conception, as Wittgenstein understands it, and the employment of a philosophical thesis is that to characterize a concept as a familyresemblance concept in Wittgenstein’s sense does not exclude the possibility of also using a precise definition (such as a definition in terms of necessary conditions) to characterize the same concept. Accordingly, immediately after having explained the notion of family resemblance in manuscript 140 (in the remark quoted above), Wittgenstein proceeds to discuss the possibility of employing precise definitions to “clear up philosophical paradoxes” relating to concepts with imprecise borders, such as family-resemblance concepts.77 Clearly, he sees no conflict here. The possibility of characterizing a concept simultaneously as a familyresemblance concept and in terms of a precise definition depends entirely on the use of these characterizations as objects of comparison. For when used as an object of comparison, a definition “only” articulates a particular way of conceiving the object of investigation or a mode of presenting it, not a truth or a “grammatical quasi-truth” about the object (see 3.4). Unlike a true/false thesis that excludes any conflicting theses, a way of conceiving things does not exclude other ways of conceiving it. It is perfectly possible, for instance, to regard language as a rule-governed practice in the context of a certain discussion and on the model of instruments in another context. What is an adequate conception in each case depends on the particular questions one is addressing. Consequently, we can now see why the gain in the flexibility of philosophical thought that Wittgenstein’s method makes possible does not come at the expense of rigor: in specific contexts and for specific purposes it is always possible to employ precise and rigorous definitions, given that this is what reaching clarity about the matters at hand requires.78 Finally, note that the two conceptions of language discussed above—language as comparable to games and as comparable to instruments—are by no means the only ones Wittgenstein develops

176

Grammar, Meaning, and Language

and deploys for the purpose of clarification. Another important conception is spelled out in his comparison between language and primitive, natural expressions.79 Here too, as in the case of the comparison with instruments, language is not regarded as simply arbitrary and based on conventions. Rather, to understand the conventional expressions of pain as extensions of primitive, prelinguistic expressions is to see linguistic conventions as a top layer in what are essentially nonconventional forms of behavior.80 In summary, Wittgenstein’s view of language is richer than an interpretation that attributes to him a thesis of language as rule-governed allows one to think. As he writes: “We may regard our language from a simple point of view. But it is a joined up kind of a formation, a river in which waters from very different sources run (together).”81 His comparisons of language with games and instruments and the idea of language as an extension of primitive behavior bring to light some such points of view. But once again, adopting a particular point of view should not be confused with stating a truth about the object of investigation. Similar observations about the richness of Wittgenstein’s views can also be made in connection with his remarks on the concept of meaning, to which I now turn.

4.6 Avoiding Dogmatism about Meaning If one does not treat the definition of meaning as constituted by rules as the definition of word-meaning, other characterizations of cases falling under the concept of the meaningful use of words become available. It is now possible to be less restrictive as to what the meaningfulness of linguistic expressions might be. Nevertheless, just as characterizing the concept of language through centers of variation does not mean that one has to accept just anything as belonging to language, orderliness and rigor are similarly retained in the case of meaning. Let us look at this more closely. Introducing the conception of meaning as use in his lectures, Wittgenstein states: “The use of a word is what is defined by the rules, just as the use of the king of chess is defined by the rules. And just as the shape and material of the king of chess are irrelevant, so are the shape and sound of a word to its use.”82 From the point of view that the meaning of a word is defined by rules for its use, the sound of a word, therefore, is irrelevant to its meaning. This does not mean, of course, that a

Grammar, Meaning, and Language

177

philosopher who regards rule-governedness as a necessary condition of meaning cannot recognize sound as relevant to meaning at all. She may well admit that there are cases in which the sound of words is relevant, i.e., that rules alone do not account for meaning.83 According to this conception, after all, rules are necessary but not sufficient for meaning. Wittgenstein too discusses examples that fit this description.84 In the Investigations he writes: “We speak of understanding a sentence in the sense in which it can be replaced by another which says the same; but also in the sense in which it cannot be replaced by any other. (Any more than one musical theme can be replaced by another.) [New paragraph] In the one case the thought in the sentence is something common to different sentences; in the other, something that is expressed only by these words in these positions. (Understanding a poem.)”85 To ask what a sentence says or which thought it expresses is to ask what it means. Hence this remark may well be said to concern meaning. Moreover, although it may seem to concern sentence-meaning rather than word-meaning, it does say something about word-meaning, too. For clearly, to say that the thought in a sentence can sometimes be expressed “only by these words in these positions” is to comment on the role of words in expressing thoughts. It says that a word in a sentence cannot always be replaced by any of its synonyms according to linguistic conventions. However, such cases clearly indicate that the contribution of a word to the meaning of a sentence (or the thought it expresses) is not always merely a matter of rules governing the use of the word. If the word’s contribution to the sentence were simply determined by rules, then it would be possible, for example, to substitute for the word an equivalent one according to linguistic conventions without affecting the meaning of the sentence.86 But as other remarks in the context of §531 indicate, Wittgenstein seems willing to acknowledge the possibility of talking about meaning also in cases where the sounds used as expressions are not governed by any rules at all. As he writes: “It would be possible to imagine people who had something not quite unlike a language: a play of sounds, without vocabulary or grammar. (‘Speaking with tongues.’)” And “ ‘But what would the meaning of the sounds be in such a case?’—What is it in music?”87 What these remarks bring to light is that although Wittgenstein does not claim that we must regard “speaking with tongues” as a language, he also makes no attempt to deny the possibility of talking about meanings in the case of such sounds. Indeed, in manuscript 180b,

178

Grammar, Meaning, and Language

where §528 and §529 are drafted, he simply notes as a comment: “And if one wants to talk about meaning here, it does not reside in the use of words.”88 It is clear from the context of Wittgenstein’s remarks in manuscript 180b that under the concept of “speaking with tongues” he includes, for instance, expressing feelings and sensations with sounds or conveying emotion with tone. But do expressions like cries of pain and sounds that people make to express contentment, satisfaction, joy, and so on have meaning? As in the case of “boo,” the relevant sounds are not conventional or are conventional only in part. Unlike in the case of “boo,” however, one may well speak of understanding in connection with such expressive sounds. Whether this latter feature is enough to show the appropriateness of talk about meaning in this connection is not entirely clear. (We are at the border area of the concept here, after all.) Perhaps the answer, therefore, is that such sounds are extremely meaningful in human interaction, and if they have meaning, it does not lie in rule-governed use.89 Regardless of how one is inclined to answer the above question, it is evident that although Wittgenstein wishes to distinguish “speaking with tongues” from cases of meaning as use, he shows no inclination to claim that it would be legitimate to talk about meaning in the latter cases only. Rather, in the context of his discussion of “speaking with tongues,” he also writes: “The meaning of a word, I said, is its use. But an important supplement must be added to this.”90 Hence, far from insisting that one can only speak of meaning insofar as the use of an expression is governed by rules, Wittgenstein thinks that the view of meaning as use should be supplemented by a discussion of cases such as “speaking with tongues.” Significantly, this supplement is included in the Investigations, where its purpose is apparently to foreground this “non-rule-governed” dimension’ of meaning, in contrast to the conception of meaning as constituted by rules. Interestingly, the case of “speaking with tongues” is juxtaposed in the Investigations with another case, apparently at the other end of the spectrum, where “the ‘soul’ of the words played no part. In which, for example, we had no objection to replacing one word by another arbitrary one of our own invention.”91 Obviously, this language wholly determined by conventions is quite different from any natural language, whose expressions seem full of meaning to its users and are anything

Grammar, Meaning, and Language

179

but replaceable. Indeed, unlike Wittgenstein’s soulless language, natural languages might be characterized by saying that the history of a language, interwoven with the history of the people who speak it, constitutes yet another determinant of meaning not explainable by reference to linguistic rules. For example, the use of an expression in the context of a particular political project may result in a burden of associations that make it impossible for others to use the expression. Here the rules of language have not been changed. One might say that in principle such an expression still means what it used to before its political adoption. In practice, however, it may be impossible to ignore how its use in the context of the project has affected its meaning.92 Such changes in meaning are impossible by definition in Wittgenstein’s soulless language. But because of this, and its contrast with “speaking with tongues,” the soulless language constitutes a useful center of variation for clarifying the concept of meaning. Like the examples of “boo” and “oob,” “speaking with tongues” and the soulless language are able to throw light on different aspects of the concept of meaning precisely because of their extremity. A final example in which meaning seems governed not by rules but by sound is onomatopoeic words. As Wittgenstein says: “In the case of an onomatopoeic word the sound belongs to the symbol.”93 Although there are onomatopoeic words with established uses, I may also invent new ones and be understood. Thus the possibility of understanding seems not based on any preestablished rules. Instead, an onomatopoeic word is, so to speak, iconic; its use is based on an understanding of a similarity between the sound and whatever it stands for. And although one might wish to problematize the notion of similarity here, it is clear that onomatopoeic words are very far removed from Wittgenstein’s soulless language, where everything depends on rules. To replace a word with another arbitrary one according to rules is precisely what one cannot do in the case of onomatopoeic words. Accordingly, Wittgenstein compares onomatopoeic words with color samples. In this context he distinguishes onomatopoeic expressions from words on the basis that words, unlike samples, are not compared with anything. Characteristically, however, he again warns against turning this distinction into a doctrine about language: “With this distinction we have not established an ultimate duality in logic; but only brought to the fore two characteristic types among the means of language.”94

180

Grammar, Meaning, and Language

To sum up, as in the case of Wittgenstein’s remarks on the concept of language we discern in his discussion of meaning a tone and pattern different from what is possible for a philosopher committed to a thesis about the rule-governedness of meaning. There is no attempt to explain away or exclude cases that do not fit the model of meaning as use. Instead of emphasizing uniformity, Wittgenstein takes great interest in the various cases that might be taken to fall under the concept of meaning. Apparently, the purpose is the same as with the concept of language: to characterize the concept through examples and definitions used as centers of variation and objects of comparison. Similarly, this interpretation does more than allow one to make better sense of Wittgenstein’s remarks about meaning. With the help of this method, one also seems able to do better justice to the manifoldness of the concept of meaning than is possible by imposing a particular unifying definition upon it. Hence it seems one would indeed benefit from taking the conception of meaning as constituted by rules as an object of comparison rather than a thesis about the essence of meaning.95

4.7 Wittgenstein’s Methodological Shift and Analyses in Terms of Necessary Conditions A goal of my discussion in this chapter has been to render more concrete Wittgenstein’s methodological remarks on what it is and is not to have theses in philosophy and to demonstrate how the lessons of Wittgenstein’s turn apply to what he says about the concepts of meaning and language. A clear comprehension of the methodological import of his turn is particularly important because it allows one to avert the danger of repeating the dogmatism of the Tractatus in describing the uses of the expressions of ordinary language. As explained in 3.3, there is a risk that the descriptions one provides—or descriptions one attributes to Wittgenstein—become a novel set of philosophical requirements that language must allegedly meet. Modifying Hacker’s characterization of metaphysics as a shadow cast by grammar on reality,96 one might therefore say there is a danger that grammar becomes a shadow of metaphysics on language. To guard against this possibility, one must keep clearly in mind the methodological insights that constitute Wittgenstein’s turn.

Grammar, Meaning, and Language

181

As I have tried to illustrate, an important consequence of this methodological shift is a certain increase in the flexibility of philosophical thought. Adopting Wittgenstein’s approach, the philosopher is no longer committed to explaining manifold cases falling under a concept through a single definition. Instead she can employ several such definitions as the different cases and contexts of discourse require. Hence one is released from a certain kind of disadvantageous asceticism in philosophy based on the idea that one must look for unifying characteristic features of things or unique definitions of essences. Giving up this asceticism, one is in a certain sense making things easier for oneself. Wittgenstein alludes to this in contrasting his approach with that of Plato or Socrates (or a certain received view of them): “Our approach is contrasted with Plato’s. Socrates pulls up the pupil who when asked what knowledge is enumerates cases of knowledge. And Socrates doesn’t regard that as even a preliminary step to answering the question. [New paragraph] But our answer consists in giving such an enumeration and a few analogies. (In a certain sense we are always making things easier and easier for ourselves in philosophy.)”97 My explanation of the use of examples as centers of variation and as objects of comparison is to be understood as specifying what Wittgenstein means in this remark by his use of examples and how he intends the contrast between his approach and that of Socrates. For it is not true, of course, that philosophers following Socrates would not use examples. But to a great extent (though this might not be a particularly apt characterization of Plato), examples are employed as something from which theses about the essence of the object of investigation are derived. It is this use of examples that Wittgenstein wishes to reject and replace by the employment of examples as centers of variation and as objects of comparison. In justification of this last interpretative suggestion, note that the remark on Socrates occurs immediately after one in which Wittgenstein characterizes his approach by comparing it with the description of the contours of blurred color patches with pen and ink (quoted in 3.6). The key point is that his characterizations of concepts—in terms of examples, for instance—are not meant to capture reality exactly as it is, to state something true of all the manifold cases falling under a concept that underlies their deceptive manifoldness. Such an underlying unity is what one would be able to see if able to comprehend a Platonic form. As I have explained, however, Wittgenstein’s characterizations of

182

Grammar, Meaning, and Language

concepts are descriptions given in particular contexts where they are meant to serve certain well-specified goals of clarification: the dissolution of particular problems. I have contrasted the method of analysis in terms of necessary conditions with clarification in terms of centers of variation, suggesting that the latter method can help one avoid dogmatism. By this I do not wish to suggest that analyses in terms of necessary conditions are somehow inherently dogmatic. Rather, the threat of dogmatism arises when this mode of analysis is combined with the assumption that conceptual unity must always be determined by reference to a characteristic (or characteristics) common to all cases falling under the concept. This characteristic would then be what the analysis brings to light. But when this assumption about the unity of concepts is problematized, it can no longer be taken for granted that a given feature x, which the analysis of certain exemplary cases shows to be a necessary condition of those cases, is a necessary condition of all cases falling under the concept in question. The legitimacy of this line of reasoning depends on the assumption that a concept must have a single, unified essence. Accordingly, giving up this assumption should profoundly affect one’s perception of the status of definitions in terms of necessary conditions. Rather than becoming the basis for a thesis about what all cases must be, such a definition extracted from certain exemplary cases should be comprehended as an object of comparison. Put another way, a definition in terms of necessary conditions cannot be presumed to reveal the limit of the concept in question. Rather, it articulates a particular way of drawing limits to the concept. It should be noted, however, that I am not denying that the unity of concepts might sometimes depend on a unifying characteristic common to all cases. Being unmarried and male may be a necessary condition of bachelorhood, and in such a case an analysis in terms of necessary conditions may be able to capture something that holds universally for cases falling under this concept. But my point is that one cannot take it for granted that such a universal analysis can always be given and that what holds for some examples holds for all.98 The next chapter focuses on Wittgenstein’s remarks on the concepts of grammar, essence, and necessity and continues the attempt begun here to make Wittgenstein’s methodological ideas and innovations

Grammar, Meaning, and Language

183

more concrete by applying them to the interpretation of remarks on central topics in his later work. In addition, I try to further illustrate the potential advantages of Wittgenstein’s method by explaining how it opens up a way beyond certain established oppositions relating to the concept of necessity and beyond the positions of certain competing philosophical parties.

FIVE

The Concepts of Essence and Necessity

THIS CHAPTER DEVELOPS an interpretation of Wittgenstein’s later conception of essence and necessity as something “expressed by grammar,” with the purpose of further elucidating his conception of the method of philosophy and what it is to philosophize without theses. This interpretation of Wittgenstein on essence and necessity stands in contrast with widely accepted conventionalist or constructivist readings (described in 5.1) that attribute to Wittgenstein a conception of essences and relevant kinds of exceptionless necessities as grammatical constructions. The problem with these readings, I argue, is that they misidentify the kind of account he is offering (5.2). Rather than presenting grammar or linguistic practices as the source of necessity and providing a philosophical thesis of the nature of necessity and essence, Wittgenstein avoids any commitment to such claims. Instead (as argued in 5.3 and 5.6) he is concerned to clarify the notion of an investigation of essences and address philosophical problems relating to the concept of necessity by characterizing the investigation of essences as a grammatical investigation. Two key issues addressed in this chapter are the relation of statements about essences and necessities to factual statements and Wittgenstein’s (largely ignored) conception of the nontemporality of grammatical statements (5.4 and 5.5). On the one hand, the discussion of these issues further clarifies Wittgenstein’s conception of the status of grammatical statements, as addressed in Chapter 3. On the other hand, it also allows me to relate his conception to certain more traditional

The Concepts of Essence and Necessity

185

philosophical views of essence and necessity, from whose perspective essential or necessary relations emerge as super-rigid factual relations or as immutable. The discussion of the nontemporality of grammatical statements in particular allows me to explain how Wittgenstein’s conception makes room for what is traditionally perceived as the timeless nature of essences. Similarly to Chapter 4, this chapter aims to clarify the way in which Wittgenstein’s approach, when interpreted in accordance with his rejection of philosophical theses, leads to a richer account of the phenomenon of necessity than seems possible in terms of such theses. More specifically, as I will explain, although Wittgenstein is highly critical of attempts to explain exceptionless necessities in terms of facts, this does not imply that he wishes to explain such necessities as dependent on linguistic conventions. Rather, it is a key feature of his conception, as I interpret it, that it does not seek to reduce necessities to either the factual or the conventional, or try to explain the complex phenomenon of exceptionless necessity in terms of only one of these two factors. Similarly to his conception of language and meaning, his account of necessity is designed to do justice to both factuality and conventionality. In other words, although Wittgenstein may be read as offering a critique of an Aristotelian realist view of necessity, his conception does not constitute an affirmation of a Kantian idealist or constructivist view. Instead, it opens up a path beyond these competing accounts. My discussion of this issue (in 5.7), as well as of how Wittgenstein positions himself in relation to the dispute on essences between the Platonists and those who view essences as something historically determined, is also intended to clarify what Wittgenstein means by saying that in philosophy one should not take sides or create parties, but the only task is to be just.1

5.1 Constructivist Readings and the Arbitrariness/ Nonarbitrariness of Grammar Wittgenstein writes in the Investigations: Essence is expressed by grammar. Consider: “The only correlate in language to an intrinsic necessity is an arbitrary rule. It is the only thing which one can milk out of this intrinsic necessity into a proposition.” Grammar tells what kind of an object anything is.2

186

The Concepts of Essence and Necessity

There is a temptation (that many have found hard to resist) to interpret Wittgenstein as asserting in these remarks that essences and associated necessities are grammatical constructions. According to this view, which I will call the “constructivist reading,” what is essential or necessary depends on the rules for the use of language, that is, on grammar or grammatical rules. Such rules constitute the source of necessity and fix what is essential or necessary. Let us start by looking more closely at this interpretation and how one may be led to adopt it. A way to argue for the constructivist interpretation is to support it by a particular reading of §372 on the arbitrariness or autonomy of grammar. By the arbitrariness of grammar Wittgenstein means that facts of nature or necessities in nature do not make logically necessary any particular conceptual formations or grammatical determinations. Thus grammar cannot be derived from facts or justified by reference to them.3 When the idea that essences find their expression in grammar is combined with the observation that grammar is not derivable from the facts of reality, the conclusion may suggest itself that essences and necessities are the product of grammatical determinations. Since it is not the facts of reality, the reasoning goes, it must be our linguistic conventions that determine the criteria for the identity of things and their essential or necessary characteristics. Baker and Hacker express a view of this kind when they write: “it is our rules for the use of colour-words which create what we call ‘the nature (the essence) of colour.’”4 Correspondingly, Hacker talks about essences as something “made rather than found” and as “a product of convention.”5 According to the constructivist reading, essences are therefore not something out in the world to be discovered, but the reflections of grammatical conventions on the world. Or as Hacker sums up this view: “Necessities in the world are the shadows cast by grammar.”6 When §372 is read as entailing the view that grammar creates essences, Wittgenstein is taken to establish an order of determination, so to speak. Rather than being determined by the external world as an independent reality, necessities and the essences of things are determined by grammar. At this point, differences emerge between readings that I am characterizing as constructivist. For example, Newton Garver and Michael Forster interpret Wittgenstein as articulating an idealist position, according to which the structure of the world is a grammatical construction, the result of the imposition of grammatical

The Concepts of Essence and Necessity

187

structures onto the world by a linguistic community or communities, analogous to the projection of categories onto the world by the Kantian experiencing subject.7 In contrast, Baker and Hacker reject such a transcendental idealist reading, reserving “projection” for use as a term of criticism. For them, projections of the above type are something traditional metaphysicians succumb to, conflating grammar and reality and turning grammatical structures into a priori structures of reality.8 Accordingly, Hacker characterizes Wittgensteinian conceptual investigation by saying: “a conceptual investigation will not produce insights into the nature of the world, but only into the grammar of our descriptions of the world.”9 Despite these differences, the above interpretations all agree that essences and necessities as objects of discourse are a grammatical construction and that grammar constitutes the source of necessity. Notably, this agreement about grammar as the source of necessity extends even further among Wittgenstein’s interpreters in the form of the widely held view of Wittgenstein as a conventionalist (see references below). But to see how Wittgenstein’s presumed conventionalism differs from certain more traditional forms of conventionalism, and what is at stake in either embracing or rejecting the constructivist reading, we must discuss the sense in which grammar, according to Wittgenstein, is nonarbitrary. As Baker and Hacker (among others) note, the arbitrariness of grammar in Wittgenstein’s sense does not mean that the rules followed in language use are a matter of choice or decision, something that can be simply stipulated or easily altered.10 Rather, grammar has an aspect of nonarbitrariness. For instance, Wittgenstein characterizes our system of color concepts as both arbitrary and nonarbitrary.11 As he remarks in the same context, that we organize things in this particular way, i.e., employ these particular color concepts, is no more arbitrary than that we nourish ourselves with certain things rather than others.12 The idea of the nonarbitrariness of grammar may be briefly explained as follows. Although grammar and concepts cannot be derived from reality, this does not mean that language would not depend in various ways on regularities of both external and human nature (e.g., neural or psychological facts) or that language would not be molded by such facts of nature. Language, as Wittgenstein says, is part of a form of life that is not merely based on conventions.13 Or as he also remarks: “Concepts with fixed limits would demand a uniformity of behaviour.

188

The Concepts of Essence and Necessity

But where I am certain, someone else is uncertain. And that is a fact of nature.”14 Wittgenstein’s (alleged) conventionalism, then, is not of a voluntaristic type, so to speak, according to which language users are in principle free to decide what conventions they adopt. But taking into account these and certain other reservations, Baker and Hacker, among others, are nevertheless willing to characterize Wittgenstein’s account of necessity as a conventionalist one. They write: “if after grasping the ways in which his analysis differs from the forms of conventionalism propounded in the 1930s and 1940s, we wish to classify it, there is surely some justice in saying that it is a conventionalist account.”15 This classification of Wittgenstein as a conventionalist of a particular kind can be used as a clue in interpreting what Baker and Hacker mean by saying that essences and related necessities are a grammatical creation (which they do not elucidate at any length). Following this lead, the constructivist interpretation might be characterized as follows: It is not that nature has nothing to say with respect to what is essential or necessary, as if this were purely a matter of choice. But whatever nature has to say, grammar has the last word. Without grammar, there is no necessity in the strong, exceptionless sense connected with the concept of essence. Grammatical rules fix concepts, and therefore also what is essential and necessary.16 Nevertheless, it is by no means obvious that this is how Wittgenstein’s conception of essence and necessity should be understood, particularly as it seems to commit Wittgenstein to a philosophical thesis about the nature of essences and necessity, as I will explain. Let me next spell out some problems with the constructivist reading.

5.2 Problems with Constructivism First, one may problematize the constructivist interpretation of Wittgenstein’s conception of the arbitrariness of grammar. Wittgenstein’s point about the arbitrariness of grammar might be described as “merely negative,” i.e., as a rejection of the justification of a grammatical rule by reference to facts. But if the point is “merely negative” in this sense, then grammar not being derivable from facts does not mean that necessities expressed by grammatical rules must be the creation of grammar, or that what one might be tempted to conceive as intrinsic necessities of nature are only a projection of grammar on reality. The

The Concepts of Essence and Necessity

189

thesis about the creation of essences by grammar is a further claim, one not entailed by the conception of the arbitrariness of grammar, as I will explain.17 In this connection it is notable that Wittgenstein does not state anywhere in so many words that essences are created by grammar. In fact, he makes remarks that suggest an interpretation different from that offered by the constructivists. He writes about the necessity relating to the exclusion of reddish-green from our color vocabulary and experiential register: “I want to say that there is a geometrical gap, not a physical one, between green and red. [New paragraph] But doesn’t anything physical correspond to it? I do not deny that. (And suppose it were merely our habituation to these concepts, to these languagegames. But I am not saying that it is so.)”18 The characterization of the gap between red and green as geometrical rather than physical is based on Wittgenstein’s comparison between grammar and geometry as well as mathematics,19 which is discussed in more detail in 5.4. In his view, the gap constituted by the absence of an intermediate color is best described as geometrical because it is part of the concepts of red and green that they have no intermediate color. Accordingly, the statement about the absence of reddish-green is not a generalization from particular cases and not a factual statement. Rather than concerning any particular occasions of the use of our color concepts, the statement expresses a rule of language or describes a relation between the concepts of red and green. Just as statements about geometrical relations, according to Wittgenstein, concern internal rather than external relations, so does the statement about the red-green gap. Nevertheless, as Wittgenstein notes, to characterize the gap as geometrical does not mean denying that anything physical corresponds to it. And more specifically, as he also says, his allowing that something physical might correspond to the gap is not meant to suggest that this physical counterpart would be merely our habituation to certain concepts, where habituation may be conceived as the result of a process of training someone to use language in specific ways or of bringing someone up to participate in a particular form of life.20 Hence the remark may be read as denying that the characterization of the red-green gap as geometrical, i.e., grammatical, entails any commitment to the idea that the necessity of the exclusion of reddish-green should be explainable by reference to our adoption of certain linguistic practices or

190

The Concepts of Essence and Necessity

grammar. How this fits Wittgenstein’s conception more generally will become clearer later on. Rather than treating the necessity in question as created by grammar or language-games, Wittgenstein might, therefore, be interpreted as leaving open the source of necessity, i.e., whether it is because of our adoption of a certain grammar or because of some physical regularities in (our or external) nature that the exclusion of reddish-green appears necessary. (I return to this in detail in 5.6.) But if so, then it is not correct to say Wittgenstein treats grammar as the source of necessity. In any case, the above considerations suffice to problematize the inference that because Wittgenstein rejects the thesis that facts of nature determine grammar and what is necessary, he affirms the opposite thesis that grammar determines what is necessary. Instead, as I will argue, Wittgenstein is merely concerned to clarify the status of statements about essences and necessities by characterizing them as grammatical statements. Crucially (as I explain in 5.6 and 5.7), one can accept this clarificatory point without taking a stand on the issue of the source of necessity or adopting the thesis about essences as grammatical creations. My critical observation can also be explained as follows. According to Wittgenstein, someone “who talks about an essence is merely noting a convention.”21 Contrary to how it might appear at first sight, however, this statement cannot be taken to lend any support to the interpretation that essences are a product of conventions. What Wittgenstein says is that in making a statement about an essence one is taking note of or stating a convention. But to say that to talk about an essence is to note a convention is not yet to say that essences are the product of conventions or that conventions are the source of necessity. Second (but not unrelated to the preceding), the constructivist reading can be problematized on the grounds that it is in conflict with Wittgenstein’s remarks on his philosophical goals and practice or his methodology. For given Wittgenstein’s denial that he is putting forward any philosophical theses, it is problematic to attribute to him a thesis about the nature of essences and necessities as grammatical creations. But insofar as the constructivist interpretation seeks to provide a unified once-andfor-all account of the phenomenon of necessity, claiming that the source of necessity must always be grammar, it seems to amount to exactly such a thesis.22 More specifically, one might also say that in explaining

The Concepts of Essence and Necessity

191

essences and necessities as the product of linguistic practices, the constructivists are (in effect, perhaps contrary to their self-understanding) providing a reductive account of a manifold of phenomena, i.e., the various instances of exceptionless necessity, in terms of a single explanatory factor. For in so doing they are ultimately asserting that what one may mistakenly take to be different phenomena—for example, the necessity of the chess king moving only one square at a time and the exclusion of reddish-green—are ultimately really manifestations of one underlying regularity: grammar’s production of necessities. But this is precisely the kind of explanation that Wittgenstein seems to oppose when he writes, for example: “we do not want to say that this is really that, but only, as far as we are capable, point out similarities and differences.”23 And: “I want to say that . . . it can never be our job to reduce anything to anything, or to explain anything. Philosophy really is ‘purely descriptive.’”24 As this chapter intends to show, it is possible to spell out Wittgenstein’s conception of essence and necessity in such a way that it involves no claims about essence and necessity in the style of reductive scientific explanations that seek to account for phenomena in terms of the smallest possible number of explanatory factors. By contrast, due to its attribution of such an explanation to Wittgenstein, constructivism emerges as exemplifying just the kind of “craving for generality” and “contemptuous attitude towards the particular case” against which Wittgenstein warns in the context of the passage above. (Recall also passages quoted in Chapter 3 in which he rejects assertions about what something must always be as dogmatic, and as confusing factual and grammatical statements.) In any case, the preceding establishes that one cannot simply assume (without further evidence and argument) that Wittgenstein’s goal is to provide a thesis of what necessity must be, in which grammar figures as an explanatory factor. Similarly, as I explain in more detail in 5.7, by attributing to Wittgenstein a thesis about the creation of essences by grammatical rules, the constructivist interpretation reads him as taking sides in philosophical debates such as realism versus idealism and Platonism versus historicism. This, however, goes against his expressed view that in philosophy one should not take sides.25 Hence, here too constructivism seems to be in conflict with Wittgenstein’s methodology. These considerations suggest that an alternative to the constructivist reading is needed. The rest of this chapter outlines such an alternative.

192

The Concepts of Essence and Necessity

5.3 The Methodological Dimension of Wittgenstein’s Conception of Essence Rather than using §372 as the key to the interpretation of §371 in the manner of constructivism, I suggest using the immediately preceding remark, §370. This brings the methodological aspect of Wittgenstein’s conception of essence into the foreground. Here Wittgenstein writes: “One ought to ask, not what images [Vorstellungen] are, or what happens when one imagines anything, but how the word ‘imagination’ [‘Vorstellung’] is used. But that does not mean that I want to talk only about words. For the question as to the essence of imagination is as much about the word ‘imagination’ as my question is. And I am only saying that this question is not to be decided—neither for the person who does the imagining, nor for anyone else—by pointing; nor yet by a description of any process. The first question also asks for a word to be explained; but it makes us expect a wrong kind of answer.”26 In §370 the problem is how to approach the question, what is imagination?—according to Wittgenstein, a misleadingly formulated question about the essence of imagination. Wittgenstein suggests that this question would be best approached as concerning the use of the word “imagination,” not by pointing at anything or by describing a process. What sort of pointing and description he has in mind is not entirely clear. Apparently, however, the intended contrast is with the factual approach to the question about the essence of imagination: trying to determine it, for example, by introspection, psychologically, neurophysiologically, or through some sort of philosophical (say, phenomenological) analysis of facts. Read this way, the pointing Wittgenstein mentions in the remark would be the focusing of one’s attention on one’s imaginings, as if pointing at them mentally. By the phrase “description of any process,” he would mean describing a mental or a neural process associated with imagining.27 Importantly, however, the attempt to investigate the essence of imagination as a factual question presupposes an understanding of the concept of imagination. In such a study, one is investigating not just anything but whatever one identifies as imagining, and making use of the concept of imagination. Without the presupposition of an understanding of the concept, it would be impossible to distinguish between necessary and merely accidental characteristics of imagination. For instance,

The Concepts of Essence and Necessity

193

by investigating what happens in me when I imagine something, we will find out facts relating to me. But for all we know, these facts may be distinctively characteristic of me, not something necessarily characteristic of imagining—unless, that is, a comprehension of the difference between what is accidental and what is not is already presupposed (or stipulated on the spot). For even if the study is set up to reveal something that holds very generally of my imagination, or extended to cover a larger group of individuals as its subjects, it is still logically possible that those generalities that will be isolated are merely idiosyncratic. These observations suggest that in order to clarify to oneself the essence of imagination, one should turn one’s attention to the concept of imagination or to the use of the word “imagination.” By studying the use of the word, or the concept that the use of the word constitutes, one studies something already assumed in the factual investigation, and unlike a factual investigation, an investigation of the use of a word can bring to light necessities and possibilities. For instance, by clarifying the conditions of the application of the concept of imagination, a study of the use of the word “imagination” can establish something that holds by necessity of cases to which the concept of imagination is appropriately applied.28 The preceding considerations also explain the sense in which the question about essence expressed in the form “what is . . . ?” is concerned with words. This, I believe, is what Wittgenstein is alluding to when he writes: “the question as to the essence of imagination is as much about the word ‘imagination’ as my question is.” For provided that the question about essence concerns the necessary characteristics of the object of investigation, and that the way to comprehend such necessities is through an investigation of the uses of words, then the original question too is concerned with words. Only, since the question “what is . . . ?” is misleadingly phrased, its concern with words is covert. The questioner does not clearly comprehend what is at stake in asking her question, and this makes her expect a wrong kind of answer. If so, it would then be incorrect to say that Wittgenstein is trying to substitute an entirely different question for the original question about essence. Rather, he is reformulating the question with the purpose of helping the questioner to attain what she is after. It would also be problematic to object that whereas the question of essence concerns what words talk about, Wittgenstein’s question

194

The Concepts of Essence and Necessity

merely concerns words. Again Wittgenstein seems aware of this objection when he notes in §370 regarding his suggestion to rephrase the question: “But that does not mean that I want to talk only about words.” His approach too can be said to be concerned with what words talk about or what corresponds to words, and so is not merely concerned with words. As he explains elsewhere: “Insofar as we are dealing with the use of the word ‘five,’ we are dealing in a certain sense with what ‘corresponds’ to the word; only this way of speaking is primitive, it presupposes a primitive conception of the use of a word.”29 What is this “certain sense” in which, when dealing with a word, one is dealing with what corresponds to it? It may be that it is simplistic to conceive of whatever words speak about as something that corresponds to them. Nevertheless, to the extent that a word corresponds or refers to something, it does this in the capacity of a linguistic expression. It is its use in language that allows one to say that a word refers to something, and whatever the word refers or corresponds to, language users know it on the basis of their knowledge of the use of language. Wittgenstein expresses this point (in the context of the previous passage) by asking: “But what does it mean to say ‘We all know what phenomenon the word “thinking” refers to’? Doesn’t it simply mean: we all can play the language-game with the word ‘think’?”30 Insofar as the last question can be rephrased in the affirmative, one may say that, according to Wittgenstein, to know what the word “thinking” means, corresponds to, or refers to, and so on, is to know how the word is used. And more generally, if one wants to know what, if anything, corresponds to a word, one should examine the word’s use. In this sense an investigation of the use of a word is also an investigation of whatever the word is about. As regards the interpretation of the Investigations §371 in the light of §370, my suggestion now is that §371 provides us with a nutshell formulation of the methodological points articulated in §370. By stating that essence is expressed by grammar Wittgenstein is saying: If you want to understand the essence of something, look at grammar, try to describe the use of the relevant expressions. Grammatical observations, rather than factual investigations of instances falling under concepts, bring to light what is essential. Remark §371 therefore makes a methodological point about how to approach the question about essences or a point about the nature of statements about essences.

The Concepts of Essence and Necessity

195

Such statements articulate grammatical rules rather than asserting anything about facts. What Wittgenstein says in §370 about questions of essence provides reasons for adopting his conception and the approach he recommends. But a thesis about the order of determination or grammar as the source of necessity does not figure here as a reason to accept his view. Hence this thesis does not seem an essential component of Wittgenstein’s view. As I will argue, it is not part of his conception at all. Next I will seek to further elucidate the difference between grammatical and factual statements, as perceived by Wittgenstein, by discussing the nontemporal character of grammatical statements. This will help us get the relation between Wittgenstein’s conception of essence and certain more traditional conceptions into a sharper focus.

5.4 The Nontemporality of Grammatical Statements According to Wittgenstein, nontemporality is a characteristic mark of grammatical rules. This nontemporality is what grammatical statements share with geometrical and mathematical statements.31 The nontemporality of grammatical rules contrasts with the use of factual statements, which is temporal. Unlike a factual statement, a rule in the relevant sense is not used to assert what is the case at any particular point(s) in time. In this sense, a rule is not a statement about actualities, and unlike a factual statement’s generality, the scope of a rule’s applicability does not depend on a generalization from particular cases.32 Wittgenstein also characterizes the distinction between grammatical rules and factual statements about language as follows: “What I call a ‘rule’ is not meant to entail anything about any determinate (or also indeterminate) time or place of use, and not to refer to any determinate (or indeterminate) persons; but to be merely an instrument of (re)presentation. [New paragraph] We say now: ‘We use the words “red” and “green” in such a way that it counts as senseless (is contradictory) to say that there is red and green in the same place at the same time.’ And this is naturally a proposition, an empirical statement about our actual language.”33 Grammatical rules, as Wittgenstein conceives them (and as explained more thoroughly in 3.3–3.6), are therefore not statements about actual practices of language use but merely tools for the description of language. Language use may be described as governed by a certain rule, but to assert that it actually is used in this way is to make

196

The Concepts of Essence and Necessity

a temporal factual statement about a particular linguistic practice. Accordingly, when a statement that a sentence follows from another is used to describe a logical relation, it does not concern any particular occasions of the employment of the sentences in the relevant meaning; its use is nontemporal. As Wittgenstein writes: “When we say: ‘This proposition follows from that one’ here again ‘to follow’ is being used non-temporally. . . . [New paragraph] Compare ‘White is lighter than black.’ This expression too is non-temporal and it too expresses the existence of an internal relation.”34 What is of particular interest as regards the concept of essence is Wittgenstein’s connecting the notion of nontemporality with that of an internal relation. In the manuscripts and typescripts from which this passage originates, he also provides the following explanation of the notion of nontemporality in connection with a remark about internal properties—or essential properties, as one may also say:35 “ ‘The 100 apples in this box consist of 50 and 50’—here the non-temporal character of ‘consist’ is important. For it doesn’t mean that now, or just for a time, they consist of 50 and 50. [New paragraph] For what is the characteristic mark of ‘internal properties’? That they persist always, unalterably, in the whole that they constitute; as it were independently of any outside happenings. . . . I should like to say that they are not subject to wind and weather like physical things; rather are they unassailable, like shadows.”36 A sentence such as “The apples in this box consist of two times fifty” could be used to make a temporal statement, for instance, about how the apples will be divided between certain people, how the contents of such boxes are usually divided, and so on. When used to make a statement about a mathematical relation (that one hundred consists of two times fifty), however, the statement does not concern any particular occasion or occasions but is employed to make a statement about the concept of a hundred, or about an essential property of a hundred. What is stated holds universally for all cases in which we talk about a hundred in the relevant meaning, independently, as it were, of any contingent facts of the world. This suggests a comparison with the Platonic conception of essence. According to the Platonic conception, essences are immutable and eternally persisting, just as Wittgenstein characterizes internal properties in the passage above. Although Wittgenstein at times uses the notion of

The Concepts of Essence and Necessity

197

an internal property without alluding to any problems with it, in the above remark his depiction of internal properties, as if they were on a par with physical properties but with the remarkable characteristic of immutability, readily invites problematic questions about the mode of existence of the former. What are such shadowlike beings that exist beyond the corrosion that penetrates the whole of the physical world? As Wittgenstein remarks elsewhere, in philosophy there is a temptation to postulate such shadowlike constructions in order to explain grammatical forms that are not understood or are misunderstood: “Philosophy is (everywhere) haunted by such shadowy formations. Their picture forces itself upon us as the explanation of misunderstood//not understood grammatical forms.”37 Examples of such shadowy formations mentioned in the context of this remark are the sense of a sentence as a shadow of a fact, the sense of a wish as a shadow of what satisfies the wish, the sense of a rule as a shadow of action that accords to it, being able as a shadow of doing, and possibility as the shadow of reality.38 To this list of shadowy entities one may well add necessary, essential properties as something unassailable and immutable, unaffected by the forces of the physical world. Taking the remarks from manuscripts 115 and 117 as our lead, the relation between Wittgenstein’s conception of essence and the Platonic conception can be described as follows: What from the point of view of the Platonic conception appears as an indication of the immutable mode of existence of essences or essential characteristics, Wittgenstein wants to explain by reference to the role of statements about essences comprehended as grammatical statements. According to him, statements about essences are not true/false factual statements about a special kind of objects with an immutable or timeless mode of existence. They are statements of a different kind concerning conceptual or logical relations, and it is characteristic of such statements that they are nontemporal, i.e., not used to say how things are on any particular occasion. Such statements disregard, as it were, questions about time. But this indicates the immutability of whatever they concern only if the statements are construed as factual ones. Hence from Wittgenstein’s point of view the problem with the Platonic conception of essence arises from its mistaken interpretation of the nontemporal character of statements of essence as indicating that they concern objects or facts of an extraordinary kind. As if in the case of an

198

The Concepts of Essence and Necessity

essential property there were a bond between the object and its property that somehow cannot be broken, unlike a physical bond. Wittgenstein may thus be seen as seeking to bring down to earth the Platonic philosophy’s claim to the comprehension of eternal, immutable truths about essences. With the adoption of Wittgenstein’s conception, the need to postulate a realm of immutable essences disappears, and one is released from ontological problems pertaining to the peculiar mode of existence of immutable essences and essential properties.39 But to further clarify the contrast between Wittgenstein’s view and the conception of statements about essences and necessities as factual statements, let us turn to the idea of explaining necessity in terms of factual regularities.

5.5 Explanations of Necessity in Terms of Factual Regularities Instead of explaining necessity as a shadowy, immutable mode of existence, one might try to explain it in terms of factual regularities, attempting to avoid in this way the problem concerning the ontological status of Platonic forms. Employing Leibnizian terminology, one might say that what is necessary is what is true in all possible worlds, while what is contingent is only true in some possible worlds. Necessity is now portrayed as rigidity or robustness: a necessary relation between an object and a property is a regularity so rigid that it allows for no exceptions, persisting in all possible worlds. Let us examine this conception and Wittgenstein’s criticism of it, taking as our example the exclusion of reddish-green. For one might say that it is of the essence of red and green that they do not mix, that there is no such color as reddish-green. Wittgenstein maintains that the factual account according to which the exclusion of reddish-green lies in the nature of colors, or alternatively depends on the perceivers of colors, leads to insurmountable philosophical difficulties and is ultimately untenable. To bring out the difficulties, let us consider what a factual explanation of this exclusion might look like. Perhaps one could say the gap between red and green lies in the inner constitution of the colors; red and green are constituted in such a way that they do not mix with each other and do not produce a mixed color.40 Alternatively, one might try to explain the exclusion of reddish-green in terms of the functioning of the human

The Concepts of Essence and Necessity

199

perceptual apparatus (described either neurophysiologically or psychologically) and its exclusion of the perception of reddish-green.41 But the difficulty with these explanations is the same, and it is connected with their attempt to account for the relevant necessity by reference to facts, i.e., regularities in nature. This can be elucidated as follows: Insofar as the exclusion of reddishgreen is an exceptionless necessity, it is not just that occurrences of reddish-green are extremely rare or contingently excluded. The occurrence of reddish-green is precluded in a stronger sense: the very possibility emerges as incomprehensible. Although one can put the words together, it remains entirely unclear what kind of color should be described or identified as reddish-green and, therefore, what it would be to establish either the truth or falsity of a statement that something is colored reddish-green.42 For instance, there is no description of reddishgreen in terms of mixtures of colors. (In this respect reddish-green differs from, for instance, International Klein Blue, which, though rare, can be described.) The difficulty, then, is to explain this complete exclusion of reddishgreen, conceived either as a fact about colors or about our perceptual apparatus. Crucially, the exclusion does not seem explainable in terms of any natural regularities, because no actual regularity can exclude the possibility of exceptions. Changes and variation in natural, physical regularities are possible in principle. But the regularity that excludes the possibility of reddish-green would have to exclude exceptions not just in fact but in principle. This is where the difficulty with the factual explanation appears to become insuperable. No natural regularity seems able to meet what is required of a super-rigid regularity that excludes reddish-green by necessity and guarantees its exclusion in the required way. Wittgenstein writes about the picture of necessary truth as rigidity in a notebook from 1942: “Necessary truth” “necessary proposition”— a poor expression leads us to think about a rigid connection between certain objects (forms, numbers, etc.) in nature; a kind of natural science of these facts. That is, we construct a superlative of the rigidity of a connection, for which our mechanisms serve as a prototype.43

200

The Concepts of Essence and Necessity

The notion of a super-rigid mechanism, and the temptation to postulate such a mechanism, is also discussed in earlier notebooks (from 1937), in connection with rule-following or as an explanation of the phenomenon of a sign being meant in a certain way. For if meant in a particular way, a formula for an arithmetic series, for example, seems to determine inexorably how the series in question is to be continued, as if the right continuation were determined by some sort of superrigid mechanism. But as Wittgenstein points out in the context of this earlier discussion (also included in the Investigations): “You have no model of this superlative fact, but you are seduced into using a superexpression. (It might be called a philosophical superlative.)”44 Similarly, when one tries to explain the exclusion of reddish-green by reference to facts, the facts required for the explanation to work seem rather extraordinary ones, quite unlike any natural facts or regularities we are familiar with. But rather than explaining this appearance of a mystery as a mysterious natural phenomenon that must now become the object of our investigation, we may explain it as rooted in a failure to distinguish between statements of different kinds or between different uses of sentences. As Wittgenstein’s remark on “necessary truth” as a poor expression suggests, it is the idea of necessity as a kind of fact (the object of true factual statements) that leads to the idea of a rigid regularity or a connection between entities, as if this were something found in nature.45 Or as he says in the aforementioned earlier notebook, the idea of necessity as rigidity suggests itself to us as a result of a variety of connected images, comparisons, and/or pictures. Accordingly, the crucial point from Wittgenstein’s perspective is the following. To talk about necessity as the rigidity of a regularity or a mechanism is to adopt a particular picture—a conception or a way of thinking.46 But when one attempts to explain necessity in terms of rigidity, necessity becomes mysterious, because the super-rigid regularities that are supposed to explain necessity are themselves mysterious. Wittgenstein tries to persuade us to give up this picture by drawing attention to the distinction between kinds of statements, which the picture ignores, and to replace the picture of necessity as rigidity with an alternative one that presents essential, necessary connections as connections in grammar. As he explains: “The difficulty arises . . . through mixing up ‘is’ and ‘is called’ [heißt]. [New paragraph] The connexion which is not supposed to be a causal, experiential one, but much

The Concepts of Essence and Necessity

201

stricter and harder, so rigid even, that the one thing somehow already is the other, is always a connexion in grammar.”47 Thus in Wittgenstein’s view, the key to the solution of philosophical problems relating to the concept of necessity is the recognition of the difference between statements about what or how something is and statements about the application of concepts to things. This distinction, of course, is the one addressed also in §370 (see 5.3). But how exactly should one understand Wittgenstein’s conception of necessary connections as connections in grammar? Although the above discussion of the nontemporality of grammatical statements gives important clues to what it is to understand statements about essences and necessities as grammatical statements, in order to get clear about this we need to examine more closely what Wittgenstein means by approaching questions about essence as questions about grammar. The shift in the way of thinking about essence and necessity that Wittgenstein wants to effect is described in the following remark: Whence comes the feeling that “white is lighter than black” expresses something about the essence of the two colours?— But is this the right question to ask? For what do we mean by the “essence” of white or black? We think perhaps of “the inside,” “the constitution,” but this surely makes no sense here. We also say e.g.: “It is part of white to be lighter than . . . .” Is it not like this: the picture of a black and a white patch serves us simultaneously as a paradigm of what we understand by “lighter” and “darker” and as a paradigm for “white” and for “black.” Now darkness “is part of ” black inasmuch as they are both represented by this patch. It is dark by being black.—But to put it better: it is called “black” and hence in our language “dark” too. That connexion, a connexion of the paradigms and the names, is set up in our language. And our proposition is non-temporal because it only expresses the connexion of the words “white,” “black” and “lighter” with a paradigm.48

Wittgenstein’s train of thought might be spelled out as follows. The question is, what makes us say that a sentence such as “white is lighter than black” expresses something about the essence of the colors, something that holds of them by necessity? But he immediately pauses to reconsider whether this is a good question to ask after all, given that the

202

The Concepts of Essence and Necessity

concept of essence tempts us to think about the inner nature of colors and their constitution as if a factual issue were at stake. To avoid the difficulties relating to a factual account of the necessity that white is lighter than black, Wittgenstein suggests approaching the issue as a question about language use. Whatever is called “white” is also called “lighter than black,” and this is what it means for white to be necessarily lighter than black. When something is white it is at the same time lighter than black, because this is (part of ) what being white means. This necessity, or the necessary darkness of black and the necessary lightness of white, comes to light in our employment of one and the same patch as a paradigm for both black and dark and of another patch as a paradigm for both white and light. The statement “White is lighter than black,” then, is a nontemporal statement that describes the conceptual relation exhibited in this employment of color patches as paradigms. Accordingly, it is the nontemporal use of this sentence that makes us identify it as a statement about what philosophers have called “the essence” of black and white. But even if this might roughly explain Wittgenstein’s conception, at this point a danger of misinterpretation is imminent. The danger is that one reads Wittgenstein as offering an account according to which the necessity of the lightness of white or the exclusion of reddish-green is to be explained by reference to our having adopted certain linguistic practices. More specifically, the problem with such an explanation is that statements about our having adopted certain linguistic practices, or actually using language according to certain rules, are factual statements. Hence instead of offering a factual explanation of necessity by reference to facts about external or human nature, this interpretation explains necessity by reference to facts about language use. This explanation, however, is susceptible to the very same difficulties as the factual explanations discussed above. (It makes no difference whether the regularity one appeals to is, so to speak, a cultural or a natural one.) Ultimately, the problem with this account is that when essence or necessity is explained as dependent on contingent linguistic practices, essences and necessities are exposed to change and become themselves contingent. Hacker contrasts Wittgenstein’s conception of necessity with a more traditional metaphysical view: “The ‘necessity’ which we ascribe to such [metaphysical] ‘truths’ is the mark not of ‘a necessary fact’ (since

The Concepts of Essence and Necessity

203

there is no such thing), but of our commitment to these concepts, of our inflexibility in employing these expressions in accord with these rules.”49 Here Hacker appears to be saying that rather than indicating that there are necessities in nature, our recognition of the necessity of certain statements indicates our commitment to certain linguistic practices. This is not incorrect, but if by saying this Hacker means to suggest an alternative to the metaphysicians’ explanation of necessity, i.e., to propose that instead of reality the source of necessity is our adoption of certain linguistic practices, the inflexibility of these practices, and so on, he seems to be committing just the kind of mistake outlined above. The problem is that the existence of our linguistic practices, their inflexibility, and our commitment to them are matters of fact. Insofar as Hacker is appealing to such facts about language as the source of necessity, he is thereby trying to provide a factual account of necessity. Thus his explanation emerges as a modified, linguistic version of the very picture of necessity as rigid factual regularity that Wittgenstein rejects.50 To clarify this confusion, it is instructive to examine more closely the above remark about black and white and distinguish between the different kinds of statements Wittgenstein makes in it. That in our language black is also called “dark” and that the same patch is used as a paradigm for both black and dark are matters of fact. (It might be that we did not have the generic concept of dark but only the more specific ones of black and brown, for example.) Accordingly, Wittgenstein’s statements above about our use of the terms “white,” “light,” “dark,” and “black” are factual statements. Similarly, he is making a factual statement when he says that such a “connexion of the paradigms and the names is set up in our language.” This is a temporal statement about a particular language, identified as “ours” by the author of the remark at the time of its composition.51 By contrast, however, the statement “White is lighter than black” is not a statement of fact but, as Wittgenstein notes, a nontemporal statement. This means that it does not concern any actual instances of the use of the relevant terms but is a rule, an instrument that can be used to spell out a conceptual relation. But in this capacity, the necessity of this statement is not a “mark of our commitment to these concepts,” contrary to what Hacker says. Rather than having anything to do with facts about our commitments, the necessity of the statement is only a mark of its role as a nontemporal statement. Notably, this role

204

The Concepts of Essence and Necessity

cannot be explained by reference to our commitments, which at most can establish the truth of the (anthropological or sociological) statement that we in fact treat such and such as necessary. Here it is also important that the nontemporal statement “White is lighter than black” and our recognition of it as a grammatical statement do not indicate anything about what makes white necessarily lighter than black or about the source of necessity. No explanation of the phenomenon of necessity as rooted in grammar is implied. Rather, Wittgenstein’s point merely concerns the status of the statement in question. Accordingly, one may now say that the constructivist reading misconstrues Wittgenstein’s view when it interprets it as an account of the source of necessity. In so doing, constructivism blurs the very distinction between factual/temporal and grammatical/nontemporal statements that Wittgenstein seeks to clarify. But what kind of an account of necessity is Wittgenstein offering? I introduce my interpretation by discussing problems with an alternative interpretation of the constructivist thesis about essences as a creation of grammar.

5.6 Wittgenstein’s Account of Essence and Necessity One might feel that my critical discussion of Hacker in the previous subchapter does not quite capture the constructivist reading in its best form. The idea of this interpretation is not, one might object, to ground necessity on facts about language use (pace Kripke and Bloor). Rather than perceiving the rule-governedness of language or the inflexibility of the use of our expressions as facts, the constructivist reading regards them as necessary characteristics of language that find their expression in grammatical statements. More specifically (to outline a possible constructivist response), the rules of language are to be seen as necessary for language because they determine its correct and incorrect uses and in this capacity are constitutive of concepts. As regards the status of the statement about the creation of essences by grammar, this statement simply follows from grammatical statements concerning the necessary rule-governedness of language, and so on. For the sake of argument, I will accept this response regarding the status of the constructivist claim.52 I wish to problematize instead a picture of the determination of essences by grammar that seems to lie at

The Concepts of Essence and Necessity

205

its root. According to this picture, once grammatical rules are fixed, then what is essential is fixed in the sense that, assuming an established practice of the use of an expression or an established concept, there is only one legitimate answer to the question about its essential features. Here grammar again figures as the source of necessity and an explanatory factor that purports to explain what makes something essential or necessary (see 5.2). To refute this reading, I will seek to show that even if one assumes that constructivism can be spelled out in such a way that it does not collapse into a factual explanation of necessity or essence, it is in conflict with Wittgenstein’s remarks on rules and essential features and that grammar cannot be taken to fix essences in the presumed sense. Subchapter 5.7 explains how Wittgenstein’s conception as interpreted here leads to a richer account of the phenomenon of necessity than constructivism allows, opening a path beyond the opposition between the Aristotelian and Kantian positions. As Wittgenstein explains, it makes perfect sense to ask whether a particular rule is essential or inessential to a game, and by analogy, which rules are essential to the use of a word. Hence, that it is entailed by the definition of a concept that objects falling under it possess a certain characteristic, and accordingly that the ascription of this characteristic to these objects is not a factual statement, does not mean that the characteristic must be an essential one. Wittgenstein elucidates this by reference to games: Let us say that the meaning of a piece is its role in the game.—Now let it be decided by lot which of the players gets white before any game of chess begins. To this end one player holds a king in each closed fist while the other chooses one of the two hands at random. Will it be counted as part of the role of the king in chess that it is used to draw lots in this way? So I am inclined to distinguish between the essential and the inessential in a game too. The game, one would like to say, has not only rules but also a point. . . . But, after all, the game is supposed to be defined by the rules! So, if a rule of the game prescribes that the kings are to be used for drawing lots before a game of chess, then that is an essential part of the game. What objection might one make to this? That one does not see the point of this prescription. Perhaps as one would not see the point either

206

The Concepts of Essence and Necessity

of a rule by which each piece had to be turned round three times before one moved it. If we found this rule in a board-game we should be surprised and should speculate about the purpose of the rule. (“Was this prescription meant to prevent one from moving without due consideration?”) If I understand the character of the game aright—I might say—then this isn’t an essential part of it.53

Thus, although the distinction between the rules and moves roughly separates the essential features of a game from its accidental or contingent features, in the end things are not quite that simple. Questions about the point or purpose of a rule reintroduce questions about what is essential. Here it is crucial, however, that considerations relating to points or purposes, by reference to which questions about essential features may be decided, do not make essences any more fixed than they emerge on the basis of Wittgenstein’s observations in the above remarks. As he says (in a different but directly relevant54 context): “it is not everywhere clear what should be called the ‘point’ of an order. (Similarly one may say of certain objects that they have this or that purpose. The essential thing is that this is a lamp, that it serves to give light;—that it is an ornament to the room, fills an empty space, etc., is not essential. But there is not always a clear distinction between essential and inessential.)”55 Evidently, it cannot then be the aim of Wittgenstein’s introduction of the notions of point and purpose in §§563–568 to cover the gaps that the concept of a rule leaves open, so that, although grammar might not completely determine what is essential, this may nevertheless be conceived as something fully determinate once the notions of point and purpose are taken into account.56 For just as it might not be fixed which rules are to be treated as essential, it might not be fixed what should be taken to be the purpose or point of something. Consequently, there may be more than one legitimate answer to the question about the essential features of an object, depending on what is seen as its point or purpose. For example, although it is possible to consider the aesthetic features of lamps as inessential and only their practical features as essential, the contrary is equally possible. Here considerations of what is essential covary, so to speak, with what is conceived as the purpose of a lamp, i.e.,

The Concepts of Essence and Necessity

207

whether a lamp is treated as a source of light or an ornament to a room. Ultimately, however, the purpose of a lamp is no more determined than its essence. These observations also explain the sense, i.e., the nonemptiness, of Wittgenstein’s hypothetical formulation above (in §568): “If I understand the character of the game aright . . . then this [rule] isn’t an essential part of it.” For to understand the character of the game aright is precisely to understand what is essential in it. One’s understanding of the character (point, purpose, and so on) of a game is not something that supports one’s statement about which rules are essential. Rather, one’s comprehension of the character of the game may be expressed in a statement about rules. The point can also be explained as follows. Rather than being determined by an object of description alone, what is essential depends also on the concerns of the determiners of essence, i.e., which aspects of an object are the focus of their interests. For example, as Wittgenstein points out in the context of the above remarks,57 one might, given the overlapping uses of two words, characterize their uses as essentially the same or essentially different with equal justification, depending on one’s concerns. He illustrates this with his example of the uses of the negation terms “X” and “Y” (“non” and “ne” in the Nachlass), which are identical except that in one case double negation constitutes an affirmation and in the other a strengthened negative. But whether this difference is to be considered essential or inessential is not fixed. Nothing in the use of “X” and “Y” as such forces one to comprehend their difference rather than their similarity as the essential feature.58 These considerations bring to light an important difference between statements of fact and statements about essences, as the latter are understood by Wittgenstein. As explained, statements about points or purposes do not give independent support to statements about which rules are essential. In this sense, talk about points or purposes does not constitute evidence for a statement about an essence comparable to evidence for the truth of a statement. Similarly, statements of grammatical rules do not support assertions about essences. Rather, grammatical rules are a vehicle for articulating or spelling out what is essential or necessary. That is, even though rules for the use of a game or language do not fix what is essential, and considerations relating to points or purposes cannot make this any more fixed, it is nevertheless by means of grammatical rules that one explains what is essential or necessary.

208

The Concepts of Essence and Necessity

This, I believe, is what Wittgenstein means by saying in §371 that essence is expressed by grammar. Essences and necessities are not created by grammar but are literally expressed by grammar, i.e., find their expression in or are described by statements of grammatical rules rather than factual statements. By characterizing essence as something expressed by grammar, Wittgenstein therefore is not putting forward a thesis about the source of necessity or about what makes something necessary. His purpose is to clarify the role of statements about essences with the help of the notion of a grammatical statement. Accordingly, the constructivist view that Wittgenstein is offering an explanation of the source of necessity misunderstands the type of account he is providing. By presenting necessities and essences as the creation of grammar and grammar as the source of necessity, the constructivist interpretation in effect treats Wittgenstein as a party in the dispute between the Aristotelian and Kantian accounts of necessity, if by the former one understands a theory of necessity as grounded on facts of reality and by the latter a theory of necessity as originating in the concepts of the thinking subject(s). Interpreted in this way, Wittgenstein’s conception appears as a variant of Kantianism, similar to the conventionalism of the logical positivists, although, as explained in 5.1, perhaps more sophisticated in its nonvoluntarism. But rather than constituting such an alternative thesis, Wittgenstein’s conception should be seen as leading beyond these competing theories (or types of theory) that have come to dominate philosophical thinking about necessity.

5.7 Beyond Theses about the Source of Necessity Wittgenstein writes: “The rules of chess are not supposed to correspond to the essence of the chess king for they give it this essence.”59 As regards chess, it seems clear that the game and the identity of the chess king are indeed created by laying down the rules for the game. Applied to this example, therefore, the constructivist account of necessity appears unproblematic. But to say that chess and the chess king are human creations, i.e., brought into existence by laying down rules for the game, is to make a factual (empirical) statement. Accordingly, the unproblematic use of the statement about the nature of the chess king is probably of no great interest to the constructivist. Such an empirical statement

The Concepts of Essence and Necessity

209

cannot satisfy her philosophical aspirations. On the other hand, if we make it philosophically more potent by extending it to cover all necessities (of the relevant exceptionless type), the statement about rules giving things their essences becomes dogmatic and highly problematic. In cases such as the exclusion of reddish-green, grammar seems better understood as reflecting regularities of nature than as creating the impossibility of our recognizing anything as reddish-green.60 But if there is such variety with respect to the sources of necessity, one might ask, why put forward a uniform philosophical thesis about the source of necessity and state what its source must always be? Instead, by remaining neutral on the question of the source of necessity, Wittgenstein’s conception seems to make possible a richer and more nuanced account of the phenomenon of necessity. Now it is possible for him to acknowledge both a sense in which necessity may have its ground in reality and a sense in which it may have its ground in conventions. More specifically, as Wittgenstein explains, the borderline between statements of fact and grammatical rules is fluid. A sentence may assume either role, and the employment of a sentence may also change permanently, say, from a factual one to a grammatical one. For example, although it was originally a discovery that the chemical makeup of water is H2O, the statement “water is H2O” has nowadays assumed more or less the role of a definition. In this way, empirical discoveries may affect the determinations of concepts, and factual regularities of nature that speakers recognize may get incorporated into language in the form of grammatical rules.61 Hence there is a clear and perfectly respectable sense in which the kind of necessities philosophers have been interested in may be said to have their ground in reality. Accordingly, against the Aristotelian realist conception Wittgenstein is pointing out only that statements about essence or necessity cannot be understood as factual statements; his point concerns the status of such statements.62 As regards Wittgenstein’s method, it is also important to note the following. That the preceding chess example should be used as a model for all cases in which we talk about essences and exceptionless necessity, and that essences and the relevant kind of necessities should on these grounds be declared to be always creations of grammar, is in direct conflict with Wittgenstein’s methodological remarks on philosophy. An example is the one from manuscript 137 quoted in 5.2, according to which the goal of philosophy is not to reveal hidden unities (to state

210

The Concepts of Essence and Necessity

what things really are) but to draw attention also to differences. Similarly, the discussions of Wittgenstein’s methodology in Chapters 3 and 4 exclude the employment of examples as the basis of philosophical theses about what all cases falling under a concept must be. It is then plausible that the chess example is not intended to reveal what all cases really are or must be but to be an object of comparison and to illuminate both differences and similarities between cases in which we talk about essence and necessity. Thus understood, the chess example does not constitute the basis for a philosophical thesis about what must always be the source of necessity. Wittgenstein’s declining to offer a theory of the source of necessity then explains the sense in which he does not take sides in traditional disputes about essences between realists and idealists (constructivists, conventionalists). For insofar as he is not putting forward a thesis about grammar as the source of necessity, it is misleading to characterize his conception of necessity as a variant of the Kantian conception of exceptionless necessity as dependent on our concepts. Rather, Wittgenstein’s approach to the investigation of essences is neutral with respect to the question of the order of determination or the source of necessity. Whatever the source(s) of necessity might be in particular cases, if Wittgenstein is right, questions about essence and necessity are best approached as grammatical questions. But here grammar constitutes merely the level, so to speak, at which questions about essences and necessities are to be addressed, and Wittgenstein’s point is in this sense methodological, not ontological. Similarly, when explaining grammar or linguistic practices as the source of necessity, the constructivist reading takes sides against the Platonist with the historicist (for example, the Marxist), who explains essences and necessities as historical constructs and the product of historically situated practices. Crucially, however, looking at the issue from this angle the constructivist does not seem able to do justice to the nontemporal character of grammatical statements, which emerge as factual statements about actual linguistic practices. Hence an important feature of the use of statements of essence, which the Platonist observes and which seems to motivate the Platonic conception of the immutability of essences, is lost from the constructivist’s sight. As a result, the constructivist is likely to be unable to fully appreciate the Platonic philosopher’s

The Concepts of Essence and Necessity

211

temptation to construe statements about essence as statements about immutable objects or properties, and consequently unable to release her from her confusion. For this requires taking the view seriously, that is, understanding its attraction. As Wittgenstein also notes: “In a certain sense one cannot take too much care in handling philosophical mistakes, they contain so much truth. [New paragraph] It is never a matter of simply saying, this must be given up.”63 But the point of not taking sides with historicism is not to deny that there is an important insight contained in the historicist account, that is, to assert that there is no sense at all in which what is essential or necessary may be characterized as historically relative. The point is only that comprehended in the manner of the constructivist reading, this historicity of essences is explained in a misleading way. Comprehending essences as a creation of linguistic practices, constructivism collapses into a factual account of necessity in the sense that statements about essence are now regarded as statements about rules that are actually followed, such statements being, as explained, temporal statements.64 But an appropriate clarification of the issue must explain both the temptation to treat essences as timeless and the temptation to see them as historically determined. As then regards the historicity of essences comprehended as finding their expression in grammatical statements, the crucial point is that although grammatical statements are not historical (temporal, factual, or superfactual) statements, they are nevertheless statements made in specific historical contexts. More specifically, clarifications in terms of grammatical statements are responses to particular philosophical problems that arise in certain historical contexts and in connection with particular actual practices of language use. Although this does not make grammatical statements factual statements about historical linguistic practices, it does make them relative to specific historical contexts in a different sense. The historicity of grammatical statements is, as it were, built into their employment for the clarification of historically determined problems, i.e., problems arising in particular historical contexts.65 Besides being neutral with respect to the question concerning the source of necessity, Wittgenstein’s conception is therefore also neutral with respect to whether a particular thing’s essence is historically variable or not. Although the essence of the chess king is clearly a cultural,

212

The Concepts of Essence and Necessity

historical phenomenon, the exclusion of reddish-green appears historically more constant, apparently having to do with the development of our biological ancestor’s brain to suit a certain kind of environment. Again, however, such questions are factual ones, and Wittgenstein’s conception, being designed to clarify the role of statements about necessities, does not say anything about such factual differences. In conclusion, Wittgenstein’s position can be described as follows. The concept of grammar does not explain what makes something essential or necessary, but questions about essences and necessities can be clarified by way of grammatical observations. Accordingly, the justification of Wittgenstein’s conception of questions of essence as grammatical questions does not presuppose the thesis about the creation of essence by grammar or grammar as the source of necessity. His view is not that it is by virtue of their being linguistic constructions that essences can be investigated by describing the use of language. Rather, Wittgenstein’s conception of essence is justified firstly insofar as it can be used to solve philosophical problems relating to the concepts of essence and necessity, such as the problem with the ontological status of Platonic forms or the problems with attempts to explain exceptionless necessities in terms of regularities in nature. Secondly the comprehension is justified insofar as the method of the investigation of essences with which it belongs together enables one to get clear about questions of essence in particular cases, i.e., allows one to address successfully questions such as, What is the essence of revolution? or What is the essence of a refined sense for music?66 Regarding the question of the status of Wittgenstein’s conception of statements about essences as grammatical statements, here it is again important that in actuality the borderline between rules and statements of fact is blurred. Actual language use is fluid and complex enough to allow one at times to describe the use of a sentence either way.67 By now such observations about the complexity of actual language use should sound familiar, and so it should come as no great surprise that Wittgenstein’s conception of statements of essence as grammatical statements, i.e., his model designed to explicate the role and logic of such statements, is ultimately itself to be understood as an object of comparison. Its status is the same as that of any other clarificatory model he articulates. This then explains why Wittgenstein’s remark that essence is ex-

The Concepts of Essence and Necessity

213

pressed by grammar does not itself constitute a thesis about the essence of statements about essences. Rather than claiming that for a statement to be recognizable as a statement about an essence it must express a grammatical rule, and that his conception covers everything one might wish to include in the notion of essence, Wittgenstein articulates a conception of essence, designed for the clarification of certain particular philosophical problems.68 Nevertheless, this conception is supposed to accommodate a significant part of what philosophers have wanted to express by the concept of essence, including its intimate connection with the notion of exceptionless necessity and the timelessness of essences. (Wittgenstein’s conception of philosophical statements as grammatical statements also retains the idea of essences as a central concern of philosophy, the traditional view of philosophical statements as statements about essence.) Correspondingly, his remark that necessary connections are always grammatical connections69 is to be understood as articulating a clarificatory model, not as a thesis. The modality expressed by “always,” that is to say, is to be seen as internal to the model and is not to be projected onto language as a claim about an alleged necessity in actual language use. In reality, actual use is more fluid than the model suggests. As in the case of the concepts of meaning and language discussed in Chapter 4, Wittgenstein’s conception of essence and necessity seems able to accommodate various seemingly conflicting aspects of these concepts that philosophers have been paying attention to. The commitment of the more traditional philosophers to philosophizing in terms of theses, however, has made their views one-sided in a problematic way, different schools being able to do justice to some aspects of the concept or phenomenon of necessity but not to others. Thus, the Kantian appears to be successful in explaining the exceptionlessness of the relevant kind of necessity but encounters difficulties in dispelling the impression that exceptionless necessity is a human creation. In this latter respect the Aristotelian seems to do better, but she has problems with exceptionlessness. In comparison, Wittgenstein’s conception seems advantageous in that it is able to accommodate both these aspects, leading to an increase in the flexibility of philosophical thought and to a richer view of necessity than it seems possible to spell out in terms of theses.

214

The Concepts of Essence and Necessity

Chapter 6 covers a variety of outstanding issues, including Wittgenstein’s dissolution of philosophical hierarchies, the concept of perspicuous presentation, the sense in which philosophy only states what everyone admits, and what the correctness of clarificatory statements means. This treatment both completes in certain relevant respects the interpretation of Wittgenstein’s method that I have been concerned to develop in this book and clears away some potential objections and misunderstandings.

SIX

Philosophical Hierarchies and the Status of Clarificatory Statements

the conception of philosophy I have attributed to Wittgenstein, I now return to more general methodological issues. I begin (in 6.1) by addressing what might seem like a crippling problem with the present interpretation: The account of clarificatory rules as objects of comparison might seem to make it impossible to explain the fundamental role that the notion of the rule-governedness of language plays in Wittgenstein’s philosophy. The purpose of this discussion is to draw attention to yet another way of misunderstanding Wittgenstein’s philosophy as a form of metaphysical philosophy. According to this view, the foundation of his philosophy is constituted by a thesis, or a “grammatical truth,” about language as a rule-governed activity, which serves to guarantee the applicability of the method of clarification by tabulating rules for language use. Problematically, however, Wittgenstein’s philosophy is here seen as hierarchically structured. By way of a response, I seek to elucidate the rejection of philosophical hierarchies in his later work and explain why foundational theses of the above type are neither necessary nor possible. Herein lies a radical aspect of Wittgenstein’s thought that reveals its distance from metaphysical philosophy. Next I discuss (in 6.2) the concept of perspicuous presentation and the danger of relapsing into metaphysics in this connection. For interpreted in a particular way (as exemplified by Baker and Hacker), perspicuous presentation emerges as an ultimate philosophical point of view or a position beyond all positions. But this reading, I argue, rests on confusion and a metaphysical projection of the philosopher’s mode of

TO FURTHER ELUCIDATE

216

Philosophical Hierarchies and the Status of Clarificatory Statements

examination onto language. Having shown where I believe the mistake lies in this interpretation, I develop a reading of Wittgenstein’s concept of perspicuous presentation consistent with the conception of the status of philosophical statements developed in the preceding chapters. Subsequently I turn to other matters relating to the status of grammatical remarks: whether there is such a thing as being forced to accept a grammatical remark (6.3), and the notion of agreement in Wittgenstein’s philosophy (6.4). My purpose is to articulate an interpretation of the notion of agreement that brings it down to earth and thus makes comprehensible Wittgenstein’s remark that philosophy only states what everyone admits. I also address (in 6.5) questions about the relation between grammatical remarks and truths, in particular the question of what the correctness of a grammatical remark means. For if grammatical remarks are not true/false statements, then apparently their correctness cannot be understood on the model of truth either—for instance, as consisting of a correspondence between a representation and an object of representation. Subchapter 6.6 discusses Wittgenstein’s idea of making a novel use of “old dogmatic claims” and the possibility of giving a new employment to metaphysical theses. This explains why Wittgenstein’s philosophy, although nonmetaphysical, is not correctly characterized as antimetaphysical. In this subchapter I also address the question of whether the conception of philosophy I have attributed to Wittgenstein leads to a loss of philosophy’s strength. I conclude the chapter by discussing the possibility of what I call “multidimensional descriptions,” that is, a particular kind of complex characterizations of language use made available by Wittgenstein’s conception of the status of philosophical statements.

6.1 Philosophical Hierarchies and Wittgenstein’s “Leading Principle” One might feel that there is something fundamentally problematic in how I have presented Wittgenstein’s concept of grammar and his conception of philosophy. According to the interpretation proposed in this book, the use of grammatical rules in philosophy is their use as objects of comparison. But has something crucial been overlooked? Consider the following argument by an imagined interlocutor, which seeks to articulate a problem pertaining to the present interpretation:

Philosophical Hierarchies and the Status of Clarificatory Statements

217

When one uses language, one follows the rules of grammar and in this sense presupposes the grammar of language. But the constitutive function of grammatical rules (the constitution of meaning and concepts through grammatical rules) presupposed in the use of language cannot be explained through the function of rules as objects of comparison. This means that the role of rules as objects of comparison is not fundamental. Indeed, the possibility of employing rules as objects of comparison itself depends on the constitution of language through grammatical rules and in this sense on the rule-governedness of language. The account of the role of grammatical rules as objects of comparison therefore fails to capture the fundamental function of such rules as constitutive of language and as a condition of the possibility of philosophy as Wittgenstein conceives it. The conception of philosophy described in this book and attributed to Wittgenstein fails to grasp its own presuppositions.1

There may be philosophically unproblematic ways of taking (most of ) the sentences in this argument. It is not unproblematic, however, to read them as an argument to the effect that the justification of grammatical investigation requires establishing the necessary rule-governedness of language and that this—or perhaps some philosophical account of what it is to follow a rule—constitutes a first (logically or philosophically prior) step of the investigation, without which the investigation lacks proper grounding. I will argue that this way of thinking constitutes a misunderstanding of Wittgenstein, missing a radical aspect of his philosophy. What is lost from view is his dissolution of philosophical hierarchies, a task intimately connected with his view of how to reach the goal of philosophy, namely, thoughts at peace (see 1.5). At the same time, my sample argument signals blindness to the real starting point of grammatical investigation. Let me first note one clearly unobjectionable point in the above argument. My imagined interlocutor is obviously right in distinguishing the function of constitutive rules from that of rules as objects of comparison. To be sure, when someone follows a rule, the rule is not functioning as an object of comparison (whatever that would mean). Accordingly, when a rule is used as an object of comparison, its role is not that of a constitutive rule of language. But it is one thing to observe this difference and another to assume that the possibility of philosophy and

218

Philosophical Hierarchies and the Status of Clarificatory Statements

the use of rules as objects of comparison can and must be grounded on an account of the constitutive function of rules. The distinction between constitutive rules and rules as objects of comparison can be elucidated by reference to the following difference. A change in the rules constitutive of the use of a word may change the meaning of the word. Changing a rule used as an object of comparison, on the other hand, does not affect the meaning of the word that is the object of description. But this merely says that to change a description is not to change the reality that is the object of description. Rules used as objects of comparison are a means of description. An account of their function in this role is not intended to explain the possibility of language, and not doing so is not a shortcoming. Moreover, as I will argue, the possibility of grammatical descriptions of language neither stands in need of justification in terms of the necessary rule-governedness of language nor can be justified in this way.2 Let us begin by examining certain relevant remarks by Wittgenstein. Consider the following remark from the early 1930s: If a man points out that a word is used with several different meanings, or that a certain misleading picture comes to mind when we use a certain expression, if he sets out (tabulates) rules according to which certain words are used, he hasn’t committed himself to giving an explanation (definition) of the words “rule,” “proposition,” “word,” etc. I’m allowed to use the word “rule” without first tabulating the rules for the use of the word. And those rules are not super-rules. Philosophy is concerned with calculi in the same sense as it is concerned with thoughts, sentences and languages. But if it was really concerned with the concept of calculus, and thus with the concept of the calculus of all calculi, there would be such a thing as metaphilosophy. (But there is not. We might so present all that we have to say that this would appear as a leading principle.)3

This remark characterizes Wittgenstein’s resistance to a particular philosophical move as a leading principle of his philosophy (or as something that could be presented as a leading principle). His point can be explained as follows. The conception of the description of language as the tabulation of rules for its use involves an account of language as something whose use is governed by rules. It might then seem that a proper

Philosophical Hierarchies and the Status of Clarificatory Statements

219

comprehension of this conception of language and its description requires an explanation of what we mean by the word “rule” or “rulegovernedness.” (Whoever adopts the approach of tabulating rules seems committed to such an explanation.) Wittgenstein, however, rejects this requirement. It is not the case that one is only justified in employing the word “rule” insofar as one can give a general explanation or definition of the concept and that one therefore must be able to provide such an account. (The ability to use a word in particular cases and the ability to describe its use are two different things, as noted in 3.3.) Moreover, according to Wittgenstein, we should not understand the explanations of our terms as constituting hierarchies of more and less fundamental concepts in an absolute sense, but should view the role of a concept in the hierarchy as relative to the questions that are being asked. An example of the hierarchical structuring of philosophy would be the view that the concept of a rule, given its employment to explain the concepts of meaning, language, and so on, should be seen as the central concept of philosophy, upon which the foundation of philosophy as the grammatical description of language rests. But—this is central to Wittgenstein’s outlook—the rule that governs the use of the word “rule” is not, as it were, a rule above all rules, a super-rule. Philosophical questions and explanations concerning the concept of a rule do not constitute a metaphilosophy that lays out a framework for the rest of philosophy. Essentially the same point is made in Wittgenstein’s lectures of 1932–33 by reference to the concept of meaning. Here he rejects the idea that meaning would be a concept of “more general importance than chairs etc.,” that questions concerning meaning would be “the central questions of philosophy,” and that the word “meaning” would have a “metalogical” status and a “higher place” than other words.4 Correspondingly, in a notebook from the same period Wittgenstein denies that the concepts of rule and game have the status of metalogical concepts.5 Hence for Wittgenstein all philosophical considerations and concepts are on the same level, including those concerning the form of the philosophical account or the concepts with which the philosophical approach is defined. As he writes in the Investigations, rejecting the idea of a metaphilosophy: “One might think: if philosophy speaks of the use of the word ‘philosophy’ there must be a second-order philosophy. But it is

220

Philosophical Hierarchies and the Status of Clarificatory Statements

not so: it is, rather, like the case of orthography, which deals with the word ‘orthography’ among others without then being second-order.”6 Ultimately, the problem with metaphilosophy and the hierarchical organization of philosophy is that it makes philosophy vulnerable to a devastating, all-encompassing critique directed at its foundations, as explained in 1.5. As a consequence, it becomes impossible to find peace in philosophy, that is, to reach the goal of philosophy as Wittgenstein conceives it. For under a hierarchical organization, all results that have been reached in the context of a particular philosophical approach can be questioned at once by problematizing the idea— definition, determination—on which the approach is founded. For instance, if understood as exhibiting a hierarchical structure with the conception of the rule-governedness of language as its foundation, the whole enterprise of grammatical investigation, including all its results, can be called into question by problematizing the notion of the rulegovernedness of language. At this point it is helpful to relate Wittgenstein’s rejection of philosophical hierarchies to his position in the Tractatus. As explained in 1.5 and 3.1, in the Tractatus the conception of the essence of propositions constitutes, in effect, a theoretical presupposition upon which its conception of philosophy as logical analysis rests—or with which it falls. Thus Wittgenstein’s early thought evidently possesses the kind of hierarchical structure he later rejects. More specifically, in the Tractatus considerations relating to the method of philosophy (including those concerning the essence of propositions) are regarded as a more fundamental undertaking (distinct from the activity of logical analysis) whose goal is first to establish a framework for (future) logical analyses. This separation, as noted earlier (in 3.1), is evident in that Wittgenstein does not rely on the method of logical analysis in laying down the framework for logical analysis, but on a different method (the one he describes as not strictly correct). By contrast, when interpreting Wittgenstein’s later philosophy as having a hierarchical structure founded on rule-following, there seems to be no particular need or temptation to assume that he would rely on any other method than that of grammatical investigation to establish the rule-governedness of language. (One can assume that he reaches this result by tabulating rules for the use of language.) Nevertheless, in the context of this hierarchical interpretation the conception of lan-

Philosophical Hierarchies and the Status of Clarificatory Statements

221

guage as rule-governed practice clearly does assume a special role. Given that it constitutes the foundation of grammatical investigation, the conception of language as a rule-governed practice becomes unquestionable, accessible to grammatical reexamination only at the risk of the investigation destroying its own foundation and with it all its results. In this sense the conception of language as a rule-governed practice may well be characterized as having a privileged or a higher status than results relating, for instance, to the concepts of a chair, music, or God. Thus, in the context of the hierarchical interpretation, questions relating to rules would be the central questions of philosophy and have more general importance than any other questions, in just the sense that Wittgenstein excludes above in the case of meaning. Wittgenstein’s rejection of the idea of philosophical hierarchies is intriguingly expressed in the following remark from 1941, which also employs or refers to the comparison with orthography seen in the Investigations §121: “Thoughts are to be arranged in such a way that the investigation can be interrupted at any point without the sequel being able to put into question what was said up to that point. Here we come again to the thought that spelling the word ‘spelling’ is not higherorder spelling.”7 The significance of the observation that spelling the word “spelling” is not a higher- or second-order activity lies in its bringing to light just the kind of nonhierarchical organization that Wittgenstein thinks philosophy should have and that would make possible the interruption of philosophy at any point. The arrangement of thoughts that enables the interruption of philosophical work without the danger of it being subsequently thrown into question is an arrangement where nothing figures as a theoretical presupposition upon which other parts of philosophy rest. Arguably, this idea of avoiding conceptual hierarchies is what Wittgenstein calls “the leading principle,” the abandonment of higher-level superconcepts or “super-rules.”8 (An example of a super-rule would be the rule about the rule-governedness of language assumed as the foundation of grammatical investigation.) Although Wittgenstein does not speak explicitly of any such leading principle in the Investigations, the principle clearly informs his approach there as well. What can be identified as the very same methodological idea, namely the shift from one single, fundamental problem to multiple problems, is characterized in §133 as making it possible to interrupt philosophy at any time, without the threat that what was said

222

Philosophical Hierarchies and the Status of Clarificatory Statements

up to that point will be put into question.9 Hence, far from being discarded, Wittgenstein’s leading principle from the early 1930s is merely expressed in a different way in the Investigations. Whereas manuscript 114 explains the leading principle rather dogmatically and in negative terms as the rejection of the hierarchical structuring of philosophy, the Investigations expresses it in positive terms by reference to what Wittgenstein regards as the goal of philosophy: thoughts at peace. With this critique of philosophical hierarchies in mind, let us return to my imagined interlocutor’s argument. The problem with the argument I imagined, with respect to the issue of how to read Wittgenstein, is that the requirement that grammatical investigation (as the tabulation of rules for language use) be given an abstract justification in terms of the necessary rule-governedness of language exemplifies just the kind of approach Wittgenstein rejects. Unlike Wittgenstein, my interlocutor conceives philosophy as hierarchically structured. In her account the notion of the rule-governedness of language figures as a foundational definition upon which grammatical investigation rests. But as I will now try to explain, the legitimacy of grammatical investigation does not require an abstract justification of this kind. Moreover, a closer examination reveals such a justification to be illusory. The justification I imagine my interlocutor to demand for the grammatical mode of examination would show that grammatical investigation is concerned with essential rather than merely accidental characteristics of language. That is, given the aim of grammatical investigation to clarify meanings, concepts, or language use, the justification would have to establish that meanings, concepts, or language use can really be clarified by tabulating rules. This would be accomplished if it were shown that grammatical rules are constitutive of language and concepts and that rule-governedness is necessary for meaningful language use. Such a demonstration would establish once and for all (with respect to our current concepts of language and meaning) that to talk about meaning and language use in terms of rules is justified, i.e., that by modeling his philosophical practice on the nonphilosophical linguistic practice of explaining meanings with rules, Wittgenstein is not modeling philosophy on something that is merely accidental to language use. How would one provide a justification of the above kind? It seems true that the possibility of grammatical investigation (or logic) presupposes

Philosophical Hierarchies and the Status of Clarificatory Statements

223

that actual language use is governed by rules. (As noted in 3.7, agreement in definitions ultimately means agreement in judgments.) But empirical facts about language do not show that language is necessarily rule-governed, and referring to such facts is not enough to establish the legitimacy of the grammatical mode of explanation. On the other hand, it cannot simply be stated that it is a “necessary fact” about language or a “grammatical truth” that rules are constitutive of meaning. The important question is, how does one know or judge a grammatical statement to this effect to be correct? The answer to this question explains what it is for a grammatical statement to be regarded as justified. Ultimately, as I explain in 6.5, the justification of grammatical statements lies in the clarificatory work they can do. A grammatical statement is justified to the extent that it meets the purpose such statements are meant to serve, i.e., clarifies language use to someone. Granted this, the following observation suffices to complete the argument against my interlocutor: Insofar as the justification of a grammatical statement depends on its clarificatory force, the idea that grammatical investigation should be given a general and abstract justification dissolves. For just as the justification of the alleged foundational statement about the constitution of meaning by rules depends on the clarificatory work this statement can do, it is so for any grammatical statement. But if the justification of each clarificatory rule depends on its own clarificatory force, then their justification does not require referring back in an abstract way to the rule-governedness of language.10 More specifically, it is unnecessary to provide a justification of the kind my interlocutor envisages for two reasons. First, the rulegovernedness of language as such does not yet show anything about the justification of any particular description in terms of rules. Establishing the justification of an individual description therefore requires the examination of the uses of language that this particular description aspires to describe. Second, the justification obtained in this latter way is fully sufficient. No additional, second-order justification of the practice of tabulating rules is required—as if an individual description could be wrong in some way besides failing to describe whatever it aspires to describe.11 The emptiness of my interlocutor’s picture is the emptiness of this second-order justification. Such a justification is an idle wheel in the machinery of language.

224

Philosophical Hierarchies and the Status of Clarificatory Statements

Note also that if grammatical descriptions require the kind of secondorder justifications my interlocutor envisages, so does the characterization of language as a rule-governed practice itself. For there is no reason to give this description a special status in that it would be able to justify itself while others cannot. Therefore, the consequence of my interlocutor’s requirement of justification, strictly thought through, would be an infinite regress. The alleged second-order justification (the description of language as necessarily rule-governed) itself stands in need of higherorder justification (of the practice of giving such descriptions), which in turn requires a higher-order justification, and so on. This regress can be halted only through the realization of the emptiness of the requirement of the second-order justification. There is no regress because there is no need for the kind of higher-level justification my interlocutor wants. We are thus released from philosophical hierarchies and the hierarchical structure of philosophy is dissolved. Now every grammatical description stands or falls “on its own,” including rules used as objects of comparison. A clarificatory rule is required to be actually capable of clarificatory work, but such rules do not need any backing from a thesis or a “grammatical truth” about the necessary rule-governedness of language. My interlocutor, therefore, is incorrect in her claim that the conception of the rule-governedness of language must be regarded as more fundamental than other descriptions of language in terms of rules. This characterization of the concept of language is on the same level as any other characterization. Hence the practice of grammatical description, as conceived by Wittgenstein, is free from any theoretical presuppositions. There are no great once-and-for-all determinations of fundamental concepts or superconcepts constituting the foundation of Wittgenstein’s philosophy. Rather than to establish a foundation for his philosophy, the purpose of Wittgenstein’s discussions of the constitution of meaning by grammatical rules is just to make more perspicuous the relations between the concepts of rules, meaning, language, and so on. His clarifications of these conceptual relations might be motivated partly by more widely shared philosophical problems relating to these concepts, partly by a desire to lay out clearly his own conception of philosophy as the description of language by tabulating rules. But these clarifications are not motivated by foundational concerns. Similarly, the purpose of considerations of rule-following in the Investigations is not to provide a foundation for Wittgenstein’s philosophy. Again, the fact that what he

Philosophical Hierarchies and the Status of Clarificatory Statements

225

says about the relevant concepts is already an exercise in philosophical clarification shows the futility of attempting to provide a foundation for philosophical clarification.12 The way I imagined my interlocutor’s requirement of justification, its central point was that grammatical investigation should capture what is essential in language use. On the basis of the discussion above, her aim can now be reformulated as the requirement that grammatical investigation should capture what is essential in each particular case, where what is essential in the use of an expression may vary depending on which philosophical problems one focuses on.13 What is essential, in other words, cannot be decided once and for all on the basis of a prior determination of the essence of language. One cannot assume with my interlocutor that what is essential in the use of a word can always be described by tabulating rules or that other types of descriptions can always be reduced to descriptions in terms of rules. Rather, the correct method of description is to be determined by reference to the particular case at hand. Or as Wittgenstein puts it: “In philosophy it is not enough to learn in every case what is to be said about a subject, but also how one must speak about it. We are always having to begin by learning the method of tackling it.”14 It is difficult to say what kinds of variations or differences in method Wittgenstein has in mind in this passage. Nevertheless, it is important to remember that by focusing on rules, grammatical investigation is abstracting from other features of language, such as its instrumentality and its rootedness in nonlinguistic and nonconventional human practices. This focus on rules involves a simplification, as explained in 4.3, and in order to avoid turning this simplification into a source of confusion, one should regard rule-governedness not as the aspect of language that must always be the focus of philosophy, but as one aspect. That is to say, as argued in Chapter 4, for Wittgenstein rule-governedness is not the essential characteristic of language. Correspondingly, the definition of grammatical investigation as the tabulation of rules should not be regarded as the definition of “grammatical investigation.” Rather, grammatical investigation is a set of techniques, one of which is the tabulation of rules.15 Instead of being treated as the essential characteristic of language across the board, rule-governedness, therefore, is the essential characteristic of language from the perspective of the method of tabulating rules. It is characteristic of grammatical investigation as the tabulation

226

Philosophical Hierarchies and the Status of Clarificatory Statements

of rules that it regards language as necessarily rule-governed, but the necessity in question is characteristic of the mode of presentation employed in the investigation, not a necessary truth about language. Consequently, one might say, pace my interlocutor, that what is fundamental is not a “grammatical truth” about the rule-governedness of language that provides a foundation for grammatical investigation. Fundamental is the adoption of the method of grammatical description, from the point of view of which language is conceived as a game according to rules—and the fact that it is possible to resolve philosophical problems this way. (The latter is not something one can decide at will, like the adoption of a particular approach. This eliminates what might otherwise seem like a voluntaristic feature of my interpretation.) This (nonrandom) adoption of the method, then, is the true starting point of grammatical investigation and a precondition one should not miss when reflecting back on this philosophical practice. Accordingly, it is not correct that grammatical investigation, as I have described it, involves a failure to address its own presuppositions. In fact, a fundamental problem with my interlocutor’s argument is that it forgets that the conception of language as governed by grammatical rules is characteristic of the grammatical mode of examination. This lack of awareness of the perspective she has adopted makes the rulegovernedness of language appear to her as a necessary fact about language on which philosophy can be founded. But here my interlocutor relapses into metaphysical projection. To understand the necessity of the rule-governedness of language as a “necessary fact” or a “grammatical truth” about language is to project onto language a conceptual determination characteristic of the perspective of grammatical investigation as the tabulation of rules. To counteract this fall back into metaphysics, one must keep up the perspective of Wittgenstein’s turn, as explained in Chapter 3. This point could also be explained by saying, as in Chapter 4, that the conception of language as rule-governed is itself an object of comparison, not something language must fit. This is not to deny that it is also a fact about language that its use is rule-governed. It is only to say that rule-governedness is not a “necessary fact” about language. Accordingly, the interpretation I am suggesting is not a form of instrumentalism that regards the rule-governedness of language only as a useful fiction,

Philosophical Hierarchies and the Status of Clarificatory Statements

227

while maintaining that language is not really rule-governed or remaining agnostic on this point. Instead, my purpose has been only to emphasize the difference between statements of fact and grammatical statements and to argue that first, contingent facts about language cannot provide grammatical investigation with the kind of grounding my interlocutor is looking for, and second, neither can claims about “necessary facts” or “grammatical truths.” There is no justification for grammatical descriptions beyond their justifying themselves by the clarificatory work they can do. Consequently, grammatical investigation is description all the way down. Or as Wittgenstein also says: “Philosophy is purely descriptive and, more precisely, it describes language. That is: it gives no foundations.”16 It seems plausible that concerns such as those discussed here are also what Wittgenstein has in mind in the following remark (from the early 1930s, like the one just quoted): “Why do we sense the investigation of grammar as being fundamental? [New paragraph] Insofar as the word ‘fundamental’ has a meaning at all it cannot mean anything metalogical or philosophical. [Ms110 adds: We are chasing metaphysics out of all its hiding-places.] [New paragraph] The investigation of grammar is fundamental in the same sense in which we may call language fundamental—say, its own foundation.”17 To begin unraveling this remark from the last paragraph, language may be said to be its own foundation insofar as its grammar cannot be derived from or justified by reference to objects in the world, purposes external to it that it serves, and so on. Similarly, grammatical investigation is its own foundation in the sense that it requires no justification beyond the justification of individual clarifications. The investigation, as it were, supports itself— provided there is something to clarify. To explain the fundamentality of grammatical investigation in metaphysical or metalogical terms would be to take this investigation as establishing that grammar—language having a grammar, the rulegovernedness of language—is the foundation of language and that grammatical investigation, therefore, is concerned with what is fundamental to language. Here the investigation is seen as laying down a foundation for itself once and for all, where the rule-governedness of language is imagined as a background that explains the possibility of individual instances of clarification. But now grammar itself becomes the last hiding place of metaphysics. The notion of the rule-governedness

228

Philosophical Hierarchies and the Status of Clarificatory Statements

of language becomes a requirement placed on language, a thesis about what language must be.

6.2 The Concept of Perspicuous Presentation The problem of the philosopher’s loss of awareness of the particular perspective involved in grammatical investigation also arises in the context of interpretations of Wittgenstein’s concept of the perspicuous presentation of language. As Baker notes critically about the interpretation he had propounded with Hacker: “Whereas Platonism and Cartesian dualism might be described as ‘ways of seeing things’ . . . a perspicuous presentation does not ex officio embody any point of view.”18 The problem is that now perspicuous presentation emerges as a point of view beyond all points of view, the ultimate point of view. In this subchapter, I first show how Baker and Hacker’s forgetfulness of the fact that grammatical investigation involves adopting a particular point of view can be explained in terms of the notion of metaphysical projection and how their interpretation of the concept of perspicuous presentation therefore constitutes a relapse into metaphysics. I then spell out an alternative interpretation that, in addition to being more faithful to Wittgenstein’s remarks on the topic, avoids the problems with Baker and Hacker’s reading. When sufficient attention is paid to Wittgenstein’s formulation in the Investigations §122 of what perspicuous presentation is meant to achieve and its significance for him, this remark can be understood as summing up his approach in exactly the kind of way the interpretation developed in earlier chapters would lead one to expect. Let us begin by looking more closely at Baker and Hacker’s interpretation. According to them, the notion of perspicuous presentation in Wittgenstein’s later philosophy is “a direct descendant of the Tractatus’ conception of ‘correct logical point of view.’”19 The difference is that in the later philosophy, “the correct logical point of view” is not sought “by means of ‘depth analysis’ ” but by “surveying all the uses and applications of words, phrases and sentences in a given domain of thought that give rise to philosophical perplexity.”20 Hacker also characterizes a perspicuous presentation as “a rearrangement of the rules of grammar that lie open to view, but are not readily taken in as a whole. They become surveyable by such a rearrangement.”21 Consequently, a perspicuous

Philosophical Hierarchies and the Status of Clarificatory Statements

229

presentation “enables us to grasp the structure of our mode of representation” and “will produce an understanding of logical connections which will dissolve confusion.”22 “This arrangement of the familiar rules for the use of words into a perspicuous presentation of the grammar of a part of our language will enable us to get an overview of our entanglement in the rules.”23 According to Baker and Hacker, therefore, grammatical investigation is the investigation of the rules of language. It seeks to dissolve philosophical problems by establishing a perspicuous order among such rules, clarifying in this way the logical relations of expressions. Here it is particularly important that this interpretation takes grammatical rules to be the object of philosophical investigation,24 not a mode of presentation that the philosopher may adopt, employing rules for specific clarificatory purposes. Thus, for Baker and Hacker, grammatical investigation is an activity whereby the rules of grammar themselves (conceived as rules we actually follow) become perspicuous through their rearrangement. Presumably, this idea of language itself becoming perspicuous and revealing its logic also explains their statement of a kinship between perspicuous presentation of language and the Tractatus’s idea of the logic of language showing itself. For perspicuous presentation, according to them, is not only a descendant of the “correct logical point of view” in the Tractatus but “also a descendant of the ineffability of ‘the Mystical.’”25 Baker and Hacker’s view that perspicuous presentation of language does not involve the employment of a mode of presentation is an important component of their interpretation. Apparently, their account of why Wittgenstein has no theses is based on this very idea. According to this view, Wittgensteinian grammatical investigation involves no theses because it merely records the grammar of language (collects reminders of language use) without any mediation by the instruments of presentation or the interference of opinions.26 In the capacity of such a neutral activity, philosophy really has no point of view of its own and accordingly is in principle incapable of doing injustice to anything or anybody. That is, given that philosophy employs no modes of presentation and is free from any opinions, there is no danger that the philosopher would come to impose any preconceptions of her own onto reality or her interlocutors’ language use. Hence, from the point of view of Baker and Hacker’s interpretation, it is impossible for the philosopher to be dogmatic insofar as she abides by the task of perspicuous presentation.

230

Philosophical Hierarchies and the Status of Clarificatory Statements

This alleged disappearance of the problem of dogmatism, however, raises suspicion. The worry is that instead of occupying no point of view, the philosopher might have simply lost awareness of one, and if that is the case, the danger of dogmatism is even more imminent than before, since now dogmatism is not even recognized as a threat. It is also notable that, contrary to Baker and Hacker’s denial of the problem of dogmatism, Wittgenstein himself not only acknowledges dogmatism as a logical possibility but regards it as an actually present risk. This is evident in the warnings he directs at himself in the Nachlass against being dogmatic in describing language.27 Indeed, in manuscript 136 from 1947—about ten years after his formulation of the idea of the turn as a strategy for avoiding the philosophical prejudice of crystalline logic—being free of prejudices is said to be the main difficulty of philosophy. As Wittgenstein observes in this context, what makes prejudices particularly difficult to avoid is that they themselves constitute forms of understanding, i.e., of rendering something comprehensible.28 More specifically, one may argue on the basis of Chapters 3 and 4 that the apparent disappearance of the problem of dogmatism in Baker and Hacker’s reading is only the result of their projection of the point of view of grammatical description onto language. In effect, they turn the idea of envisaging language as a rule-governed practice, which is a characteristic feature of the philosopher’s mode of examination, into a thesis about rule-governedness as a necessary characteristic of language. Similarly, the possibility of describing the use of an expression in terms of certain grammatical rules now emerges as a fact about the expression’s use as governed by these rules, which can simply be recorded. But if this is correct, rather than managing to leave behind metaphysics and dogmatism, Baker and Hacker’s interpretation of the concept of perspicuous presentation constitutes a relapse into metaphysics. These issues may be further elucidated as follows. There is a misleading aspect to the view that in philosophy one seeks to obtain a perspicuous view of the grammatical rules, as if such rules were one’s object of investigation. Although it is in a certain sense unproblematic, and even unavoidable, to talk about grammatical rules in such an objectifying way, the importance of a clear comprehension of what this means and does not mean cannot be overestimated. The sense in which grammatical rules are not the object of investigation but a mode of description can be clarified with the help of a comparison with measuring.

Philosophical Hierarchies and the Status of Clarificatory Statements

231

If I want to know how long a stick is, I measure it. I may express my interest by saying that I want to know how many inches long the stick is. But this does not mean that inches are my object of investigation. Rather, (the length of ) the stick is. Analogously, I may say that I want to know the grammar of a particular word, or which rules govern its use. As in the case of measurement, however, this just means that I will employ a particular mode of examination, namely description in terms of rules; that I will investigate the use of the word by tabulating rules for it. Accordingly, from the point of view of Wittgenstein’s approach to philosophy, the point of grammatical rules is ultimately that they constitute a means of presenting language use.29 Or as he says, the point of such rules is that they provide us with “an easily surveyable [leicht übersehbares] characteristic picture”30 of language use. Thus, even though one may indeed talk about the use of language in terms of grammatical rules, one should not allow oneself to get confused about the sense in which grammatical rules are and are not the object of grammatical investigation.31 Accordingly, although it is possible to say, as Hacker does (see above), that perspicuous presentation enables one to grasp the structure of our mode of presentation or language, it is crucial to remember that this talk about structures is merely a particular way of talking about language as the object of philosophical investigation. It is not a truth that language has a structure constituted by grammatical rules, though one may, for particular purposes, describe it as having such a structure. Put another way, Wittgenstein’s point in his later philosophy is not simply to argue that the Tractatus was looking for the structure of language in the wrong place. It is not that the structure of language is to be found on the surface of language rather than beneath it. Rather, the idea of a structure is to be recognized as a particular picture of language, and descriptions of language as possessing a structure constituted by rules are to be identified as employing a particular mode of presenting language. Parallel considerations apply to the notion of the “bounds of sense” that perspicuous presentation, according to Hacker, reveals.32 Again it is of great importance to recognize the relation between the bounds that language is taken to possess and the bounds of the mode of presentation the philosopher is using in describing language. For instance, to regard the concept of meaning from the point of view of the conception of meaning as rule-governed use is to envisage meaning as something constituted by rules. According to this conception, nothing counts as a

232

Philosophical Hierarchies and the Status of Clarificatory Statements

meaningful word unless its use is governed by rules (or parasitic on rulegoverned uses). Here then is a “bound of sense.” But it is crucial to see that this bound is characteristic of a particular conception or picture of meaning. It is not to be projected onto language and dogmatically claimed to be the bound of the concept of meaning, where meaning serves as the object of description. Rather than disclosing the limits of the concept of meaning, as if stating a fact about language, the conception of meaning as constituted by rules is a particular way of drawing limits to the concept of meaning, that is, of presenting the concept. The purpose of all these observations has been to elaborate on Wittgenstein’s strategy for avoiding dogmatism. Here the bottom line is ultimately that only if grammar is comprehended as the means of description is it possible to come to terms with the fact that philosophical descriptions have a grammar of their own and to counteract the problem of dogmatically projecting that grammar onto language. This possibility is lost from sight in Baker and Hacker’s interpretation, which consequently leaves one defenseless against dogmatism. But keeping these observations in mind, let me now outline an alternative to Baker and Hacker’s interpretation of the concept of perspicuous presentation. I start from what might be perceived as a difficulty regarding my suggestion that grammar is more appropriately characterized as the means than as the object of perspicuous presentation. Wittgenstein writes about the concept of perspicuous presentation in the Investigations: “A main source of our failure to understand is that we do not command a clear view [übersehen] of the use of our words.—Our grammar is lacking in perspicuity [Übersichtlichkeit]. A perspicuous representation produces just that understanding which consists in ‘seeing connexions.’ Hence the importance of finding and inventing intermediate cases. [New paragraph] The concept of a perspicuous representation is of fundamental significance for us. It earmarks the form of account we give, the way we look at things. (Is this a ‘Weltanschauung’?)”33 Do the first two sentences of this remark indicate that grammar is the object, not the means of grammatical investigation? This impression may arise if “grammar” in the second sentence is understood as synonymous with “the use of our words” in the first. (Now the second sentence provides the empty explanation that we do not have a perspicuous view of the use of our words because the use of our words

Philosophical Hierarchies and the Status of Clarificatory Statements

233

lacks perspicuity. Or it repeats the thought of the first sentence in different words.) Such an interpretation, however, is not the only possible one. The second sentence might also be read as specifying what problem Wittgenstein is talking about. Read in this way, the sentence describes the lack of perspicuity as a problem with our descriptions of language use. “Our grammar”, i.e. the grammatical descriptions of language we provide, fails to make language use perspicuous, and does not enable us to command a clear view of the use of words.34 As regards this alternative, however, the interpretational claim is not that the second sentence must be read in the way just suggested. Because talk about grammar, according to this interpretation, constitutes a particular way of characterizing language as the object of philosophical investigation, the interpretation is in principle compatible with both readings of the second sentence. Consequently, it suffices that the alternative reading is possible, not in conflict with what Wittgenstein says about these matters in the Investigations or elsewhere. Besides the ambiguous formulation in §122, it is easy to find additional textual evidence in support of the view that grammatical rules are a means of the perspicuous presentation of language. Wittgenstein writes about the notion of ordering, for instance: “ ‘Organise these things!’—what does this mean? Something like: ‘arrange them!’ [New paragraph] It could mean: bring about an order among them;—or also: learn to know your way about them, learn to describe them; learn to describe them through a system, through a rule.”35 Here a rule is clearly meant to be the means of description (mode of presentation) and a means of arranging. Similarly, Wittgenstein talks about obtaining a perspicuous view as a matter of finding a scheme through which to look at a case. The difficulty is: “The most important aspects of the case are not accessible to me because I do not have a perspicuous view of the possibilities.” And the solution: “Look at it in a new way, through a new scheme. [New paragraph] Make a different kind of comparison!”36 Comparisons are mentioned also elsewhere as means of creating perspicuity: “The purpose of a good expression [and] a good comparison is that it makes possible an immediate overview.”37 Finally, he writes about modes of presentation and perspicuous presentation as follows: “A philosophical problem, to be sure, arises from our not knowing our way about the grammar of our mode of expression. And one means of perspicuously framing the grammar [insertion: the use of our

234

Philosophical Hierarchies and the Status of Clarificatory Statements

words] to us is the introduction of a novel mode of expression particularly suited for this purpose.”38 On the basis of these passages, and taking into account the results of Chapter 3, it seems evident that perspicuous presentation, as Wittgenstein conceives it, does indeed involve the use of modes of presentation, such as rules and comparisons by means of which language use is made perspicuous. Perspicuous presentation, therefore, cannot be regarded as “a descendant of the ineffability of ‘the Mystical,’ ” insofar as this means that its goal would be to enable grammar to show itself without any modes of presentation being employed. Rather, perspicuous presentation involves, among other things, as Wittgenstein says, the introduction of novel expressions with the purpose of making it easier to achieve a clear comprehension of conceptual relations. (An example of such a novel expression would be a concept redefined in a simplifying way in order to highlight a certain aspect or aspects of the actual concept. More specifically, on the basis of the interpretation developed in Chapter 4, meaning defined as rule-governed use can be regarded as an example of such a simplified concept. Another example is reading as defined in the Investigations §156 and discussed in 3.6.) Furthermore, as indicated by Wittgenstein’s emphasis that the introduction of novel expressions is only one means of perspicuous presentation, there are also other means of rendering language use perspicuous. Perspicuity can be created, for instance, by using examples as centers of variation, as explained in 4.5, or more generally by arranging the relevant material, i.e., particular cases and examples of language use. (The employment of centers of variation may be considered a particular technique of arranging and ordering.) Wittgenstein writes about arranging facts as a means of creating perspicuity in the context of his discussion of (the anthropologist) Frazer in manuscript 110, where §122 of the Investigations is first drafted: “ ‘And so the chorus points to a secret law’ one feels like saying to Frazer’s collection of facts. I can present this law, this idea . . . by means of a schema of a religious ceremony, but also by means of the arrangement of facts [Tatsachenmaterials] alone, in a ‘perspicuous’ presentation.”39 Here arranging facts without deploying a schema is presented as an alternative way of articulating an organizing principle or a law and producing a perspicuous presentation. Notably, that Wittgenstein regards arranging with and without a schema as alternatives may also be

Philosophical Hierarchies and the Status of Clarificatory Statements

235

taken to speak against Baker and Hacker’s interpretation. Rules, such as definitions, exemplify schemas in the relevant sense. But they are not something to be arranged, as Baker and Hacker maintain; they are a means of arranging language use into a perspicuous order.40 Baker and Hacker’s characterization of perspicuous presentation therefore seems to conflate things that are best kept apart. That is, on the one hand, grammatical investigation does indeed involve arranging by means of rules. On the other hand, it involves alternative techniques of arrangement. But the methods of perspicuous presentation do not include arranging rules. At any rate, this would constitute a special, peripheral case, which cannot be used to characterize the activity of perspicuous presentation generally. It is also worth noting that the possibility of producing a perspicuous order simply by arranging relevant material need not be taken to imply that there are ways of creating perspicuity that do not involve any kind of mode of presentation. That arranging things in a particular way can make them perspicuous indicates that arranging can change the face of things, i.e., that arranging things can make them appear in a particular light. (For instance, museums are based on this idea: the way the exhibits are ordered plays a crucial role in ascribing meaning to them. Other examples are easy to find.) Consequently, such arrangements of things are best characterized as constituting modes of presentation in their own right. But let us now turn to the issue of how the Investigations §122 can be seen as summing up Wittgenstein’s conception of philosophy as interpreted in the preceding chapters. The remark just quoted from manuscript 110 is followed by the first draft of §122, including the following characterization of perspicuous presentation: “A perspicuous presentation produces just that understanding which consists in ‘seeing connections.’ Hence the importance of finding intermediate cases.”41 Here the second sentence about intermediate cases might again be read as referring to the technique of arranging examples without employing a separate scheme of organization. At least insofar as this technique is at stake, the role of intermediate cases seems clear. Such cases can be employed, for instance, to show how distant members of a family-resemblance concept with no common features are connected. Accordingly, it is plausible to think that by “ ‘seeing connections,’” Wittgenstein does not only mean seeing how different

236

Philosophical Hierarchies and the Status of Clarificatory Statements

concepts relate to each other. “Seeing connections” is also a matter of understanding how different cases falling under a single concept relate to each other and in this sense a matter of obtaining a perspicuous view of a particular concept or the use of a word. Wittgenstein is concerned with “seeing connections” in this latter sense, for instance, in the Investigations §92, where he talks about the essence of language becoming “perspicuous by arranging.” In this case the goal of arranging is evidently to make the concept of language perspicuous by enabling one to have a clear view of the various kinds of cases that fall under this concept and how they relate to each other. In §92 Wittgenstein also contrasts the method of perspicuous presentation by means of arranging with the search for a unifying characteristic that provides a key to the essence of language. This contrast between “comprehending an essence by seeing connections” and “comprehending an essence by getting hold of a common defining feature” can now be put to work in interpreting the last sentences of §122 (in a slightly modified translation): “The concept of a perspicuous representation is of fundamental significance for us. It earmarks our form of presentation, the way we look at things. (Is this a ‘Weltanschauung’?)” The question awaiting an answer, which the above contrast can help us provide, is how exactly perspicuous presentation “earmarks” Wittgenstein’s approach. What is distinctive about this notion and why does he employ it to characterize what is unique to his approach? An answer can be outlined as follows: The characterization of perspicuous presentation as a matter of “seeing connections” is meant to differentiate Wittgenstein’s “way of looking at things” from philosophy as a search for unified definitions or theses of essences in terms of universally shared essential features. Accordingly, even when a general scheme such as a rule is employed for the purpose of perspicuous presentation in Wittgenstein’s sense, it is not used to provide a unified, once-and-for-all account. Rather the scheme is employed to enable one to “see connections,” that is, to bring order to the cases falling under a concept by making it clear how such cases relate to each other, to relevant cases falling under other concepts, and so on. Similarly, Wittgenstein does not use examples as the basis for theses of essences but as objects of comparison intended to make it possible to “see connections.” Hence his approach is fundamentally different from metaphysical philosophy as the articulation of theses of essences.

Philosophical Hierarchies and the Status of Clarificatory Statements

237

In this connection, the exact wording of Wittgenstein’s statement that the concept of perspicuous presentation “earmarks our form of presentation, the way we look at things” is also important. This formulation of the significance of the concept of perspicuous presentation can be seen as crafted so as to leave no unclarity about the status of his conception of philosophy as perspicuous presentation. As explained, to determine a form of presentation or to articulate a way of looking at things is logically distinct from putting forward a thesis or a true/false statement. Hence when Wittgenstein writes in §122 that the aspiration to perspicuity “earmarks our form of presentation” (i.e., is a characteristic feature of his mode of examination), it would clearly be problematic to read him as putting forward a thesis about what philosophy must be (or what philosophy really is, correctly understood, and so on). Rather, he is merely providing a characterization of his approach and articulating a particular conception of philosophy. In connection with the idea of “a particular conception,” two things need emphasizing. First, to take Wittgenstein’s remark as a thesis about what philosophy must be, i.e., that it must be understood as perspicuous presentation in the sense of “seeing connections,” would be to fall into the confusion about modal notions that, according to him, is characteristic of metaphysical philosophy. It would turn a necessity that is a defining characteristic of the concept of philosophy as he defines it into a claim about what philosophy must be. Second, to say that Wittgenstein’s conception of philosophy is a particular conception does not mean that (nonsubjective) reasons could not be given in support of the adoption of this conception. As I have explained, Wittgenstein’s conception of the status of philosophical statements is motivated by problems relating to the practice of philosophy and intended to (dis)solve such problems, including the problem of dogmatism and confusions relating to modal concepts. Naturally, not just any conception of philosophical practice would count as a solution to these problems, and in this sense the choice of a philosophical approach is not a matter of subjective preference, even though no claim is made about Wittgenstein’s approach being the approach one must adopt.42 Finally, this interpretation also provides us with a way of reading the question, “Is this a ‘Weltanschauung?’” in parentheses at the end of §122. Perhaps the point of the question is to urge us to think about the notion of perspicuous presentation in relation to the notion of a

238

Philosophical Hierarchies and the Status of Clarificatory Statements

Weltanschauung and observe similarities and differences. From the point of view of the reading of §122 outlined above, Wittgenstein’s conception of philosophy is comparable to a Weltanschauung—a worldview or “way of looking at the world.” Like a Weltanschauung, Wittgenstein’s conception of philosophy is not a true-or-false thesis about anything. Rather, it is, just as he says, a way of looking at things. On the other hand, it might be a narrow interpretation of Wittgenstein’s reference to worldviews to take it as having to do with the status of his conception. The question in parentheses might also be read as urging the reader to think about the extent to which the aspiration to clarity characteristic of his philosophy already involves a commitment to a certain set of values, or is expressive of a particular approach to life not compatible with all other outlooks. Wittgenstein’s conception of philosophy might then also be regarded as expressing a particular worldview in this sense.43 The purpose of 6.1 and 6.2 has been to further clarify the status of grammatical statements. Let us now examine the question of whether grammatical statements are statements one must accept.

6.3 The (Alleged) Necessity of Accepting Philosophical Statements Had my interpretation been susceptible to the allegedly fundamental problem discussed in 6.1, this would have provided a way of arguing that Wittgenstein’s conception of language and meaning as constituted by rules is not “merely” an object of comparison. But even if one accepts my preceding attempts at clarification, one might still feel unsatisfied. If philosophical statements are objects of comparison, does this not mean that philosophy becomes an idle game of coining possible comparisons and ways of looking at things? This impression, as I will try to explain, is incorrect. It is true that there is no such thing as being forced to accept a grammatical remark. More specifically, and contrary to the suggestions of commentators such as Glock (see below), it is not the case, for instance, that insofar as one uses “meaning” in the normal way one has no option but to adopt the conception of meaning as constituted by rules, to recognize it as true on pain of irrationality. Thus this conception of meaning is not something one must agree on, as is the case with philosophical theses

Philosophical Hierarchies and the Status of Clarificatory Statements

239

conceived as truths. The same goes for other grammatical remarks. Nevertheless, the fact that philosophy does not force anybody to accept anything does not imply that philosophy loses its strength in any way. (Its true strength does not lie in a capacity to force anyone to accept anything, but in not needing to do so, one might say.) In particular, this does not open the door to an “anything goes” attitude or to irrationalism. I also argue in 6.6 for caution about underrating the work comparisons can do or the difficulty of finding enlightening ones. To distinguish the different senses in which one might argue for the necessity of accepting particular grammatical remarks, I will use as my example the conception of meaning as constituted by rules. Intimately connected with Wittgenstein’s notion of the arbitrariness of grammar and his emphasis on the distinction between true/false statements and conceptual determinations is his view that it is misguided to aspire to demonstrate by reference to the nature of things that language must be used in some particular way. Our concepts are what they are; there is no such thing as giving grounds why they must be what they are. This thought is summarized in the remark: “I would like to say: ‘I must begin with the distinction between sense and nonsense. . . . I cannot give it grounds.’”44 One way to explain Wittgenstein’s point is to say that there is no justification for a grammatical rule or a concept similar to the justification of a statement of fact by reference to the state of affairs that makes it true. Rules are, as explained, arbitrary. Wittgenstein’s argument for the arbitrariness of grammatical rules can be summed up as follows. Granted that “the language-game of correct/incorrect description of reality” is a rule-governed activity, any descriptions of reality one might appeal to in attempting to justify grammatical rules already presuppose grammatical rules. Hence, either the alleged justifying descriptions presuppose the grammar they are intended to justify, or they presuppose a different grammar. But this either means that the alleged justifications are circular or raises the question, why is this grammar given a privileged status?—to which one cannot answer that its privilege derives from its leading to the correct description of reality.45 But this explanation fits some cases better than others. It is well suited for concepts such as color, which one might indeed feel tempted to justify by reference to reality. (Accordingly, an example that Wittgenstein

240

Philosophical Hierarchies and the Status of Clarificatory Statements

examines critically in this connection is the sentence “There are really four primary colors.”) Nevertheless, in the case of other more abstract concepts, such as meaning and language, the temptation is not so much to talk about a match between reality and concepts, but rather to refer to the “inner nature” of the phenomenon or the concept. Here one might feel inclined to say something like: “The meaningfulness of words necessarily presupposes the rule-governedness of language because their meaningfulness requires the possibility of distinguishing between correct and incorrect uses of words. That is why language must be rulegoverned.” Let us look at this type of explanation more closely. Assume that the meaning of a particular word depends on the rules according to which it is used. In this case one might be tempted to argue for the necessity of using the word according to these particular rules by saying that this word could not mean whatever it means unless it were used according to them. However, this alleged explanation for why the word must be used in this particular way lacks any explanatory power. For if the word’s meaning depends on the rules that actually govern its use, then were those rules different, there would be nothing that could be referred to as what the word fails to mean. The alleged account of why the word must be used the way it is used, or the apparent claim about necessity, simply evaporates, because there is no way to specify the aim or task that the rules are supposed to serve independently of whatever their aim or task actually is. (Curiously, the appearance that the conditional “could not” has an explanatory function here depends on one not taking seriously the idea that the word’s meaning depends on rules.) Wittgenstein expresses this point in the Investigations as follows: “If someone says ‘If our language had not this grammar, it could not express these facts’—it should be asked what ‘could’ means here.”46 Accordingly, one might say that grammar does not enable language to function the way it functions. Grammatical rules only spell out how it functions. They describe the use of language.47 Regarding the concept of language itself, then, the above points imply that there is no argument to the effect that language must be rulegoverned on the grounds that it could not function the way it functions unless its use were governed by rules. Presumably, Wittgenstein has this sort of consideration in mind when he says of our color concepts and the concept of agreement: “ ‘If humans were not in general agreed about the colours of things, if undetermined cases were not exceptional,

Philosophical Hierarchies and the Status of Clarificatory Statements

241

then our concept of colour could not exist.’ No:—our concept would not exist.”48 The crucial difference between statements with “would” and those with “could” is that the former are descriptive. They give no reason or explanation for why language use must be what it is—governed by rules, for example, or characterized by agreement. Wittgenstein’s statement with “would” just highlights what one may regard as a constitutive feature of our color concepts. To say that agreement is constitutive of them is to say that, given our color concepts, we would not recognize a concept as a color concept in their absence. As to the status of grammatical remarks, these considerations exclude an unconditional or absolute sense of arguing for the necessity of a grammatical remark such as “The meaning of a word is its use” or “Language is a rule-governed practice.” One cannot argue for the necessity of accepting the conception of meaning as constituted by rules or the conception of language as a rule-governed practice on the grounds that language must be rule-governed by inner necessity or because this corresponds to the true nature of language. Provided that it is impossible to establish the necessity of the conception of meaning as constituted by rules in an absolute or unconditional sense, could it nevertheless be established in a conditional sense? There might be a temptation to argue along the following lines: No reason can be given for why word-meaning must be what it is. But given that it is what it is (or that our concept of word-meaning is what it is), anyone who talks about word-meaning in the normal sense must accept the conception of meaning as constituted by rules on pain of contradicting themselves. That is, even though a concept can never be justified on the grounds that it must be whatever it is, individual cases of the use of the concept may be critiqued on the basis of an analysis of its necessary characteristics.

This conforms broadly to the form of Kantian critique. Arguably (although it is not possible to discuss this here), Kant’s analysis of the conditions of possibility of knowledge in the Critique of Pure Reason cannot—and is not intended to—justify knowledge claims generally, i.e., establish the possibility of knowledge as such. Rather, the analysis presupposes the possibility of knowledge. Provided, however, that knowledge is possible, and assuming the correctness of Kant’s analysis,

242

Philosophical Hierarchies and the Status of Clarificatory Statements

the legitimacy of individual claims to knowledge may be examined on the basis of the criteria that Kant has established. The function of grammatical remarks in philosophy, one might suggest, is analogous. Here we come to issues already addressed in Chapters 3 and 4. As regards the method of analysis in terms of necessary conditions, some important consequences of Wittgenstein’s conception of the status of philosophical statements should be noted. As explained in 4.7, the necessity of a condition for certain cases falling under a concept does not imply its necessity for all cases falling under the concept. This latter universal statement follows only on the assumption that the concept has a single, unified essence. But as we have seen, Wittgenstein questions this assumption. Accordingly, one should beware of projecting the characteristics of the examples that one has actually analyzed onto other cases falling under the concept and making a dogmatic claim about what they all must be. Instead, statements concerning necessary conditions are to be understood as objects of comparison. Hence the proposed Kantian analogy is misleading. Here it is helpful to contrast my interpretation of the status of grammatical remarks with that of Glock,49 according to whom such remarks indeed must be accepted and are in this sense indisputable. According to Glock, grammatical propositions have a “normative function.” This means that grammatical propositions “constitute standards for our use of words” and “remind us of our own rules.” As for the role of grammatical reminders in philosophy, “Such reminders point out that outside philosophy we use certain words in such-and-such ways. That we do so could of course be the subject of an empirical statement. . . . But the special function of grammatical reminders is to draw attention to the violation of linguistic rules by philosophers, a violation that results in nonsense.”50 More specifically: “It may be granted that mere deviance from ordinary linguistic practice does not constitute a philosophical mistake. But the idea is to demonstrate a certain kind of inconsistency in the philosophical positions or questions attacked, an inconsistency concerning the uses of words.”51 The idea is that grammatical propositions or reminders are perfectly indisputable because they articulate standards of sense that the speaker already tacitly acknowledges in her language use. Consequently, grammatical reminders must be accepted on pain of inconsistency or

Philosophical Hierarchies and the Status of Clarificatory Statements

243

irrationality. Their violation would result in nonsense because one would thereby be violating one’s own standards of sense. Hence, “we should expect” “a universal assent” to grammatical propositions or reminders.52 Glock’s interpretation of the role of grammatical remarks seems to conform to the Kantian model described above.53 I agree that philosophical problems can sometimes be understood as arising from inconsistencies in language use and that philosophical questions may sometimes be nonsensical. Nevertheless, I dispute the conception that grammatical remarks may be put forward as something one must accept or suffer irrationality: that it is possible, as Glock says, to demonstrate an inconsistency or put forward arguments that “philosophers will have to accept.”54 Insofar as this means that a person may be forced to admit an inconsistency in her use of language on the basis of her use of language in a way that contradicts a grammatical reminder that she acknowledges, I do not think that this view can be attributed to Wittgenstein. We may examine this issue more closely by asking: what is the status of the statement that there are propositions of a special kind that command “universal assent” and would necessarily be agreed upon? One possibility would be to maintain that this is an empirical statement or a hypothesis concerning language. Given Wittgenstein’s explicit denial that his observations about language are empirical statements, however, I will not consider this possibility any further.55 Alternatively, one might maintain that it is necessary that there be such propositions. The statement that there are propositions necessarily agreed upon might then itself be accorded the status of a grammatical proposition. Indeed, does not Wittgenstein say that language is characterized by agreement in definitions?56 But one must take great care in interpreting this and similar statements. As I will argue, Wittgenstein’s point that the possibility of language use (for communication) presupposes an agreement on definitions does not imply the necessity of agreement on descriptions of language use. Let us accept that language use is characterized by agreement to the extent that it is rule-governed. As Wittgenstein notes, the concepts of rule and agreement are closely connected.57 Even so, this does not justify the expectation of universal assent to the philosopher’s clarificatory rules. This expectation presupposes that clarificatory rules stated by the philosopher are the rules by which actual language use is governed. Provided there is agreement on the rules actually followed in

244

Philosophical Hierarchies and the Status of Clarificatory Statements

language use and granted the assumption that clarificatory rules are the rules of actual language use, the agreement on the actual use of language would extend to cover rules used to describe it—barring problems with the identification of correct descriptions. But as I have explained, it is central to Wittgenstein’s outlook that he does not take clarificatory rules to be rules by which language is actually governed. Rather, clarificatory rules are a means of describing actual language use by way of comparison. Consequently, agreement on actual language use does not entail agreement on descriptions of language use. As regards the demonstration of contradictions in an interlocutor’s language use, a conflict between the interlocutor’s actual language use and the rules she acknowledges as descriptive of it implies a contradiction in her language use only on the assumption that the descriptive rules are rules she actually follows. Granted this assumption (and excluding any mistaken acknowledgments of descriptions), one could infer from a conflict between her actual language use and the rules she acknowledges that there is an inconsistency in her position. However, because clarificatory rules as conceived by Wittgenstein are not rules we actually follow, there are no demonstrations of inconsistencies of this sort. The problem with Glock’s account of Wittgenstein’s method, therefore, is that it is based on a conception of the role of grammatical reminders that Wittgenstein rejects as leading to dogmatism (and so on), as explained in Chapters 3 and 4. Abandoning the assumption that clarificatory rules are rules actually followed in language use, the practice of clarification can be redescribed. Rather than capturing the interlocutor’s language use exactly as it is, a clarificatory rule supplies an approximate description designed to capture a particular aspect of a concept. Because the rule does not describe the interlocutor’s language use exactly as it is, no inconsistencies can be deduced from conflicts between the interlocutor’s actual language use and the descriptions of it to which she has assented. For instance, one may accept the rule “The meaning of a word depends on rules for its use” and still maintain that meaning does not always depend on rules—a position Wittgenstein himself adopts, as argued in Chapter 4. Crucially, insofar as the statement “Meaning depends on rules” is not put forward as an exclusive thesis about actual language use but as an object of comparison, there is no conflict in

Philosophical Hierarchies and the Status of Clarificatory Statements

245

maintaining that meaning both is and is not constituted by rules. The contradiction arises only insofar as it is assumed that one must be committed either to the exclusive definition of meaning as constituted by rules or to its negation.58 Consequently, grammatical investigation allows for demonstrations of inconsistencies only insofar as the interlocutor acknowledges each step of the demonstration as well as its conclusion. Thus she might agree that an inconsistency has been demonstrated, but she does not have to agree to this. The critical point is that grammatical rules, as Wittgenstein conceives them, are instruments that can be used to describe language use—and so to help a person detect an inconsistency. For example, descriptions of inferential patterns from what the interlocutor says to apparent inconsistencies in her position may be used to bring to light problems in her position. But because grammatical reminders do not state truths about people’s language use, in grammatical investigation there are no deductions of inconsistencies according to the model of deducing further truths from already established ones.59 One cannot infer: “She admits this and she admits that, hence she must admit that too and her position is inconsistent.” In this sense grammatical remarks, or the interlocutor’s acknowledgment of such remarks, do not function in Wittgensteinian clarification as premises in a deductive argument. Rather, the last inference is replaced by something of the form: “You admit this and you admit that, so would you admit that too? If you do, would you not agree that there is an inconsistency?” In conceptual investigation there are demonstrations of inconsistencies only in this latter, clarificatory sense. Importantly, this redescription of the practice of clarification does not imply tolerance of incoherence or irrationalism. Logic is not disregarded and philosophy is not weakened in any way when the conception of philosophical statements as necessary to accept on pain of irrationality is rejected.60 Here it is notable that the conception that philosophy could necessitate an agreement on pain of irrationality is itself idealized in an important sense. It is then crucial that this idealization not be confused with the actual reality of clarification. Such confusion may be described as a variant of the metaphysical projection of philosophical descriptions onto language, as the following example illustrates.

246

Philosophical Hierarchies and the Status of Clarificatory Statements

Imagine that an interlocutor has acknowledged that a rule proposed by the philosopher aptly captures her use of a term in a certain meaning. Nevertheless, as things turn out, the interlocutor then also insists on using the term (in the relevant meaning) in a way that does not conform to the rule she has acknowledged. Must she admit that this is inconsistent? Only if the rule is taken to cover all instances of the use of the term (in the relevant meaning) and to show what they all must be. Thus, the demonstration of inconsistencies assumes the projection of clarificatory rules onto language as theses about necessities in language use. The possibility of employing rules as instruments of clarification is not recognized. The sense in which the conception of philosophy as coercive is idealized can now be explained. Let us accept that an ideally rational person cannot knowingly hold an inconsistent view without thereby compromising her rationality. However, when an actual person insists on a view that appears contradictory, it might not be because she is irrational or mistaken. It might also be that the view she holds is too complex to be captured neatly in a unified description. Rather, doing justice to the various aspects of her view might require description in terms of apparently contradictory definitions used as objects of comparison. Accordingly, when the philosopher describes the interlocutor’s position as either inconsistent or irrational, the description is in effect treated as ideally valid, i.e., as capturing the interlocutor’s language use exactly as it is and without any incongruity. Clarification, however, is an activity that takes place in an actual, nonideal world. To avoid doing injustice to the interlocutor under such conditions, the philosopher’s descriptions cannot therefore be assumed to possess ideal validity, nor, consequently, the kind of coercive force Glock assumes. But it does not follow from the above that it can never be concluded that the interlocutor is irrational or self-contradictory. The interpretation I am proposing does not make the interlocutor infallible, as Glock makes the philosopher when the latter declares that the interlocutor has to accept something. Rather, such charges that the interlocutor is irrational are to be seen as instances in which the clarificatory activity breaks down. They constitute an acknowledgment that clarity cannot be reached: that either the philosopher or the interlocutor (or both) are falling short of clarity without being able to do anything about it at that point.61

Philosophical Hierarchies and the Status of Clarificatory Statements

247

Another way of characterizing the practice of clarification as redescribed here would be to say that what is at stake is a rejection of the metaphor of philosophy as a battle for the winning argument and position. For it is characteristic of this redescribed practice that in its context, an individual’s consent to a depiction of her position may never be used as a mere starting point for an attack that aims to demonstrate the unviability of her position and that no longer requires her consent.62 These points may be further elucidated by discussing Wittgenstein’s concept of agreement and its role in philosophy.

6.4 The Concept of Agreement and the Problem of Injustice According to Wittgenstein, “Philosophy only states what everyone grants to it.”63 From the point of view of Glock’s reading, this remark would be read as stating that there is something (grammatical statements) that philosophy reveals and that everyone (participating in certain linguistic practices) must grant to it.64 However, the conception of grammatical statements as instruments of clarification designed to help one get clear about philosophical problems leads naturally to an alternative interpretation that brings Wittgenstein’s surprising and difficult statement down to earth. Wittgenstein’s point in §599 can be taken to be that, given that philosophy is an activity of clarification, something not agreed upon simply does not qualify as a philosophical, clarificatory statement. This is because what counts as clarification is ultimately defined by its clarificatory effect. As he says: “An explanation of words has clarificatory value for the person to whom it clarifies something, upon whom it has a clarifying effect. Independently of that it is not an explanation.”65 According to Wittgenstein, therefore, something qualifies as a clarification only insofar as it actually clarifies something. “To clarify,” one might say, is a success-verb. What does not clarify, despite intentions to do so, is only an attempt at clarification. This conception of what counts as a clarificatory and a philosophical statement implies that in philosophy the very possibility of putting forward theses, in the sense of controversial statements to be debated, is excluded. As Wittgenstein writes in the Investigations: “If one tried to advance theses in philosophy, it would never be possible to debate them,

248

Philosophical Hierarchies and the Status of Clarificatory Statements

because everyone would agree to them.”66 Similarly, he remarks in the same manuscript where this remark is first drafted: “only when I say what is self-evident is it philosophy.”67 Numerous statements can also be found in Wittgenstein’s lectures in which he is willing put to the side anything not be agreed upon. For instance: “I won’t say anything that anyone can dispute. Or if anyone does dispute it, I will let that point drop and pass on to say something else.”68 But can one really take seriously Wittgenstein’s statement that he would drop any point not agreed upon? I believe one can, but only if grammatical statements are understood as instruments of clarification. For as long as grammatical reminders are conceived according to the model of “grammatical truths,” a feeling persists that Wittgenstein’s statement about dropping anything not agreed upon is merely rhetorical. Secretly he knows that insofar as one uses the word “meaning” in the normal way, for example, one will have to acknowledge the rule “meaning is use” as the description of the word’s use. Or as Glock says above, “we should expect” “a universal assent” to grammatical statements. Similarly to Glock, Hacker writes about Wittgenstein’s notion that an inner process stands in need of outer criteria and the conception of mathematical propositions as rules: “If challenged, they can be withdrawn, and painstaking description of the use of words resumed until it becomes clear that the way we use psychological predicates in the third person involves behavioural criteria for their application, and that the characteristic employment of mathematical propositions is normative.”69 According to Hacker’s interpretation, therefore, the end result, or what will be agreed upon, is already known in advance. Discussion will be resumed until agreement about the grammatical reminders is reached. But the contrast with Wittgenstein could hardly be starker when he says: “I have no right to want you to say that mathematical propositions are rules of grammar.”70 Wittgenstein does indeed propose that the matter be investigated, but unlike Hacker, he does not suggest in any way that the outcome is already decided. Rather, Wittgenstein denies himself the right to want a specific outcome.71 However, the suspicion that Wittgenstein cannot really mean that he would give up his statements about grammar is dissipated as soon as such statements are understood as instruments of clarification. This adjustment—which is essentially a switch from one picture of grammatical statements to another—transforms the whole setting of clarification.

Philosophical Hierarchies and the Status of Clarificatory Statements

249

For from the point of view of the instrumental conception, dropping a statement not agreed upon emerges not as equivalent to abandoning an important “grammatical truth” but as analogous to exchanging one tool for another more suitable for a task at hand. When viewed in the light of this picture, Wittgenstein’s readiness to drop any grammatical remarks not agreed upon begins to sound not only comprehensible, but sensible. More specifically, which clarificatory statements or conceptions one can and must employ in tackling a problem depends on what exactly those to whom the clarification is offered find problematic. This also implies that in some cases, one might not be able to employ certain clarificatory statements or conceptions because the participants in the discussion take those conceptions themselves to be problematic. In such a case, the clarifier has to find other means to address the issue at hand. Hence, as this sketch illustrates, in selecting the right instruments for clarification one must also take into account the interlocutors’ intellectual backgrounds, i.e., the philosophical convictions that provide a background for the problem and constitute a context in which it is to be resolved. Now a readiness to drop clarificatory statements if they are not agreed upon emerges as a readiness to respond to a clarificatory task as required by its specific circumstances. It does not amount to holding no views about grammar or philosophical matters, or to overconfidence about others’ acknowledgment of certain descriptions of grammar. These are false options.72 Importantly, this account of Wittgenstein’s clarificatory practice manages to bring down to earth the notion of agreement in his philosophy. According to the proposed reading, his claim is not that there are statements that everyone who adheres to a certain practice of language must agree upon. It is not, in this sense, of the essence of grammatical statements that they will—eventually, necessarily—be recognized and agreed upon by all rational beings equipped with certain concepts. Indeed, since neither Wittgenstein nor any Wittgensteinians have actually produced any statements that are both readily recognizable as philosophical and agreed upon by everyone, it is hard to see why the idea that there must be such statements should not be characterized as a theoretical postulate of certain Wittgensteinians, and a controversial and implausible one at that.73 Instead of postulating such extraordinary statements, however, one may simply maintain that to agree on a grammatical statement is to

250

Philosophical Hierarchies and the Status of Clarificatory Statements

agree in the context of some specific issue that the statement is capable of clarifying some aspect of the matter at hand. Thus one is not expected to accept grammatical statements in the abstract; they are only offered in response to particular actual unclarities. Accordingly, it is misleading to conceive grammatical reminders as truths about concepts that must be recognized by all users of those concepts, as suggested by the picture of their necessary acceptance on pain of irrationality. Rather, grammatical reminders are instruments employed to dissolve actual philosophical problems that particular people have. The interpretation of the concept of agreement in Wittgenstein’s philosophy proposed here seems to remove a certain air of implausibility or disingenuousness that surrounds it. Nevertheless, it is important to notice that the preceding description of the practice of clarification remains idealized in several respects. Rather than indicating a problem with Wittgenstein’s conception of philosophy (or my interpretation of it), however, this is indicative of the status of his conception. It is clearly not true that in real life the roles of the participants in a philosophical discussion would be as neatly delineated as portrayed above. In reality the roles of philosopher and interlocutor cannot be identified with specific participants in a discussion. It is crucial, then, that the preceding depiction is not intended as a correct factual description but as a simplified model to be used as an object of comparison. The purpose of the model is to spell out a way of conceiving philosophical discussion and the participants’ roles in it and so to introduce order and perspicuity into the actual blurred reality of philosophical discourse. The point of conceiving philosophical discussion in this way is that it dissolves certain problems relating to the philosopher’s role, in particular the problem of the philosopher doing injustice to the interlocutor by imposing views on her as something she allegedly must accept. As Wittgenstein says about the task of philosophy: “Our only task is to be just. That is, we must only point out and resolve the injustices of philosophy, and not posit new parties—and creeds.”74 It is plausible that Wittgenstein is referring to philosophical parties as defined by their theses about how things must be. He himself, on the other hand, rather than trading in such theses, merely employs philosophical definitions (and so on) as instruments of clarification. Consequently, because it is

Philosophical Hierarchies and the Status of Clarificatory Statements

251

characteristic of this use of definitions that they can be withdrawn if not agreed upon, there is no need to insist on any statements of grammar in the context of Wittgenstein’s approach—or even to insist on continuing the examination until an agreement is reached. With the disappearance of the need to insist on the acceptance of any particular grammatical statements, the problem of injustice also disappears. In this sense the requirement of agreement is to be regarded as part of Wittgenstein’s design for a method that can dissolve the problem of injustice in philosophy. To describe the problem of injustice in more general terms, one might speak of a problem of conceptual domination by the philosopher. This type of domination arises in the context of certain more traditional conceptions of philosophy that assume the philosopher to have an access to universal theses of essence or to universally valid conceptual determinations. Assuming that the philosopher and her interlocutor share a common system of concepts, the problem is that the philosopher seems to be in a position to legitimately coerce the interlocutor to accept whatever conceptual clarifications the philosopher comes up with, because of her presumably more advanced capacity to clarify those shared concepts. Glock’s view that the philosopher is able to tell her interlocutors what they must accept or suffer irrationality seems to exhibit features of this traditional view of the philosopher’s role. In an analogy to the traditional philosopher’s access to insights about essences, for Glock it is the philosopher’s access to grammatical rules that puts her in the position of intellectual power. Once such a rule is agreed upon, the philosopher is, according to this view, licensed to apply it to any further cases that may arise without having to consult the interlocutor and is allegedly able to tell the interlocutor what she must think about such further cases. But if language is complex enough not to fit neatly any such systems of rules, as Wittgenstein suggests, then a philosopher with Glock’s approach is in constant danger of doing injustice to her interlocutors.75 Another sense in which Wittgenstein’s conception of the status of philosophical remarks is idealized is this: If one takes his claim that only what is agreed upon or self-evident qualifies as a philosophical statement to be a thesis about what a philosophical statement must be, very little of

252

Philosophical Hierarchies and the Status of Clarificatory Statements

what is said when philosophy is actually practiced qualifies as philosophical. His statement, therefore, seems to express a very strict view of philosophy unable to accommodate what most philosophers would call “philosophizing.” However, this problem disappears, just like the problem of the unrealism of his characterization of the philosopher’s and the interlocutor’s roles, when the statement is understood as articulating a philosophical model to be used as an object of comparison. This means that the conception of philosophical statements as self-evident and agreed upon constitutes not a thesis that every (legitimate) philosophical statement must fit but a defining characteristic of Wittgenstein’s approach, a principle according to which the activity of clarification is to be carried out. It is also notable that according to Wittgenstein, the activity of clarification or philosophy consists not only of making statements but also, for instance, of asking questions. As he remarks: “In philosophy it is always good to put a question instead of an answer to a question. For an answer to the philosophical question may easily be unjust; disposing of it by means of another question is not.”76 If agreement is interpreted as a requirement that everything we say in philosophy must meet, asking questions does not qualify as part of philosophizing. (Or at most, rhetorical questions will qualify.) Hence it cannot be—would be contradictory to say—that Wittgenstein intends the requirement of agreement as something all philosophical utterances must meet.

6.5 The Criteria of the Correctness of Grammatical Remarks But if grammatical remarks are not to be conceived on the model of truths, how should one understand the correctness of philosophical, clarificatory statements? To clarify the status of philosophical statements and what it means for a grammatical remark to be correct, we must reenter the crosscurrent encountered in Chapter 3, where I discussed the distinction between rules and factual descriptions and the notion of describing language with the help of rules. We need to ask, what is the criterion for the correctness of a philosophical description of language with the help of a grammatical rule? Here it is as if pressure were directed at the notion of a grammatical rule from different directions. On the one hand, one wants to talk about the correctness

Philosophical Hierarchies and the Status of Clarificatory Statements

253

and incorrectness of descriptions and philosophical clarifications. On the other hand, rules are not correct or incorrect about anything. I argue that, contrary to what one may be tempted to think, clarificatory rules cannot be taken to be any different from other rules with respect to their arbitrariness. Arbitrariness here means that grammatical rules cannot be derived from the reality they are used to describe, i.e., that language does not necessitate its description in terms of any particular rules. In this sense language is at the same level as any other object of description. Similarly, the empirical fact that language use is governed by rules does not allow one to claim that grammatical rules are nonarbitrary. After all, grammatical rules are not empirical descriptions of language use. But in order to get clearer about this, let me begin by recalling the cases in which we talk about rules of language and the occasions where philosophical clarification is needed. According to Wittgenstein, we talk about rules of language in the following kinds of situations. In one kind of case, rule formulations play a role in the language-game itself. For instance, the game is taught with the help of statements of a rule. Or an expression of a rule, such as a table, is employed in the game for some purpose (say, as a memory aid). In a second kind of case, rule formulations play no role in the game. A distinction is drawn between correct and incorrect ways of playing the game—some “moves” are barred as incorrect, and so on— but the distinction is not articulated in the form of statements of rules. Nonetheless, the game can be characterized by an observer as proceeding according to certain rules.77 Philosophical clarification addresses cases where it is not clear how to characterize and understand the use of expressions. For if their use is clear, there is no need or room for clarification, and clarificatory statements have no role to play. This observation may sound trivial but is of great importance. It means that philosophy is concerned with cases of the second rather than the first kind distinguished above. When a game is taught according to explicit rules or when expressions of rules play a role in the game, it is obvious what the rules are. They are already tabulated, so to speak, and anyone uncertain about the rules need only consult a rule book or any competent player (speaker). Similarly, it would be odd to say, for example, that citing a well-known definition of the word “bachelor” or reading aloud a rule from a rule book constitutes an announcement of the results of an investigation concerning rules. Rather,

254

Philosophical Hierarchies and the Status of Clarificatory Statements

an investigation is needed—or possible at all—only when the rules have not been tabulated, or when one is not satisfied with how they are tabulated—when, for instance, the available definitions lead into paradoxes.78 It is therefore misleading to model philosophical clarification on the first kind of case distinguished above, where rules play a role in the game. When judging the correctness of a rule formulation that purports to describe a linguistic practice, there is no canonical formulation against which it could be compared for aptness. In particular, although one might characterize a clarificatory rule by saying that it brings to light a rule implicit in a linguistic practice, it is not that the philosopher’s statement of a rule is compared with such an implicit rule for correspondence. Rather, the philosopher’s statement gives expression to a rule implicit in language, thus making it explicit. The philosopher’s object of description when she states a rule, in other words, is not a rule of language. Rather, the statement of a rule, as I have argued, is a means of describing a linguistic practice or uses of language. More specifically, to describe the uses of language by stating a rule is to describe instances of language use as conforming to a certain schema, i.e., to organize the uses of language in a systematic way.79 Here the statement of a rule articulates an organizational principle according to which instances of language use are ordered in the sense that out of the numerous regularities that language use exhibits, it brings a certain regularity to the fore and to our attention. Accordingly, the correctness of the philosopher’s statement of a rule is not judged on the basis of whether it corresponds to a schema, principle, or rule implicit in a linguistic practice. Rather, its correctness is decided on the basis of how the instances of language use fit it. This, of course, is something quite familiar in philosophy, as exemplified by the practice of giving definitions and responding with counterexamples. This practice shows what it is, generally speaking, to examine the correctness of a rule as a description of language use. Here the function of counterexamples, of course, is to show that the rule the use of a word is supposed to fit does not capture all relevant instances after all. (This characterization does not explain what it is to evaluate a rule used as an object of comparison, but this is a different point, to which I return shortly.) Now, insofar as the purpose of a statement of a rule employed to describe language use is not to describe a rule implicit in language but to

Philosophical Hierarchies and the Status of Clarificatory Statements

255

serve as a means of organizing instances of language use according to a rule (as forming a certain kind of pattern), such descriptive rules seem arbitrary indeed in a certain sense. It is logically possible to arrange instances of language use into general patterns in more than one way, apparently in an infinite number of ways (limited only by one’s imagination). Hence rules used to describe language cannot be derived from language as the object of description, and language does not make necessary any particular characterization of its use with the help of a rule.80 But if rules used to describe language are arbitrary in this sense, what does the correctness of clarificatory rules mean? Naturally, I am not suggesting that any clarificatory rule is as good as any other. The justification and significance of such rules derive from the purpose they serve. As Wittgenstein says about philosophical description: “this description gets its light, that is to say its purpose, from the philosophical problems.”81 Accordingly, clarificatory rules are to be evaluated on the basis of the clarificatory work they can do, i.e., whether they allow one to make sense of the uses of language with which one is concerned, rather than giving rise to conflicts and paradoxes. Wittgenstein characterizes the role of clarificatory rules as follows: “The difficulty lies only in understanding how laying down a rule helps us. Why it calms us after we have been so profoundly uneasy. Obviously what calms us is that we see a system which (systematically) excludes those constructions that always made us uneasy, those we were unable to do anything with, and which we still thought we had to respect.”82 Examples can clarify this. The referentialist model of meaning, which maintains that for a word to have meaning is for it to stand for something, seems to capture something important about meaning. But although this model might fit some cases well, it is problematic in others. For instance, in the case of numerals it invites the postulation of abstract objects for which these words stand but whose nature is highly problematic. (What is the ontological status of such objects, and what is it to refer to them and to have knowledge of them and their relations?) It is not clear what one is to make of such objects, and this may create uneasiness. At the same time, however, it seems that one has to respect these objects, as they are apparently needed to explain the meaningfulness of the relevant kind of words. The rule “The meaning of a word is its use in language” may

256

Philosophical Hierarchies and the Status of Clarificatory Statements

then be offered as a response to the problem with abstract objects. As a consequence the need to postulate such objects disappears. Just as Wittgenstein says in the passage above, the rule excludes those constructions that made one uneasy, thus dissolving the problem relating to them. The problem about abstract objects does not arise at all in the context of the conception of meaning as use.83 The notion that the meaning of a word is its use in language seems similarly able to dissolve philosophical problems with mentalist accounts of meaning. According to mentalism, the meaning of a word depends roughly on an interpretation given to the word by the speaker or hearer. The problem with this account is that it does not allow one to distinguish between the meaning of a word and the meaning a particular speaker or hearer takes the word to have. Consequently, the explanation does not leave room for the possibility of misunderstanding the meaning of a word. In contrast, if one regards meaning as use, the possibility of misunderstandings is readily understood. On this account, the meaning of a word does not depend on an individual act of understanding but is a function of the use of the word in language. Clearly, one can be mistaken about the use of a word in language.84 The conception of meaning as use, then, seems capable of dissolving problems to which certain other accounts of meaning lead. If it can really achieve this, one may justifiably characterize it as a correct account. But what exactly does that mean? Naturally, the conception’s ability to dissolve particular problems such as the above does not suffice to show that we must always understand meaning according to this model. Such a conclusion would be extremely hasty. Similarly, it would be misleading to maintain that the dissolution of a problem constitutes, as it were, inductive evidence for the correctness of the clarificatory statement about meaning. Now the suggestion would be that we cannot be sure whether or not the conception of meaning as use has really dissolved the problem with referentialism.85 Rather, given Wittgenstein’s conception of philosophy as the clarification of particular problems, a clarificatory statement or a model is to be understood as correct insofar as it can dissolve those problems it was meant to dissolve. This is what it means for a philosophical account to be correct. Consequently, it only makes sense to talk about the correctness of grammatical remarks in the context of actual philosophical problems, not in the abstract—a point that coincides with Wittgenstein’s conception of philosophy as a concern with particular actual problems.86

Philosophical Hierarchies and the Status of Clarificatory Statements

257

Accordingly, as regards the status of the definition of meaning as use, it might be characterized as one possible way of arranging “our knowledge of language use”—“an order with a particular end in view . . . not the order.”87 It is characteristic of this ordering of language that it connects the concept of meaning closely with the concept of rule, and that the sound of words, for instance, is regarded as irrelevant to meaning. Nevertheless, as we can say on the basis of Chapter 4, Wittgenstein is not committed to this order as the philosophical order and arrangement but is prepared to also recognize cases of the meaningful use of words that do not fit it. However, insofar as the arrangement is a response to particular philosophical problems, there is no need to try to make all cases of word-meaning match the definition of meaning as use. In this connection it is also important that insofar as philosophical orderings are relative to particular problems and there is no ultimate, definitive list of “all philosophical problems,” there is no room for speculating about whether some order might not solve all problems relating to a concept after all. All such speculation is idle as long as the meaning of “all problems” is not determined. In this sense there is no ultimate philosophical order of our concepts that must be accepted. Here it is important to note, however, that the account of the correctness of philosophical statements I have provided is not meant to constitute a sufficient definition. Presented as such, the definition is open to objections such as: “If knocks on the head made our philosophical problems disappear, we would have to recognize knocks on the head as instances of philosophical clarification.” As Glock points out, what is missing here is an internal relation between the problem and its solution.88 From the point of view of the present interpretation, this internal relation can be characterized as follows. A philosophical problem, according to Wittgenstein, arises when one construes misleadingly the use of one’s expressions or looks at familiar facts through a misleading form of expression.89 Correspondingly, the solution to the problem is the articulation of an alternative mode of presenting language use or facts that allows one to get rid of the problems and confusions one fell into. It is also important that this characterization of the disappearance of philosophical problems “in the right way” cannot be completed with further characterizations in such a way that the result would be a universally applicable criterion for the dissolution of a philosophical problem.

258

Philosophical Hierarchies and the Status of Clarificatory Statements

Rather, what counts as the dissolution of a problem can be determined only in particular cases, and there is no criterion or definition for what it is to get grammar right independently of the recognition of (dis)solutions in particular cases. The task is analogous to that of distinguishing sense from nonsense, for which there is, according to Wittgenstein, no universal criterion.90

6.6 Multidimensional Descriptions and the New Use of Old Dogmatic Claims I have characterized the conception of meaning as constituted by rules as a way of conceiving the concept of meaning.91 More generally, one may say that the purpose of grammatical remarks is the articulation of ways of conceiving things or clarificatory models. (As explained, in addition to rules, clarificatory models can also be articulated, for instance, with the help of real or invented examples or various kinds of comparisons.) Importantly, deployed to articulate ways of conceiving the uses of words rather than to state facts, grammatical remarks are nonexclusive. An example is provided by Wittgenstein’s characterization of the concept of language, comparing it with instruments on the one hand and games constituted by arbitrary rules on the other.92 That is, although it would be contradictory to maintain that a particular word is (as a matter of fact) an instrument designed for an externally determined purpose and that its use is governed by arbitrary rules, the situation is different when the comparisons are understood as articulating ways of conceiving the function of the word. Let us look at this more closely. In a posthumously published paper (completed by Katherine Morris), Baker examines the grammar of the concepts of conception and way of looking (Auffassung and Betrachtungsweise) in Wittgenstein. He argues that conceptions or ways of looking at things are “globally pluralistic” in the sense that acknowledging one does not render others illegitimate. For instance, conceiving meaning as a matter of a word standing for an object on one occasion does not exclude the possibility of conceiving meaning as use on another occasion.93 Nevertheless, conceptions are “locally incompatible” in the sense that it is not possible to see a concept simultaneously in two different ways—as if seeing a drawing both as a picture of a rabbit and of a duck at the same time. Thus, philosophical

Philosophical Hierarchies and the Status of Clarificatory Statements

259

conceptions in Wittgenstein’s sense are noncombinatory or nonadditive: “attempting to combine [conceptions] produces, not a more comprehensive way of looking at a concept, but muddle.”94 This seems correct as far as it goes, although the correctness of what Baker says also depends on the more precise definition of “nonadditivity” (which Baker does not provide). At any rate, Baker’s view seems to require complementing. As regards local incompatibility, there is indeed a sense in which the comparisons of language with games and instruments, for instance, do not mix. Put forward as exhaustive accounts of a word’s functioning, the explanation of its use as governed by arbitrary rules and the explanation of its use as nonarbitrary, serving some external purpose, exclude each other. On the other hand, beyond local incompatibility the possibilities of comparing language with instruments on the one hand and games on the other are complementary.95 Crucially, however, it seems that in the case of complex concepts to be described, and in the context of particular philosophical problems, the complementarity of different comparisons needs to be recognized as not merely a possibility of characterizing a concept in different ways. In such a case local incompatibility may give way, and apparently incompatible ways of characterizing a concept may be used simultaneously to capture different aspects of a concept. Here the descriptions become, so to speak, multidimensional. At this point we reach a limit of the aptness of the comparison between philosophical statements and different visual aspects of the duck-rabbit picture. For instance, as regards the concept of language, the aspects of the arbitrariness and nonarbitrariness of language may be considered equally important for comprehending the concept in its manifoldness. For although the function of language cannot be explained simply by reference to the external purposes it serves, ignoring its nonarbitrariness means treating as accidental the fact that language has any utility. Even if this may sometimes be harmless, sticking onesidedly to the conception of language as governed by arbitrary rules— thereby ignoring the interwovenness of language with nonconventional practices and forms of life—may sometimes create philosophical problems of its own. Thus although one cannot say that either aspect or both aspects must always be regarded as relevant for philosophical discussions concerning the concept of language, it may be important in some contexts to characterize the concept of language as genuinely

260

Philosophical Hierarchies and the Status of Clarificatory Statements

manifold, as exemplified by its two “incompatible” aspects. Here the descriptions may still be said to be nonadditive in the more narrow sense of the exclusion of the requirement that both aspects always be regarded as relevant. But nonadditivity in this sense does not mean that it could not sometimes be important to keep in mind several “incompatible” aspects at once. Failing to do so may create confusion, and in this sense Baker’s statement quoted above seems misleading. The need for such a simultaneous recognition of apparently incompatible aspects comes to light in the following passage, where Wittgenstein remarks on the manifoldness of the concept of mathematics:96 “A rule qua rule is detached, it stands as it were alone in its glory; although what gives it importance is the facts of daily experience. [New paragraph] What I have to do is something like describing the office of a king;—in doing which I must never fall into the error of explaining the kingly dignity by the king’s usefulness, but I must leave neither his usefulness nor his dignity out of account.”97 Here we have two different aspects: the king’s dignity and his usefulness—or a mathematical definition’s nonderivability (detachment) from the empirical world and its practical applications in describing that world. According to the above remark, one should not try to reduce one of these aspects to the other (to explain one in terms of the other). Rather, such different aspects are to be taken into account in their own right in order not to simplify matters in misleading ways. Similarly, as we saw in the case of the concept of meaning, the conception of meaning as use needs to be supplemented with the recognition of the relevance of tone to meaning, an aspect excluded by the first characterization.98 Although the inclusion of incompatible aspects in a single account is clearly impossible when different philosophical characterizations are put forward as theses or truths, it becomes a possibility when philosophical statements are comprehended as objects of comparison. In this sense the conception of philosophical statements as objects of comparison makes possible multidimensional descriptions of manifold concepts. To give one more illustration of this idea of multidimensional descriptions, imagine a calculus originally based on arbitrary stipulations that turned out to have practical applications and was developed further in a particular direction better suited to these practical purposes. Should one say that the current use of the calculus is arbitrary or determined by external purposes? One should answer that it is both, as

Philosophical Hierarchies and the Status of Clarificatory Statements

261

could easily be explained by reference to the history of its development. But as regards grammatical investigation’s concern with the present employment of the calculus, the history of the calculus is, of course, in a certain sense irrelevant. Any conjectures regarding its present application on the basis of historical knowledge have logical gaps and bring hypothetical elements into the description. Crucially, however, the use of grammatical characterizations as objects of comparison permits a multidimensional description of the calculus that can acknowledge both its arbitrariness and its nonarbitrariness. Accordingly, the possibility of multidimensional descriptions should be recognized as an important logical characteristic of philosophical statements as conceived by Wittgenstein. The conception of philosophical statements as objects of comparison and articulations of ways of conceiving things also makes possible a new use of philosophical theses, or as Wittgenstein says, of his dogmatic claims.99 This possibility can be explained by reference to the Tractatus’s view of propositions as pictures. Wittgenstein says in his lectures: “I once said that a proposition is a picture of reality. This might introduce a very useful way of looking at it, but is nothing else than saying, I want to look at it as a picture.”100 As he notes elsewhere, the characterization of propositions as pictures brings to light certain features in the grammar of “proposition.”101 The problem with the Tractatus, then, is that it stretches the analogies between pictures and propositions too far.102 Comprehended as an object of comparison, however, the conception of propositions as pictures could take on a new life. The same goes, in principle, for any philosophical theses. To the extent that they do capture aspects or features of concepts, a novel use can be made of philosophical theses as objects of comparison. Hence Wittgenstein’s turn away from metaphysics is not simply a rejection of what has been said in metaphysical philosophy, and it would be misleading in this sense to characterize him as an antimetaphysical philosopher. This attitude toward metaphysics is also reflected in Wittgenstein’s statement to his pupil Drury, who reports him as saying: “Don’t think I despise metaphysics or ridicule it. On the contrary, I regard the great metaphysical writings of the past as among the noblest productions of the human mind.”103

262

Philosophical Hierarchies and the Status of Clarificatory Statements

The difference between theses and philosophical statements used as objects of comparison can be characterized as follows: In contrast to a thesis concerning a necessity pertaining to reality that one must allegedly accept (on pain of irrationality or ignorance),104 a philosophical statement used as an object of comparison presents a possible way of conceiving things. Such a way of conceiving things, or a philosophical model, may consist in regarding or presenting something as a necessary characteristic of the objects of investigation. Being used as an object of comparison, however, the necessity remains internal to the model and is not projected onto reality in the guise of a statement about a necessary fact. Consequently, the model does not exclude other ways of conceiving things, as explained above. For instance, propositions may be characterized as pictures in the context of one discussion, and the concept of proposition may be treated as a family-resemblance concept in another context. Although it might seem that philosophy loses something of its strength when we move away from statements about how things must be to the articulation of possible ways of conceiving things, this impression is misleading. Consider the following remark: It is often quite enough for us to show that one does not have to call something that; but that it can be called that. For already this changes our perception of the things [insertion: the face of things]. In this sense my dogmatic statements were unjust [insertion: incorrect]. But they can be set straight, if instead of saying “one has to regard it thus” one says: “one can [insertion: also] regard it thus.” And it would be wrong to think that the proposition had thereby lost its real strength [insertion: // real point //].105

First, the point or strength of philosophical statements is not lost insofar as the introduction of a new way of conceiving a matter can release one from a philosophical problem. To the extent that the purpose of philosophical statements is to dissolve philosophical problems, and that statements articulating alternative ways of conceiving things can achieve this purpose, such statements can serve the purpose for which philosophical statements are intended. (The disappearance of the problem about abstract objects made possible by the transition from referentialism to the conception of meaning as use provides an example.) Second, even if it might be true that there is always some similarity

Philosophical Hierarchies and the Status of Clarificatory Statements

263

between two things and, therefore, that anything can be compared with anything or regarded as anything, it does not follow that philosophy, as I have described Wittgenstein’s conception, becomes an idle game of coining possible comparisons and ways of conceiving things. Here the mere existence of philosophical problems shows that not every way of conceiving things is satisfactory in allowing one to understand what one is concerned to understand. Not every comparison is illuminating, but nobody—especially not Wittgenstein—ever claimed they were. It would then be hasty to object to the view I have attributed to Wittgenstein on the grounds that it is always possible to look at things in various ways or to compare some two things in some respect. This objection disregards the goal of philosophical clarification: the dissolving of philosophical problems. Third, it is important to not confuse the disappearance of a philosophical problem with the disappearance of our interest in the relevant issue, as if the aim of philosophy were to stop one from thinking rather than to enable one to think more clearly. For instance, the purpose of the conception of meaning as use is to make the concept of meaning perspicuous. And although philosophical clarification does not provide one with a thesis about meaning, it does not leave one empty-handed either. What one is left with when the problem about meaning has been dissolved is a way of thinking about meaning, a model (potentially) applicable in a variety of (further) contexts. Thus the dissolution of a philosophical problem may well leave one with something whose significance reaches beyond the particular problems one was initially concerned with. The articulation of a way of conceiving things can have far-reaching consequences—for instance, by giving a new direction to further investigation. Similarly, such conceptions may have many practical applications and consequences.106 Finally, one might worry that Wittgenstein’s conception of philosophy (as presented here) implies relativism. Although a thorough discussion of relativism is beyond the scope of this book, there is no reason to think that relativism as the plurality of incompatible truths follows from Wittgenstein’s conception. For rather than a plurality of truths, his conception allows for a pluralism of ways of conceiving things or of modes of presentation. This kind of pluralism is not threatened by “the problem of irrationalism” as other forms of relativism may be, because

264

Philosophical Hierarchies and the Status of Clarificatory Statements

different modes of presentation are not exclusive in the way true/false statements are. Accepting two incompatible modes of presenting or conceiving things is not a matter of accepting two contradictory truths, from which, as a matter of logic, anything follows. Hence a grasp of the difference between the status of grammatical statements and that of truths is crucial to the dissolution of the worry of relativism.

SEVEN

Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics

for a certain interpretation of what Wittgenstein means by not having theses in philosophy. In the course of my argument I have attributed to him various views concerning the methodology of philosophy but have not yet explicitly explained their status. Do they make up a thesis about the essence of philosophy, and is Wittgenstein’s philosophy therefore incoherent after all? The problem might be explained as follows. Although the Tractatus too aimed only to spell out a way to approach philosophical questions, its claims about the correct method of philosophy and its universal applicability constitute, in effect, a thesis about the nature of philosophy. In the Tractatus a metaphysics of philosophy and language found a refuge in methodological statements, as explained in 3.1 and 3.2. The purpose of 7.1 is to explain how Wittgenstein avoids this problem in his later work and why his conception of philosophy, as described in the preceding chapters, does not constitute a thesis about the essence of philosophy, i.e., a claim about what philosophy must be or what must be understood by “philosophy.” Subchapter 7.2 then situates Wittgenstein’s considerations regarding philosophy in a broader context by discussing the historicity of philosophy, that is, his conception of how philosophical questions and views arise in particular historical settings and in connection with particular historical languages. Subchapter 7.3 takes up the question of the relation of philosophical language use to everyday language. I explain here why Wittgenstein’s notion of “bringing words back to everyday use”

I HAVE ARGUED

266 Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics

does not entail a view of everyday language as a standard of sense to which philosophical uses must conform, contrary to how Wittgenstein is often interpreted. This then leads me to some issues concerning the development of Wittgenstein’s philosophy, notably the sense in which the Tractatus was already concerned with clarifying everyday uses of language. This completes my discussion of the relation between Wittgenstein’s early and later philosophy. In connection with the relation of philosophy to everyday life, I also address the relation of the present account of Wittgenstein to certain aspects of Cavell’s reading of his later work. I conclude by discussing Wittgenstein’s notion of an ethical dimension to philosophy that pervades it in its entirety, and the sense in which the problem of dogmatism is also an ethical problem.

7.1 Metaphysics Disguised as Methodology The point that the disappearance of problems is not evidence but the criterion for the correctness of a philosophical conception1 also applies to Wittgenstein’s conception of philosophy and its method. That it is possible to dissolve philosophical problems by employing the method he develops in his later work does not constitute evidence for the correctness of this method. (Not, in any case, in the sense in which one can give evidence for the correctness of a factual claim, establishing its truth as more or less certain.) Rather, that problems can be dissolved by employing this method is what it means for it to be correct, to qualify as correct. As Wittgenstein says in the Investigations, he aims to “demonstrate a method, by examples.”2 Taking the examples by which his method is demonstrated as evidence for the correctness of his conception of philosophical practice, there seem to be two ways to fill in the details. The examples could be seen as inductive evidence for the correctness of an empirical claim about how to approach philosophical questions. But since Wittgenstein’s purpose is not to make an empirical statement about philosophy, this interpretation can be put aside. Alternatively, one might take the examples as support for a thesis about what philosophy must be. But the problem—in addition to the interpretative problem of inconsistency in Wittgenstein’s philosophy—is that the examples in the Investigations cannot justify a claim of this kind. A gap remains between what the examples can establish and Wittgenstein’s presumed thesis about the essence of philosophy, including his alleged claim

Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics 267

about the universal applicability of his method.3 To overlook or ignore this problem and insist that these examples (somehow) do bring the essence of philosophy into view is to relapse into metaphysical projection and dogmatism. This is where the Tractatus’s mistake lies.4 These problems can be avoided, however, if the examples are interpreted as Wittgenstein’s way of explaining what he means by “philosophy,” by “dissolving philosophical problems,” and so on. Read this way, the examples are first of all means employed to articulate a particular conception of philosophy. They are a way of explaining what it is to philosophize in Wittgenstein’s way, not vehicles for anything like the definition of the concept and essence of philosophy. Their role is to make understandable and bring to light the essence of Wittgenstein’s conception—which may then be compared and contrasted with other conceptions, say, with respect to how successful each actually is in solving philosophical problems. Accordingly, the Investigations provides many examples not to accumulate the maximum amount of evidence in support of Wittgenstein’s approach but to render as clear as possible the nature of his approach to philosophy, i.e., to elucidate the multifaceted aspects of this complicated activity.5 Notably, not all Wittgenstein’s examples serve to introduce the method in the same way.6 While some discussions—concerning the concept of pain, for instance—might be seen as contributing to the introduction of his conception of philosophy by exemplifying the application of his method, other discussions—such as those concerning the concepts of meaning and language—contribute also in a different way. They clarify concepts that are more directly relevant to Wittgenstein’s ideas of how to approach philosophical questions as conceptual questions. Recall, for instance, the contrast drawn in 4.1 between two conceptions of clarification, one connected with a mentalist conception of meaning and the other with the conception of meaning as constituted by rules. This brings to light an intimate connection between the method of clarification by tabulating rules and the conceptions of meaning as use and of language as a rule-governed practice. Nevertheless, as explained in 6.1, the point of Wittgenstein’s discussion of meaning and the rulegovernedness of language is not to provide an abstract, once-and-for-all justification of the method of tabulating rules. Instead, the purpose is to elucidate the relevant concepts and thus also the approach to philosophy they articulate.

268 Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics

As for the justification of Wittgenstein’s conception of philosophy and its method, ultimately this depends entirely on the philosophical work they can do. A second function of Wittgenstein’s examples, in addition to characterizing his approach to philosophy, is to make it apparent to his reader that by practicing philosophy in Wittgenstein’s way, one can indeed get clear about one’s philosophical problems. This is also meant to dissolve any worries relating to the apparent arbitrariness of the choice of philosophical approach that may arise when Wittgenstein’s approach is described (as I have done on the basis of considerations relating to modal notions), as merely his approach or a particular approach, not the approach one must adopt. For there is, of course, nothing arbitrary in choosing a philosophical approach on the grounds that it is able to solve one’s philosophical problems. Rather, insofar as one’s aim is to find solutions to such problems and unclarities, this is the only nonarbitrary ground.7 The difference between this account and one that takes Wittgenstein’s examples to support a claim about what philosophy must be is that in the present account, the generality of Wittgenstein’s claim about philosophy is a function of its justified applications. No claim is made about the universal applicability of his method, but the generality of its application is ultimately left open. Thus the status of his conception of philosophy and its method is exactly the same as that of any philosophical model in the context of his approach.8 Consequently, the problem of Wittgenstein having a thesis about the essence of philosophy, and of a gap between what his examples can establish and his allegedly universal thesis about philosophy, is dissolved. The unity of Wittgenstein’s method, on the other hand, does not seem to depend on anything common to all its instances, except perhaps that they are all instances of the clarification of language or of the perspicuous presentation of language use. But these characterizations do not suffice to determine the identity of his later method. There is no one way to comprehend philosophy as the clarification of language, or as perspicuous presentation, as the contrast between Wittgenstein’s early and later philosophy illustrates. Rather, the identity of his later method is exhibited by the various techniques of clarification that make it up. As Wittgenstein says, there is not just one method, but many: “There is not a philosophical method, though there are indeed methods, like

Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics 269

different therapies.”9 The observation that his method is built up of a variety of methods or techniques also explains why Wittgenstein is not contradicting himself in §133 when he talks about demonstrating a method and, at the same time, denies that there is just one method. Given Wittgenstein’s emphasis in the Investigations §122 on the fundamentality of the notion of perspicuous presentation, one might explain the contrast in §133 between one method and multiple methods by saying that although Wittgenstein’s method, generally speaking, is that of the perspicuous presentation of language, there are many techniques for creating perspicuity. Thus perspicuous presentation, as conceived in Wittgenstein’s later philosophy, is not just one method but a set of techniques or methods. Similarly, grammatical investigation might be generally characterized as an activity of clarifying and dissolving philosophical problems, without being identified with any particular method of Wittgenstein’s, such as the method of tabulating grammatical rules.10 At this point an important difference surfaces between Wittgenstein’s later philosophy and the Tractatus. As explained, the Tractatus identifies logical clarification or perspicuous presentation with analysis in terms of the formulae of the concept-script, presenting this method as a privileged one whose adoption is supposed to enable one to dissolve any philosophical problem whatsoever. The Tractatus’s solution to the question of the methodology of philosophy therefore has the form of a great once-and-for-all answer, part of which is, as explained in 3.1 and 3.2, a once-and-for-all determination of the essence of language. This type of solution, however, is exactly what Wittgenstein wants to avoid in his later work.11 Thus he does not aim to show that there is some particular method, and a corresponding conception or thesis of language, that one must adopt—for example, that one must regard philosophy as essentially concerned with the rules of language. Rather, different conceptions of language correspond to different methods. For instance, his quasi-ethnological or quasi-anthropological method (see below) belongs together with a conception of language as a form of life, in which the focus is not on the rule-governedness of language but on its intertwinedness with other activities and aspects of life, including primitive, nonconventional forms of behavior. The emphasis here is on the factual, not the conventional.12 This plurality of methods means that for the later Wittgenstein the concept of philosophy is not defined through a particular methodology

270 Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics

in the sense of a doctrine of the correct method. Consequently, although one may say that in the Tractatus a metaphysics of philosophy and language finds a refuge in the methodology of logical analysis, Wittgenstein’s later philosophy does not suffer from the same problem. A brief list13 of Wittgenstein’s methods in addition to the tabulation of grammatical rules might include: 1. The use of comparisons and analogies. 2. The use of examples as centers of variation. 3. The finding and constructing of intermediate cases.14 4. The use of simple invented language-games as objects of comparison. 5. Quasi-ethnology or quasi-anthropology.15 6. The use of invented history of ideas.16 7. The use of invented natural history.17 8. The use of questions.18 9. The use of jokes and humor.19 One might be tempted to respond that all these apparently different methods are ultimately reducible to a concern with the rules of language and that philosophy, therefore, is always—essentially—a concern with the rules of language, if not explicitly then implicitly. This response, however, is problematic in the sense that it assumes that philosophy should have a hierarchical structure founded on the thesis that grammar constitutes the essence of language. Accordingly, one might say that in the end, the idea that clarification is necessarily a concern with rules merely constitutes a variant of the Tractatus’s idea that there is an ultimate, privileged form of expression that captures the essence of language. For reasons already explained, this view cannot be attributed to the later Wittgenstein.20 Note also that although all the above methods may be characterized as methods for the clarification of language use, it does not follow that Wittgenstein is putting forward a thesis of philosophy as necessarily concerned with language. His way of dissolving this implication comes to light in a remark he makes about the fact that clarification “is carried out completely in language” as a consequence of the grammatical conception of meaning (see 4.1): “What is it supposed to mean that it is all carried out in language? For us everything is carried out in language.”21 Here Wittgenstein is formulating the same point that was

Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics 271

made above, namely that his conception of philosophy is a particular conception of philosophy, not the conception. Everything for Wittgenstein is carried out in language, but he is not putting forward a thesis that this is the essence of all philosophy. That everything is carried out in language is simply the essence of his conception.22 Finally, Wittgenstein’s explanation of his method of philosophy by means of examples is to be considered open-ended, in that it is not given in terms of a single characteristic mark that provides unity to his concept of a philosophical method and determines its limits. Consequently, there seems to be no reason to think that the set of methods actually used by Wittgenstein could not be expanded or that one could not come up with new methods that could be identified as part of a Wittgensteinian approach.

7.2 The Historicity of Philosophy Philosophy for Wittgenstein is a struggle with language: “We are engaged in a struggle with language.”23 Perhaps to illustrate the kind of difficulties that are at stake, he formulates this point in the Investigations as follows: “Philosophy is battle against the bewitchment of our intelligence by means of language.”24 Curiously, it is the ambiguity of this sentence that seems to make it so acutely descriptive of the situation. Language is both the source of our confusions and the means by which we seek to address them—and all these possibilities might be expressed by (or contained in) the very same string of words. Because language is a historically changing phenomenon,25 philosophy is a historically situated undertaking. Philosophical problems arise, and re-arise, in particular historical settings: “The obsessions of philosophers vary in different ages because terminologies vary. When a terminology goes some worries may pass only to arise in a similar terminology.”26 Accordingly, philosophical problems have to be dealt with in historically determined contexts. Moreoever, given the embeddedness of language in forms of life, to be engaged in a struggle with language may mean struggling with a whole culture and era—including oneself as a product of a culture and its traditions.27 In this sense also the problems with which Wittgenstein is dealing are rooted in more general tendencies of thinking embedded in the Western culture and philosophical tradition. Accordingly, he sees himself on the one hand as swimming

272 Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics

against the main current of Western civilization, characterized by the idea of progress and belief in science and technology as solutions to all problems.28 On the other hand, he characterizes philosophy as “work on oneself ”: “Work on philosophy is . . . a kind of work on oneself. On one’s own conception. On the way one sees things. (And what one demands of them.)”29 We may assume that the conceptions and requirements one imposes on the objects of investigation in philosophy are at least partly inherited from the philosophical tradition and internalized through a process of indoctrination. Therefore there need not be any clear-cut distinction between struggling against an age and with oneself. More specifically, as explained in the preceding chapters, Wittgenstein is wrestling with the attitude of “craving for generality” or the “contemptuous attitude towards the particular case.”30 He finds an early and important expression of this attitude in Socrates or Plato, but evidently it also pervades the Tractatus. This attitude, as Wittgenstein says, “is the resultant of a number of tendencies connected with particular philosophical confusions.”31 Here two central tendencies, as discussed in the preceding chapters, are (1) “the tendency to look for something in common to all the entities which we commonly subsume under a general term”, which finds its expression in attempts to articulate the definition of a concept, and (2) a “preoccupation with the methods of science,” understood as a search for a unified and simple account of phenomena, i.e., an explanation in terms of certain primitive laws or general explanatory factors.32 We have seen both how Wittgenstein falls prey to these tendencies of thinking in his early philosophy and how he attempts to respond to them in his later work, trying to modify his approach accordingly.33 This conception of philosophical problems as historical phenomena has an important bearing on how one should understand the significance of philosophical work. Insofar as philosophy is concerned with problems that arise in certain cultural settings, one is not automatically justified in claiming for philosophical views a universal and transhistoric value that transcends the cultural settings in which the work originated. Wittgenstein remarks: “Someone who must fight (against) (swarms of ) mosquitoes finds it an important matter to have chased some away. But that is quite unimportant to those who are not concerned with mosquitoes. When I solve philosophical problems I have a feeling as though I had done something of utmost importance for all of

Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics 273

humanity [and] don’t think that these matters appear so immensely important to me (or shall I say are so important to me) because they plague me.”34 Nevertheless, these observations do not exclude the possibility of philosophical work having universal and transhistoric value. Whether it has such value depends on whether the problems—or tendencies of thinking—one is dealing with really are universal or universally human. Naturally, the above does not mean that the misunderstandings and confusions Wittgenstein discusses would not be very old, originating in ancient Greece, for instance, or inherited from (through) the Greeks.35 What Wittgenstein says about the concept of philosophy is no exception. His conception of philosophy arises as a response to problems in the philosophical tradition in which he operates and gets its significance from its contrast with the tradition. Thus, were the tradition different, things that needed saying might be different too. In this sense, for instance, Wittgenstein’s discussion and rethinking of the role of rules in philosophical clarification is a response to philosophers’ concern with rules, i.e., their attempts to provide definitions and philosophical theories. His conception of clarificatory rules as objects of comparison, that is to say, is an adjustment of these philosophical practices, whereby his aim is to solve the problems of dogmatism and injustice to which these practices give rise. But had we not inherited from our philosophical tradition the problematic conception of the unity of concepts that provides the ground for philosophical theories and, subsequently, the motivation for Wittgenstein’s methodological adjustment, there might not be any need for the adjustment. In this sense it would then be a misunderstanding of Wittgenstein’s intentions to think that his conception of philosophy is meant to be a doctrine of what philosophy must be.36 He addresses this issue in his lectures: “Why do I wish to call our present activity philosophy, when we also call Plato’s activity philosophy? Perhaps because of a certain analogy between them, or perhaps because of the continuous development of the subject. Or the new activity may take the place of the old because it removes mental discomforts the old was supposed to.”37 Wittgenstein is not claiming here that to the extent Plato deserves the name of a philosopher, he must have been doing the same thing as Wittgenstein (though perhaps clumsily and without proper self-understanding) because this is what philosophy really is or must be. Without assuming that to call both Plato’s and Wittgenstein’s activities “philosophy” requires

274 Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics

that they share a certain characteristic or characteristics, we may call Wittgenstein’s activity “philosophy” because of a certain historical continuity and similarities with Plato’s. And most importantly, Wittgenstein’s activity may be called “philosophizing” insofar as it is capable of removing mental discomforts, that is, problems that philosophy traditionally was supposed to remove. Here the justificatory basis of Wittgenstein’s approach surfaces again: his activity may be called philosophy if it can deliver what traditional philosophy was meant to deliver, or at least a significant part of this, while explaining why certain aspirations of traditional philosophy are misplaced, most significantly the aspiration to establish philosophical theses. The problem of transhistorical claims about what philosophy must be also appears in the following remark concerning the idea of philosophy as a struggle with language: “Do we have to do with mistakes & difficulties that are as old as our language? Are they, so to speak sicknesses that are bound together with the use of a language, or are they special, characteristic of our civilisation? . . . Or also: is the preoccupation with the means of language that penetrates our whole philosophy an age-old feature of all philosophy, an age-old struggle? Or is it new like our science. Or like this as well: does philosophizing always waver between metaphysics & a critique of language?”38 I take Wittgenstein’s point to be that although it may be said of our philosophical tradition that its questions are bound together with old misunderstandings concerning language and that it is a struggle with language, this still might not justify the claim that this is what philosophy must always be. The tendencies to misunderstand language that Wittgenstein sees as characteristic of philosophy might typify only the Western tradition going back to the ancient Greeks (and so on), and not be a necessary characteristic of everything recognizable as philosophy. I do not think, that is to say, that Wittgenstein’s question in the remark is rhetorical, obviously to be answered affirmatively: “Yes, philosophy is always a struggle with language; it always wavers between metaphysics and a critique of language.” Rather, it seems important that he leaves the question unanswered. We cannot say that philosophy is always a struggle with language simply because it has to be that for us now. Here we come again to the problem of projection. On the basis of our condition (if one accepts Wittgenstein’s view), it may very well seem that philosophy is by its very nature a struggle with language. But even

Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics 275

assuming that this is a correct description of our situation (or of the Western tradition), this claim might only be a projection of our condition, or of a conception that forces itself on us, onto reality as a universal truth about philosophy. On the other hand, neither does it follow from this possibility of projection that one should answer Wittgenstein’s question negatively. Our entitlement to that negative claim is no less problematic, and the crucial question concerns the justification of this pair of claims. The problem, in other words, is the tendency apparently inherent in our philosophical tradition to exceed the grounds for our statements in our craving for generality, and to exaggerate, to state once and for all how things are or what all cases must be. For we seem to have an immensely strong urge to draw definitive conclusions instead of being satisfied by understanding possibilities and remaining open to them. Here a whole field of work on oneself and philosophical therapy as a way of dealing with problematic tendencies of thinking opens up.

7.3 Philosophy and the Everyday It remains to consider Wittgenstein’s conception of the relation of philosophical language use to everyday language. The preceding chapters have shown that his philosophy is not to be seen as constituting an affirmation of the truth of everyday conceptions or commonsense beliefs. Such an interpretation fails to take into account the distinction between conceptual and factual questions, which is at the heart of Wittgenstein’s thought throughout his career.39 It is less clear, however, whether or not his philosophy involves a requirement that philosophers not depart from everyday uses of language (assuming that everyday language is the basis of their questions) and what exactly this would mean. I begin with this question and move on to further issues relating to philosophy and the everyday and the role of ethical considerations in philosophy. Baker and Hacker take everyday or ordinary language to constitute a ground of intelligibility for Wittgenstein in a very particular sense. As explained earlier, from their point of view deviation from ordinary language means either speaking nonsense or falling into irrelevance, that is, losing contact with what we speak about when we employ terms such as “language” and “meaning,” and so on, that have their roots in everyday life and language. To avoid this fall into nonsense or irrelevancy,

276 Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics

philosophers must use their expressions in accordance with everyday language, which in this sense constitutes a standard for correct or meaningful language use and also for the correctness of philosophical views relating to everyday concepts.40 But although there is such a thing as losing contact with the everyday, and something like this may often or sometimes happen in philosophy, it is not evident that Wittgenstein’s view of the relation between philosophical and everyday language use is correctly captured by characterizing the latter as a standard of correct use. We must examine the issue in more detail to settle this matter. I will argue that although there is a sense in which the dissolution of a philosophical problem, insofar as it concerns everyday concepts, involves finding one’s way back to the everyday, there is no reason to assume that this would be, for Wittgenstein, a matter of finding one’s way back to any clearly definable standard. Wittgenstein famously writes in the Investigations: “When philosophers use a word—‘knowledge,’ ‘being,’ ‘object,’ ‘I,’ ‘proposition,’ ‘name’—and try to grasp the essence of the thing, one must always ask oneself: is the word ever actually used in this way in the language-game which is its home ground?— [New paragraph] What we do is to lead the words back from their metaphysical to their everyday use.”41 Here a contrast is drawn between traditional and Wittgensteinian philosophers, whereby it is characteristic of the latter that they lead words back to everyday use from their metaphysical use. Clearly, the Investigations §116 is urging a return to everyday language in some sense. But in what sense? Let me first note some problems with Baker and Hacker’s account. To begin with, Baker and Hacker’s view presupposes that an (ultimately) uncontroversial way of identifying what counts as conformity with everyday language is available. For if the alleged standard of correct use and sense cannot be uncontroversially identified, then it cannot be appealed to in order to settle philosophical disputes. (When competent language users have a dispute about the definition of an everyday concept, who speaks with the authority of everyday language?) More precisely, there seem to be two principal ways of envisaging the identification of what counts as everyday language, which need to be examined in turn. One might seek to provide a general definition of the concept of everyday language (and hence conformity with it), or alternatively, rejecting the idea of such a general definition, one might maintain that

Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics 277

what counts as everyday use is only to be identified in particular cases. As I will argue, however, on neither account can ordinary language be understood as providing a standard of sense that one could appeal to in philosophy. With respect to the first option, it is notable that no attempts to give a general definition of what counts as everyday language use can be found in Wittgenstein’s writings. Rather, as Baker has argued in his later work in connection with the Investigations §116, it is metaphysical use that Wittgenstein defines in more precise terms, and everyday use only by way of contrast with this.42 That is, instead of providing a positive, general characterization of the kinds of uses that constitute everyday or ordinary language, Wittgenstein undertakes the much more manageable task of specifying certain problematic philosophical uses of language with the purpose of excluding them. But then the notion of conformity with everyday language cannot be said to constitute a general criterion of sense that could be appealed to in the course of philosophical discussion. To define the notion of metaphysical use is not yet to say anything specific about everyday use. As regards the option that everyday uses are to be identified in particular cases rather than in terms of a general definition, it is important that for Wittgenstein there are no readily identifiable everyday uses of language in this sense either. What counts as an everyday use of an expression, in contrast to a philosophical or metaphysical use, is obscured in everyday language by a plethora of misleading pictures, analogies, and so on engraved in this language itself. As Wittgenstein writes: “Why are the grammatical problems so tough and seemingly ineradicable?— Because they are connected with the oldest thought habits, i.e., with the oldest pictures that are engraved into our language itself. . . . [New paragraph] People are deeply embedded in philosophical, i.e., grammatical confusions. And to free them from these presupposes pulling them out of the immensely manifold connections they are caught up in. . . . —But this language came about // developed // as it did because people had—and have—the inclination to think in this way.”43 As this passage makes clear, Wittgenstein is not committed to an assumption about the separation of everyday language from philosophical ideas (perhaps one could say more generally, from various ideologies). Far from being uninfluenced by philosophically problematic tendencies of thinking, the development of everyday language, according to him, is

278 Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics

affected by just the kind of tendencies of thinking that also lie at the root of philosophical problems. But if everyday language itself is molded by philosophically problematic thought habits in this sense, then one evidently cannot appeal to it in any straightforward way to settle philosophical disputes. Rather, when describing ordinary uses one is in constant danger of producing descriptions that are themselves informed by philosophical prejudices and pictures. Accordingly, Wittgenstein’s conception of everyday language seems radically different from that of the ordinary language philosophers, who seem to have more or less assumed that ordinary language constitutes a reliable ground for deciding philosophical questions.44 That Wittgenstein rejects the idea of everyday language as a readily identifiable standard for what counts as philosophically unconfused complicates the issue for Baker and Hacker and again raises the suspicion that the standard they assume for correct use is ultimately a theoretical postulate connected with their conception of language use as based on a structure of implicit grammatical rules that fix what makes sense. As already argued, this conception of language cannot be attributed to the later Wittgenstein. Instead, it is characteristic of his later thought that he rejects this essentially Tractarian idea of there being a certain definite logical order of language that contains the solution to philosophical disputes, including the differentiation of the metaphysical from the everyday.45 These difficulties with Baker and Hacker’s reading constitute a sufficient reason for considering the possibility of an alternative interpretation. But how then should one understand Wittgenstein’s conception of philosophy as leading words back from metaphysical to everyday use? I suggest approaching the issue from a different angle. The philosophical deceptiveness of everyday language raises a methodological question: what is one to do to enable oneself to navigate past the misleading pictures engraved in language and to avoid the misleading tendencies of thinking they express? Central to answering this question is the realization that the describer of language is not a neutral collector of data or passive receiver. Rather, those problematic thought habits that have been engraved in language in the form of misleading pictures are the describer’s own thought habits too—assuming that the language in question is her own. And provided that she uses language to describe language, these habits are

Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics 279

likely, if not bound, to inform her attempts to describe language. This is what makes philosophy so difficult. To paraphrase (fairly freely) the previous quotation, to liberate herself from her confusions she must pull herself out of the manifold conceptual connections she is caught up in. One such tie she must cut herself loose from is the assumption about conceptual unity that informs metaphysical thinking, according to which a unified definition of cases falling under a concept should or can always be given. Accordingly, when Wittgenstein talks about leading words back from metaphysical to everyday use, he is talking about the exclusion of this kind of usages from philosophy. He is talking about uses or pictures that prevent one from finding one’s way back to the everyday. (One must find one’s own way back before one can think about leading others, even though we may, of course, search together.) Crucially, when Wittgenstein is interpreted in this way, the need to evoke everyday uses as a clearly specifiable standard of sense disappears, as I will explain. Consequently, the problem with Baker and Hacker’s interpretation, namely the theoretical assumption that everyday uses constitute a determinable standard of sense, is dissolved. More specifically, assuming that the point of paying attention to everyday language at all was sufficiently elucidated in 2.31 for it to lose its dogmatic appearance,46 the preceding chapters (especially Chapter 3) can be read as an explanation of what Wittgenstein means by leading words back from metaphysical use to everyday language. This can now be spelled out. At stake in bringing words back from metaphysical to everyday use is how to avoid mystifying the objects of philosophical investigation by imposing a metaphysical requirement of a unified essence on them, as Wittgenstein did in the Tractatus.47 As explained, in the Tractatus the concept of a proposition was turned into a metaphysical concept intended to capture the underlying sublime essence common to every possible proposition. Furthermore, this concept was also made the center of an ultimate philosophical conceptual order: the characterization of the essence of proposition was intended at the same time to capture the essence of language and thought, and even more generally the essence of representation. In this capacity it was also taken to capture the essence of reality as something experienced or as an object of

280 Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics

representation. Thus, the concept of proposition was given a truly allencompassing philosophical employment. Wittgenstein writes about this, contrasting such metaphysical uses of words with what he calls “a humble use”: “We are under the illusion that what is peculiar, profound, essential, in our investigation, resides in its trying to grasp the incomparable essence of language. That is, the order existing between the concepts of proposition, word, proof, truth, experience, and so on. This order is a super-order between—so to speak—super-concepts. Whereas, of course, if the words ‘language,’ ‘experience,’ ‘world,’ have a use, it must be as humble a one as that of the words ‘table,’ ‘lamp,’ ‘door.’ ”48 The difference between metaphysical and humble nonmetaphysical uses of words can be elucidated as follows. In the Nachlass Wittgenstein often characterizes the problem with the Tractatus by saying that the solution it offered to philosophical problems was not plain or homespun enough.49 Although the Tractatus emphasized the difference of philosophy from the sciences, its author had not wholly released himself from thinking about philosophy in terms of discoveries. As he comments: “In my earlier book the solution to the problems is not yet presented nearly plainly [hausbacken] enough[;] it still makes it seem as though discoveries were needed to solve our problems and not enough has been done to bring everything in the form of grammatical truisms in ordinary mode of expression. Everything gives too much the appearance of discoveries.”50 What he is objecting to is the impression created by the Tractatus that the clarificatory work undertaken in it had brought to light something extraordinary. If the Tractatus is right, every proposition can upon analysis be revealed as a truth function of elementary propositions, which are concatenations of simple names and stand for the simple objects of reality. But although Wittgenstein’s laying out the logical structure of language in anticipation of such analyses was merely supposed to clarify what his readers already knew, all this looks very much like he had made a great discovery entirely comparable to the great discoveries of the sciences. Instead, the later Wittgenstein maintains, philosophical concepts are to be employed in a more homespun or humble manner. The notion of such an employment can be elucidated through the concept of an object of comparison. As explained, in this capacity a definition of a concept or a conception is not presented as something to which all cases falling under the concept must correspond. Rather, a definition as an

Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics 281

object of comparison is simply one piece of language used to represent another, and the generality of the definition depends on its justified applications in particular cases.51 Moreover, because clarification concerns particular philosophical problems and the clarification of different problems relating to a concept may require different clarificatory models, there are no ultimate, fundamental conceptual definitions in philosophy as conceived by Wittgenstein, i.e., definitions presupposed every time a particular concept is employed. Hence there is no ultimate superorder of superconcepts. To lead words back to everyday language, according to the interpretation I am proposing, is therefore to refrain from assuming with the metaphysician that there must be some unified definition of the cases falling under a concept and to adopt instead a more humble employment of clarificatory concepts. As Wittgenstein says in §116, confronted with a metaphysical use one must ask whether the concept is ever actually employed in this way in the language-game that is its home ground. Certainly, it would seem extremely dogmatic of Wittgenstein—and the reader—to take the answer to be a categorical no. The question, therefore, cannot be a rhetorical one with “no” as the obvious answer. Instead, the question must concern the more general methodological issue of how to do justice to language use when describing it, i.e., how to describe the uses of expressions with due complexity rather than false simplicity and consequently make possible the resolution of philosophical problems. The answer is to abandon the metaphysical straitjacket of assertions about what all things falling under a concept must be and to use definitions (and so on) as objects of comparison. Hence, by characterizing what “we”—the Wittgensteinian philosophers—do as “leading words back,” Wittgenstein is referring to a particular approach or to a style of philosophy. But he is not committed to any claims or assumptions about a clearly circumscribable concept of everyday language under which philosophical uses must be subsumed or to the idea that the uses of everyday expressions are clearly definable and that philosophers have to match their concepts with such uses.52 In this connection it is also important that humble or homespun applications of philosophical models in the sense explained above do not exclude something like a technical use of philosophical concepts. That is, although the employment of a simplified philosophical definition as an object of comparison may be a technical one,

282 Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics

there is no reason to think Wittgenstein would have any objections to such an employment. There is no ban on technical concepts in philosophy as long as they are not put forward dogmatically as something to which the objects of investigation must conform.53 Notably, due to the Wittgensteinian philosopher’s opposition to widespread but problematic tendencies of thinking engraved in language, what she says about language use need not sound familiar at all, even if intended to elucidate everyday concepts. Rather, it is familiar-sounding reflections on the everyday that should arouse our suspicion. For this reason, it should not be surprising that many of Wittgenstein’s characterizations of the use of language sound unfamiliar—for instance, his denial that first-person utterances of pain are knowledge claims, or his remark that “An ‘inner process’ stands in need of outward criteria.”54 Whether or not Wittgenstein is right about these issues, his remarks hardly reflect common tendencies of thinking about sensations and inner processes. But the idea that Wittgenstein’s descriptions should be readily recognizable as descriptions of everyday uses depends on the false—or in any case un-Wittgensteinian—premise that everyday uses of language are readily identifiable. For related reasons it would be misleading to characterize the development of Wittgenstein’s philosophy by saying that whereas in the Tractatus the object of investigation was the sublime and hidden essence of language, for the later Wittgenstein it is everyday language.55 Not that this is untrue. However, insofar as the Tractatus’s goal was already merely to remind the reader of what she already knows, in principle it took the everyday concept of proposition as the object of philosophical clarification. (The reader could not be reminded of something she did not already have knowledge of.) But the description of the logic or grammar of the everyday concept of propositions is not an easy task. Consequently, instead of saying what everyone knows, the Tractatus actually developed a myth of symbols hidden under the surface of language and ended up giving a description of something sublime.56 The point is that by describing Wittgenstein’s development in terms of a switch in the object of investigation, the crucial difficulty of philosophy that his early work and the development of his thought illustrate—the difficulty of depicting the uses of everyday language

Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics 283

without being dogmatic and falsifying them—is eclipsed. In this sense such a characterization of Wittgenstein’s development is misleading (rather than untrue). It disguises the very difficulty that motivated the development of Wittgenstein’s method. In other words, from the point of view I suggest, the Tractatus’s notion that to describe the logic of language one must get hold of something sublime is to be seen first and foremost as a confused view of what is required to get clear about—to perspicuously present—the logic or grammar of everyday language. Hence although there are problems relating to the details of the Tractatus’s account of logic,57 such problems, even if they may be important, are not the key to understanding Wittgenstein’s later method and the development of his philosophy. As explained in 6.6, it is possible to give the Tractatus’s model of language a novel, nonmetaphysical employment as an object of comparison, that is, as a partial description of the grammar of relevant expressions. This indicates that the key problem with the Tractatus is the background assumption that there must be a unified essence of language, which if not discernible on the surface is to be found underneath it, i.e., that there must be a definitive, once-andfor-all description of the essence of language. Problems relating to this assumption rather than the details of the Tractatus’s model of language motivate the methodological idea of leading words back to ordinary use. This methodological idea, that is to say, is a response to the metaphysical conception of the nature of philosophical statements that prevents one from achieving a clear comprehension of whatever one wants to understand. Wittgenstein’s later methods might, therefore, be characterized as designed to enable one to find one’s way back to the everyday. They are methods for loosening the grip of misleading pictures, analogies, and so on that hold one’s thinking in a cramp and stand in the way of the recognition of the ordinary—the familiar, that which one already knew—for what it is.58 Why such pictures have an appeal for us, however, does not go without saying. Apparently, the attraction of particular pictures (and so on) can be made comprehensible by taking a closer look at how inclinations to think in particular ways are rooted in the human form of life, i.e., connected with various human aspirations, desires, fears, and so on. For example, philosophers’ craving for generality and the desire to describe

284 Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics

reality as a neatly ordered system—as Wittgenstein did in the Tractatus—might be partly an expression of a wish to gain control over an arbitrary and chaotic world.59 Similarly, the traditional view of statements about conceptual relations as statements of immutable truths60 might be seen as expressing a desire to transcend the limits of corporeal human existence. This suggestion that the roots of philosophical views might sometimes be found in practical desires may also be extended to include, for instance, aesthetic preferences and political convictions as a source of (or motivation for) philosophical views. (Surely the wish for systematic, elegant structures is also an aesthetic aspiration—though maybe the beauty found in philosophy will ultimately turn out to be something different from that of mathematics.) This means that the sources of philosophical conceptions—and consequently also difficulties with them—are not purely intellectual, so to speak. Wittgenstein writes: “What makes a subject difficult to understand—if it is significant, important—is . . . the contrast between the understanding of the subject and what most people want to see. Because of this the very things that are most obvious can become the most difficult to understand. What has to be overcome is not a difficulty of the intellect but of the will.”61 This observation might go some way toward explaining the persistence of particular philosophical conceptions and difficulties. But more importantly perhaps, the remark also brings to light an ethical dimension in philosophical thinking, or a sense in which philosophy requires self-examination. This point can best be explained by reference to certain connections between the interpretation of Wittgenstein developed in this book and that of Cavell. Among Wittgenstein’s commentators, the difficulty in philosophy of coming to grips with the everyday, or the dissatisfaction of philosophy with the ordinary, has been addressed especially by Cavell, who discusses skepticism as an expression of a relevant kind of dissatisfaction. In this context skepticism is defined as “any view that takes the existence of the world to be a problem of knowledge,”62 the skeptical philosopher’s task thus being the examination of the foundations of knowledge. The task of the Wittgensteinian philosopher, in turn, according to Cavell, is not to refute skepticism by determining what makes or does not make sense and demonstrating that the skeptic has overstepped the bounds of sense. Rather, the task is therapeutic: to discover the causes of philosophy’s disappointment with the ordinary and

Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics 285

its desire to renounce everyday language and the form of life with which it belongs together.63 Characteristic of Wittgenstein’s treatment of skepticism as presented by Cavell is, therefore, an emphasis on the intertwinedness of philosophy with other aspects of human life, or a recognition of philosophical views as deeply rooted in everyday life. This is illustrated by Cavell’s identification of a certain feeling of separateness from others—the feeling that others are enigmatic or do not understand one—as one source of skepticism about other minds. Skepticism gives this feeling of separateness a philosophical interpretation as metaphysical finitude, i.e., principled ignorance about others.64 The recognition of the intertwinedness of philosophical views with the rest of life has the important methodological consequence that achieving clarity about what tempts one to think about philosophical matters in such and such ways may require clarity about various aspects of everyday life and about oneself as a creature of that life. In this sense resolving philosophical problems may require the examination of one’s life and self, and philosophical clarification, as Wittgenstein envisages it, is also an exercise in self-examination. Accordingly, his methods can be characterized as methods of attaining self-knowledge, as Cavell also observes.65 This “practical” aspect of Wittgenstein’s approach to philosophy may be further elucidated as follows. As explained in earlier chapters, a key aspiration of Wittgenstein’s later thought is to do justice to the various aspects of philosophy’s objects of investigation in their complexity and to the thinking of those to whom clarifications are offered. In this respect Wittgenstein’s philosophy is informed by an ethical goal, namely justice. Accordingly, the core of his critique of metaphysics as well as the motive behind his struggle to develop an alternative approach to philosophy might ultimately be seen as ethical. As I have argued, the fundamental problem with metaphysics for Wittgenstein is that it involves dogmatic and falsifying projections of philosophical views onto reality, thus leading to injustice. The struggle against dogmatism, in turn, is a struggle with oneself and one’s prejudices (wherever they might have been inherited from). Thus, from the goal of justice ethical demands follow. For certainly the requirement of self-examination, or more concretely, the demand to address one’s prejudices and other tendencies of thinking that may falsify one’s perception of reality and lead to injustice, is an ethical demand.

286 Wittgenstein’s Conception of Philosophy, Everyday Language, and Ethics

More specifically, as Wittgenstein notes, what makes the critical examination of one’s views particularly difficult is the attachment one may develop to them, perhaps fueled by ambition, pride, and the like: “What [earlier variant: where] is the borderline between a judgment and a prejudice [Urteil & Vorurteil]?— [New paragraph] The edifice of your pride must be demolished. And that is a dreadful work.”66 This remark illustrates acutely the sense in which the problem of dogmatism is not merely a “theoretical” problem but an ethical (“practical”) problem.67 The preceding also explains Wittgenstein’s emphasis on the need for courage in philosophy, which otherwise seems merely an idiosyncratic concern. He writes: “I believe that what is essential is for the activity of clarification to be carried out with courage; without this it becomes a mere clever game.”68 That is, one reason courage is needed is that serious self-examination requires it. It takes courage to look into oneself, to face one’s shortcomings, to examine one’s prejudices, and so on. Similarly, to question seriously and critique one’s own thought—rather than merely going through the motions—is no easy undertaking. Such a critique may require one to dismantle conceptions that have taken years to develop—as in the case of the Tractatus—and to face the insecurity and stress of giving up what one may have considered one’s life achievement. Wittgenstein writes in a related remark: “You could attach prices to ideas. Some cost a lot some little. . . . And how do you pay for ideas? I believe: with courage.”69 Ethics then emerges here not as a branch of philosophy but as a dimension that pervades it in its entirety.70 Not only is the goal of philosophy ethical, but philosophy also places ethical demands on those who practice it. In describing the relation of Wittgenstein’s philosophy to the tradition of philosophy, one might, therefore, speak of a switch from metaphysical demands to ethical ones. Corresponding to Wittgenstein’s redescription of the goal of philosophy as not truth but justice (although conceptual bounds are not clear-cut here), his philosophy turns from the imposition of metaphysical demands onto reality to the acknowledgment of the ethical demands that philosophy places on its practitioners.

Notes

INTRODUCTION

1. Wittgenstein does not assume, at least not in his later work, that either of these two classes of statements should be unified or that their borderline is always clear. See for example, Ms174, 35v/OC §167. 2. G. E. Moore reports Wittgenstein to have said in his lectures “that there was now, in philosophy, a ‘kink’ in the ‘development of human thought’ comparable to that when Galileo and his contemporaries invented dynamics; that a ‘new method’ had been discovered, as had happened when ‘chemistry was developed out of alchemy’; and that it was now possible for the first time that there should be ‘skillful’ philosophers, though of course there had been in the past ‘great’ philosophers” (PO, 113). These are not modest assertions. The intriguing question is whether they could be taken seriously. And if the answer is yes, what is the method Wittgenstein thought he had discovered? 3. George Pitcher, The Philosophy of Wittgenstein (Englewood Cliffs, NJ: Prentice-Hall, 1964), 323. 4. Jaakko Hintikka and Merrill B. Hintikka, Investigating Wittgenstein (Oxford: Blackwell, 1986), 189. 5. Anthony Kenny, Wittgenstein, rev. ed. (Oxford: Blackwell, 2006), 126. 6. Anthony Kenny, “Philosophy States Only What Everyone Admits,” in E. Ammereller and E. Fischer, eds., Wittgenstein at Work: Method in the Philosophical Investigations (London: Routledge, 2004), 181. 7. Kenny, “Philosophy States Only What Everyone Admits,” 181–182. 8. See, for instance, Norman Malcolm, Nothing Is Hidden: Wittgenstein’s Criticism of His Early Thought (Oxford: Basil Blackwell, 1986), 148, 149.

288

Notes to Pages 6–11

9. Saul Kripke, Wittgenstein on Rules and Private Language (Oxford: Blackwell, 1982), 73, 74. Kripke’s attitude is more ambiguous than Malcolm’s. This seems clear from his suggestion that Wittgenstein’s nonstraightforward way of expressing his views might just be designed to cover contradictions in his philosophy. Had Wittgenstein expressed himself more straightforwardly (as Kripke himself is inclined to do on his behalf ), it would have become obvious that he is stating theses that not everyone would agree upon, contrary to what he says in PI §128 (See ibid., 69). 10. Cora Diamond, The Realistic Spirit (Cambridge, MA: MIT Press, 1991), 179. 11. Marie McGinn, “Saying and Showing and the Continuity of Wittgenstein’s Thought,” Harvard Review of Philosophy 9 (2001): 26. 12. See for instance essays in Alice Crary and Rupert Read, eds., The New Wittgenstein (London: Routledge, 2000); and in Erich H. Reck, ed., From Frege to Wittgenstein: Perspectives on Early Analytic Philosophy (Oxford: Oxford University Press, 2002). 13. For two different accounts, see Steven Gerrard, “One Wittgenstein?” in Reck, ed., From Frege to Wittgenstein; and Kenny, Wittgenstein, chap. 12. 14. See especially 1.1, 1.5, 2.1, 3.1, and 3.2. 15. See TLP 4.112. 16. PI, preface. 17. Newton Garver, This Complicated Form of Life: Essays on Wittgenstein (Chicago: Open Court, 1994), 13. A more sophisticated version of this kind of interpretation is developed by Hacker. See Peter Hacker, “Was He Trying to Whistle It?” in Wittgenstein: Connections and Controversies (Oxford: Oxford University Press, 2001), 98–140. 18. This approach is adopted by James Conant and Cora Diamond, whose work on the early Wittgenstein has inspired much recent work on the Tractatus. See, for example, James Conant and Cora Diamond, “On Reading the Tractatus Resolutely: Reply to Meredith Williams and Peter Sullivan,” in M. Kölbel and B. Weiss, eds., Wittgenstein’s Lasting Significance (London: Routledge, 2004), 46–99. 19. Books on Wittgenstein’s early conception of philosophy and logic include Mathew B. Ostrow, Wittgenstein’s Tractatus: A Dialectical Interpretation (Cambridge: Cambridge University Press, 2002); Denis McManus, The Enchantment of Words: Wittgenstein’s Tractatus Logico-Philosophicus (Oxford: Oxford University Press, 2006); and Marie McGinn, Elucidating the Tractatus: Wittgenstein’s Early Philosophy of Logic and Language (Oxford: Oxford University Press, 2006). For recent books on Wittgenstein’s later conception of philosophy, see S. Stephen Hilmy, The Later Wittgenstein:

Notes to Pages 14–19

289

The Emergence of a New Philosophical Method (Oxford: Basil Blackwell, 1987); Wolfgang Kienzler, Wittgensteins Wende zu seiner Spätphilosophie 1930–1932 (Frankfurt am Main: Suhrkamp, 1997); Beth Savickey, Wittgenstein’s Art of Investigation (London: Routledge, 1999); Gordon Baker, Wittgenstein’s Method: Neglected Aspects, ed. Katherine Morris (Oxford: Blackwell, 2004); and Alois Pichler, Wittgensteins Philosophische Untersuchungen: Vom Buch zum Album (Amsterdam: Rodopi, 2004). For discussions of the relation between the early and later philosophy, see Peter Hacker, Insight and Illusion, rev. ed. (Oxford: Oxford University Press, 1986); James Conant, “Why Worry about the Tractatus?” in B. Stocker, ed., Post-Analytic Tractatus (Aldershot: Ashgate, 2004), 167–192; Cora Diamond, “Criss-Cross Philosophy,” in Ammereller and Fischer, eds., Wittgenstein at Work, 201–220; and McGinn, Elucidating the Tractatus, chap. 12. 20. Hilmy, The Later Wittgenstein, 9–10. 21. For the notion of passage-hunting, see Hans-Johann Glock, “Perspectives on Wittgenstein: An Intermittently Opinionated Survey,” in G. Kahane, E. Kanterian, and O. Kuusela, eds., Wittgenstein and His Interpreters: Essays in Memory of Gordon Baker (Oxford: Blackwell, 2007), 47. 22. For different notions of “context” and Wittgenstein’s way of working, see Joachim Schulte, “Ways of Reading Wittgenstein: Observations on Certain Uses of the Word ‘Metaphysics,’” in Kahane, Kanterian, and Kuusela, eds., Wittgenstein and His Interpreters, 154–159. 1. WITTGENSTEIN ON PHILOSOPHICAL PROBLEMS

1. Problems with attempts to use Wittgenstein’s remarks for unintended purposes—for example, the construction of philosophical theories— often arise from a failure to see how the philosophical conceptions he offers are meant to engage with the problems he is addressing. An example is Kripke’s interpretation of Wittgenstein’s remark (PI §244) that characterizes verbal expressions of sensations (such as screams of pain) as replacements of their natural expressions, which, according to Kripke, amounts to the theory—clearly false—that every sensation has (must have) a corresponding natural expression. See Kripke, Rules and Private Language, 102–104, n. 83. See 2.32 for further discussion. 2. TLP 27, translation modified. 3. TLP 4.003. 4. TLP 3.323–3.324. 5. Some examples are the more technical problems of Frege or of Russell and Whitehead’s Principia Mathematica addressed in the Tractatus, for

290

Notes to Page 20

instance, the solution of Russell’s paradox or the question of whether the logical connectives represent something. 6. TLP 3.31–3.313. 7. See TLP 3.315–3.317. 8. TLP 3.321, 3.323, 3.326, 3.327. 9. One could also explain the distinction between meaning and mode of signification thus: given that a sign has a particular mode of signification, the meaning it has is contingent. What is not contingent, however, is that in order to have a particular meaning a sign must have a particular mode of signification. Logic as a discipline, according to the Tractatus, is concerned with this condition (see 2.1). It is concerned with what is necessary for sense and meaning. 10. To explain this use/interpretation of “logical syntax” in more detail, a distinction is to be drawn between using a sign according to logical syntax—which means symbolizing, in contrast to not symbolizing—and Wittgenstein’s notion of a language being “governed by logical syntax” (for this notion, see 2.1). Thus, every instance of symbolizing accords with logical syntax. In contrast, not every language is governed by logical syntax. Languages that are governed by logical syntax are tools that make it possible to avoid logical confusions resulting from our failure to use signs according to logical syntax. The notion of logical syntax in Wittgenstein has recently been an object of dispute among his interpreters. According to a view defended by Hacker, rules of logical syntax constitute a condition of the possibility of the expression of sense and in this capacity fix what makes sense. (See Hacker, “Trying to Whistle,” 117–121; see also “Wittgenstein, Carnap and the New American Wittgensteinians,” Philosophical Quarterly 53 (2003): 1–23. In contrast, Conant and Diamond deny that the notion of logical syntax in Wittgenstein can be understood in this way, and that one can, consequently, characterize nonsense as the result of violating the rules of logical syntax. (See James Conant, “The Method of the Tractatus,” in Reck, E. G., ed., From Frege to Wittgenstein, Perspectives on Early Analytic Philosophy (Oxford: Oxford University Press, 2002), 414ff.; Cora Diamond, “Criss-cross Philosophy” and “Logical Syntax in Wittgenstein’s Tractatus,” Philosophical Quarterly 55 (2005): 78–89). My interpretation accords with that of Conant and Diamond. To speak about logical syntax is to describe or clarify the use of a sign and how it symbolizes, but there seems to be no reason to assume that Wittgenstein would be committed even in the Tractatus to a philosophical explanation according to which the meaningfulness of a sign depends on it being used in conformity with a system of rules that fixes what makes sense. Rather, in the

Notes to Pages 21–23

291

Tractatus he is committed to the methodological idea that to analyze or describe the use of an expression is to lay out every possible logical relation of an expression to every other expression, and to the corresponding view about language that to use and expression in a meaningful way always involves the establishing or assuming such a set of definite connections (see 2.21). In this sense he is committed to an idea of there being an implicit system of rules in language, i.e., that language has a definite implicit order. This, however, is not to assume that the possibilities of expression depend on the rules of logical syntax but only that expressions have a determinate sense—that anyone who utters an expression is playing a definite game, as he later puts it (see Ts213, 253v; PI §81). This dispute has a connection with the subsequent discussions in this book. Assuming that Wittgenstein does not take the meaningfulness of a sign to depend on its use according to any established rules, a question arises: What is the point of talking about rules of language in philosophy or logic? A way to read Chapter 3 is to see it as spelling out an answer to this question, including the question why Wittgenstein thought it necessary to revise the Tractatus’s conception of the role of rules. 11. TLP 5.473. 12. TLP 5.4733. 13. As noted, Wittgenstein might be assuming here a specific notion of properly philosophical problems as ones concerning necessities rather than empirical matters. 14. TLP, 29. 15. Wittgenstein’s conception of a critique of language is also discussed in 2.1. 16. TLP, 27. 17. See TLP 4.5, 5.471; see 2.1 and Chapter 3 for further discussion of the notion of general propositional form. 18. For Kantian interpretations of this kind, see for example Erik Stenius, Wittgenstein’s Tractatus: A Critical Exposition of Its Main Lines of Thought (Oxford: Blackwell, 1960), chap. XI; Hacker, Insight and Illusion, chap. 1.4; Heikki Kannisto, Thoughts and Their Subject: A Study of Wittgenstein’s Tractatus. Acta Philosophica Fennica, vol. 40 (Helsinki: Philosophical Society of Finland, 1986). 19. See TLP 4.12–4.124. 20. See Hacker, Insight and Illusion, 20, 21, 25, 26; “Trying to Whistle,” 99, 101, 105, 117, 121. Hacker represents here a traditional line of interpretation, other representatives of which include, Elisabeth Anscombe,

292

21.

22.

23.

24. 25. 26.

Notes to Pages 24–26

An Introduction to Wittgenstein’s Tractatus (London: Hutchinson, 1971); Norman Malcolm, Nothing Is Hidden; and Kenny, Wittgenstein. See 2.1 for further discussion of the problem of the Tractatus’s “nonsensical doctrine.” The contrast between thought and language here turns on the idea that the concept of thought only includes real thoughts that can be made sense of, i.e., entertained, while the concept of language in a broad sense also incorporates nonsensical sign formations, i.e., signs that do not symbolize. The contrast between the two approaches here reflects a dividing line in Wittgenstein interpretation. Since the 1980s, the so-called resolute readers of the book have sought to contest the traditional line of interpretation that attributes to Wittgenstein an ineffable theory of the essence of language and reads him as accepting a notion of truths (or quasi-truths) that cannot be put into words. It is characteristic to resolute readings that they interpret Wittgenstein as aspiring to draw a limit to language use not by reference to a technical notion of nonsense about which the Tractatus (paradoxically) seeks to inform its reader, but by relying on its reader’s natural logical capacity. For resolute readings, see for example Conant, “The Method”; Diamond, Realistic Spirit, chap. 6; Michael Kremer, “The Purpose of Tractarian Nonsense,” Noûs 35 (2001): 39–73; Ostrow, Wittgenstein’s Tractatus; and McManus, The Enchantment. My interpretation of the Tractatus accords mostly with that of the resolute readers. In 2.1 I explain how I deviate from (at least some of ) the authors mentioned here. Consider the remarks at the end of the Tractatus. Here Wittgenstein’s point about silence is also explained by reference to methodological considerations rather than doctrinal commitments. Remarks 6.53 and 6.54, before the final declaration of silence in remark 7, concern the method of philosophy. TLP 4.112. TLP 4.115; see also 4.113–4.114. Marie McGinn contrasts her reading with that of Conant and Diamond by characterizing the latter as interpreting the Tractatus in negative terms, as not offering any positive insights about the logic of language. See McGinn, Elucidating the Tractatus, 6, 8. As I seek to explain in 2.1, there is something misleading in characterizing the dissolution of philosophical nonsense as a negative task in contrast to a positive task of the clarification of logic. Accordingly, Wittgenstein’s conception of philosophical problems as misunderstandings does not entail that philosophy has no “positive insights” to offer.

Notes to Pages 27–31

293

27. See PI §§90, 93, 109, 111, 194. In many contexts Wittgenstein seems to use the expressions “forms of expression,” “forms of language,” and “modes of expression” interchangeably (see for instance Ms114, 223; Ms116, 245; Ms120, 68v; and Ms157b, 12r). These terms do not seem to have any technical meaning in his writings but simply refer to words and the kinds of expressions we use. Thus he talks, for example, of “I have an image [Vorstellung]” as an expression of ordinary language, as a mode of expression, and as a form of expression (PI §402). 28. PI §90. 29. Ts213, 421/ PO 180, and Ms140, 10, respectively; see also Ms183, 65. 30. PI §122. 31. PI §203 and §123, respectively. 32. LC 2. See also 1.4 below. 33. BB 28; see also Ms116, 62, 63. 34. Ts220, §106; see also Ms142, 110; translation based on Ts213, 410/ PO 165. 35. See BB 26 and quotations below. 36. TLP 4.003, quoted in 1.1. 37. BB 46. 38. Ms115, 10 / PG 169; see also PI §93. 39. BB 31; see PI §§89, 109. As Wittgenstein notes, however, confusions or misunderstandings similar to philosophical problems might arise in various representational media (Ms114, 173ff. / PG 141ff.; see also Ms156b, 2r). In this sense, this kind of problem is not specific to “word language.” 40. Ms115, 36 / PG 193; see also Ms145, 41. 41. Ms179, 17v, 18r. 42. The problem about measuring time—and what it is for time to be short or long—is raised and discussed by St. Augustine. See The Confessions, rev. trans. E. B. Pusey (Oxford: John Henry Parker, 1853), secs. 11, 15. The same conception of time, however, is already present in Aristotle. See The Physics, ed. T. E. Page, E. Capps, W. H. D. Rouse, trans. P. H. Wicksteed, The Loeb Classical Library (London: William Heinemann, 1929), 217b33–218a2, 218a6–a8. It should be noted, however, that since it is up to the person who has the problem to acknowledge its description and the dissolution offered (see above), examples like the present one are only suggestive illustrations of the activity of solving a philosophical problem. 43. BB 26; Ms302, 24. See also Friedrich Waismann, The Principles of Linguistic Philosophy (London: Macmillan, 1997), 41–43.

294

Notes to Pages 31–40

44. See BB 26; Ms302, 24; Ms116, 185, 186; Ms120, 82v; Ms115, 110; Ms157a, 24r. 45. In both his early and later philosophy, Wittgenstein characterizes the failure to distinguish between these two kinds of sentences as the central confusion and unclarity of philosophy. See TLP 4.122, 4.126 and RPPi §949. The notion of necessary truth is discussed more closely in Chapters 3 and 5. 46. BB 57. 47. See PI §§303, 402; Ms148, 32r / PO 208, 209. 48. BB 57 and 56–57, respectively. 49. See BB 9, 10 and discussion in 2.31 and 7.3. 50. BB 7; see also Ms111, 107; PI §90. 51. Ms154, 76r; see also Ts213, 409. 52. PI §112. 53. BB 35; see also 43, 44. 54. See PI §§143–201, 665–693 and BB 141–143 for discussion of the problem of meaning. 55. Many commentators take a different line here. For instance, Kripke suggests that the solution Wittgenstein offers to the problem of meaning is a communal account of meaning according to which the agreement of the linguistic community decides what linguistic expressions mean. See Kripke, Rules and Private Language, 96, 97. 56. BB 169. 57. PI §305; see also §§103, 115. 58. TLP 2.02–2.0212, 3.23; see also Ms102, 161r / NB 63, 64. 59. See Baker, Wittgenstein’s Method, 266–269 for a discussion of the concept of picture in Wittgenstein. 60. See PI §115. 61. PI §425. 62. PI §425. 63. PI §295. 64. See PI §308. 65. PI §1. 66. PI §423; see also §§374, 424. 67. See PI §§258, 293 and discussion in the next subchapter. 68. PI §117; see Ms127, 73. 69. PI §350. 70. See PI §351; Ms121, 91r, 91v. 71. See BB 48, 49, 53; PI §302. 72. See PI §349. 73. See, for example Kenny, Wittgenstein, 130 and discussion in 2.31 and 7.3.

Notes to Pages 41–48

74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84.

85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97.

98.

295

Ms116, 216–218; see also Ms120, 39v. Ms121, 59v. BB 28. The concept of meaning is discussed in more detail in Chapter 4. Ms120, 143v. PI §3. Ms115, 262. Ms154, 31v; a notebook from 1932. See PI §§81, 130. Ts213, 406–407/ PO 161. See 7.3 for further discussion. The characterization of philosophical problems as expressions of certain kinds of tendencies of thinking is, as it were, formal. It says nothing about how such tendencies are rooted in human life—in its desires, fears, disappointments, and so on—or connected with nonphilosophical aspects of human life. Stanley Cavell has discussed such themes under the theme of “skepticism.” See for example Cavell, “Something Out of the Ordinary,” in Philosophy the Day after Tomorrow (Cambridge, MA: Belknap Press / Harvard University Press, 2005), 11, 12; and The Claim of Reason: Wittgenstein, Skepticism, Morality, and Tragedy (Oxford: Oxford University Press, 1979). I return to these issues in 7.3. PI §254; see also §§90, 299. Ms179, 22v, 23r. PI §255. See Ms110, 229–230; Ts213, 410 / PO 165; Ts239 §139. Ms110, 58 / PO 119. PI §133. Ms116, 185, 186; also in Ms120, 82v; see also Ms115, 110; Ms157a, 24r. Ms118, 59r, 59v. Ms107, 208, 209. Ms127, 82 / CV 50e; see Ms117, 95 for a similar remark. PI §133. See also Ms163, 41r, 41v (quoted in 6.1). See also 2.1 and 6.1 for discussion. Ms102, 63r, 64r / NB 39; italics in the original. Given that the Tractatus conceives language as the totality of propositions, the question about the nature of the proposition is simply a more specific formulation of the question about the essence of language (see TLP 4.001). See Diamond, “Criss-Cross Philosophy” for a discussion of the contrast between a single problem and multiple problems. For discussion of the Tractatus’s concern with language as a great problem, see McGinn, Elucidating the Tractatus, chap. 1; and “Wittgenstein’s Early Philosophy of Language and the Idea of ‘Single Great Problem,’” in A. Pichler and

296

99. 100.

101.

102. 103.

104. 105. 106.

Notes to Pages 48–50

S. Säätelä, eds., Wittgenstein: The Philosopher and His Works. Working Papers from the Wittgenstein Archives at the University of Bergen, no. 17 (Bergen: Wittgenstein Archives at the University of Bergen, 2005). See BB 44. Ms142 §134 / Ts220 §116 / Ts213, 431, 432 / PO 195/ Z §447; first drafted in 1931 in Ms112, 71v. This remark may be read as supporting the argument that Wittgenstein’s aim of finding peace in philosophy does not translate into abandoning philosophy altogether. Rather than a distinction between philosophizing and not philosophizing, the distinction he draws in §133 and elucidates here is between two approaches to philosophy: a piecemeal approach and an approach that aims at once-and-for-all answers to philosophical questions. Notably, once-and-for-all answers of this type may be conceived as relative to particular languages and need not involve “an absolutist view of essences” in which essences and concepts are conceived as something ahistorical or transhistorical. “Once and for all,” that is to say, may be taken to mean merely “once and for all in the context of a particular language” or “once and for all insofar as we mean by the concept what we mean by it now” or “insofar as we mean by the concept what it normally means,” and so on. A once-and-for-all answer need not make a stronger claim than this. I also employ the term “universal” in a correspondingly “weak” or contextualized sense (unless otherwise noted). A statement, that is to say, is universal insofar as it covers all cases falling under a concept, although the concept in question may be conceived as historically contingent. The important point is that Wittgenstein’s critique of universal philosophical theses about essences, as I describe it in the following chapters, applies whether such theses are interpreted in a strong (absolutist) or a weak (historically contextualized) sense. I discuss the concepts of philosophical thesis and essence more closely in 3.1–3.2 and Wittgenstein’s later conception of essence in Chapter 5. See 6.1 for further discussion. More recently, the question of being has been taken up by Heidegger, whose early fundamental ontological approach in Being and Time, with the analysis of Dasein at its center, seems to exhibit a relevant kind of hierarchical structure. The early Wittgenstein too conceives his question about the nature of the proposition as a key to the question of being. See Ms102, 63r, 64r / NB 39. See also Ms110, 200; Ts213, 407. Ts213, 431/ PO 195. Ibid.

Notes to Pages 50–56

297

107. A more concrete description of the kind of disquietude caused by the conception of philosophy as attempting to provide once-and-for-all answers to philosophical questions is given in the following remark, in which Wittgenstein recalls his work with Russell: “In the course of our conversations Russell would often exclaim: ‘Logic’s hell!’—And this fully expresses what we experienced while thinking about the problems of logic; namely their immense difficulty. Their hardness—their hard & slippery texture. . . . The primary ground of this experience, I think, was this fact: that each new phenomenon of language that we might retrospectively think of could show our earlier explanation to be unworkable. But that is the difficulty Socrates gets caught up in when he tries to give the definition of a concept. Again and again an application of the word emerges that seems not to be compatible with the concept to which other applications have led us. We say: but that isn’t how it is!—it is like that though!—& all we can do is keep repeating these antitheses” (Ms119, 60 / CV 35). 108. This is the germ of the explanation of why and in what sense Wittgenstein is not committed to a thesis about the essence of philosophy, as discussed in Chapter 7. See 6.1 for an explanation of the sense in which Wittgenstein’s later approach is not based on a thesis about the essence of language and a discussion of the rejection of philosophical hierarchies in his later work. 109. Ms116, 185–186; see quotation in 1.4. 110. See 2.31, where I seek to further clarify what is at stake in Wittgenstein’s shift from great problems to particular problems. 111. As he says in the quotation from Ms142, §134 / Ts220 §116 / Ts213, 431, 432 / PO 195/ Z §447 above. 112. Ibid. 113. Since we are not dealing with an empirical generalization, more examples can merely serve as a means of persuasion, but this is a psychological matter. See Chapter 3 for a discussion of problems with the philosophical notion of deriving theses of essence from examples and of Wittgenstein’s conception of the role of examples in philosophy. 114. See 7.1 for an explanation of the sense in which Wittgenstein seeks to “demonstrate a method by examples.” 115. See 7.1 for discussion. 2. TWO CONCEPTIONS OF CLARIFICATION

1. TLP 4.112. 2. TLP 3.325; see RLF 29–30 for another statement of this idea.

298

Notes to Pages 56–58

3. TLP 5.5563. 4. Ms102, 85r / NB 45. McManus characterizes the concept-script as a “mechanical expedient” that “prevents ‘mechanically’ ” the formulation of meaningless strings of signs (McManus, The Enchantment, 85) Although this is correct in certain respects, it also seems potentially misleading, since analyses in terms of a concept-script, which are essentially translations from other modes of expression into the concept-script (see below), cannot be a merely mechanical matter. But certainly a very significant part of clarification in terms of a concept-script seems to consist of figuring out appropriate translations. Nevertheless, insofar as questions of translation have been settled in particular cases, then the concept-script may be said to mechanically prevent the formulation of meaningless statements. The case is analogous to how Finnish, for example, “mechanically prevents” the expression of gender distinctions in the sense that it utilizes no he/she distinction and assigns no gender to substantives. 5. Juliet Floyd contests the idea that there is, according to the Tractatus, such a thing as the logical order of language which a correct conceptscript is meant to capture. (See her “Number and Ascriptions of Number in Wittgenstein’s Tractatus,” in J. Floyd and S. Shieh, eds. Future Pasts: The Analytic Tradition in Twentieth Century Philosophy (Oxford: Oxford University Press), 179–180). For a recent discussion of this issue, see “Wittgenstein and the Inexpressible,” in A. Crary, ed., Wittgenstein and the Moral Life: Essays in Honor of Cora Diamond (Cambridge, MA: MIT Press, 2007), 177–234.) This view seems to be in conflict with the one presented in this book, although, as I explain in 3.1, I understand Wittgenstein’s commitment to the idea of a correct concept-script in a wide sense. He is committed to an idea of it being possible in principle to capture the logic of language in terms of some concept-script, not to a claim about the correctness of any particular actual notation. As becomes clearer later on, Wittgenstein’s commitment is intimately connected to his assumption that a satisfactory philosophical method must be the universally applicable to any philosophical problem whatsoever. 6. TLP 3.34. 7. TLP 3.341. 8. TLP 3.3411. 9. TLP 3.3421. 10. See TLP 3.343, 4.0141. 11. See TLP 4.002. 12. See Gottlob Frege, Begriffschrift und andere Aufsätze, ed. I. Angelelli (Hildesheim: Georg Olms Verlagsbuchhandlung, 1964), xi, xii, 97, 98. For Leibniz’s project of constructing a universal language, see Adolf

Notes to Pages 58–60

13. 14.

15. 16. 17. 18. 19. 20. 21. 22. 23.

24.

25.

299

Trendelenburg, Historische Beiträge zur Philosophie, Dritter Band, Vermischte Abhandlungen (Berlin: Verlag von G. Bethe, 1867). TLP 4.122–4.124. See Diamond, Realistic Spirit, 119; McGinn, “Saying and Showing,” 24–36. This is a contested point at which my interpretation differs both from the mainstream of traditional interpretations and from at least some resolute readings. The latter difference is evident in the case of Ostrow, who writes: “But while it is unquestionable that the notion of canonical Begriffsschrift plays an important (if extremely unclear) role in the Tractatus, it is equally certain that Wittgenstein has not actually provided us with any such language” (Wittgenstein’s Tractatus, 9). Interestingly, Russell on the other hand seems to take it for granted in his introduction to the Tractatus that Wittgenstein does provide us with such a notation (whatever misunderstandings he might have of Wittgenstein’s project otherwise). Thus, Russell characterizes the Tractatus as concerned with “conditions which would have to be fulfilled by a logically perfect language” (TLP, 7) and talks about “Wittgenstein’s theoretical logical language” (16). Perhaps Russell’s famously mistaken belief that Wittgenstein assumes everyday language to be logically imperfect is due to his focus on Wittgenstein’s scheme for a logically perfect notation, i.e., one that (unlike everyday language) is governed by logical syntax in the sense explained above (see TLP, 8). For a discussion of Russell’s mistake, see Anscombe, An Introduction, 91–92. See TLP 4.22, 4.221, 4.5, 5. See for instance Diamond, Realistic Spirit, 115, 184. TLP 3.261. TLP 3.343. See TLP 4.221. See TLP 3.201, 3.202, 3.26, 3.261. See Hacker, Insight and Illusion, 24. TLP 3.25. See Ms102, 88r-v / NB 46–47 for the notion of analysis of propositions. See 2.2 for further discussion of the Tractatus’s conception of analysis and problems with it. Traditional interpretations might be characterized as taking Wittgenstein to be putting forward such a speculative doctrine, a philosophical fairy tale, as Kannisto aptly characterizes this approach. See Kannisto, Thoughts and Their Subject, 15. See TLP 3.14, 3.21, 3.22, 4.031, 4.0311. I will not go into the details of the picture theory or the Tractatus’s conception of language more generally. For expositions and discussions, see Anscombe, An Introduction;

300

26.

27.

28.

29. 30. 31.

Notes to Pages 61–62

Stenius, Wittgenstein’s Tractatus; Hacker, Insight and Illusion; McGinn, Elucidating the Tractatus; and McManus, The Enchantment. I cannot discuss examples of such problems here, but they include problems relating to the status of logical constants and the propositions of logic. These issues are discussed in detail by McGinn in Elucidating the Tractatus. At this point, the awkwardness of the division of the Tractatus’s task into a negative (critical) and a positive component (perspicuous presentation of logic, offering insights into the logic of language) should become evident. Wittgenstein’s critique of the philosophical tradition as engaged in the discussion of nonsensical questions (his “negative achievement”) is entirely dependent on the correctness of the clarifications of the logic of language he undertakes in the form of the introduction of his concept-script (i.e., on his “positive achievement” in logic). But one could not be said to have correctly or fully understood the principles of Wittgenstein’s concept-script without comprehending its critical import, in particular, how one’s recognition of those principles as justified excludes from the realm of sense things one might otherwise be inclined to say. (Indeed, a comprehension of his book’s critical import is what Wittgenstein mentions as a criterion for understanding him correctly in TLP 6.54.) In this sense, introducing this notation, thereby offering something one might call “insights into the logic of language,” is not simply a “positive” undertaking either. (See below for discussion of Wittgenstein’s method of introducing his concept-script.) See TLP 6.53. As mentioned in 1.1, a current leading proponent of the traditional interpretation is Hacker. See Hacker, “Trying to Whistle,” 101–102 for a statement about difficulties that the nonsensicality of the Tractatus poses for its interpreters. See discussion in 1.1. Hacker, “Trying to Whistle,” 124. This problem is discussed by Warren Goldfarb, “Metaphysics and Nonsense: On Cora Diamond’s The Realistic Spirit,” Journal of Philosophical Research 22 (1997): 57–73. More precisely, an interpretation is “purely therapeutic” in the relevant sense if it denies that a goal of Wittgenstein’s is to introduce a concept-script to be used for the purpose of clarification—perhaps arguing instead that his purpose is merely to reveal the nonsensical nature of such a script. Taking into account the interpretational possibility that the Tractatus might have intended the concept-script not as universally applicable to every possible philosophical problem but merely as (what the later Wittgenstein calls) an object

Notes to Pages 62–64

32. 33.

34. 35. 36. 37. 38.

39. 40.

301

of comparison (see Chapter 3), there might not be that many actual representatives of the “purely therapeutic” reading. An interpretation according to which a concept-script is to be seen as an object of comparison is adopted by Rupert Read (personal communication; see also Rob Deans and Rupert Read, “ ‘Nothing Is Shown’: A Resolute Response to Mounce, Emiliani, Koethe and Villhauer,” Philosophical Investigations 26 [2003]: 239–268.) It seems also compatible with the reading of Juliet Floyd (see “Number and Ascriptions of Number”). An argument against this kind of reading may be developed on the basis of Wittgenstein’s critique of his early philosophy, as described in Chapter 3, but I leave this for another occasion. Ms105, 10, 12. The present interpretation is evidently in conflict with the resolute reading of Ostrow, who denies that Wittgenstein introduces a conceptscript (see above). It seems to me compatible with the interpretation of Conant and Diamond, but they have not developed their reading in this direction, and consequently the matter is somewhat unclear. (See Diamond, Realistic Spirit, 183 and discussion below.) For a more detailed discussion of this interpretative issue with respect to Conant and Diamond, see Oskari Kuusela, “Nonsense and Clarification in the Tractatus—Resolute and Ineffability Readings and the Tractatus’ Failure,” in Sami Pihlstöm, ed., Wittgenstein and the Method of Philosophy. Acta Philosophica Fennica 80 (Helsinki: Philosophical Society of Finland, 2006), 35–65. See Diamond, Realistic Spirit, 183 and chap. 6 for discussion. See also Conant, “The Method.” See TLP 4.12-4.1212. See TLP 4.122, 4.126. I discuss this issue in more detail in 3.1. TLP 6.53. For discussion, see Diamond, Realistic Spirit, 197 and McGinn, “Saying and Showing,” 27–28. For the role that Wittgenstein ascribes to tautologies in this regard, see TLP 6.124. For discussion, see Richard McDonough, The Argument of the Tractatus, Its Relevance to Contemporary Theories of Logic, Mind and Philosophical Truth (Albany: State University of New York Press, 1986). For a case study of how some of the Tractatus’s sentences dissolve into nonsense, see Stephen Mulhall, “Words, Waxing and Waning: Ethics in/and/of the Tractatus Logico-Philosophicus,” in G. Kahane, E. Kanterian, and O. Kuusela, eds., Wittgenstein and His Interpreters: Essays in Memory of Gordon Baker (Oxford: Blackwell, 2007), 221–247. See TLP 6.54. TLP 5.472.

302

Notes to Pages 64–75

41. TLP 4.5, 4.53, 5.47. See also 3.31–3.317, where the relevant sense of “description” is explained. Consider also Wittgenstein’s remark in his notebook: “Does the general propositional form exist? Yes, if by that is understood the single ‘logical constant!’” (Ms102, 85r / NB 45). 42. Both what I am calling here the “initial task” and the “task” of philosophy fall under the broader notion of the critique of language. I return to the distinction between these tasks in 3.1. 43. TLP 3.2. 44. PI §90. 45. PI §91. 46. See 1.5. 47. PI §92. 48. One might distinguish a third strand in Wittgenstein’s later critique of the Tractatus, the critique of the idea that analysis reveals something hidden or involves a discovery. I do not discuss this in any detail, however, as it is not particularly important to my overall argument. 49. See PI §91, quoted above. 50. Ms116, 80–81/ PG 211. 51. AWL 21. 52. Ts220 §114 / Z §440; see Ms142 §132 and also Ms115, 50–52. 53. As Wittgenstein says: “Whatever is possible in logic is also permitted” (TLP 5.473). See discussion in 1.1. 54. See Ms115, 50. 55. PI §87. See Ms115, 54, 55 and 2.3 below for further discussion. 56. See PI §88; Ms115, 51; Ts213, 256v. 57. Ms115, 55; see also Ts213, 257. 58. PI §64. 59. PI §61. 60. PI §62. 61. Related issues are discussed in Chapter 5 on Wittgenstein’s conception of essence. 62. PI §63. 63. See 6.5. 64. See 2.1. 65. Ms110, 221; see also Ms136, 31a / CV 73, 74. 66. Ms110, 222, 223; see also Ts230, 1–3. 67. PI §108. The second of these two paragraphs of §108 is originally drafted in Ms110, 221 (from 1931) in the immediate context of the previous quoted passage (see Ts211, 259, 260). The first paragraph is originally drafted in Ms114, 108 (from 1932/33). See Chapter 3 for further discussion of §108.

Notes to Pages 75–79

303

68. PI §§109, 124, 126. 69. Ms130, 35. See 6.1 for a discussion of the notion of philosophical, metaphysical explanations. 70. Ms130, 51, 52. 71. Rush Rhees, “Wittgenstein’s Builders,” in K. T. Fann, ed., Ludwig Wittgenstein: The Man and His Philosophy (New York: A Delta Book, 1967), 253. See also Rhees, Wittgenstein and the Possibility of Discourse, ed. D. Z. Phillips (Cambridge: Cambridge University Press, 1998). 72. Oswald Hanfling, Wittgenstein’s Later Philosophy (Basingstoke: MacMillan, 1989), 42. See also Newton Garver, This Complicated Form of Life, 228. 73. The sense in which the Tractatus put forward such a thesis about the essence of language is discussed in Chapter 3. 74. For Wittgenstein’s conception of essence, his account of the appearance of philosophical statements concerning something immutable or timeless, and the historicity of statements about essence, see Chapter 5. 75. As will become apparent in subsequent chapters, the methodological shift that constitutes Wittgenstein’s turn to his later philosophy is not a shift from strong, universal philosophical theses to weaker, historically contextualized theses. Clearly, the latter would still count as philosophical theses, even if only weak ones. I discuss the historicity of philosophy at more length in 7.2. 76. PI §§111 and 119, respectively. See also §118. 77. PI §120. 78. It remains entirely unclear how such a claim to universality would be justified without falling into dogmatism; see Chapter 3 for further discussion. 79. Ts245, 232 / RPPi §548. 80. See 1.3. 81. See RPPi §549. 82. See AWL, 27, 97 and 6.4 for further discussion. 83. For interpretations according to which everyday language does constitute a standard of sense to which one must conform, see Hintikka and Hintikka, Investigating Wittgenstein, 138, 147, 161; Avrum Stroll, Moore and Wittgenstein on Certainty (New York: Oxford University Press, 1994), and passages from Baker and Hacker in 7.3. Similarly, Kenny’s idea that philosophical problems come about when we use an expression “in a language-game other than the one appropriate to it” (Wittgenstein, 130) seems to assume that sensible language use must conform to some clearly definable standard of sense provided by one language-game or another. See 7.3 for further discussion.

304

Notes to Pages 79–81

84. Savickey, Wittgenstein’s Art, 105. 85. The view that deviation from ordinary language as such is not philosophically problematic, and consequently that appealing to ordinary language as a standard of sense cannot settle philosophical disputes, distinguishes Wittgenstein from Oxford ordinary language philosophers such as J. L. Austin and Gilbert Ryle (at least according to a certain received interpretation of these philosophers). Cavell writes about problems with such appeals: “The appeal to ordinary language cannot directly repudiate the skeptic (or the traditional philosopher generally) by, for example, finding that what he says contradicts what we ordinarily say or by claiming that he cannot mean what he says: the former is no surprise to him and the latter is not obviously more than a piece of abuse” (Stanley Cavell, Must We Mean What We Say? (Cambridge: Cambridge University Press, 1976), 240; see also 60–62). The relation of Wittgenstein’s philosophy to everyday language use is discussed in more detail in 7.3. 86. Ms156a, 19v, 20r; from 1932. 87. Ms121, 59r, 59v; see also PI §182. This remark occurs in the context of remarks on the verbs “to be able” and “to understand.” 88. AWL 97. 89. Here relevant problems are any philosophical problems one might need to address in the course of discussing a philosophical issue, regardless of whether one sets out from these problems, they emerge in the course of the investigation, or somebody points them out after one had already closed the case. Apparently, the point about the relativity of completeness of descriptions to particular problems applies to Wittgenstein’s “perspicuous presentations” as well. As he writes (remarking on a problem about infinity in the philosophy of mathematics): “The aim is a perspicuous comparative presentation of all the applications, illustrations and conceptions of the calculus. The complete survey of everything which may produce unclarity” (Ts233a, 56 / Z §273, my emphasis; see also Ms116, 54, 55). Here again the criteria of completeness are determined by actual unclarities in a specific historical context. (The remark is a response to a statement by the mathematician Hardy.) I return to the notion of perspicuous presentation at various points in subsequent chapters and at length in 6.2. 90. PI §133. 91. In addition, other conceptions of philosophical practice may, of course, fit the description. 92. Ts220 §114 / Z §440. 93. Connections between these issues are also witnessed in Wittgenstein’s discussion in Ms115, 50–55; cf. Ms152, 88–92.

Notes to Pages 83–89

305

94. The same point is made in a manuscript of the same year, and again in a slightly different way in manuscripts and typescripts from 1944. (See Ms114 195, 196 / PG 119, 120; Ts229, 334 / RPPi §633.) These other occurrences make manifest that this quotation describes a constant feature of Wittgenstein’s later methodology. 95. Ms145, 39, 40. I have replaced a semicolon in the first sentence with a comma. 96. The notion of a theory of essence, as well as Wittgenstein’s later conception of the use of examples, is discussed in detail in Chapter 3. 97. This also explains what Wittgenstein means by saying that the procedure of examining the uses of words is not “another method of giving meaning” (see the passage from AWL 97 above). A philosophical account that concerns some particular way a word is used does not constitute an account that one would give to anyone to explain the meaning of a word. Philosophy as conceived by Wittgenstein therefore has very little to do with the composition of dictionaries. 98. PI §109. 99. See 1.2. 100. See 6.5 for further discussion. 101. See PI §§127, 89. 102. But see also the discussion of what I call “multidimensional descriptions” in 6.6. The possibility of such descriptions is a consequence of Wittgenstein’s conception of the status of philosophical statements, as described in Chapter 3. 103. PO 114; cf. Ms108, 238; Ms117, 140. 104. Ts213, 421/ PO 181; see also Ms112, 117v; Ms153b, 34r; Ms172, 23. 105. PI §109. 106. Ms117, 220–221. 107. See PI §92. 108. I discuss Wittgenstein’s concept of perspicuous presentation at length in 6.2, where I will also be in a position to draw on the results of the preceding chapters on Wittgenstein’s method. 109. See Ms137, 9b-10a. 110. Ms133, 9v. The relevant notion of ad hoc concepts and their employment in clarification become clearer in Chapter 3. 111. Ms134, 154 / Ts229 §1617/ RPPi §950. 112. See PI §§244, 245, 310; RPPi §§151, 313. 113. See PI II v; RPPi §§287, 288; LWi §964. 114. See RPPi §§137, 141. 115. See PI §345. 116. Ms117, 264; see also Ms120, 143v, quoted in 1.3.

306

Notes to Pages 89–91

117. Ms118, 73v; cf. 77v; Ms134, 146 / CV 70e; PI §144. 118. See PI §360 for an example of a suggestion of this kind. The conception of philosophy as ordering can also help us comprehend Wittgenstein’s view that philosophy does not provide new knowledge (see PI §89). With respect to this view, one might wonder, for instance, whether Wittgenstein means that philosophy does not teach us anything new. The notion of philosophy as ordering suggests a response. When, as a consequence of facts being ordered in a new way, one comes to see connections that one did not see before, or an aspect dawns on one, one comes to realize something. Although such a realization may be aptly described as coming to see something new, it is not a case of acquiring new knowledge in the sense of discovering new facts. One already knew the facts; now they are only arranged in a new way. 119. PI §132. 120. Gordon Baker and P. M. S. Hacker, Wittgenstein: Understanding and Meaning. Vol. 1 of An Analytical Commentary on the Philosophical Investigations (Oxford: Blackwell, 1980), 484. Anthony Kenny proposes a similar interpretation in his The Legacy of Wittgenstein (Oxford: Blackwell, 1984), 43. 121. See Baker and Hacker, Commentary, 1: 484 and Gordon Baker and P. M. S. Hacker, Wittgenstein: Rules, Grammar and Necessity. Vol. 2 of An Analytical Commentary on the Philosophical Investigations (Oxford: Blackwell, 1985), 54, 55. See also Ts213, 408–411, 413 and AWL 31, where Wittgenstein discusses the differences between his investigations and empirical studies of language. 122. See Baker and Hacker, Commentary, 1: 531. Baker and Hacker’s view of Wittgenstein’s notion of ordering is also discussed in 6.2 in connection with the interpretation of Wittgenstein’s concept of perspicuous presentation. 123. TLP 5.5563. 124. Ts213, 253r; see Ms112, 95r; 211, 491; 212, 728. This remark is ultimately developed into PI §81; see also §97, which is an explicit comment on TLP 5.5563. 125. See Ms112, 95r and Ms152, 95. 126. One difference between the view attributed by Baker and Hacker to the later Wittgenstein and that of the Tractatus is that Baker and Hacker conceive Wittgenstein’s order of language as specific to particular languages rather than underlying every possible language. (See Baker and Hacker, Commentary, 2: 40.) Significantly, however, this seems merely to make clarification in Wittgenstein’s later philosophy a

Notes to Pages 91–97

127. 128.

129. 130. 131.

132. 133. 134. 135.

307

localized, weaker version of clarification in the Tractatus, while at the local level of particular languages the aim of clarification is the old one: to establish the logical or grammatical order. Many of the problems relating to the Tractatus’s notion of logical analysis are then repeated at this local level. I also discuss similarities (and differences) between Baker and Hacker’s interpretation of the later Wittgenstein and the Tractatus in 3.3 and 6.2. See also 1.1, n. 10, and 6.2 and 6.5 for a discussion of the notion of a logical order implicit in language. This programmatic claim effectively constitutes a thesis about the nature of philosophy. For an explanation of why Wittgenstein’s later philosophy does not involve such a thesis, see 7.1. LFM 142. Ms163, 55r-v. The problem of dogmatism is given a more detailed characterization in Chapter 3. See the discussions of expressions and descriptions of inner states in PI §585–587., II ix, 187–189. and LWi §§27, 43–50. Neither is Wittgenstein claiming that for every sensation there is a natural expression that the verbal expression of the sensation replaces, contrary to Kripke’s interpretation of his view as a general theory of sensationtalk. See Kripke, Rules and Private Language, 102–104, n. 83. Rather than putting forward a thesis about what expressions of sensations must be that covers all such cases, Wittgenstein is articulating a possible way of conceiving them. (For this contrast, see Chapter 3 and 6.6.) Baker, Wittgenstein’s Method, 69. Ts220, §113; see Ms142, §131. PI §17; see also Ms141, 3. See the quotation from Ms107 in 1.4.

3. FROM METAPHYSICS AND PHILOSOPHICAL THESES TO GRAMMAR

1. Ms120, 136r, 136v. 2. Conceptual and grammatical investigation are terms Wittgenstein uses to refer to his investigations. See PI §90 and RPPi §§949, 950, quoted in 2.32 and 3.2. 3. Such “governing principles” might be sought, for instance, from the world (as in realism), the mind (as in idealism), or language (sometimes a form of “linguistic idealism”). How Wittgensteinian grammar relates to these isms is discussed in Chapter 5.

308

Notes to Pages 97–100

4. Below I simply assume that a theory or a doctrine consists of theses, though perhaps not exclusively of theses. 5. The difference between universality and generality is, briefly: since there might be only one x, a universal thesis about all x’s might not have any generality in the sense that it only applies to one object. (Generality in this sense is a contingent matter, whereas universality in the relevant sense is not.) In his discussions concerning the status of philosophical statements, Wittgenstein normally uses the German word “allgemein” (in its different forms), not distinguishing between universality and generality. In his English notes as well as in Rhees’s translation of the Investigations (checked by Wittgenstein), the word “universal/ity” is sometimes used in relevant contexts, but also “general/ity” (see esp. BB). I distinguish universality from generality when relevant in my discussion in order to enhance clarity. As explained in 1.5, universality may be understood here in a “weak,” historically contextualized sense, if one wishes. The problem Wittgenstein addresses with philosophical theses (as explained below) applies equally to such theses interpreted in a strong trans- or ahistorical sense and in a weaker historical sense. 6. The possibility of a state of affairs, for instance, is simply made plain in the concept-script by the possibility of formulating a statement describing this state of affairs. See TLP 3.02. 7. See TLP 4.12–4.124, 6.54 and discussion in 2.1. 8. TLP 4.122, 4.123. 9. TLP 4.126. 10. LR 71/ CL 124; see also 2.1. For a detailed discussion of this letter as an expression of the Tractatus’s view, see my “The Development of Wittgenstein’s Philosophy,” in Marie McGinn, ed., The Wittgenstein Handbook (Oxford University Press, forthcoming). 11. TLP 4.111, 4.112. 12. TLP, preface. See 1.1. 13. See TLP 4.5, 5.471. 14. Diamond provides a similar account of the early Wittgenstein’s commitment to a metaphysics of language, writing on the Tractatus: “What is metaphysical there is not the content of some belief but the laying down of a requirement, the requirement of logical analysis. . . . The metaphysics of the Tractatus—metaphysics which is not ironical and not cancelled— is in the requirements which are internal to the character of language as language, in there being a general form of sentence, in all sentences having this form” (Realistic Spirit, 19; see also “Criss-Cross Philosophy”). 15. As is well known, upon his return to philosophy in 1929, Wittgenstein attempted to mend his concept-script with respect to the so-called color

Notes to Pages 101–102

16.

17.

18.

19.

309

exclusion problem, which shows that inferential relations between propositions cannot be analyzed in the way the Tractatus assumes. (See Gordon Baker, Wittgenstein, Frege and the Vienna Circle [Oxford: Blackwell, 1988].) Hacker presents Wittgenstein’s attempt to fix the Tractatus’s conception of language as evidence that Wittgenstein did not treat it as nonsense and as supportive of the view that the Tractatus puts forward an ineffable doctrine (“Trying to Whistle,” 132, 134–136). It should be clear on the basis of 2.1, however, that this argument is not valid as it stands. There are good reasons for thinking that what Wittgenstein wanted to fix was not a doctrine but a notation to be used for the purpose of philosophical clarification. Hence, although one might well ask in the case of a “purely therapeutic” reading why Wittgenstein would have tried to fix something that was meant to be mere nonsense from the start, from the point of view of the interpretation developed in this book (the kind of resolute reading I wish to promote) there is nothing surprising in his attempt to address the color exclusion problem. See Kuusela, “Nonsense and Clarification.” The same problem of a relapse to theses also seems to arise in connection with the philosophical order of language that Baker and Hacker attribute to the later Wittgenstein, which was discussed in 2.32. As noted in 1.1, the question of the method and the question of the essence of language are not to be seen as separate problems but as different aspects of one problem. The question of the essence of propositions is, as it were, a particular (more detailed) formulation of the question regarding the correct method. This point might be expressed by saying that for Wittgenstein, the question of the essence of propositions is about developing a framework for the analysis of language. (Similarly, as McGinn observes, the question of the essence of propositions breaks into smaller problems that Wittgenstein conceives as aspects of this great problem. See McGinn, Elucidating the Tractatus, 15.) Questions relating to the translation of particular statements into the formulae of the concept-script always remain a “piecemeal” task in Diamond’s sense (see Diamond, “Criss-Cross Philosophy”). What is not decided piecemeal, however, is that translatability into Wittgenstein’s concept-script should be the means to determine whether something makes sense. This is a question about adopting or abandoning the Tractatus’s scheme of analysis as a universally applicable instrument of analysis. See TLP 6.53, 6.54. Given that from the point of view of the strictly correct method only factual statements emerge as having a sense, as

310

20.

21. 22.

23. 24. 25. 26. 27.

Notes to Pages 102–104

Wittgenstein suggests in 6.53, this method seems safely identified with the employment of Wittgenstein’s concept-script. As explained, his notation treats any sensible proposition as analyzable into elementary sentences that are contingent factual statements. For the distinction between establishing and applying the method of logical analysis, see also TLP 5.555–5.557. In accordance with what I said in Chapter 2, n. 42, both the strictly correct and the not strictly correct method might be characterized as methods of clarification and the critique of language. In 2.1, I used the expressions “initial task” and “task” of philosophy. The initial task is to be identified as the more fundamental one of the introduction of the method, including the solution of the problems of the essence of language. The notion of philosophical hierarchies is also discussed in 1.5, 4.1, and 6.1. See PI §109. Ms134, 153/ RPPi §949; see BB 18, 35. One might well ask why Wittgenstein’s statement about what is essential to metaphysics is not itself a metaphysical thesis about the essence of metaphysics. The short answer is that Wittgenstein’s remark is not meant to constitute a claim that the unclarity in question must always be considered as the essential characteristic of metaphysics. Thus he is not making a statement about the essence of metaphysics in an absolute sense. Rather, the characteristic in question is essential from the point of view of the particular philosophical issues he is dealing with. For Wittgenstein’s later conception of statements about essence, see also the contrast drawn between his and the Platonic conception of essence in 5.4. Ms148, 32r / PO 208, 209; see BB 59, 60. BB 55 and 57, respectively; see also BB 56 and AWL 22. See PI §251; Ms114, 122 / PG 129, 130. Ts220 §91/ Z §442; see also Ms157a 65r, 65v; Ms157b, 3v. Kripke has contested the idea that a priori statements are necessary. See his Naming and Necessity (Oxford: Blackwell, 1981), 54–56, 63. This does not affect Wittgenstein’s point, which is to be understood as concerning so-called necessary truths independently of whether or not a priori propositions are necessary. Apparently, Wittgenstein’s remark is originally meant as a diagnosis of the view held by many philosophers at his time, according to which a priori propositions are necessary, but the remark has wider relevance. (Simply substitute “necessary truth” for “a priori statement” in my discussion.) Incidentally, Kripke’s argument (that a priori propositions are not necessary) seems to involve an unclarity concerning the distinction, to be discussed below, between factual statements and statements of rule. Briefly, Kripke seems problematically to as-

Notes to Pages 104–109

28.

29. 30. 31. 32. 33. 34. 35. 36.

37. 38.

311

sume that a (token) sentence could play the roles of a factual statement and a rule simultaneously. Accordingly, my discussion could be taken to contain an implicit critique of Kripke’s argument. But I will not try to discuss Kripke explicitly any further. See Cora Diamond, “How Long Is the Standard Meter in Paris?” in T. G. McCarthy and S. C. Stidd, eds., Wittgenstein in America (Oxford: Clarendon Press, 2001), 104–139. Ms157b, 1r, 1v. In the context of an earlier version of the above remark from Ts220 / Z,Wittgenstein characterizes the locution “a priori” by saying: “The ‘a priori’ is a mode of presentation for a mode of presentation [eine Darstellungsform für eine Darstellungsform]” (Ms142, 86). This may be read as follows: The locution “a priori” gives expression to a particular way of characterizing the role of certain kinds of sentences. From its point of view, sentences that do not state contingent facts are to be seen as stating facts known to be true independently of experience. Clearly, Wittgenstein does not regard this as a happy characterization of the role of sentences that are not statements of contingent facts. According to him, it would be better not to conceive such statements as statements of fact at all. In this sense the locution “a priori” might be said to embody and epitomize what is problematic in metaphysics for Wittgenstein. It is an expression rooted deeply in philosophical language, but it might be better avoided or grasped metaphorically rather than literally. Ts220, §110; see also Ms142, §125; PI §104, 114. See 2.1 and 3.1. Ts220, §93/ Ms142, §105; modified from translation in Z §444. Ms142, 88; see also PI §103; Ms157a, 57v. Ts220, §92. Ts220, §93/ Ms142, §106 / Z §444; see PI §97. See also PI §89. To review: according to this more favorable interpretation, the Tractatus seeks to elucidate the distinction between sense and nonsense by examining examples of language use. This way, it seeks to introduce a program for philosophy as clarification and a notation to be used in clarification. But it does not put forward a “nonsensical theory” about a general criterion of meaningful language use. See AWL 108, quoted in 6.6. The statement “Every proposition says: This is how things stand” might be compared with the characterization of the concept of “game” through that of “board games.” Whereas such a definition is not universally valid for games, it may well be good enough for one’s purposes if one is merely concerned with features of games that are present in board games. See PI §3, quoted in 1.3.

312

Notes to Pages 109–113

39. Ms115, 56, 57; see also Ms142, §122; Ts220, §107. 40. The Tractatus’s conception (dis)solves, for example, the problem of negative propositions, discussed by philosophers since antiquity, and other problems. So it certainly has advantages that recommend its adoption. 41. See PI §89. For a discussion of this term, see Stephen Mulhall, Inheritance and Originality: Wittgenstein, Heidegger, Kierkegaard (Oxford: Clarendon Press, 2001), 87–92. and P. M. S. Hacker, “Turning the Examination Around: The Recantation of a Metaphysician,” in E. Ammereller and E. Fischer, eds., Wittgenstein at Work: Method in the Philosophical Investigations (London and New York: Routledge, 2004), 3–21. 42. See PI §101, 102; Ms157a, 53r, 53v; Ms183, 163, 164. 43. Ms142, 88; see also §107; PI §105. 44. See Ms142, §107; Ts220, §94; and Hacker, “Recantation” for a discussion of Wittgenstein’s motives for assuming the “ideal of crystalline logic.” Subchapters 3.4 and 3.5 further discuss the notion of the sublimation of logic and problems relating to it. 45. PI §108. This remark is discussed in more detail in 3.5. 46. The interpretation I am identifying as Baker and Hacker’s has since the late 1980s been discussed by Hacker in various single-author publications. Throughout this book I do not seek to distinguish between Baker and Hacker and Hacker alone but mostly talk about the interpretation of Baker and Hacker. 47. These problems may be characteristic to metaphysical philosophy but may also occur in other connections. (The projection of expectations, fears, prejudices, and so on onto their objects as if they were characteristics of those objects is far from uncommon.) On the other hand, perhaps not all thinking that calls itself metaphysical suffers from the above problems. This must be decided, in principle, case by case. 48. Hacker Insight and Illusion, 196; see also Baker and Hacker Commentary 2: 266, 267. 49. Hacker, Insight and Illusion, 151 and 152, respectively. 50. Ibid., 156; reference omitted; see also P. M. S. Hacker, “Philosophy,” in H.-J. Glock, ed., Wittgenstein: A Critical Reader (Oxford: Blackwell, 2001), 338. 51. See Hacker, Insight and Illusion, 151, 152, 180, 181. 52. Baker and Hacker, Commentary 1:516; see also Hacker, “Recantation.” 53. Ms140, 15r / PG 60; see also PI §496. 54. The remark just quoted and other aspects of Wittgenstein’s concept of grammar are discussed in more detail in Chapter 4. 55. The concept of a statement of a rule is also discussed in 5.4 with respect to the nontemporality of rules.

Notes to Pages 114–121

313

56. See for example Ms113, 22r-23v.; Ts211, 569–570; Ts213, 240–242r.; and Ms115, 57f-59 for discussions of the notions of rule and factual statement. 57. Whether or not the distinction between these two types of statements is sharp, as Hacker claims in the first quoted passage above, is another issue. Arguably this distinction is not sharp, if its sharpness means that one should always in principle be able to clarify which role—that of a conceptual or a factual statement—a sentence plays and that something important always depends on this. Consider, for example, the employments of the sentence “Humans are mortal.” But I will not discuss this issue any further here. See n. 143 below and 5.7 for discussion. 58. See PI §50. 59. See Ms113, 27v, 28v. See quotation below. 60. See also the quotation in 5.4 from Ms113, 29v / Ts212, 716 / Ts213, 246r, where Wittgenstein characterizes the difference between grammatical and factual statements about language. 61. Ms113, 28v / Ts211, 576 / Ts213, 245r; see also Ms115, 59. 62. Hacker, Insight and Illusion, 160, 161; see also Baker and Hacker, Commentary, 1:479, 480. 63. Hacker, Insight and Illusion, 161. 64. See PI §109; Ms152, 93–95; and discussion toward the end of 2.31. 65. See Hacker, Insight and Illusion, 152. 66. See also my discussion of Baker and Hacker on Wittgenstein’s notion of ordering in 2.32. 67. See PI §43 and Ms140, 15r / PG 60. 68. See Baker and Hacker, Commentary 1:250, 251; 2:36, 37. I return to Wittgenstein’s conception of meaning and Baker and Hacker’s interpretation in more detail in Chapter 4. 69. See PI §§128, 599. I return to the notion of agreement in 4.1, 6.3, and 6.4. 70. Hacker, Insight and Illusion, 161. 71. Z §§111, 114, 115, respectively; see also §121/ RPPi §556; Ms132, 198–199; Ts233a, 24–25. 72. I return to this issue in 7.3. 73. BB 17; see PI §5. Authors who date Wittgenstein’s turn to 1930–32 include Baker and Hacker, Commentary, 1:93–95, 451; Hilmy, The Later Wittgenstein, 86, 89–91; and Kienzler, Wittgensteins Wende, 13, 28, 29. See also Wolfgang Kienzler, “About the Dividing Line between Early and Late Wittgenstein,” in Gianluigi Olivieri, ed., From the Tractatus to the Tractatus and Other Essays. Wittgenstein Studien, vol. 2 (New York: Peter Lang, 2001). Diamond, “Criss-Cross Philosophy,” section 5 sug-

314

74. 75.

76. 77. 78. 79. 80. 81. 82. 83.

Notes to Pages 121–124

gests that in the Big Typescript (from 1932) Wittgenstein continued to “think in terms of Big Questions,” implying that a crucial shift took place only later. Pichler, Vom Buch zum Album, chap. 4.3 suggests that the methodological shift that constitutes Wittgenstein’s answer to the problem of dogmatism takes place only in 1936. A similar suggestion is made by Joachim Schulte in his “Wittgenstein’s Method,” in Rudolf Haller and Klaus Puhl, eds., Wittgenstein and the Future of Philosophy (Vienna: Öbv&hpt, 2001). But as I will argue, although an important reorientation does indeed take place around 1931, there is no need to choose either the early or the late 1930s as the date for Wittgenstein’s turn. The methodological shift in his philosophy is a process stretching through the 1930s, during which period Wittgenstein becomes gradually clearer about certain issues and finds ways of expressing himself that better match his aspirations. In this book I do not discuss the post-1936 changes in the presentation of his philosophy or the style of the Investigations, as discussed by Schulte and Pichler. However, I would argue that such changes are to be understood within the framework of what I describe as Wittgenstein’s turn. It also seems to me problematic to place too much emphasis on changes in “Wittgenstein’s style” (from about 1936 on), the danger being an overemphasis on his form of expression at the expense of attending to the use he makes of language. For instance, seen in their intended role, Wittgenstein’s dogmaticsounding statements about language and meaning from the early 1930s arguably lose their dogmatism (see Chapter 4). See also Ms131, 48, 49/ CV 55. See Ms157a, 46r–71r and Ms157b, 1r–17v. These parts of the notebooks might be described as a sketchbook for the relevant remarks in the Investigations and its 1936 manuscript, Ms142, as well as its 1937 typescript, Ts220. Ms183, 163–164 / PPO 97. Ms183, 162; see also PI §59. Ms183, 163, 164. Ms115, 56–57. Ms142 §122; Ts220 §107; Ts238 §141. This remark is repeated in various forms in the Nachlass; see Ms116, 220; Ms120, 43v, 44r; Ms157b, 13r. See Ms111, 118 and discussion in 1.4 about examining the genesis of a philosophical problem. Ms111, 119–120; also in Ms211, 72; Ms212, 745; Ms213, 259r; modified from translation in CV 21–22. Empiricism is often thought of as offering a way to avoid dogmatism. As is well known, however, empiricism gives rise to difficulties of its own,

Notes to Pages 124–130

84. 85. 86. 87. 88.

89. 90. 91. 92. 93. 94. 95.

96. 97. 98.

315

especially as regards explaining logical necessity. Wittgenstein’s way out of dogmatism provides an alternative not susceptible to these problems. Briefly, from his point of view the acknowledgment of something as logically necessary expresses an adherence to particular forms of expression or particular modes of presentation. (See 3.2 and below.) This does not commit Wittgenstein to any views about why one holds something as necessary, for example that this is simply a matter of convention. Wittgenstein’s view therefore does not imply conventionalism. See Chapter 5 for a detailed discussion of this issue. See quotations from Mss 111 and 115 above, Ms142 below. See also PI §131. Ms156a, 50r. See also Ms113, 45v. Ms157b, 15v / CV 31. Ms142 §122. Virtually the same formulation is given in PI §131, only the reference to Spengler is omitted and the last sentence occurs in brackets in a shortened form. See Ms113, 45v, 46r. Ts219, 22, 23. Ms111, 86; Ts211, 50 / Ts212, 1162 / Ts213, 420. See Baker, Wittgenstein’s Method, 98 for the interpretation of the term “emptiness.” See 3.2 and below. BB 46; see also Ms111, 185/ Ms 153a, 107r, 107v for the notion of emptiness. Apparently the problem of emptiness may take more than one form, and there seems to be no need to take the above as the interpretation or definition of what Wittgenstein means by this notion. Alternatively, the problem may arise when the definition of a philosophical notion—in response to the problem of injustice—is loosened in order to accommodate broader variety among the objects of description. For example, if the concept of picture is extended enough, anything will count as a picture of anything, the concept becoming quite empty. (See Ts213, 188v; Ms116, 123.) Ms142, §122 / Ts220, §107. See also Ts220, §115. It is not entirely correct to say that a comparison between philosophical statements and units of measurement is first presented in Ms115. In Ms113, 22v (Ts211, 569/ Ts212, 704 / Ts213, 240) a “logical comparison” is characterized as one making use of a grammatical rule whereby the determination of a grammatical rule is compared with the determination of

316

99. 100. 101.

102.

103. 104.

105.

106. 107. 108. 109. 110. 111. 112.

Notes to Pages 130–134

a unit of measurement. These remarks may be taken to lend further support to the present interpretation, but I will not discuss them here. Ms111, 118; Ms115, 55, 56; Ms142, §122; Ts220, §107; Ts238 /9, §142. Ms157a, 54r. Examples of such detailed examinations of his earlier views employing the new methods are to be found in his manuscripts from 1930–32 and have been discussed by Kienzler in Wittgensteins Wende. A similar account of the idea of Wittgenstein’s turn is given by Hilmy, The Later Wittgenstein, chap. 3, although he approaches the issue from a slightly different angle, as a question concerning the emergence of the method of hypothetical (constructed) language-games. However, there seems to be also a misleading aspect to Hilmy’s characterization of objects of comparison and models in philosophy as heuristic devices (for example, on 74). For being a mode of presentation, such a model is not merely a heuristic device, something “an expert” can dispense with. Although talk about heuristic devices captures well the idea that the models are not final and ultimate determinations of the nature of the objects of investigation, it fails to capture the indispensable role of the models as modes of presenting the objects of investigation. PI §108. The satisfaction of the real need is directly connected with the problem of dogmatism, i.e., the question of how to avoid being misled by pictures one creates or adopts, and so on. For dogmatism and falling prey to prejudices are instances of failing to reach clarity. In light of Wittgenstein’s conception of language as a form of life (see PI §23), one might also say that what is at stake is clarity about one’s life—i.e., what one’s life is like with the language and concepts one employs, and whether this life really fits the pictures one makes of it and how one wants to describe it. These latter aspects of the matter are discussed in 7.3. Baker and Hacker, Commentary 1:516; see also Robert Fogelin, “Wittgenstein’s Critique of Philosophy,” in H. Sluga, H. Stern, and D. G. Stern, eds., The Cambridge Companion to Wittgenstein (Cambridge: Cambridge University Press, 1996), 51. Baker and Hacker, Commentary 1:516. Ibid. See also Hacker, “Recantation.” TLP 5.5563. Ts220, §95; see Ms142, §102. Ms157a, 66v–67v. Ms157b, 5r.

Notes to Pages 135–138

317

113. Ms114, 103/ PG 115; also in Ms110, 206 (1931). 114. Ms152, 93, 94. 115. Both in Ms114 and Ms152, Wittgenstein’s reply to the worry about the collapse of philosophy is that philosophy does not dissolve because its task is to clarify particular philosophical problems, not to reveal to us as if for the first time what language really is. (Ms114, 103, 104 / PG 115; Ms152; see also Ms110 and the corresponding remarks in Mss211–213.) Thus, a crucial component of his reply is the move to particular problems discussed in 1.5 and Chapter 2. And indeed, the idea that the clarifications of what is essential or makes sense are relative to particular problems seems to successfully eliminate the need to postulate sublime universal essences. Rather, logic clarifies such aspects of the use of our words as are required for solving the actual philosophical problems at hand, and therefore it does not need to unearth a sublime essence containing the solution to every possible unclarity. Moreover, in this capacity—as the clarification of unclarities concerning something we are already familiar with—logic does not become an empirical investigation. However, at this point the question of how to understand logic’s aspiration to exactness and how to do justice to this aspiration (without relapsing into dogmatism) still remains unanswered. It is later explicated in the discussion of the turn in 1936–37. 116. One might say that the origin of the ideal of the crystal purity of logic lies in the logical investigation’s aspiration to exactness and clarity. In the Tractatus this ideal, which is descriptive of the goal of the logical investigation, gets projected onto language in the form of a metaphysical thesis about the essence of language. Thus logic’s aspiration to clarity is, as it were, reified in the form of a claim about its object of investigation. 117. Ms157b, 2v, 3r. 118. Ms115, 46 / Z §467. Now it should also be clearer what Wittgenstein means by the remark quoted in 2.32: “The investigation of language is a description and comparing of concepts, also with ad hoc constructed concepts” (Ms133, 9v). Such ad hoc concepts are concepts used as objects of comparison and designed to highlight particular aspects of the uses of language in response to particular philosophical problems. 119. This is particularly evident in Hacker, “Recantation.” 120. Hacker, Insight and Illusion, 162. 121. Ibid., 150; see also Baker and Hacker, Commentary 2:158. 122. See Mulhall, Inheritance, 44.

318

Notes to Pages 138–142

123. According to a widely accepted thesis about Wittgenstein’s development, he moved from a calculus view of language in the early 1930s to a view of language as language-games; see Hacker, Insight and Illusion, 132; Kenny, Wittgenstein, chap. 9.; Joachim Schulte, Wittgenstein: An Introduction (Albany: State University of New York Press, 1992), 84, 103, 104. Although it may be correct that Wittgenstein assumed a thesis of language as calculus in the Tractatus, on the basis of the discussion in this chapter it is clearly not correct that he held such a thesis in the early 1930s (for instance, at the time of Ms140 and Ms114 or after 1931). For the idea of the comparability of language with calculi is not a thesis about what language must be. Accordingly, it is misleading to say that Wittgenstein moved from the conception of language as a calculus to a view of it as “a motley of language-games” (as Hacker puts it), as if this were yet another thesis about language. Rather than a thesis, Wittgenstein’s conception of language as games is another way of conceiving the uses of language for the purposes of clarification. It does not exclude the conception of language as a calculus (see 6.6). Hilmy, in The Later Wittgenstein, chap. 4 argues against the above developmental thesis on the basis of a detailed study of remarks from Wittgenstein’s Nachlass. 124. BB 25, 26. 125. Observe also the following remark from a dictation to Moritz Schlick: “We specifically do not claim that language is a game played according to rules, (for in that case we would assert an untruth) but we compare the phenomena of language with such a game, and the former is more or less similar to the latter. Our investigation of language is one sided” (Ms302, 14). 126. BB 27; see also BB 28, PI §125. 127. See AWL 48; Ms140, 24 / PG 68; Ts220, §99. 128. BB 27. 129. Ms111, 87/ Ts211, 51; see also Ts213, 249v–251. 130. Hacker, Insight and Illusion, 163. 131. Ibid., 150–151. 132. Ms157b, 15v, 16r. 133. See the quotation from Ms115 in 3.4. 134. See Wittgenstein’s remark about the goodness of units of measurement and dogmatism in the aforementioned passage from Ms115 and corresponding remarks in Ms142 and Ts220. 135. PI §81. See also discussions in the context of earlier versions of this remark in Ms112, 94v; Ts212, 728; and Ts213, 252v, 253r that facilitate its interpretation in important ways.

Notes to Pages 142–146

319

136. Ms140, 33/ PG 77. 137. “Wir betrachten die Sprache unter dem Gesichtspunkt des Spieles nach festen Regeln.” 138. Ms113, 24r / Ts211, 571; see Ms111, 67; Ms112, 94v; Ms140, 18, 24; Ts211, 40. 139. Ms140, 32 / PG 76. 140. Ms114, 107–108 / PG 120. 141. The color patch comparison occurs also in Ts213, 68v, where drawing sharp pictures of blurred reality is said to be the task of logic. In Ms113, 45v, 46r and Ts211, 594 the comparison is used to characterize the role of simple language-games. Wittgenstein seems to use this comparison for the first time for this purpose in Ms111, 83, where he raises the question of presenting vague language use in a calculus. This, he says, is similar to talking about contours of colors in the case of a picture consisting of blurred color patches. The remark in question may again be taken to support the dating of the beginning of Wittgenstein’s methodological shift to 1931. A slightly different but related characterization is given in the 1934–35 lectures: “Suppose I draw a curve, and ten oscillating circles which come near to describing the curve. This is the way I can describe the use of the word ‘can.’ I shall give you a number of usages regulated by rules which will oscillate the actual use. There is no exact description” (AWL 95). 142. PI §77. 143. The distinction between statements of a rule and factual statements, as I presented it in 3.3, may be taken as an example of a simplifying distinction to be used as an object of comparison. In reality the relation between rules and factual statements is not always as clear as presented there, my example of a rule being a stipulative rule, which is a special case. For complexities regarding the distinction between rules and factual statements, see for instance OC §§318–321 and Chapter 5. For the purposes of the current discussion, however, the complexities concerning the relation between statements of a rule and factual statements are not directly relevant. (Similarly, as noted, my definition of metaphysical philosophy in 3.1 is intended to be used as an object of comparison.) 144. PI §156. 145. PI §162. In Ms152, 14–16 the color patch comparison is used in connection with the concept of reading. 146. See the discussion of the notion of the completeness of analysis in 2.21. 147. Ms183, 163, 164; quoted in 3.4. 148. Ts220, §110, quoted at more length in 3.3.

320

Notes to Pages 146–148

149. TLP 5.5563, quoted at more length in 2.32. 150. As explained in 2.32 (see also 3.3), Baker and Hacker seem to give in to this temptation when they postulate the grammatical order of language that philosophy allegedly aims to reveal. This is also apparent in their interpretation of the concept of perspicuous presentation, discussed in 6.2. 151. Recall also Wittgenstein’s metaphor of grammar as the account books of language, i.e., a record of linguistic transactions (Ms114, 58 / PG 87; see Ts213, 58, 526; Ms116, 55). I return to these issues at length in Chapter 6. 152. See 3.3. 153. See 2.32 and 6.5 for discussion of this point. 154. See Ms174, 32v / OC §149. This sort of view of Wittgenstein’s conception of language is propounded most famously by Cavell. See Cavell, Must We Mean, 52. Presumably as a result of his rejection of the conception of language use as based on rules, however, Cavell is much more dismissive of the picture of language as a calculus than I would be on the basis of my account of Wittgenstein’s method. Cavell writes: “Whether the later Wittgenstein describes language as being roughly like a calculus with fixed rules working in that way is not a question which can seriously be discussed” (Must We Mean, 48). Seriousness here, I would say, depends entirely on the point of the comparison between language and calculi. Similarly, in his own work Cavell does not seem particularly fond of rigorous or strict definitions. However, on the basis of Wittgenstein’s methodological considerations, this is to be seen as constituting an available possibility also for philosophers who do not assume language use to be determined by rules in any straightforward sense. See Chapter 4 for further discussion. 155. For a discussion of this issue, see Thomas Wallgren, Transformative Philosophy: Socrates, Wittgenstein, and the Democratic Spirit of Philosophy (New York: Lexington Books, 2006), 388–391. 156. See Chapter 5 for further discussion of this issue. 157. PI §242; see also Ms129, 128, 129. 158. Wittgenstein characterizes the agreement presupposed in communication (and in many other tasks that language serves) as an agreement not on opinions but on a form of life, i.e., not an agreement on what is true or false but on concepts. See PI §241 and Ms164, 118–121, 151. 159. See in particular 4.3–4.5 for how Wittgenstein’s conception of language accommodates the sense in which the functioning of language is based on facts about its users and their environment in addition to the sense in which it is conventional. The concept of exceptionless

Notes to Pages 149–152

321

necessity and its relation to the factual is discussed in detail in Chapter 5. See also 6.1 and 6.5 for Wittgenstein’s views on the relation of the logical/grammatical to the factual. 4. GRAMMAR, MEANING, AND LANGUAGE

1. As quotations in the Introduction show, there is widespread doubt about whether this can be done. 2. These manuscripts date from 1932–34. Ms140 is a revision of Ts213 and pages 1–56 of Ms114. Most of the remarks discussed here also occur in Ms116, from 1937/38. 3. Given that each sentence in the following passage forms a separate paragraph, I will call them “remarks.” 4. Ms140, 15r / PG 59–60; see also Ms114, 43–45, 179–180; Ms116, 30, 32–33. 5. PI §560. 6. BB 1. 7. See also Ms114, 44, 45; Ms116, 32, 33. 8. Ms140, 24 / PG 68; see also Ms111, 111. 9. See Ms145, 62; BB 1. 10. From this point of view one might talk about switching the places of two words in a language, and thereby their meanings—for example, “yes” and “no.” See Ms114, 40 / PG 59; Ms116, 29. 11. Ms140, 20, 23/ PG 63; Ms114, 156 / PG 184. But as Wittgenstein observes, not all rules for the use of language determine meanings: “One can say: the explanation of meaning gives the meaning & the explanation is a rule for the use of the word. But we do not say of every rule that it determines the meaning, only of some of them” (Ms147, 4v). 12. One might also adopt a broader notion of use (still consistent with Wittgenstein’s method as described in this book). In this case the rulegovernedness of use would not be taken as a given, but the degree of the rule-governedness of use would be seen as something variable, artificial calculi representing one end of the scale of variation and instrumental uses that do not depend on rules (as described in 4.3) representing the other end. Now the degree of rule-governedness would essentially constitute an adjustable parameter of models employed for the description of language. (This possibility should be more understandable at the end of this chapter.) Below I will nevertheless assume the rule-governedness of use. For one important goal of my discussion is to illustrate Wittgenstein’s method of making a nondogmatic and easily agreed-upon use of strict definitions of complex concepts such as meaning.

322

Notes to Pages 152–153

13. See Ms114, 43–45; Ms116, 3, 4. Hence Wittgenstein’s conception stands in contrast with mentalism, for instance, according to which understanding the meaning of a sign is a matter of giving it an interpretation in the mind, an appropriate mental state allegedly constituting the correct interpretation. From this point of view clarification would ultimately be an investigation of mental states and processes that give meanings to signs, i.e., an investigation of the causes and effects of signs vis-à-vis our mental states. (See Ms140, 15r) Similarly, if one took seriously a dispositional account of rule-following of the kind that Kripke considers in his Rules and Private Language (see Saul Kripke, Wittgenstein on Rules and Private Language [Oxford: Blackwell, 1982], 73, 74.)) and tried to clarify the meanings of words in terms of people’s dispositions, this would not qualify as grammatical clarification. (Note also that to maintain in Wittgenstein’s sense that clarification is carried out in language is, naturally, not to deny that in language we talk about things external to it. It means that whatever language is used to talk about can be clarified through an investigation of language. See 5.3 for further discussion of this point in connection with Wittgenstein’s conception of essence.) 14. Ms110, 160, 189; Ms114, 2, 15, 27, 83; Ms116, 2; Ms140, 8. Here Wittgenstein can be seen as rejecting the idea of philosophical hierarchies. This is (roughly) to reject the idea that there would be a separate task of establishing a foundation for the investigation besides the clarification of particular unclarities, corresponding to the Tractatus’s division between the tasks of establishing a framework for logical analysis and undertaking logical analyses of particular statements. (See 3.1.) I return to the concept of metalogic in 6.1, in particular to the question of whether the concept of rules (or rule-following) should be seen as providing a foundation for Wittgensteinian grammatical investigation. 15. PI §43. 16. Baker and Hacker, Commentary 1:250, 251. 17. Baker and Hacker, Commentary 2:36–37, italics in the original; see also P. M. S. Hacker, Wittgenstein: Mind and Will. Vol. 4 of An Analytical Commentary on the Philosophical Investigations (Oxford: Blackwell, 1996), 214. 18. Hans-Johann Glock, A Wittgenstein Dictionary (Oxford: Blackwell, 1996), 376; reference omitted. The quotations from both Glock and Baker and Hacker come from contexts where the view of meaning as use is contrasted with the Tractatus’s view of meaning as an object that a name stands for. Thus although both Glock and Baker and Hacker talk about signs and expressions rather than words in these passages, they might not wish to attribute to Wittgenstein a claim about the meaningfulness

Notes to Pages 153–156

19. 20.

21.

22. 23. 24.

25.

323

of signs or expressions generally. But clearly, they do attribute to him a view of what word-meaning must be. According to this interpretation, Wittgenstein offers, essentially, an alternative account to that of the Tractatus, switching from one universal account of word-meaning to another in his later philosophy. See Hacker, Insight and Illusion, 206. As noted in Chapter 3, Hilmy interprets the idea of Wittgenstein’s turn similarly to how I am suggesting. Nevertheless, after his discussion of the turn, when Hilmy moves on to discuss Wittgenstein’s conception of language, he too seems to attribute to Wittgenstein a thesis of grammar as a necessary condition of meaning. According to Hilmy’s Wittgenstein: “A sign exists, is meaningful qua sign, only as embedded in a ‘grammar’ qua system or locus of linguistic practice” (Hilmy, The Later Wittgenstein, 179; see also 161, 163, 180). Thus Hilmy appears to fail to keep up consistently the perspective of the turn and to fall back into theses about language. Similarly, Kripke assumes that Wittgenstein’s discussion of the notion of a rule and rule-following concerns the problem of explaining the possibility of all meaningful language use. See Rules and Private Language, 62. See Baker and Hacker, Commentary 1:490, 491; Hacker, Commentary 4:244, 245; Hacker, “Philosophy,” 333, 335. At best deviation from grammar leads to irrelevancy, according to this view. Alternative rules determine alternative concepts, but by spelling out such alternatives one passes by the original problems that arise in the context of the original concepts. In a sense this account is certainly ingenious: it ascertains that any objection to the Wittgensteinian philosopher’s statement of a grammatical rule fails to constitute a contribution to philosophical discourse, being either nonsense or irrelevant! Nevertheless, as I argue, Wittgenstein has little to do with this view of grammar as conversationstopper. I discuss the issue of whether grammatical statements are something one can be required to agree upon in 6.3 and 6.4. Ms140, 33/ PG 77. See PI §132 and VW 279. In actuality, of course, whether one is dogmatic in one’s philosophizing will be a matter of one’s philosophical practice. Wittgenstein offers a strategy for dealing with the problem of dogmatism, but to put this strategy into use is up to the individual philosopher. See 6.2 for a discussion of the notion of perspicuous presentation and dogmatism. Given that, according to Wittgenstein, philosophy only states what everyone admits, he would apparently maintain the uncontroversiality of his conception of meaning as well. (See PI §§128, 599.) Accordingly,

324

Notes to Pages 157–160

the goal of this chapter might be described as that of illustrating how the method described in Chapter 3 can account for this uncontroversiality. Wittgenstein’s concept of agreement on philosophical statements is the topic of 6.4. 26. See Cavell, Must We Mean, 48, 52. 27. BB 69. 28. BB 69. If I label an object, this might mean many different things, for instance that the object is a chair, but also that it is a chair of a particular style. Or perhaps the label marks its size or color, reserves it for someone, indicates that it should be painted, considered for next year’s auction, and so on. 29. See PI §§40–45. Insofar as the meaning of a name is its use, it does not matter whether the bearer of the name exists or not. A name can have a use in the absence of a bearer. 30. AWL 44. See also 6.5 for discussion of the problem with referentialism. The later Baker has proposed a methodological maxim of “minimalism” in the interpretation of Wittgenstein, according to which “given our strong cravings for generality and the inclination to extract generalizations from [Wittgenstein’s] investigations of specific concepts or language-games, we should try to relate every remark to the specific arguments which support it and to the particular purpose which informs the surrounding remarks, and we should attach to each expression the interpretation which gives it the minimum generality compatible with the context” (Wittgenstein’s Method, 67). I subscribe to this (or in any case, something close to this) and would express the point by saying that one cannot assume that Wittgenstein is putting forward universal theses, and that to extract from him or attribute to him such statements requires a separate justification. 31. Instead of focusing on how Wittgenstein expresses himself, commentators who attribute to Wittgenstein a thesis of the essence of meaning typically formulate his points in a much more direct way than he does. Such formulations might be understood as an attempt to rescue Wittgenstein’s thought from his odd ways of expressing himself. However, the result may also be that Wittgenstein’s thought is lost. 32. AWL 48. 33. Ibid. 34. Ibid. 35. See 4.4. 36. See Glock, Dictionary, 153. 37. See AWL 46–48. 38. Ibid., 46, 47.

Notes to Pages 160–163

325

39. Ibid., 48. 40. Note the parallels between the ways Wittgenstein explains the use of exact rules and clarificatory pictures as objects of comparison in his lectures and in the quotation from Ms140, 33 above. In the lectures it is the “fluid use” of “understanding” and “meaning,” and in Ms140 the “fluctuating” “actual use of a word,” that exact rules or simple structures are set beside in order to achieve a clearer comprehension of the former. 41. Ms147, 4r; quotation marks added around “meaning.” 42. See also Ms147, 11r. 43. PI §49; Katherine J. Morris, “The ‘Context Principle’ in the Later Wittgenstein,” The Philosophical Quarterly 44 (1994): 294–310, 295, 296, 307. 44. Morris, “The ‘Context Principle,’ ” 295. 45. Ms114, 58 / PG 87. 46. Ms140, 25/ PG 69. 47. Ms140, 15r / PG 60; see also Ms114, 45. 48. Ms140, 15r / PG 60. 49. See Ms140, 17, 18. 50. Baker has also drawn attention, in his later work, to the kind of qualifications discussed above and suggests, on this basis, an interpretation of Wittgenstein’s conception of meaning similar to the one argued for here (Wittgenstein’s Method, 269–271). There seems to be an important difference between Baker’s suggestion and my interpretation, however. According to Baker, Wittgenstein draws (and we would do well to draw) a distinction between conceptions and pictures on the one hand and descriptions of language on the other. Baker then uses this distinction to make the point that Wittgenstein’s conception of meaning, for instance, is not intended as the correct description of the ordinary use and to contrast Wittgenstein’s philosophical practice with other philosophers’ descriptions of language, such as Carnap’s and Ryle’s (Ibid., 52, 67ff., 273ff.). Clearly, I agree with Baker on many points here. On the basis of my interpretation of Wittgenstein’s turn in Chapter 3, however, I would not draw a distinction between the articulation of pictures or conceptions and description. It seems more appropriate to say that Wittgenstein has articulated a particular conception of the description of language with rules as objects of comparison, or more generally pictures and conceptions as objects of comparison. Certainly, the concept of description seems flexible enough to accommodate Wittgenstein’s conception of description. It is also notable that Wittgenstein does in fact talk about, for instance, Augustine’s picture or conception of language (PI §§1, 4) as a description (§3).

326

Notes to Pages 164–170

51. Ms140, 24 / PG 68. Recall also the quotation from Schlick in 3.5, n. 125. 52. Other examples a few paragraphs earlier include the expressions “oh,” “hurrah,” and “hm.” Consider also the remark: “Are there explanations for every word which could be called grammatical explanations of meanings? What if the purpose of the word is to transfer a particular mood in the other person? Let us imagine that we said this of the word, then I would have to say: this is not what I call explanation of meaning, it does not have anything to do with an explanation of meaning” (Ms156a, 35r, 35v). 53. Ms140, 25/ PG 69. 54. See Ms140, 25. 55. See Ms145, 62. 56. PI §492. 57. PI §497. For a discussion of the arbitrariness of grammar, see Ms116, 133–135, where this remark is first drafted. See also Ms113, 33v–34r and Ms117, 139–141; I return to the arbitrariness of grammar briefly in 4.4, at more length in 5.1, and again in 6.3. 58. As becomes clearer below, this is also how Wittgenstein understands the status of his characterization of language as a rule-governed practice. 59. See 6.6 for a discussion of what I call “multidimensional descriptions” made possible by Wittgenstein’s method. 60. Ms110, 199–200 (1930–31); also in Ts213, 193. 61. See PI §497 (quoted in 4.3) and references given there. These issues are also discussed in Ms114, 172, 173/ PG 192, 193; see also Ms114, 156–158 / PG 184, 185; Ms156b, 10r. 62. Ms114, 172, 173. Ms114 is a revision of parts of Ts213, the so-called Big Typescript. 63. Hacker, Commentary, 4:253; see also 252. 64. PI §498; drafted in Ms114, 166 / PG 189. See also Ts209, 8 / PR 64 for an earlier discussion in the context of which Wittgenstein presents this point as a critique of Russell. 65. A problem with the thesis that the essence of language lies in its rulegovernedness is that it makes it purely accidental that language should have any utility. See 6.6. 66. See 4.5 for further discussion of the role of examples. 67. See Hacker, Commentary 4:253–256. 68. Similarly, one might remark about Glock’s interpretation that although one may well say, as he does (Dictionary, 151), that Wittgenstein has “a normativist picture of language,” one must not project this picture onto language as a requirement it must meet. This, however, is just what Glock seems to do when he claims that rule-governedness is necessary for word-meaning (see 4.1).

Notes to Pages 171–177

69. 70. 71. 72.

73. 74. 75. 76. 77. 78.

79. 80.

81. 82. 83.

84. 85. 86.

327

For further discussion of the notion of an essential feature, see Chapter 5. See Ms111, 119/ CV 21–22 (quoted in 3.4). See PI §§65–67. Notably, therefore, Wittgenstein does not only differentiate the idea of family resemblance from definitions by necessary and sufficient conditions. The notion of a family-resemblance concept stands in contrast more generally to the idea that there must be some shared feature or features that define a concept. Clearly, such a shared feature could also be a necessary condition. Ms140, 31–32 / PG 75. Ms152, 16–17; see Ms115, 220; Ms150, 6. PI §66. See PI §132 and the discussion of the notion of arranging in 2.32. Ms140, 32, 33/ PG 76, 77; see discussion in Chapter 3. As I explain in 6.6, it does not follow from the possibility of using various definitions to characterize a particular concept that any definition is as good as any other. It is simply not the case that all characterizations of a concept are equal when it comes to resolving philosophical problems relating to it. See PI §244; Z §545; Ms131, 127. This aspect of Wittgenstein’s view of language is discussed in Norman Malcolm, Wittgensteinian Themes: Essays 1978–1989, ed. G. H. von Wright (Ithaca: Cornell University Press, 1995), chap. 6; and Nothing Is Hidden, chap. 8. Malcolm does not address the question of why this view of language is not a thesis about language. The discussions in Chapter 3 and here, however, provide the outline for an answer. The conception of language as an extension of primitive behavior, and the examples Wittgenstein uses to introduce it, are to be comprehended as objects of comparison. Notably, in the context of relevant remarks, Wittgenstein expresses reservations with regard to the status of his conception similar to those discussed in 4.2. Ms147, 1r. AWL 48. Similarly, one might sometimes want to talk about the relevance of the shape of words to meaning, for instance, in the case of specific fonts and layouts in literature. For instance in the context of his remarks on “boo,” Ms140, 25/ PG 69. PI §531. It seems quite generally true of poetry that the role of a word in the poem cannot be reduced to linguistic conventions. This is evident, for instance, in the fact that poems cannot usually be translated simply by

328

87. 88. 89. 90. 91. 92. 93. 94. 95.

96. 97. 98.

Notes to Pages 177–185

looking up the relevant words from a dictionary, i.e., by choosing words according to rules of translation provided in dictionaries. Rather, translating a poem is often a matter of rewriting it in another language. PI §§528 and 529, respectively. Ms180b, 3v. See also Ms129, 173–175. See PI §§543, 544. Ms180, 4r. §530. As examples, consider the terms Endlösung and Lebensraum (final solution, living space), which have become, as it were, icons of Nazi ideology. Ms109, 109; Ts211, 388. Ms141, 3; see also BB 84, 85. Interpreting PI §43 as a thesis about the nature of meaning, Garver is led into a problem about Wittgenstein’s conception of meaning analogous to Hacker’s problem about the concept of language, as discussed in 4.4 (Garver, This Complicated Form of Life, chap. 12). The pattern is the same in both cases. First Wittgenstein is attributed a universal thesis about x, whereupon the problem arises how the different things that he says about x can be accommodated consistently into the proposed interpretative framework. The solution to Garver’s problem is the same as to Hacker’s: we are not to attribute to Wittgenstein a thesis of the essence of meaning in the first place. That is, the interpretative framework of Wittgenstein’s remarks on particular philosophical topics needs to be changed so that it agrees with his methodology. This chapter is put forward as an example of how this can be achieved. Hacker, Insight and Illusion, 179. Ms114, 108 / PG 120–121; the deleted sentence does not occur in PG. See Ms111, 70 for a very interesting formulation of the same point. See further discussion of these issues in 6.3 and 6.4 and Oskari Kuusela, “Transcendental Arguments and the Problem of Dogmatism,” International Journal of Philosophical Studies, forthcoming. 5. THE CONCEPTS OF ESSENCE AND NECESSITY

1. See Ts213, 420 / PO 181, quoted in 6.4, where the goal of justice is discussed in more detail. See also other references given in 5.2. 2. PI §§371–373. Remarks 371 and 373 are first drafted in Ms116 in 1945 (pp. 340 and 341, respectively). Remark 372 dates from a much earlier time; it was first drafted without quotation marks in Ms110, 115 in 1931. The quotation marks and the word “consider” are added in Ts213 (p. 235r; see Ms114, 157 and Ms115, 55).

Notes to Pages 186–188

329

3. Ms114, 157, 159–161; Ts228, 108; PG 184–186. See 6.3 for further discussion of the arbitrariness of grammar and Wittgenstein’s argument for the conception of its arbitrariness. 4. Baker and Hacker, Commentary 2:331. 5. Hacker, Insight and Illusion, 195. 6. Ibid., 199; see also 187. See Baker and Hacker, Commentary 2:286. 7. Michael N. Forster, Wittgenstein on the Arbitrariness of Grammar (Princeton: Princeton University Press, 2004), 15–17; Garver, This Complicated Form of Life, chap. 14. 8. Hacker, Insight and Illusion, 186; see also Hacker, “Recantation.” See my discussion of the metaphysical projection in 3.2. 9. Hacker, Insight and Illusion, 206. Although Hacker (see esp. “Recantation”) denies that grammar is meant to reveal an a priori structure of the world, the idea that the criteria for the identity of things are decided by grammar nevertheless appears to imply that empirical reality is a linguistic construction. For if the criteria for the identity of things depend on grammar, then so do possible (though not actual) states of affairs. Interestingly, Anscombe describes a position she calls “linguistic idealism” in exactly the terms Baker and Hacker use to characterize their interpretation: “what I mean by ‘linguistic idealism’ would . . . say ‘Essence is created by grammar.’” G. E. M. Anscombe, “The Question of Linguistic Idealism,” in From Parmenides to Wittgenstein: Collected Philosophical Papers Volume I (Oxford: Basil Blackwell, 1981), 112. For discussions of Wittgenstein and idealism, see also Bernard Williams, “Wittgenstein and Idealism,” in Moral Luck: Philosophical Papers 1973–1980 (Cambridge: Cambridge University Press, 1983); Norman Malcolm, “Wittgenstein and Idealism,” in Wittgensteinian Themes; and David Bloor, “The Question of Linguistic Idealism Revisited,” in H. Sluga, H. Stern, and D. G. Stern, eds., The Cambridge Companion to Wittgenstein (Cambridge: Cambridge University Press, 1996). 10. Baker and Hacker, Commentary 2:330. 11. Ms137, 6a, 6b / Z §358. 12. Ms137, 5b. 13. See PI §§19, 23, 520. 14. Ms137, 65b / Z §374. 15. Baker and Hacker, Commentary 1:347. See also Hacker, Insight and Illusion, 194. Stroud and Glock also describe Wittgenstein as a conventionalist. See Barry Stroud, Understanding Human Knowledge: Philosophical Essays (Oxford: Oxford University Press, 2000), 14, 15; Glock, Dictionary, 131ff. Baker comes to question the usefulness of this characterization

330

16. 17.

18. 19.

20. 21. 22. 23. 24. 25. 26.

27.

28.

29. 30. 31. 32.

33. 34.

Notes to Pages 188–196

in his later work. See Baker, Wittgenstein, Frege and the Vienna Circle, chap. 7. See Baker and Hacker, Commentary 2:269. When the interpretation that the conception of the arbitrariness of grammar entails a thesis about the creation of essences by grammar is rejected, the significance of this conception can be seen in that it guarantees the autonomy of Wittgensteinian grammatical investigation. Given the arbitrariness of grammar, questions about concepts cannot be reduced to factual questions. Ms137, 5b / Ts232, 712–713/ Z §§354, 355. See also the remark: “ ‘There is no such thing as a reddish-green’ is akin to the sentences that we use as axioms in mathematics” (Ms133, 25r / Ts245, 244 / Z §346). That “it” in the remark refers to the word “physical” is evident in the German original. Ms117, 96 / Ms119, 82 / Ts221, 203/ Ts222, 52 / RFM I §74. See 3.1 for the concept of a philosophical thesis. Ms137, 9b; see PI §130. BB 18. See Ts211, 50; Ts213, 414, 420 (quoted in 6.4); see also PI §128. PI §370. Like §371 and §373, but unlike §372, this remark is first drafted in 1945 in the same notebook (Ms116), about twenty pages earlier than §371 and §373. Section 370 is just one remark of several in the Investigations that contrast an introspectionist approach to the investigation of concepts and essences with a grammatical one. See PI §§314, 316, 327, 547, 551. This does not mean that there must be something that holds by necessity of all cases falling under the concept of imagination, i.e., that there is something common to all such cases. As explained, Wittgenstein rejects this assumption about the unity of concepts (appropriately termed “essentialism”), proposing that the unity of concepts may also be understood in terms of family resemblance. See Ms140, 31–32 / PG 75 (quoted in 4.5). Ms132, 195/ Ts245, 233/ RPPi §551; see also §§548–550; PI §120. Ms132, 195/ Ts229, 335/ RPPi §550. See Ms117, 37, 38; Ms118, 27r; Ms138, 8a; Ms164, 4; OC §57. However, as Wittgenstein notes, a rule is a concept with blurred contours. See Ms113, 29v / Ts212, 716 / Ts213, 246r. I return to this bluriness in 5.7. Ms113, 29v / Ts212, 716 / Ts213, 246r. Ms118, 23v / Ts221, 161/ RFM I §§103, 104.

Notes to Pages 196–203

35. 36. 37. 38. 39.

40. 41. 42. 43. 44. 45.

46. 47. 48. 49.

331

See Ms117, 24. Ms117, 25/ Ms118, 18r / Ts221, 156, 157/ RFM I §102. Ms115, 259. Ms115, 259, 260. According to Baker and Hacker, “contrary to the opinion that has held sway from Plato to the present, arithmetic and geometry (and formal logic) are no closer to philosophy than astronomy or physiology” (Commentary 2:266). On the basis of my discussion, this seems misleading. The nontemporality of grammatical, geometrical, and mathematical statements does reveal something common to philosophy, geometry, and mathematics, even if this characteristic of grammatical and mathematical statements may have been misleadingly conceived by philosophers and mathematicians alike, due to their tendency to conceive necessary statements as true factual statements. See Ms117, 34; Ms137, 5b. See Ms137, 6a / Z §357. See Ms137, 5b. Ms126, 18. PI §192; see Ms119, 34, 35, 36; RFM I §124. The skeptical problem attributed to Wittgenstein by Kripke, according to which there is no fact that determines and justifies one’s following a rule one way rather than another, and Kripke’s “sceptical solution” to this problem in terms of rule-followers’ communal agreement about how to follow a rule, presuppose that the necessity expressed by a rule should be explainable (and justifiable) by reference to facts. (See Rules and Private Language, 21, 60, 66, 70–71, 92, 97). Thus one may say that Kripke’s whole discussion of the skeptical problem and its solution (in terms of communal agreement) is founded on a nonrecognition of the very distinction (between factual statements and statements of a rule) that Wittgenstein is concerned to clarify. This means that Kripke’s efforts are devoted to spelling out a positive solution to a problem that Wittgenstein regards as a confusion to be dissolved, and the discussion of which he arguably intends only as an illustration of the muddles that are created by ignoring the distinction between facts and rules. See also n. 50 below. See Ms119, 26–29. / RFM 84–85; Ms126, 18. Ms119, 43–44 / Ts221, 233/ RFM I §§127–128. I return to this remark (and the word “always”) in 5.7. Ms117, 34–35/ Ms118, 24v / Ts221, 162 / Ts222, 77/ RFM, 75–76. Peter Hacker, Wittgenstein’s Place in Twentieth-Century Analytical Philosophy (Oxford: Blackwell, 1996), 199.

332

Notes to Pages 203–206

50. This is not the only way one can interpret this quotation from Hacker. I outline a second possibility in 5.6. Kripke’s account of rule-following in terms of communal agreement (see Rules and Private Language, 92, 97) may be seen as a prime example of the relevant kind of confusion about necessity. See also Bloor, “Linguistic Idealism,” for an account of Wittgenstein’s conception of normativity in terms of facts about linguistic practices. Despite his very different take on Wittgenstein, Cavell too characterizes Wittgenstein’s conception of necessity in similar terms: “His philosophy provides, one might say, an anthropological, or even anthropomorphic, view of necessity. . . .” (The Claim of Reason, 118), it being characteristic of this conception that the a priori is here seen as something historical (119). What I find problematic in this—and the same or a very similar assumption seems to inform the accounts of Hacker, Kripke, and Bloor—is the idea that the fact that the concept of necessity is a (human) concept, i.e., something that comes into existence in the context of certain human practices, should be part of the explanation (clarification) of the concept, as if it were part of the meaning of the word “necessity” that it is employed by humans. That this concept is employed by humans is a fact, but this fact does not enter into the definition or characterization of the concept. If it did, it should apparently be part of the explanation of all our concepts, but it is hard to see how this would help to understand the function of any particular concept in contrast to another one. 51. This also means that the expression “set up” cannot be used to ascribe to Wittgenstein the thesis that necessities are set up in or created by language. 52. Recall, however, my discussion and rejection of this kind of interpretation of Wittgenstein’s conception of language in chap. 4. 53. PI §§563–564 and 567–568, respectively. See Ms115, 69–71; Ms147, 13r14v; Ts221, 264 / RFM, Appendix I, §§19–24 for sources and other occurrences of these four remarks. 54. This remark occurs in the context of Wittgenstein’s remarks on the issue of whether there is a single, ultimate end to logical analysis and was quoted in 2.22 in connection with my discussion of this matter. Given that the purpose of logical analysis, as explained, is to reveal the essential logical features of language, there is a close connection between the remarks in these two contexts in the Investigations. 55. PI §62. 56. Hacker seems to adopt this problematic view, maintaining that claims about what is essential may be supported by referring to purposes (and so on). See Hacker, Commentary, 4:387–388.

Notes to Pages 207–217

57. 58. 59. 60.

61.

62. 63. 64. 65. 66. 67. 68.

69.

333

See PI §§556, 562. See Ms137, 116b / Ms169, 2r; see also Ms138, 14b; Ms115, 59–64. Ms117, 140 / Ms160, 4v. Correctly understood, as long as the statement about the exclusion of reddish-green is not thought of as justifiable by reference to reality in the manner of a factual statement, this point about grammar reflecting general facts (regularities) of nature does not conflict with the idea of the arbitrariness of grammar. See Ms174, 21v, 22r / OC §§96–99. Did English-speaking people, according to this account, mean something different by “water” before the discovery of its chemical makeup? There is no need to claim this, unless one assumes that the term “water” is always used according to a precisely defined meaning. See PI §§87–88. See also discussion in 3.2. Ms112, 99r / Z §460, except the last sentence. See 3.3 and 5.4. See 7.2 for further discussion of the historicity of philosophy. See Ms152, 16–17, quoted in 4.5. I return to the issue of the justification of Wittgenstein’s conception of philosophy in 7.1. See Ms175, 40v, 41r / OC §§318–320. See also n. 57 in Chapter 3. Baker reaches a similar conclusion about Wittgenstein’s conception as a particular conception, not the conception of essence, on the grounds of his investigation of Wittgenstein’s use of italics. See Baker, Wittgenstein’s Method, 246–247. See Ms119, 43–44 / Ts221, 233/ RFM I §127–128; quoted in 5.4.

6. PHILOSOPHICAL HIERARCHIES AND THE STATUS OF CLARIFICATORY STATEMENTS

1. I imagine this as an argument that an interpreter with a “Hackerian” orientation might present against my reading. From this point of view the necessary rule-governedness of language emerges as the foundation of grammatical investigation, as illustrated by the way Hacker explains Wittgenstein’s notion of agreement, i.e., the statement that in philosophy nothing disputable is said. According to this interpretation, language itself in the capacity of a rule-governed activity excludes as nonsense (or irrelevant) any philosophical statements except a certain well-defined set of grammatical statements concerning the grammatical order of language that everyone will have to agree about. (See 4.1, esp. n. 21, and 6.3–6.4.) Clearly, the possibility of grammatical investigation in this sense depends on the thesis that rules determine

334

2.

3. 4. 5.

6.

7. 8. 9. 10.

11.

Notes to Pages 218–223

meaning. Many other interpreters also assume that the notion of rulegovernedness plays a fundamental role for Wittgenstein in the sense that it is supposed to explain the possibility of language and meaning. See for example Saul Kripke, Rules and Private Language, 62, 78 (See Saul Kripke, Wittgenstein on Rules and Private Language (Oxford: Blackwell, 1982), 73, 74). Here again Cavell represents a dissenting voice, maintaining that rules do not constitute a foundation of language on which the possibility of the correct application of words in further contexts rests. See Cavell, Must We Mean, 52. See also Cavell, The Claim of Reason, 34. Note that in the following my purpose is not to deny that there are rules that are constitutive of language. It is merely to argue against a metaphysical use of the notion of such rules in philosophy. Ms114, 104 / PG 115, 116; see Ms111, 110; Ts211, 66; Ts212, 222, 223; Ts213, 67; Ms116, 57, 58. AWL 31. Ms153a, 159v, 160r (1931). See also Hilmy, The Later Wittgenstein, chap. 2. for a discussion of the term “metalogic” in Wittgenstein. In his discussion Hilmy focuses on Wittgenstein’s critique of the idea that psychological concepts such as understanding or meaning could be regarded as metalogical concepts in the sense of providing a foundation for language. My discussion can be regarded as complementary to Hilmy’s. A metalogical role is not something reserved for psychological concepts; concepts such as rule, game, or rule-following may also be mistakenly assigned such a role. PI §121, first drafted in Ms156b, 23r, 23v. According to Pichler this notebook dates from 1932–33. See Alois Pichler, Untersuchungen zu Wittgensteins Nachlaß. Working Papers from the Wittgenstein Archives at the University of Bergen, no. 8 (Bergen: Wittgenstein Archives at the University of Bergen, 1994), 121, n. 48. Ms163, 40v–40r. See PI §97. See 1.5. This is not meant to indicate that it should be possible to examine the correctness of clarificatory statements individually. The correctness of such a statement might be decidable only as part of a larger set of clarificatory statements designed to resolve a particular problem or set of problems. There is an analogy at this point to what Wittgenstein says in PI §402 about idealists’ and realists’ disputes concerning the justification of our ordinary forms of expression. My point is that descriptions of language

Notes to Pages 225–231

12. 13. 14. 15.

16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28.

29. 30.

335

in terms of rules require as little extra justification (besides the description’s justification in particular cases) as ordinary true/false statements require justification of the practice of making true/false statements. See PI §120 and discussion in 2.31. For the issue of the dependence of what is essential on one’s actual concerns and interests, see 2.22 and 5.6. Ms173, 11v / ROC 23. See PI §133 and 7.1 for further discussion and a list of different methods. The identification of grammatical investigation with the technique of tabulating rules in this chapter serves the specific purpose of responding to my imagined interlocutor’s critical argument, but it is ultimately a simplification. Ms111, 179; Ms153a, 100v; see Ts219, 10; PI §124. Ts213, 412, 413/ Ms110, 194; see Ts211, 245; Ts212, 1129. Translation modified from PO 169. Baker, Wittgenstein’s Method, 27. Baker and Hacker, Commentary 1:531; see Hacker, Insight and Illusion, 151, 152. Baker and Hacker, Commentary 1:531; see Hacker, “Philosophy,” 340; Hacker, Insight and Illusion, 151–153. Hacker, “Philosophy,” 341; italics in the original. Hacker, Insight and Illusion, 151 and 154, respectively. See also Baker and Hacker, Commentary 1:542. Hacker, “Philosophy,” 341; see also 344. Glock, Dictionary, 278–283; and Kenny, The Legacy, 42, 43 offer similar readings. See Baker and Hacker, Commentary 1:542. Ibid., 532; see also 544. See Hacker, Commentary 4:244, 245. See Ms117, 192; Ms122, 27r. Ms136, 11a, 18b; see also Ms137, 77b. Rather than something that could be left behind once and for all, dogmatism and the tendency to metaphysical thinking seem to be something that one has to continuously resist and struggle with in philosophy (at least in the present era). Indeed, the idea of a decisive trick that could somehow separate one from metaphysics and dogmatism in one blow and for the rest of one’s philosophical career has a suspiciously metaphysical flavor to it. Instead of such a once-and-for-all liberation, Wittgenstein offers, as noted in 4.1, a particular strategy for avoiding dogmatism, it being up to each individual philosopher to put this strategy into use. See 3.4–3.6. Ms147, 4r (quoted in 4.2).

336

Notes to Pages 231–240

31. To the extent that concepts are used on the basis of criteria, one might also say that grammatical rules are a means of spelling out or reminding ourselves of the criteria for the employment of our concepts. (Not that rules always have to remind us about our concepts in any straightforward way. A definition can also be employed to spell out how language is not used, i.e., to counteract an inclination to describe language in a misleading way.) This characterization of the role of rules allows me to establish an affinity with Cavell, who characterizes Wittgenstein’s notion of a criterion by saying: “Wittgensteinian criteria are appealed to when we ‘don’t know our way about,’ when we are lost with respect to our words” (The Claim of Reason, 34). Just as I have characterized grammatical rules as instruments of clarification, so Cavell here characterizes statements about criteria (as employed in Wittgenstein’s philosophy) as clarificatory statements. 32. See Hacker, Insight and Illusion, 152; Commentary 4: 238–239. 33. PI §122. 34. A reading along these lines is also defended by Baker (Wittgenstein’s Method, 56–59), who provides a variety of observations in support of it. See also Wittgenstein’s Method, chap. 1. 35. Ms135, 137, 138 (1947). Recall also the passage from Ms120, 143v quoted in 1.3, in which Wittgenstein talks about pictures as a means of organizing. 36. Ms162b, 66v, 67r (1940). 37. Ms112, 112r; see Ms153a, 156r. See also Ts220, §99 and §100, where the latter is the equivalent of PI §122. 38. Ms152, 91 (1936). Note also how Wittgenstein’s revision of this remark—his substitution of “use of our words” for “the grammar”— might be taken to support what I said about PI §122 above: it is not grammar that one seeks to present perspicuously but the use of words. 39. Ms110, 256, 257; translation modified from PO 133. 40. See 6.5 for further discussion. 41. Ms110, 257/ PO 133; instead of “finding,” PI §122 has “finding and inventing” (see above). 42. I return to this issue and to Wittgenstein’s lack of a thesis about the essence of philosophy in 7.1. See also 6.6 for why Wittgenstein’s approach should not be seen as weakening philosophy. 43. See the foreword to PR and Ms109, 204–209/ CV 8–11. See 7.3 for a discussion of the ethical dimension of Wittgenstein’s philosophy. 44. Ts213, 79r / Ms114, 118; translation modified from PG 126–127. 45. See Ms114, 159–161/ PG 184–186; Ts228, §§386, 387/ Z §§331, 320. 46. PI §497; see Ms110, 176; Ts211, 239; Ms116, 134; Ms117, 140.

Notes to Pages 240–247

337

47. PI §496. 48. Ms136, 121b / Z §351. A similar point is made about the concept of argumentation in Ms118, 22r. 49. Essentially the same interpretation is propounded by Baker and Hacker, but I will restrict my discussion mainly to Glock and only occasionally quote from Baker and Hacker. Recall, however, the description of their position in 4.1, n. 21, and see Baker and Hacker, Commentary, 1:480, 490, 491; Hacker, “Philosophy,” 334, 335; Commentary, 4:214, 215, 242–245; and Insight and Illusion, 159. 50. All quotations from Hans-Johann Glock, “Philosophical Investigations Section 128: ‘Theses in Philosophy’ and Undogmatic Procedure,” in R. L. Arrington and H.-J. Glock eds., Wittgenstein’s Philosophical Investigations: Text and Context (London: Routledge, 1991), 78. All italics in the original. 51. Ibid., 84; italics in the original. 52. Ibid., 79. Similarly, Kenny characterizes Wittgensteinian philosophical statements as “assured of universal assent.” Kenny, “Philosophy States Only What Everyone Admits,” 178. 53. Glock himself characterizes it as exemplifying the method of Hegelian immanent critique. See “Undogmatic Procedure,” 86. 54. Ibid., 86. 55. See PI §109. 56. See PI §242, quoted at the end of 3.7. 57. See PI §§224, 225. 58. See 6.6 for further discussion. 59. See Ts220, §102. 60. Note that I am not denying that a logical argument could be described as coercing someone or as fitting the form “You admit this, you admit that, so you must admit that too.” Wittgenstein himself describes it in these terms, for instance in Ms117, 27. The crucial difference is that the acknowledgment of a grammatical rule is not unconditional in the same sense as the acknowledgment of a true premise. Rather, agreement on a clarificatory rule concerns something specific: a description of a concept is agreed upon with some particular concerns in mind, and the agreement is relative to such concerns. See below and 6.5. 61. As Cavell emphasizes, none of the participants in a philosophical discussion can privilege their knowledge of the use of language over that of others. See Cavell, Must We Mean, 239–241. Accordingly, there is indeed such a thing as failure to reach agreement and something one might call “an intellectual tragedy.” See The Claim of Reason, 19. 62. See also Baker, Wittgenstein’s Method, 269.

338

Notes to Pages 247–253

63. PI §599 (modified translation). 64. Similarly, Baker and Hacker maintain that it is not possible to disagree with the philosopher’s statement of a grammatical rule. Whoever disagrees is either failing to make sense or stating something irrelevant to the philosophical issue at stake. See 4.1, n. 21. 65. Ms123, 18r (1940). 66. PI §128 / Ms110, 259; see WVC 183. 67. Ms110, 214 / Ts211, 256 (1931). A similar point is made in Ms136, 41b (1948). 68. LFM 22; see 55, 102, 103, AWL 97. 69. Hacker, “Philosophy,” 334. 70. LFM 55. 71. In this connection, Wittgenstein’s exact wording is important. As Kenny points out (“Philosophy States Only What Everyone Admits,” 173), according to Wittgenstein philosophy only says what everyone agrees upon (PI §599), not what everyone should agree upon, as Glock and Hacker seem to maintain. 72. Recall the discussion in 1.4 of Wittgenstein’s notion that to deal with a philosophical problem is to deal with its genesis, i.e., problematic tendencies of thinking that give rise to philosophical problems. 73. See 7.3 for further discussion. I suspect that any appearance of plausibility the idea that there should be universal assent to grammatical statements may have derives from a picture of philosophy as a science. This picture—not, for instance, the facts of the history of philosophy!—suggests that eventually an agreement can be reached about philosophical theses, just as scientists are able to reach agreement. However, perhaps the wide disagreement in philosophy should be understood as symptomatic, as something essential rather than accidental—as indicating that philosophy is not to be understood on the model of the sciences after all. 74. Ts213, 420 / PO 181; see PI §131. 75. I return to the problem of injustice in 7.3. 76. Ms122, 14r / RFM 147; see AWL 97. 77. PI §§53, 54, 82, 83. In an important sense it makes no difference whether “the observer” is the user of language herself or an outsider. Insofar as meaning is conceived as constituted by rules, I do not have “immediate privileged access” to the concepts of the language that I use. The meaning of expressions is independent of what I think their meaning is, and it is possible for me to characterize as well as use words and concepts incorrectly. I take this to follow from Wittgenstein’s discussion of the notion of essentially—not merely accidentally—“private rules,” but I cannot go into this here (see PI §§202, 258).

Notes to Pages 254–258

339

78. See Ms112, 99r / Ts213, 426 / PO 189; Ts213, 415/ PO 172; PI §125. 79. See passages from Wittgenstein to this effect in 6.2. 80. It is important to recall that using an expression and describing its use are two distinct things (see Z §§111, 114, 121, quoted in 3.3). It does not follow, therefore, from what I am saying that one must decide between an infinite number of possible interpretations when using language—as suggested by the skeptical paradox construed by Kripke in Rules and Private Language. (See PI §201.) Here it is important that Wittgenstein declares the paradox to be based on a misunderstanding in the very same remark in which he formulates it—a point Kripke simply ignores. In the case of philosophical descriptions of language in terms of rules, on the other hand, the number of possible descriptions is limited by the purpose those descriptions serve. Not just any characterization can resolve philosophical problems. 81. PI §109; see §§90, 118, 127. 82. Ms112, 119v / Ts213, 416; translation modified from PO 173–175. 83. See AWL 44. 84. See discussion in 1.3. 85. My discussion of this problem is brief sketchy and is not intended as a proper discussion or dissolution of this problem, only as an illustrative example. 86. Due to the complexity of philosophical problems, it might only be possible to evaluate the correctness of a particular grammatical statement in connection with others, as part of a more complex account designed to solve many connected problems. As noted in 1.3, solving one philosophical problem may require solving many. 87. PI §132; see 2.32 and 6.2. 88. Glock, “Undogmatic Procedure,” 84. 89. See 1.2–1.3. 90. See AWL 20, 21. Naturally, a problem being dissolved does not mean merely that one thinks it has been dissolved. The criterion is whether one now is actually able to find one’s way about the problematic concepts and is released from conflicts. It may well happen in a particular case that one comes to realize later that, contrary to what one thought, the problem was not dissolved after all but the conceptual relations are still unclear, the problem is repeated in a different form, and so on. Or perhaps some progress was made, but one is not quite in the clear yet, and so on. 91. Subchapters 3.4 and 3.5 and Chapter 4. 92. PI §492, discussed in 4.3. 93. Baker, Wittgenstein’s Method, 283, 284; he makes the same point about pictures in Wittgenstein’s sense on 268.

340

Notes to Pages 259–266

94. Ibid., 284; see also 43, 268. 95. Baker notes the complementarity of different aspects in his discussion of the concept of aspect-seeing (Wittgenstein’s Method, 280) but does not discuss complementarity in the case of philosophical conceptions. Apparently, complementarity is meant to be covered here by “global plurality.” 96. On the one hand, he suggests looking at mathematical propositions as arbitrary rules; on the other hand, it is crucial to mathematics that it is not just a game but has practical application. RFM 99, 257. 97. Ms124, 13, 14 / RFM 357. 98. See 4.4; Ms180b, 4r. 99. See the quotation from Ms163 below. 100. AWL 108. 101. Ms110, 164 / Ts213, 83. 102. Ms111, 107. 103. DC 105. 104. See 3.5, 3.6, and 6.5. 105. Ms163, 58r–58v (1941). 106. See RPPi §950 and Z §372 for an example. I do not think that this is incompatible with statements of Wittgenstein’s such as the following, quoted in 1.2: “As I do philosophy, its entire task consists in expressing myself in such a way that certain problems disappear” (Ts213, 421/ PO 180). As noted, the disappearance of problems does not imply the disappearance of our interest in the matters at hand. Rather, the disappearance of problems just means we are able to think about the matter without entanglement in confusions. Recall also Wittgenstein’s characterization of the goal of philosophy as the state of “thoughts at peace” (Ms127, 82 / CV 50; see 1.5). Surely thinking peacefully (without disquietude) would still be a form of thinking. As for the practical applications of philosophical conceptions, the application of philosophical conceptions outside of philosophy is, of course, no longer (purely) the responsibility of philosophy. But one way philosophy influences the sciences, for instance, is by articulating conceptions and ways of looking at things that scientists may adopt and employ in their work.

7. WITTGENSTEIN’S CONCEPTION OF PHILOSOPHY, EVERYDAY LANGUAGE, AND ETHICS

1. See 6.5. 2. PI §133.

Notes to Pages 267–270

341

3. See 1.5 for discussion of this issue. 4. See 3.2. 5. I am ignoring for the moment any purposes Wittgenstein’s examples and discussions may serve besides the introduction of the method. But there is no need to assume that each discussion serves some single purpose exclusively. 6. It would also be misleading to assume some kind of fundamental contrast between Wittgenstein’s examples, which help to demonstrate the method, and his explicit statements about philosophy. Such explicit statements are part of the demonstration of a method through examples, and are intended to elucidate the import of the examples. 7. Metaphysical attempts to justify an approach once and for all will, after all, also be judged on the basis of whether they work—granted one’s interests are philosophical and not, say, primarily aesthetic, political, or religious. 8. See 3.4 for the status of philosophical models. 9. PI §133. 10. See PI §90. 11. For his reasons for avoiding this kind of approach, see 1.5 and 6.1. 12. As explained in the introduction, my focus in this book has been on the method of clarification with the help of rules because it seems particularly suitable for explaining what a transition from “traditional modes of philosophizing” to Wittgenstein’s approach would amount to. Nevertheless, this focus—which is essentially a focus on particular philosophical problems relating to the concept of philosophy—is not meant to disparage the importance of Wittgenstein’s other methods of clarification. 13. This list has no pretension to completeness. The borderlines between its items might not be clear-cut, and the methods may also overlap. See Savickey, Wittgenstein’s Art for a discussion of some of these methods. 14. As Wittgenstein explains, hypothetical intermediate cases serve to draw attention to similarities between cases. (See Ms110, 257; PI §122.) 15. “When I describe language, I describe the way people behave, so to speak, ethnologically” (Ms124, 253). See Ms162b, 67v / CV 45; Ms117, 256. 16. “One of my most important methods is to imagine a historical development of our ideas different from what has actually occurred. If we do that the problem shows us a quite new side” (Ms162b, 68v / CV 45). 17. By imagining the facts of nature to be different, one can highlight dependencies between language and the world and make different conceptual formations conceivable. See PI II xii.

342

Notes to Pages 270–273

18. See Ms122, 14r / RFM 147; quoted in 6.4; see also AWL 97. 19. “In philosophy it is significant that such-and-such a sentence makes no sense; but also that it sounds funny” (Ms137, 73a / Z §328); see also PI §111. “I parody a conception in order to show a mistake in it” (Ms112, 33r). 20. See 2.22. 21. Ms156b, 6r, 6v. 22. Consider, for instance, the question of whether Zen Buddhism is a form of philosophy. Clearly, the practice of Zen through archery, for instance, is not concerned with language (except perhaps as an attempt not to employ language). Crucially, however, from the point of view of my interpretation of Wittgenstein, there is no need to either deny or affirm that Zen is a form of philosophy. The important point is that one can decide separately what one wants to say about this particular case. Wittgenstein’s conception of philosophy does not force one to adopt any particular view about this matter and thus allows one to avoid dogmatism. 23. MS 153a, 35r / CV 13. This conception goes back to the pre-Tractatus writings. As Wittgenstein remarks in the Notes on Logic: “Distrust of grammar is the first requisite for philosophising” (NB 106). 24. PI §109. 25. See PI §108 and 2.3. 26. AWL 98. 27. See Ms135, 139. 28. See prefaces to PR and PI; Ms109, 204–210 / CV 8–11; Ms126, 129/ CV 49; Ms134, 146 / CV 70. 29. Ts213, 407/ PO 161, 163; see Ms112, 24r. I return to the idea of philosophy as work on oneself in 7.3. 30. BB 17, 18, 20; see also Ms114, 108 / PG 120; TLP 3.3421. 31. BB 17. 32. Ibid., 17 and 18, respectively. 33. Recall what was said in 1.4 on tendencies of thinking as the proper object of Wittgensteinian philosophical therapy. These passages in the Blue Book may be characterized as identifying a main target of Wittgensteinian philosophical therapy or a main area where philosophers schooled in the Western tradition, according to Wittgenstein, need to engage in (self-)therapy. See also Ms133, 87r. 34. Ms183, 125, 126 / PPO 133, 135. 35. See Ts213, 424. 36. Accordingly, von Wright seems exactly right when he writes: “I do not think that Wittgenstein would have claimed that his conception of

Notes to Pages 273–277

37. 38. 39. 40.

41. 42.

43.

343

philosophy was valid for all the historical phenomena which we heap under the label ‘philosophy.’ . . . His way of seeing philosophy was not an attempt to tell us what philosophy, once and for all, is, but expressed what for him, in the setting of his times, it had to be.” Georg Henrik von Wright, Wittgenstein (Oxford: Basil Blackwell, 1982), 216. AWL 27–28. See also BB 28, PO 113. Ms132, 7–8 (1946). See also BB 58, 59; AWL 108, 109. Baker and Hacker explain metaphysical use as a misuse on the grounds that “it is in the language-game where the expression is at home that it has an established and intelligible sense.” Gordon Baker and P. M. S. Hacker, Wittgenstein: Understanding and Meaning, An Analytical Commentary on the Philosophical Investigations, vol. 1, rev. ed. (Oxford: Blackwell, 2005), 253. On page 254 ordinary use is characterized as constituting a “standard of correctness” for language use. See also Glock, “Undogmatic Procedure” for a similar view. PI §116, translation modified. See also Ts211, 155. See Baker, Wittgenstein’s Method, chap. 4. As explained in Chapter 3 above, metaphysics in a sense relevant for Wittgenstein is concerned with essences and the necessary characteristics of things. (My interpretation agrees here with Baker’s.) At this point the narrowness of this definition should not be seen as a defect, provided that Wittgenstein’s purpose is not to give a once-and-for-all definition of the concept that covers every possible instance of metaphysics. Rather, the definition is expressive of certain particular interests: it is to be employed in the context of certain specific problems relating to metaphysical philosophy. (This is what I meant when I said in 3.1 that the definition of metaphysics there is to be taken as an object of comparison.) For this reason, Schulte’s and Putnam’s critiques of Baker’s allegedly narrow interpretation of Wittgenstein’s definition of metaphysics remain ineffective. The definition may be narrow, one may respond, but that is how it is meant, and other definitions can be employed when necessary. See Joachim Schulte, “Ways of Reading Wittgenstein: Observations on Certain Uses of the Word ‘Metaphysics,’” and Hilary Putnam, “Metaphysical/Everyday Use: A Note on a Late Paper by Gordon Baker,” in G. Kahane, E. Kanterian and O. Kuusela, eds., Wittgenstein and His Interpreters, Essays in Memory of Gordon Baker (Oxford: Blackwell, 2007), 145–168 and 169–173, respectively. Ts213, 422–423/ PO 183–185 (1932). See also Ms113, 23v. Recall Wittgenstein’s formulation in PI §109 according to which we have an urge to misunderstand the workings of language.

344

Notes to Pages 278–281

44. Cavell writes about Oxford ordinary language philosophy: “This, I take it, is what the phrase ‘ordinary language’ meant to its Oxford coiners: a view of words free of philosophical occupation” (Cavell, Must We Mean, 238; see Cavell, The Claim of Reason, pt. 1 for a discussion of certain relevant differences between Austin and Wittgenstein.) In a related sense, the later Baker talks about a “genre misidentification” in the case of Wittgenstein’s association with Oxford ordinary language philosophy. (See Baker, Wittgenstein’s Method, 92, 103.) 45. See 2.32, Chapter 4, 6.2, and 6.5. 46. To recap: an examination of everyday language is required insofar as it is the language in which the philosophical questions at hand were formulated, and to the extent that a clear grasp of questions is desirable in philosophy. 47. Referring to the passage about the philosophical deceptiveness of ordinary language, however, one might say the metaphysical tendency of thinking is ultimately only one misleading expectation about how our concepts should work. Other problematic thought habits may require a different treatment. 48. PI §97; see PI §92; Ms152, 92, 93; TLP 5.4711. 49. The term Wittgenstein uses is hausbacken, literally “home-baked.” Later he substitutes “trivial,” “ordinary,” and “humble” for it. (See Ms142, 83, 84; Ms157a, 63r, 63v; cf. Ms110, 34; Ts211, 155; Ts213, 412; Ts220, 90; Ts239, 80; PI §§94, 97.) In the early 1930s the term grob, “rough,” is sometimes used as an alternative to hausbacken (Ms111, 132 / Ms155, 21v, 22r). Hausbacken is contrasted with something sublime (Ms152, 96), extraordinary or unique (Ms117, 138), ideal (Ms114, 109), abstract (Ts213, 71v), or something lying under the surface to be revealed by analysis (Ms157b, 8r, 8v). 50. Ms109, 212–213; translation based on Rush Rhees, “On Wittgenstein IX,” Philosophical Investigations 24 (2001): 154. See also Ms111, 132. 51. See 3.4 and, for instance, Ms114, 109 on the generality of logical statements. 52. The latter interpretation would be in conflict with Wittgenstein’s repeated statements about the fluidity of everyday uses of language. (See Chapters 3 and 4.) Here the requirement of precision is not optional: to assume, since everyday uses are fluid, that it is good enough to establish a rough match between them and philosophical uses would be to give up on the rigor of logic. My rejection of the idea that §116 implies a conception of everyday uses as something fixed seems to agree with Cavell’s interpretation, according to which Wittgenstein’s leading words back means not “returning to a place, but . . . returning words to the

Notes to Pages 282–285

53.

54. 55. 56.

57.

58. 59.

60. 61. 62. 63.

345

circulation in language,” where the idea of circulation includes the possibility of unpredictable employments of words in new contexts, in contrast to their limited use in certain fixed roles and services. Stanley Cavell, “The Wittgensteinian Event,” in Philosophy the Day after Tomorrow (Cambridge, MA: Belknap Press / Harvard University Press, 2005), 199. Examples of technical notions in this sense are the concept of reading as defined in PI §156 (see 3.6) and meaning as constituted by grammatical rules (see 4.2). Rupert Read argues that a proper comprehension of Wittgenstein’s later approach involves the overcoming of any technical philosophical notions. His use of “technical notion,” however, differs from mine, standing roughly for what I have called a “philosophical thesis.” Hence the conflict seems only a terminological one. See Rupert Read, “Throwing Away ‘the Bedrock,’” Proceedings of the Aristotelian Society 105 (2005): 81–98. PI §246 and §580, respectively. McGinn suggests such a characterization of Wittgenstein’s development in her Elucidating the Tractatus, chap. 12; see esp. 286. As Wittgenstein remarks: “In philosophy one is constantly tempted to invent a mythology of symbolism or of psychology, instead of simply saying what everyone knows and must acknowledge.” (Ms116, 95/ PG 56). This remark is first drafted in Ms108 in 1930 and repeated many times in the Nachlass. (See for example Ts209, 9/ PR §24; Ts213, 145; Ts233, 44 / Z §211.) For instance, relating to the notion of generality and the assumption of the mutual independence of the truth-values of elementary propositions. See PI §93. Wittgenstein writes late in his career (in 1948): “When one philosophizes one has to descend into the old chaos & feel at home there.” Ms136, 51a / CV 74; translation modified. See 5.4. Ts213, 406, 407/ PO 161; see Ms112, 111v, 112r, 153, 155v. See also Ms174, 6r for a related point. Cavell, The Claim of Reason, 46. See Cavell, Must We Mean, chap. 9, 11–12, 15; “The Wittgensteinian Event,” 195, 210–211; “The Uncanniness of the Ordinary,” in In Quest of the Ordinary: Lines of Skepticism and Romanticism (Chicago: University of Chicago Press, 1994), 154, 163–164, 170–171. From the point of view of the notion of philosophical therapy presented in 1.4, according to which the aim of Wittgensteinian philosophical therapy is to treat the tendencies of thinking that give rise to philosophical problems, the

346

64.

65.

66. 67.

68. 69. 70.

Notes to Pages 285–286

Wittgensteinian approach outlined by Cavell clearly qualifies as therapeutic in the relevant sense. And although Wittgenstein only rarely characterizes himself as responding to skepticism or discusses a philosophical position by that name, what Cavell understands by skepticism can be seen as an example of the (more general) tendency to give explanations in the sense of providing foundations rather than descriptions, which is a central target of Wittgenstein’s critique. Accordingly, as Cavell notes, “skepticism for Wittgenstein is the intellectual twin of metaphysics” (“Wittgensteinian Event,” 195). See Cavell, Must We Mean, 263; see also Stanley Cavell, Conditions Handsome and Unhandsome: The Constitution of Emersonian Perfectionism. The Carrus Lectures 1988 (Chicago: University of Chicago Press, 1990), 23. Cavell, Must We Mean, 66. For a discussion of why one need not assume, at this point, any fundamental division or opposition between individual and social emancipation, see Wallgren, Transformative Philosophy, chap. 7. Ms157a, 58r. As noted in 6.2, there is no once-and-for-all liberation from the problem of dogmatism. This indicates a need for continuous selfexamination in philosophy. Ms154, 16v / CV 16. Ms132, 75, 76 / CV 60. For a discussion of this issue, see James Conant, “On Going the Bloody Hard Way in Philosophy,” in John H. Whittaker, ed., The Possibilities of Sense (Houndmills, UK: Palgrave, 2002).

Index

Agreement, Wittgenstein notion of, 13, 156–57, 216, 243–45, 247–52, 337nn 52, 60, 338n73 Analysis, logical, 11, 23, 54, 55, 59–60, 65, 67, 69, 71, 72, 75, 82, 98–102, 108, 110, 220, 249, 270, 307n125, 308n14, 310n20, 322n14, 332n54; completeness of, 54, 66–69, 80–82, 85, 91, 319n146 Anscombe, G.E.M., 291n20, 299nn14, 25, 329n9 A priori, 104, 310n27, 311n28 Arbitrariness, of grammar, 166, 168, 171, 185–89, 239, 253, 255, 259, 261, 326n57, 329nn3, 7, 330n17, 333n60 Aristotle, 293n42 Aristotelian, 13, 49, 101, 185, 205, 208, 209, 213 Augustine, 37, 42–43, 293n42, 325n50 Austin, J. L., 304n85, 344n44 Baker, Gordon, 8, 93, 94, 288, 258–60, 289nn19, 21, 294n59, 301n38, 307n131, 309n15, 315n92, 324n30, 325n50, 329n4, 15, 333n68, 335n18, 336n34, 337n62, 339n93, 340n95, 343n42, 344n44 Baker [Gordon] and Hacker [Peter], 7–8, 55, 86, 89, 90–92, 96, 111, 113, 116–20, 132–34, 137, 141, 145–46, 149, 153–56, 186–88, 215, 228–30,

232, 235, 275–79, 303n83, 306nn119, 121, 125, 309n16, 312nn46, 313nn62, 73, 320n150, 322n18, 329n9, 331n39, 337n49, 338n64, 343n40 Bloor, David, 204, 329n9, 332n50 Calculus, language as, 42, 67–69, 130, 137–38, 141–42, 152, 169, 218, 260–61, 304n88, 318n123, 319n141, 320n154 Carnap, Rudolf, 290n10, 325n50 Cavell, Stanley, 8, 157, 266, 284–85, 295n82, 304n85, 320n154, 324n26, 332n50, 334n1, 336n31, 337n61, 344nn44, 52, 345nn52, 62, 63, 346nn63, 64, 65 Center of variation, 172–75, 179–81, 182, 234, 270 Conant, James, 62, 290n10, 288n18, 289n19, 290n10, 292n22, 301nn33, 34, 346n70 Concept-script, 54, 56–61, 63–64, 66, 69, 74–75, 92, 98–100, 130, 269, 298nn4, 5, 300nn27, 31, 301n31, 308nn6, 15, 309n18, 310n19 Conceptual investigation, 87, 96, 103–5, 111–13, 115, 119–20, 132, 146, 245 Condition of possibility, 23, 182, 217, 241–42, 290n10. See also Necessary condition

348 Constructivism, 12, 184–91, 204–5, 208, 210–11 Conventionalism, 12, 184, 185, 187–88, 208, 210, 315n83, 329n15 Correctness, of grammatical rule/statement, clarificatory remark/statement, 13, 85, 109, 216, 223, 252–57, 334n10, 339n86 Crystalline purity, of logic 111, 121, 137; ideal of 133, 134, 135, 140, 317n116 Deans, Rob, 301 Diamond, Cora, 7, 62, 63, 288nn10, 18, 288–89n19, 290n10, 292nn22, 26, 295n96, 298n5, 299nn13, 16, 300n31, 301nn33, 34, 37, 309n18, 310–11n27, 313–14n73 Doctrine, philosophical, 1–3, 7, 9–10, 13, 24–27, 53, 55, 58, 64, 75, 96–97, 100, 102, 179, 270, 273, 291–92n20, 299n24, 308n4, 308–9n17. See also Theory, philosophical; Thesis, philosophical Dogma, 109–111, 118–20, 125, 139, 144; dogmatic 109, 119, 122–24, 130, 140–41, 144–46, 162, 168, 174, 182, 191, 209, 216, 222, 229, 232, 242, 258, 261–62, 279, 281–83, 285, 323n24; dogmatism, 3, 12, 92–93, 95, 96–97, 109, 111, 118–19, 121–23, 126, 128–32, 140–42, 145, 154, 156, 163, 168, 169–71, 176, 180, 182, 230, 232, 237, 244, 266–67, 273, 286, 303n78, 307n129, 313–14n73, 314–15n83, 316n104, 317n115, 318n134, 323n24, 328n97, 335n28, 342n22, 346n67 Drury, M. O. C, 261 Ethical, 13, 142, 143, 266, 275, 284–86, 336n43 Essence: traditional conception of, 97–98; Wittgenstein’s conception of, 12, 26, 37–38, 47–49, 64, 66, 76–77, 83–84, 93, 98–103, 106–10, 135–36, 149–50, 154, 163–64, 167, 169, 170, 172–74, 180–82, 184, 184–213, 220, 236, 242, 292n22, 296n99, 297n111, 303n74, 317n115, 324n31, 330nn17, 27, 343n42

Index Emptiness, 126–27, 207, 223–24, 315n92, 315n95 Everyday language, 13, 19, 40, 56, 59, 65, 77, 79, 110, 126–27, 146, 265–66, 275–83, 299n14, 303n83, 304n85, 344n46, 344n52 Facts of nature, 165, 167, 186, 187, 190, 198–200, 202, 209 341n17 Factual statement, 3, 12, 33, 63, 98, 103–4, 114, 147–48, 184, 189, 192, 195–98, 200, 202–3, 205, 208–11, 309–11, 313nn56, 57, 319n143, 331nn39, 45, 333n60; statement of a fact, 113–14 Family-resemblance concept, 144, 149, 172–73, 175, 235, 262, 327n72 Flexibility, 3, 12, 150, 171, 175, 181, 213 Floyd, Juliet, 298n5, 300–301n31 Fogelin, Robert J., 316n105 Form of life, 28, 147, 187, 189, 269, 283, 285, 316n104, 320n158, Forster, Michael N., 186, 329n7 Frege, Gotlob, 56, 58, 62, 289n5, 298n12 Fundamental problem, task, 50–51, 101–2, 220, 226–27, 238, 296n101, 310n20 Garver, Newton, 9, 186, 288n17, 303n72, 328n94, 329n7 Generality, 98–99, 105–6, 124–25, 131, 153, 158, 191, 195, 268, 272, 275, 281, 283, 308n5, 324n30, 344n51, 345n57 Gerrard, Steven, 288n13 Goldfarb, Warren, 300n31 Glock, Hans-Johan, 7, 153, 238, 242–44, 246–48, 251, 257, 289n21, 322n18, 326n68, 329n15, 337n53 Grammatical rule, 103, 113, 116, 146, 150, 152–56, 161, 162, 166, 168, 186, 188, 191, 195, 205, 207–9, 216–17, 222, 224, 226, 229–31, 233, 239–40, 245, 251, 269, 278, 315n98, 323n31, 336n31, 337n60; grammatical truth, 195, 215, 223–24, 226–27, 248–49; grammatical statement, 147, 154–55, 184–85, 190–91, 195, 197, 201, 204,

Index 208, 210–12, 223, 227, 248–50, 264, 323n21, 333n1, 338n73 Hacker, P. M. S., 62, 93, 111–14, 116–20, 137–46, 169–70, 180, 186–87, 202–4, 228–31, 248, 288n17, 288–89n19, 290n10, 291–92n20, 300n28, 308–9n15, 312n46, 313n57, 318n123, 328n94, 329n9, 332nn50, 56, 333–34n1. See also Baker and Hacker Hanfling, Oswald, 76–77, 303n72 Heidegger, Martin, 296n101, 312n41 Hierarchy, 101, 219; hierarchical, 11, 48–49, 215, 219–22, 224, 270, 296n101 Hilmy, Stephen, 13–14, 288n19, 289n20, 313n73, 316n102, 318n123, 323n20, 334n5 Hintikka, Jaakko, and Merril B., 5, 287n4, 303n83 Historical, 76–78; historicity, 76–78, 210–12, 265, 271–75, 296n95, Idealism, 185–87, 191, 210, 307n3, 329n9 Immutable, 77–78, 185, 196–98, 211, 284, 303n74 Ineffable, 299, 234, 292n22, 301n33, 308–309n15 Introspection, 192, 330n27 Instrument, language as, 163–68, 169–71, 247–50 Instrumentalism, 147, 226 Justice, injustice, 3, 126–27, 150, 168–71, 180, 185, 213, 229, 246, 250–51, 273, 281, 285–86 Kannisto, Heikki, 299n24 Kant, Immanuel, 22, 23, 25, 241–42 Kantian, 153, 185, 187, 205, 208, 210, 213, 241–42, 243 Kienzler, Wolfgang, 289n19, 313n73, 316n101 Kremer, Michael, 292n22 Kripke, Saul, 6, 7, 8, 204, 288n9, 289n1, 294n55, 307n130, 310–11n27, 322n13, 323n20, 331n45, 332n50, 339n80

349 Language, essence of, 22–24, 47–48, 64, 66, 102, 106, 156, 163–67, 169, 171–72, 220, 225, 251, 279–80, 282–83, 295n95, 309n17, 317n116, 326n65; Language-game, 6, 43, 70–73, 88, 120, 128–29, 131–32, 138, 190, 253, 270, 276, 281, 303n83, 316n102, 318n123, 319n141, 324n30, 343n40 Leibniz, 58, 298n12 Leibnizian, 198 Logical syntax, 20, 56, 58, 60, 63, 90, 290–91n10 McDonough, Richard M., 301n38 McGinn, Marie, 7, 292n26, 300n26, 309n17, 345n55 McGuinness, Brian, 15 McManus, Denis, 288n19, 292n22, 298n4, 300n25 Malcolm, Norman, 6, 287n8, 288n9, 291–92n20, 327n80, 329n9 Meaning, essence of, 149, 154, 161, 164, 176–80, 322n13, 324n31, 328n94 Metaphysics, metaphysical, 3, 10, 12, 13, 60, 64, 65, 96, 97–105, 106, 108, 111–12, 118–24, 137, 146, 163, 180, 202–3, 215–16, 226–27, 228, 230, 236–37, 245, 261, 265, 267, 270, 274, 276–81, 283, 285–286, 308n14, 310n22, 311n28, 312n47, 317n116, 334n2, 335n28, 341n7, 343n42. See also Projection, metaphysical Mode of presentation, 70, 105–7, 109, 122, 125, 129, 132, 136, 141, 154, 226, 229, 231, 233, 235, 311n28, 316n102 Moore, G. E., 86, 287n2, 303n83 Morris, Katherine J., 161–62, 258, 288–89n19, 325nn43, 44 Mulhall, Stephen, 301n38, 312n41, 317n122 Multidimensional description, 13, 216, 258, 260–61, 305n101, 326n59 Necessary condition, 23, 25, 119, 140, 150, 153–54, 159, 160, 163, 165, 172, 175, 177, 180, 182, 242, 323n20, 327n72

350 Necessary truth, 3, 32, 97, 99, 102–4, 111, 114, 123, 146, 199–200, 226, 310n27 Necessity, exceptionless 3, 12, 125, 145, 148, 156, 163, 184–85, 187–91, 199, 208–10, 212, 213, 320, 321n159; factual, 118, 185, 198–204, 211, 331n45, 332n50 Nontemporal, 12, 75, 184–185, 195–197, 201–4, 210, 312n55, 331n39 Object of comparison, 97, 122–26, 129, 132, 140, 145, 155–56, 160, 169, 174–75, 180, 182, 210, 212, 217–18, 226, 238, 244, 252, 261–62, 280–81, 283, 301n31, 319n143, 343n42 Once-and-for all answer, 48, 51–52, 74, 84, 93, 269, 296nn98, 99; definition, 91, 343n42; determination 54–55, 75, 81, 107, 172, 224, 269 Ordering, 86–95, 233, 234, 257, 306nn117, 121, 313n66 Ostrow, Matthew, 288n19, 292n22, 299n14, 301n33 Perspicuous presentation, 112, 161, 214–16, 228–37, 268, 269, 300n27, 304n88, 305n107, 306n121, 320n150, 323n24 Phenomenological, 192 Philosophy, essence of, 265–71 Philosophical Investigations, 6, 13–15, 45, 46, 48–50, 52, 65–66, 81, 96, 121, 128–31, 143–44, 221–22, 233, 266–67 Pichler, Alois, 288–89n19, 285n96, 313–14n73, 334n6 Picture, philosophical, 35–38, 41–42, 89, 122, 142–43, 158–61, 169, 171, 200, 231–32, 277–79, 283, 316n104, 319n141, 325n50; theory of proposition as, 60–62, 107–9, 118–19, 125, 261–62, 299n25, 315n95 Pitcher, George, 5, 287n3 Plato, 181, 272–74, 331n39 Platonic, 37, 110, 181, 185, 196–98, 210, 212, 310n22 Projection, metaphysical, 105, 108, 109, 111–12, 118, 121–24, 128, 132, 145, 187, 188, 215, 226, 228, 245, 246, 267, 274–75, 285, 312n47, 329n8

Index Read, Rupert, 288n12, 300–301n31, 344–45n53 Relativism, 13, 263–264 Resolute reading, 61, 62, 292n22, 299n14, 301n33, 308–9n15 Rhees, Rush, 76–77, 303n71, 308n5, 344n50 Rule, statement of a, 113–18, 254, 312n55 Russell, Bertrand, 56, 62, 98, 289–90n5, 296–97n105, 299n14, 326n64 Ryle, Gilbert 304n85, 325n50 Savickey, Beth, 79, 288–89n19, 304n84, 341n13 Schlick, Moritz, 140, 318n125, 326n51 Schulte, Joachim, 289n22, 313–14n73, 318n123, 343n42 Spengler, Oswald, 126, 130, 315n88 Stenius, Erik, 291n18, 299–300n25 Stroud, Barry, 329n15 Style, 281, 313–14n73 Sublimation, of logic, 110, 312n44 Super-expression, 200; super-factual, 103, 146, 148, 211; super-rigid, 200 Superlative, philosophical, 199–200 Theory, philosophical, 2, 6–7, 23, 24, 16, 35, 37, 39, 40, 60–64, 83, 95, 100, 106, 108, 208, 210, 292n22, 305n95, 308n4, 311n36. See also Doctrine, philosophical; Thesis, philosophical Therapy, philosophical, 18, 43–46, 51, 95, 275, 342n33, 345–46n63 Thesis, philosophical, 2, 6–7, 12, 24, 50, 97–98, 99–100, 102, 105–6, 118, 120, 123–25, 136–37, 153–55, 172, 175, 190, 191, 209, 237, 308n5. See also Doctrine, philosophical; Theory, philosophical Tractatus logico-philosophicus, 3, 8–12, 15, 17–26, 29, 30, 35, 46–51, 54–57, 59–69, 71, 74–75, 79–82, 90–92, 96–111, 117–19, 121–23, 126–27, 130–37, 145–46, 180, 220, 229, 261, 265–66, 269–70, 279–80, 282–84, 290–91nn9, 10, 292nn22, 23, 26, 295–96nn95, 96, 298n5, 299n14,

Index 302n48, 306–7n125, 308–9nn14, 15, 18, 311n36 Transcendental, 187, 328n97 Turn, Wittgenstein’s turning the investigation around, 96–97, 102, 105, 109–13, 120–41, 148, 169, 170, 180, 230, 261, 303n75, 313–14n73, 316n102, 323n20, 325n50 Universal, universality 47, 51, 57, 65, 76–78, 98–99, 101, 106–7, 110, 123, 127, 135, 137, 157, 170, 182, 196, 236,

351 242, 251, 267–68, 272–73, 296n99, 303nn75, 78, 308n5, 317n115, 328n94 Waismann, Friedrich, 393n43 Wallgren, Thomas, 320n155, 346n65 Way of looking, 62, 138, 162, 236–38, 259, 261; seeing, 35, 48, 162, 228, 232, 235–37 Williams, Bernard, 329n9 Von Wright, Georg Henrik, 237n80, 342–43n36